d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
c9003bc7fa
freebsd-dev/sys/compat/ndis/subr_ndis.c

3545 lines
84 KiB
C
Raw Normal View History

Start each of the license/copyright comments with /*-
2005-01-05 22:34:37 +00:00
/*-
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
* Copyright (c) 2003
* Bill Paul <wpaul@windriver.com>. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by Bill Paul.
* 4. Neither the name of the author nor the names of any co-contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY Bill Paul AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL Bill Paul OR THE VOICES IN HIS HEAD
* BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF
* THE POSSIBILITY OF SUCH DAMAGE.
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
/*
* This file implements a translation layer between the BSD networking
* infrasturcture and Windows(R) NDIS network driver modules. A Windows
* NDIS driver calls into several functions in the NDIS.SYS Windows
* kernel module and exports a table of functions designed to be called
* by the NDIS subsystem. Using the PE loader, we can patch our own
* versions of the NDIS routines into a given Windows driver module and
* convince the driver that it is in fact running on Windows.
*
* We provide a table of all our implemented NDIS routines which is patched
* into the driver object code. All our exported routines must use the
* _stdcall calling convention, since that's what the Windows object code
* expects.
*/
In subr_ndis.c, when searching for keys in our make-pretend registry, make the key name matching case-insensitive. There are some drivers and .inf files that have mismatched cases, e.g. the driver will look for "AdhocBand" whereas the .inf file specifies a registry key to be created called "AdHocBand." The mismatch is probably a typo that went undetected (so much for QA), but since Windows seems to be case-insensitive, we should be too. In if_ndis.c, initialize rates and channels correctly so that specify frequences correctly when trying to set channels in the 5Ghz band, and so that 802.11b rates show up for some a/b/g cards (which otherwise appear to have no 802.11b modes). Also, when setting OID_802_11_CONFIGURATION in ndis_80211_setstate(), provide default values for the beacon interval, ATIM window and dwelltime. The Atheros "Aries" driver will crash if you try to select ad-hoc mode and leave the beacon interval set to 0: it blindly uses this value and does a division by 0 in the interrupt handler, causing an integer divide trap.
2004-05-29 06:41:17 +00:00
#include <sys/ctype.h>
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
#include <sys/param.h>
#include <sys/types.h>
#include <sys/errno.h>
#include <sys/callout.h>
#include <sys/kernel.h>
#include <sys/systm.h>
#include <sys/malloc.h>
#include <sys/lock.h>
#include <sys/mutex.h>
#include <sys/socket.h>
#include <sys/sysctl.h>
Implement some more NDIS and ntoskrnl API calls: subr_ndis.c: NdisGetCurrentSystemTime() which, according to the Microsoft documentation returns "the number of 100 nanosecond intervals since January 1, 1601." I have no idea what's so special about that epoch or why they chose 100 nanosecond ticks. I don't know the proper offset to convert nanotime() from the UNIX epoch to January 1, 1601, so for now I'm just doing the unit convertion to 100s of nanoseconds. subr_ntoskrnl.c: memcpy(), memset(), ExInterlockedPopEntrySList(), ExInterlockedPushEntrySList(). The latter two are different from InterlockedPopEntrySList() and InterlockedPushEntrySList() in that they accept a spinlock to hold while executing, whereas the non-Ex routines use a lock internal to ntoskrnl. I also modified ExInitializePagedLookasideList() and ExInitializeNPagedLookasideList() to initialize mutex locks within the lookaside structures. It seems that in NDIS 5.0, the lookaside allocate/free routines ExInterlockedPopEntrySList() and ExInterlockedPushEntrySList(), which require the use of the per-lookaside spinlock, whereas in NDIS 5.1, the per-lookaside spinlock is deprecated. We need to support both cases. Note that I appear to be doing something wrong with ExInterlockedPopEntrySList() and ExInterlockedPushEntrySList(): they don't appear to obtain proper pointers to their arguments, so I'm probably doing something wrong in terms of their calling convention (they're declared to be FASTCALL in Widnows, and I'm not sure what that means for gcc). It happens that in my stub lookaside implementation, they don't need to do any work anyway, so for now I've hacked them to always return NULL, which avoids corrupting the stack. I need to do this right though.
2003-12-12 22:35:13 +00:00
#include <sys/timespec.h>
Back out the last batch of changes until I have a chance to properly evaluate them. Whatever they're meant to do, they're doing it wrong. Also: - Clean up last bits of NULL fallout in subr_pe - Don't let ndis_ifmedia_sts() do anything if the IFF_UP flag isn't set - Implement NdisSystemProcessorCount() and NdisQueryMapRegisterCount().
2003-12-26 03:31:34 +00:00
#include <sys/smp.h>
Implement NdisScheduleWorkItem() and RtlCompareMemory(). Also, call the libinit and libfini routines from the modevent handler in kern_ndis.c. This simplifies the initialization a little.
2004-01-04 07:47:33 +00:00
#include <sys/queue.h>
Implement NdisOpenFile()/NdisCloseFile()/NdisMapFile()/NdisUnmapFile(). By default, we search for files in /compat/ndis. This can be changed with a systcl. These routines are used by some drivers which need to download firmware or microcode into their respective devices during initialization. Also, remove extraneous newlines from the 'built-in' sysctl/registry variables.
2004-01-09 03:57:00 +00:00
#include <sys/proc.h>
Implement support for single packet sends. The Intel Centrino driver that Asus provides on its CDs has both a MiniportSend() routine and a MiniportSendPackets() function. The Microsoft NDIS docs say that if a driver has both, only the MiniportSendPackets() routine will be used. Although I think I implemented the support correctly, calling the MiniportSend() routine seems to result in no packets going out on the air, even though no error status is returned. The MiniportSendPackets() function does work though, so at least in this case it doesn't matter. In if_ndis.c:ndis_getstate_80211(), if ndis_get_assoc() returns an error, don't bother trying to obtain any other state since the calls may fail, or worse cause the underlying driver to crash. (The above two changes make the Asus-supplied Centrino work.) Also, when calling the OID_802_11_CONFIGURATION OID, remember to initialize the structure lengths correctly. In subr_ndis.c:ndis_open_file(), set the current working directory to rootvnode if we're in a thread that doesn't have a current working directory set.
2004-02-03 07:39:23 +00:00
#include <sys/filedesc.h>
Implement NdisOpenFile()/NdisCloseFile()/NdisMapFile()/NdisUnmapFile(). By default, we search for files in /compat/ndis. This can be changed with a systcl. These routines are used by some drivers which need to download firmware or microcode into their respective devices during initialization. Also, remove extraneous newlines from the 'built-in' sysctl/registry variables.
2004-01-09 03:57:00 +00:00
#include <sys/namei.h>
#include <sys/fcntl.h>
#include <sys/vnode.h>
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
#include <sys/kthread.h>
Big mess 'o changes: - Give ndiscvt(8) the ability to process a .SYS file directly into a .o file so that we don't have to emit big messy char arrays into the ndis_driver_data.h file. This behavior is currently optional, but may become the default some day. - Give ndiscvt(8) the ability to turn arbitrary files into .ko files so that they can be pre-loaded or kldloaded. (Both this and the previous change involve using objcopy(1)). - Give NdisOpenFile() the ability to 'read' files out of kernel memory that have been kldloaded or pre-loaded, and disallow the use of the normal vn_open() file opening method during bootstrap (when no filesystems have been mounted yet). Some people have reported that kldloading if_ndis.ko works fine when the system is running multiuser but causes a panic when the modile is pre-loaded by /boot/loader. This happens with drivers that need to use NdisOpenFile() to access external files (i.e. firmware images). NdisOpenFile() won't work during kernel bootstrapping because no filesystems have been mounted. To get around this, you can now do the following: o Say you have a firmware file called firmware.img o Do: ndiscvt -f firmware.img -- this creates firmware.img.ko o Put the firmware.img.ko in /boot/kernel o add firmware.img_load="YES" in /boot/loader.conf o add if_ndis_load="YES" and ndis_load="YES" as well Now the loader will suck the additional file into memory as a .ko. The phony .ko has two symbols in it: filename_start and filename_end, which are generated by objcopy(1). ndis_open_file() will traverse each module in the module list looking for these symbols and, if it finds them, it'll use them to generate the file mapping address and length values that the caller of NdisOpenFile() wants. As a bonus, this will even work if the file has been statically linked into the kernel itself, since the "kernel" module is searched too. (ndiscvt(8) will generate both filename.o and filename.ko for you). - Modify the mechanism used to provide make-pretend FASTCALL support. Rather than using inline assembly to yank the first two arguments out of %ecx and %edx, we now use the __regparm__(3) attribute (and the __stdcall__ attribute) and use some macro magic to re-order the arguments and provide dummy arguments as needed so that the arguments passed in registers end up in the right place. Change taken from DragonflyBSD version of the NDISulator.
2004-08-01 20:04:31 +00:00
#include <sys/linker.h>
#include <sys/mount.h>
#include <sys/sysproto.h>
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
#include <net/if.h>
#include <net/if_arp.h>
#include <net/ethernet.h>
#include <net/if_dl.h>
#include <net/if_media.h>
Use atomic ops for the interlocked increment and decrement routines in subr_ndis and subr_ntoskrnl. This is faster and avoids potential LOR whinage from witness (an LOR couldn't happen with the old code since the interlocked inc/dec routines could not sleep with a lock held, but this will keep witness happy and it's more efficient anyway. I think.)
2004-01-07 07:29:27 +00:00
#include <machine/atomic.h>
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
#include <machine/bus.h>
#include <machine/resource.h>
#include <sys/bus.h>
#include <sys/rman.h>
#include <machine/stdarg.h>
Big round of updates: - Make ndis_get_info()/ndis_set_info() sleep on the setdone/getdone routines if they get back NDIS_STATUS_PENDING. - Add a bunch of net80211 support so that 802.11 cards can be twiddled with ifconfig. This still needs more work and is not guaranteed to work for everyone. It works on my 802.11b/g card anyway. The problem here is Microsoft doesn't provide a good way to a) learn all the rates that a card supports (if it has more than 8, you're kinda hosed) and b) doesn't provide a good way to distinguish between 802.11b, 802.11b/g an 802.11a/b/g cards, so you sort of have to guess. Setting the SSID and switching between infrastructure/adhoc modes should work. WEP still needs to be implemented. I can't find any API for getting/setting the channel other than the registry/sysctl keys.
2003-12-21 00:00:08 +00:00
#include <net80211/ieee80211_var.h>
#include <net80211/ieee80211_ioctl.h>
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
#include <dev/pci/pcireg.h>
#include <dev/pci/pcivar.h>
#include <compat/ndis/pe_var.h>
Throw the switch on the new driver generation/loading mechanism. From here on in, if_ndis.ko will be pre-built as a module, and can be built into a static kernel (though it's not part of GENERIC). Drivers are created using the new ndisgen(8) script, which uses ndiscvt(8) under the covers, along with a few other tools. The result is a driver module that can be kldloaded into the kernel. A driver with foo.inf and foo.sys files will be converted into foo_sys.ko (and foo_sys.o, for those who want/need to make static kernels). This module contains all of the necessary info from the .INF file and the driver binary image, converted into an ELF module. You can kldload this module (or add it to /boot/loader.conf) to have it loaded automatically. Any required firmware files can be bundled into the module as well (or converted/loaded separately). Also, add a workaround for a problem in NdisMSleep(). During system bootstrap (cold == 1), msleep() always returns 0 without actually sleeping. The Intel 2200BG driver uses NdisMSleep() to wait for the NIC's firmware to come to life, and fails to load if NdisMSleep() doesn't actually delay. As a workaround, if msleep() (and hence ndis_thsuspend()) returns 0, use a hard DELAY() to sleep instead). This is not really the right thing to do, but we can't really do much else. At the very least, this makes the Intel driver happy. There are probably other drivers that fail in this way during bootstrap. Unfortunately, the only workaround for those is to avoid pre-loading them and kldload them once the system is running instead.
2005-04-24 20:21:22 +00:00
#include <compat/ndis/cfg_var.h>
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
#include <compat/ndis/resource_var.h>
Correct the implementation of NDIS_BUFFER_TO_SPAN_PAGES().
2003-12-13 09:07:35 +00:00
#include <compat/ndis/ntoskrnl_var.h>
Continue my efforts to imitate Windows as closely as possible by attempting to duplicate Windows spinlocks. Windows spinlocks differ from FreeBSD spinlocks in the way they block preemption. FreeBSD spinlocks use critical_enter(), which masks off _all_ interrupts. This prevents any other threads from being scheduled, but it also prevents ISRs from running. In Windows, preemption is achieved by raising the processor IRQL to DISPATCH_LEVEL, which prevents other threads from preempting you, but does _not_ prevent device ISRs from running. (This is essentially what Solaris calls dispatcher locks.) The Windows spinlock itself (kspin_lock) is just an integer value which is atomically set when you acquire the lock and atomically cleared when you release it. FreeBSD doesn't have IRQ levels, so we have to cheat a little by using thread priorities: normal thread priority is PASSIVE_LEVEL, lowest interrupt thread priority is DISPATCH_LEVEL, highest thread priority is DEVICE_LEVEL (PI_REALTIME) and critical_enter() is HIGH_LEVEL. In practice, only PASSIVE_LEVEL and DISPATCH_LEVEL matter to us. The immediate benefit of all this is that I no longer have to rely on a mutex pool. Now, I'm sure many people will be seized by the urge to criticize me for doing an end run around our own spinlock implementation, but it makes more sense to do it this way. Well, it does to me anyway. Overview of the changes: - Properly implement hal_lock(), hal_unlock(), hal_irql(), hal_raise_irql() and hal_lower_irql() so that they more closely resemble their Windows counterparts. The IRQL is determined by thread priority. - Make ntoskrnl_lock_dpc() and ntoskrnl_unlock_dpc() do what they do in Windows, which is to atomically set/clear the lock value. These routines are designed to be called from DISPATCH_LEVEL, and are actually half of the work involved in acquiring/releasing spinlocks. - Add FASTCALL1(), FASTCALL2() and FASTCALL3() macros/wrappers that allow us to call a _fastcall function in spite of the fact that our version of gcc doesn't support __attribute__((__fastcall__)) yet. The macros take 1, 2 or 3 arguments, respectively. We need to call hal_lock(), hal_unlock() etc... ourselves, but can't really invoke the function directly. I could have just made the underlying functions native routines and put _fastcall wrappers around them for the benefit of Windows binaries, but that would create needless bloat. - Remove ndis_mtxpool and all references to it. We don't need it anymore. - Re-implement the NdisSpinLock routines so that they use hal_lock() and friends like they do in Windows. - Use the new spinlock methods for handling lookaside lists and linked list updates in place of the mutex locks that were there before. - Remove mutex locking from ndis_isr() and ndis_intrhand() since they're already called with ndis_intrmtx held in if_ndis.c. - Put ndis_destroy_lock() code under explicit #ifdef notdef/#endif. It turns out there are some drivers which stupidly free the memory in which their spinlocks reside before calling ndis_destroy_lock() on them (touch-after-free bug). The ADMtek wireless driver is guilty of this faux pas. (Why this doesn't clobber Windows I have no idea.) - Make NdisDprAcquireSpinLock() and NdisDprReleaseSpinLock() into real functions instead of aliasing them to NdisAcaquireSpinLock() and NdisReleaseSpinLock(). The Dpr routines use KeAcquireSpinLockAtDpcLevel() level and KeReleaseSpinLockFromDpcLevel(), which acquires the lock without twiddling the IRQL. - In ndis_linksts_done(), do _not_ call ndis_80211_getstate(). Some drivers may call the status/status done callbacks as the result of setting an OID: ndis_80211_getstate() gets OIDs, which means we might cause the driver to recursively access some of its internal structures unexpectedly. The ndis_ticktask() routine will call ndis_80211_getstate() for us eventually anyway. - Fix the channel setting code a little in ndis_80211_setstate(), and initialize the channel to IEEE80211_CHAN_ANYC. (The Microsoft spec says you're not supposed to twiddle the channel in BSS mode; I may need to enforce this later.) This fixes the problems I was having with the ADMtek adm8211 driver: we were setting the channel to a non-standard default, which would cause it to fail to associate in BSS mode. - Use hal_raise_irql() to raise our IRQL to DISPATCH_LEVEL when calling certain miniport routines, per the Microsoft documentation. I think that's everything. Hopefully, other than fixing the ADMtek driver, there should be no apparent change in behavior.
2004-04-14 07:48:03 +00:00
#include <compat/ndis/hal_var.h>
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
#include <compat/ndis/ndis_var.h>
#include <dev/if_ndis/if_ndisvar.h>
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
#include <vm/vm.h>
#include <vm/vm_param.h>
#include <vm/pmap.h>
#include <vm/uma.h>
#include <vm/vm_kern.h>
#include <vm/vm_map.h>
Implement NdisOpenFile()/NdisCloseFile()/NdisMapFile()/NdisUnmapFile(). By default, we search for files in /compat/ndis. This can be changed with a systcl. These routines are used by some drivers which need to download firmware or microcode into their respective devices during initialization. Also, remove extraneous newlines from the 'built-in' sysctl/registry variables.
2004-01-09 03:57:00 +00:00
static char ndis_filepath[MAXPATHLEN];
SYSCTL_STRING(_hw, OID_AUTO, ndis_filepath, CTLFLAG_RW, ndis_filepath,
MAXPATHLEN, "Path used by NdisOpenFile() to search for files");
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisInitializeWrapper(ndis_handle *,
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
driver_object *, void *, void *);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status NdisMRegisterMiniport(ndis_handle,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_miniport_characteristics *, int);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status NdisAllocateMemoryWithTag(void **,
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
uint32_t, uint32_t);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status NdisAllocateMemory(void **,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
uint32_t, uint32_t, ndis_physaddr);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisFreeMemory(void *, uint32_t, uint32_t);
static ndis_status NdisMSetAttributesEx(ndis_handle, ndis_handle,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
uint32_t, uint32_t, ndis_interface_type);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisOpenConfiguration(ndis_status *,
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
ndis_handle *, ndis_handle);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisOpenConfigurationKeyByIndex(ndis_status *,
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
ndis_handle, uint32_t, unicode_string *, ndis_handle *);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisOpenConfigurationKeyByName(ndis_status *,
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
ndis_handle, unicode_string *, ndis_handle *);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
static ndis_status ndis_encode_parm(ndis_miniport_block *,
struct sysctl_oid *, ndis_parm_type, ndis_config_parm **);
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
static ndis_status ndis_decode_parm(ndis_miniport_block *,
ndis_config_parm *, char *);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisReadConfiguration(ndis_status *, ndis_config_parm **,
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
ndis_handle, unicode_string *, ndis_parm_type);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisWriteConfiguration(ndis_status *, ndis_handle,
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
unicode_string *, ndis_config_parm *);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisCloseConfiguration(ndis_handle);
static void NdisAllocateSpinLock(ndis_spin_lock *);
static void NdisFreeSpinLock(ndis_spin_lock *);
static void NdisAcquireSpinLock(ndis_spin_lock *);
static void NdisReleaseSpinLock(ndis_spin_lock *);
static void NdisDprAcquireSpinLock(ndis_spin_lock *);
static void NdisDprReleaseSpinLock(ndis_spin_lock *);
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
static void NdisInitializeReadWriteLock(ndis_rw_lock *);
static void NdisAcquireReadWriteLock(ndis_rw_lock *,
uint8_t, ndis_lock_state *);
static void NdisReleaseReadWriteLock(ndis_rw_lock *, ndis_lock_state *);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static uint32_t NdisReadPciSlotInformation(ndis_handle, uint32_t,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
uint32_t, void *, uint32_t);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static uint32_t NdisWritePciSlotInformation(ndis_handle, uint32_t,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
uint32_t, void *, uint32_t);
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
static void NdisWriteErrorLogEntry(ndis_handle, ndis_error_code, uint32_t, ...);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
static void ndis_map_cb(void *, bus_dma_segment_t *, int, int);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisMStartBufferPhysicalMapping(ndis_handle,
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
ndis_buffer *, uint32_t, uint8_t, ndis_paddr_unit *, uint32_t *);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisMCompleteBufferPhysicalMapping(ndis_handle,
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
ndis_buffer *, uint32_t);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisMInitializeTimer(ndis_miniport_timer *, ndis_handle,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_timer_function, void *);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisInitializeTimer(ndis_timer *,
Reorganize the timer code a little and implement NdisInitializeTimer() and NdisCancelTimer(). NdisInitializeTimer() doesn't accept an NDIS miniport context argument, so we have to derive it from the timer function context (which is supposed to be the adapter private context). NdisCancelTimer is now an alias for NdisMCancelTimer(). Also add stubs for NdisMRegisterDevice() and NdisMDeregisterDevice(). These are no-ops for now, but will likely get fleshed in once I start working on the Am1771/Am1772 wireless driver.
2004-01-26 21:21:53 +00:00
ndis_timer_function, void *);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisSetTimer(ndis_timer *, uint32_t);
static void NdisMSetPeriodicTimer(ndis_miniport_timer *, uint32_t);
static void NdisMCancelTimer(ndis_timer *, uint8_t *);
static void ndis_timercall(kdpc *, ndis_miniport_timer *,
Finally bring an end to the great "make the Atheros NDIS driver work on SMP" saga. After several weeks and much gnashing of teeth, I have finally tracked down all the problems, despite their best efforts to confound and annoy me. Problem nunmber one: the Atheros windows driver is _NOT_ a de-serialized miniport! It used to be that NDIS drivers relied on the NDIS library itself for all their locking and serialization needs. Transmit packet queues were all handled internally by NDIS, and all calls to MiniportXXX() routines were guaranteed to be appropriately serialized. This proved to be a performance problem however, and Microsoft introduced de-serialized miniports with the NDIS 5.x spec. Microsoft still supports serialized miniports, but recommends that all new drivers written for Windows XP and later be deserialized. Apparently Atheros wasn't listening when they said this. This means (among other things) that we have to serialize calls to MiniportSendPackets(). We also have to serialize calls to MiniportTimer() that are triggered via the NdisMInitializeTimer() routine. It finally dawned on me why NdisMInitializeTimer() takes a special NDIS_MINIPORT_TIMER structure and a pointer to the miniport block: the timer callback must be serialized, and it's only by saving the miniport block handle that we can get access to the serialization lock during the timer callback. Problem number two: haunted hardware. The thing that was _really_ driving me absolutely bonkers for the longest time is that, for some reason I couldn't understand, my test machine would occasionally freeze or more frustratingly, reset completely. That's reset and in *pow!* back to the BIOS startup. No panic, no crashdump, just a reset. This appeared to happen most often when MiniportReset() was called. (As to why MiniportReset() was being called, see problem three below.) I thought maybe I had created some sort of horrible deadlock condition in the process of adding the serialization, but after three weeks, at least 6 different locking implementations and heroic efforts to debug the spinlock code, the machine still kept resetting. Finally, I started single stepping through the MiniportReset() routine in the driver using the kernel debugger, and this ultimately led me to the source of the problem. One of the last things the Atheros MiniportReset() routine does is call NdisReadPciSlotInformation() several times to inspect a portion of the device's PCI config space. It reads the same chunk of config space repeatedly, in rapid succession. Presumeably, it's polling the hardware for some sort of event. The reset occurs partway through this process. I discovered that when I single-stepped through this portion of the routine, the reset didn't occur. So I inserted a 1 microsecond delay into the read loop in NdisReadPciSlotInformation(). Suddenly, the reset was gone!! I'm still very puzzled by the whole thing. What I suspect is happening is that reading the PCI config space so quickly is causing a severe PCI bus error. My test system is a Sun w2100z dual Opteron system, and the NIC is a miniPCI card mounted in a miniPCI-to-PCI carrier card, plugged into a 100Mhz PCI slot. It's possible that this combination of hardware causes a bus protocol violation in this scenario which leads to a fatal machine check. This is pure speculation though. Really all I know for sure is that inserting the delay makes the problem go away. (To quote Homer Simpson: "I don't know how it works, but fire makes it good!") Problem number three: NdisAllocatePacket() needs to make sure to initialize the npp_validcounts field in the 'private' section of the NDIS_PACKET structure. The reason if_ndis was calling the MiniportReset() routine in the first place is that packet transmits were sometimes hanging. When sending a packet, an NDIS driver will call NdisQueryPacket() to learn how many physical buffers the packet resides in. NdisQueryPacket() is actually a macro, which traverses the NDIS_BUFFER list attached to the NDIS_PACKET and stashes some of the results in the 'private' section of the NDIS_PACKET. It also sets the npp_validcounts field to TRUE To indicate that the results are now valid. The problem is, now that if_ndis creates a pool of transmit packets via NdisAllocatePacketPool(), it's important that each time a new packet is allocated via NdisAllocatePacket() that validcounts be initialized to FALSE. If it isn't, and a previously transmitted NDIS_PACKET is pulled out of the pool, it may contain stale data from a previous transmission which won't get updated by NdisQueryPacket(). This would cause the driver to miscompute the number of fragments for a given packet, and botch the transmission. Fixing these three problems seems to make the Atheros driver happy on SMP, which hopefully means other serialized miniports will be happy too. And there was much rejoicing. Other stuff fixed along the way: - Modified ndis_thsuspend() to take a mutex as an argument. This allows KeWaitForSingleObject() and KeWaitForMultipleObjects() to avoid any possible race conditions with other routines that use the dispatcher lock. - Fixed KeCancelTimer() so that it returns the correct value for 'pending' according to the Microsoft documentation - Modfied NdisGetSystemUpTime() to use ticks and hz rather than calling nanouptime(). Also added comment that this routine wraps after 49.7 days. - Added macros for KeAcquireSpinLock()/KeReleaseSpinLock() to hide all the MSCALL() goop. - For x86, KeAcquireSpinLockRaiseToDpc() needs to be a separate function. This is because it's supposed to be _stdcall on the x86 arch, whereas KeAcquireSpinLock() is supposed to be _fastcall. On amd64, all routines use the same calling convention so we can just map KeAcquireSpinLockRaiseToDpc() directly to KfAcquireSpinLock() and it will work. (The _fastcall attribute is a no-op on amd64.) - Implement and use IoInitializeDpcRequest() and IoRequestDpc() (they're just macros) and use them for interrupt handling. This allows us to move the ndis_intrtask() routine from if_ndis.c to kern_ndis.c. - Fix the MmInitializeMdl() macro so that is uses sizeof(vm_offset_t) when computing mdl_size instead of uint32_t, so that it matches the MmSizeOfMdl() routine. - Change a could of M_WAITOKs to M_NOWAITs in the unicode routines in subr_ndis.c. - Use the dispatcher lock a little more consistently in subr_ntoskrnl.c. - Get rid of the "wait for link event" hack in ndis_init(). Now that I fixed NdisReadPciSlotInformation(), it seems I don't need it anymore. This should fix the witness panic a couple of people have reported. - Use MSCALL1() when calling the MiniportHangCheck() function in ndis_ticktask(). I accidentally missed this one when adding the wrapping for amd64.
2005-03-27 10:14:36 +00:00
void *, void *);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisMQueryAdapterResources(ndis_status *, ndis_handle,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_resource_list *, uint32_t *);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status NdisMRegisterIoPortRange(void **,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_handle, uint32_t, uint32_t);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisMDeregisterIoPortRange(ndis_handle,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
uint32_t, uint32_t, void *);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisReadNetworkAddress(ndis_status *, void **,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
uint32_t *, ndis_handle);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status NdisQueryMapRegisterCount(uint32_t, uint32_t *);
static ndis_status NdisMAllocateMapRegisters(ndis_handle,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
uint32_t, uint8_t, uint32_t, uint32_t);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisMFreeMapRegisters(ndis_handle);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
static void ndis_mapshared_cb(void *, bus_dma_segment_t *, int, int);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisMAllocateSharedMemory(ndis_handle, uint32_t,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
uint8_t, void **, ndis_physaddr *);
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
static void ndis_asyncmem_complete(device_object *, void *);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status NdisMAllocateSharedMemoryAsync(ndis_handle,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
uint32_t, uint8_t, void *);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisMFreeSharedMemory(ndis_handle, uint32_t,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
uint8_t, void *, ndis_physaddr);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status NdisMMapIoSpace(void **, ndis_handle,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_physaddr, uint32_t);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisMUnmapIoSpace(ndis_handle, void *, uint32_t);
static uint32_t NdisGetCacheFillSize(void);
static uint32_t NdisMGetDmaAlignment(ndis_handle);
static ndis_status NdisMInitializeScatterGatherDma(ndis_handle,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
uint8_t, uint32_t);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisUnchainBufferAtFront(ndis_packet *, ndis_buffer **);
static void NdisUnchainBufferAtBack(ndis_packet *, ndis_buffer **);
static void NdisAllocateBufferPool(ndis_status *,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_handle *, uint32_t);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisFreeBufferPool(ndis_handle);
static void NdisAllocateBuffer(ndis_status *, ndis_buffer **,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_handle, void *, uint32_t);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisFreeBuffer(ndis_buffer *);
static uint32_t NdisBufferLength(ndis_buffer *);
static void NdisQueryBuffer(ndis_buffer *, void **, uint32_t *);
static void NdisQueryBufferSafe(ndis_buffer *, void **,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
uint32_t *, uint32_t);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void *NdisBufferVirtualAddress(ndis_buffer *);
static void *NdisBufferVirtualAddressSafe(ndis_buffer *, uint32_t);
static void NdisAdjustBufferLength(ndis_buffer *, int);
static uint32_t NdisInterlockedIncrement(uint32_t *);
static uint32_t NdisInterlockedDecrement(uint32_t *);
static void NdisInitializeEvent(ndis_event *);
static void NdisSetEvent(ndis_event *);
static void NdisResetEvent(ndis_event *);
static uint8_t NdisWaitEvent(ndis_event *, uint32_t);
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
static ndis_status NdisUnicodeStringToAnsiString(ansi_string *,
unicode_string *);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
NdisAnsiStringToUnicodeString(unicode_string *, ansi_string *);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status NdisMPciAssignResources(ndis_handle,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
uint32_t, ndis_resource_list **);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status NdisMRegisterInterrupt(ndis_miniport_interrupt *,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_handle, uint32_t, uint32_t, uint8_t,
uint8_t, ndis_interrupt_mode);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisMDeregisterInterrupt(ndis_miniport_interrupt *);
static void NdisMRegisterAdapterShutdownHandler(ndis_handle, void *,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_shutdown_handler);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisMDeregisterAdapterShutdownHandler(ndis_handle);
static uint32_t NDIS_BUFFER_TO_SPAN_PAGES(ndis_buffer *);
static void NdisGetBufferPhysicalArraySize(ndis_buffer *,
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
uint32_t *);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisQueryBufferOffset(ndis_buffer *,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
uint32_t *, uint32_t *);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static uint32_t NdisReadPcmciaAttributeMemory(ndis_handle,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
uint32_t, void *, uint32_t);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static uint32_t NdisWritePcmciaAttributeMemory(ndis_handle,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
uint32_t, void *, uint32_t);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static list_entry *NdisInterlockedInsertHeadList(list_entry *,
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
list_entry *, ndis_spin_lock *);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static list_entry *NdisInterlockedRemoveHeadList(list_entry *,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_spin_lock *);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static list_entry *NdisInterlockedInsertTailList(list_entry *,
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
list_entry *, ndis_spin_lock *);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static uint8_t
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMSynchronizeWithInterrupt(ndis_miniport_interrupt *,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
void *, void *);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisGetCurrentSystemTime(uint64_t *);
static void NdisGetSystemUpTime(uint32_t *);
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
static void NdisInitializeString(unicode_string *, char *);
static void NdisInitAnsiString(ansi_string *, char *);
static void NdisInitUnicodeString(unicode_string *, uint16_t *);
static void NdisFreeString(unicode_string *);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status NdisMRemoveMiniport(ndis_handle *);
static void NdisTerminateWrapper(ndis_handle, void *);
static void NdisMGetDeviceProperty(ndis_handle, device_object **,
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
device_object **, device_object **, cm_resource_list *,
cm_resource_list *);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisGetFirstBufferFromPacket(ndis_packet *,
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
ndis_buffer **, void **, uint32_t *, uint32_t *);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisGetFirstBufferFromPacketSafe(ndis_packet *,
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
ndis_buffer **, void **, uint32_t *, uint32_t *, uint32_t);
Big mess 'o changes: - Give ndiscvt(8) the ability to process a .SYS file directly into a .o file so that we don't have to emit big messy char arrays into the ndis_driver_data.h file. This behavior is currently optional, but may become the default some day. - Give ndiscvt(8) the ability to turn arbitrary files into .ko files so that they can be pre-loaded or kldloaded. (Both this and the previous change involve using objcopy(1)). - Give NdisOpenFile() the ability to 'read' files out of kernel memory that have been kldloaded or pre-loaded, and disallow the use of the normal vn_open() file opening method during bootstrap (when no filesystems have been mounted yet). Some people have reported that kldloading if_ndis.ko works fine when the system is running multiuser but causes a panic when the modile is pre-loaded by /boot/loader. This happens with drivers that need to use NdisOpenFile() to access external files (i.e. firmware images). NdisOpenFile() won't work during kernel bootstrapping because no filesystems have been mounted. To get around this, you can now do the following: o Say you have a firmware file called firmware.img o Do: ndiscvt -f firmware.img -- this creates firmware.img.ko o Put the firmware.img.ko in /boot/kernel o add firmware.img_load="YES" in /boot/loader.conf o add if_ndis_load="YES" and ndis_load="YES" as well Now the loader will suck the additional file into memory as a .ko. The phony .ko has two symbols in it: filename_start and filename_end, which are generated by objcopy(1). ndis_open_file() will traverse each module in the module list looking for these symbols and, if it finds them, it'll use them to generate the file mapping address and length values that the caller of NdisOpenFile() wants. As a bonus, this will even work if the file has been statically linked into the kernel itself, since the "kernel" module is searched too. (ndiscvt(8) will generate both filename.o and filename.ko for you). - Modify the mechanism used to provide make-pretend FASTCALL support. Rather than using inline assembly to yank the first two arguments out of %ecx and %edx, we now use the __regparm__(3) attribute (and the __stdcall__ attribute) and use some macro magic to re-order the arguments and provide dummy arguments as needed so that the arguments passed in registers end up in the right place. Change taken from DragonflyBSD version of the NDISulator.
2004-08-01 20:04:31 +00:00
static int ndis_find_sym(linker_file_t, char *, char *, caddr_t *);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisOpenFile(ndis_status *, ndis_handle *, uint32_t *,
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
unicode_string *, ndis_physaddr);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisMapFile(ndis_status *, void **, ndis_handle);
static void NdisUnmapFile(ndis_handle);
static void NdisCloseFile(ndis_handle);
static uint8_t NdisSystemProcessorCount(void);
static void NdisMIndicateStatusComplete(ndis_handle);
static void NdisMIndicateStatus(ndis_handle, ndis_status,
In ndis_attach(), report the NDIS API level that the Windows miniport driver was compiled with. Remove debug printf from ndis_assicn_pcirsc(). It doesn't serve much purpose. Implement NdisMIndicateStatus() and NdisMIndicateStatusComplete() as functions in subr_ndis.c. In NDIS 4.0, they were functions. In NDIS 5.0 and later, they're just macros. Allocate a few extra packets/buffers beyond what the driver asks for since sometimes it seems they can lie about how many they really need, and some extra stupid ones don't check to see if NdisAllocatePacket() and/or NdisAllocateBuffer() actually succeed.
2004-01-04 03:00:21 +00:00
void *, uint32_t);
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
static void ndis_intr(void *);
static void ndis_intrhand(kdpc *, ndis_miniport_interrupt *, void *, void *);
Fix an amd64 issue I overlooked. When setting up a callout to ndis_timercall() in NdisMInitializeTimer(), we can't use the raw function pointer. This is because ntoskrnl_run_dpc() expects to invoke a function with Microsoft calling conventions. On i386, this works because ndis_timercall() is declared with the __stdcall attribute, but this is a no-op on amd64. To do it correctly, we have to generate a wrapper for ndis_timercall() and us the wrapper instead of of the raw function pointer. Fix this by adding ndis_timercall() to the funcptr table in subr_ndis.c, and create ndis_findwrap() to extract the wrapped function from the table in NdisMInitializeTimer() instead of just passing ndis_timercall() to KeInitializeDpc() directly.
2005-03-31 16:38:48 +00:00
static funcptr ndis_findwrap(funcptr);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisCopyFromPacketToPacket(ndis_packet *,
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
uint32_t, uint32_t, ndis_packet *, uint32_t, uint32_t *);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisCopyFromPacketToPacketSafe(ndis_packet *,
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
uint32_t, uint32_t, ndis_packet *, uint32_t, uint32_t *, uint32_t);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status NdisMRegisterDevice(ndis_handle,
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
unicode_string *, unicode_string *, driver_dispatch **,
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
void **, ndis_handle *);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status NdisMDeregisterDevice(ndis_handle);
static ndis_status
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
NdisMQueryAdapterInstanceName(unicode_string *, ndis_handle);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisMRegisterUnloadHandler(ndis_handle, void *);
static void dummy(void);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
In ndis_attach(), report the NDIS API level that the Windows miniport driver was compiled with. Remove debug printf from ndis_assicn_pcirsc(). It doesn't serve much purpose. Implement NdisMIndicateStatus() and NdisMIndicateStatusComplete() as functions in subr_ndis.c. In NDIS 4.0, they were functions. In NDIS 5.0 and later, they're just macros. Allocate a few extra packets/buffers beyond what the driver asks for since sometimes it seems they can lie about how many they really need, and some extra stupid ones don't check to see if NdisAllocatePacket() and/or NdisAllocateBuffer() actually succeed.
2004-01-04 03:00:21 +00:00
/*
* Some really old drivers do not properly check the return value
* from NdisAllocatePacket() and NdisAllocateBuffer() and will
* sometimes allocate few more buffers/packets that they originally
* requested when they created the pool. To prevent this from being
* a problem, we allocate a few extra buffers/packets beyond what
* the driver asks for. This #define controls how many.
*/
#define NDIS_POOL_EXTRA 16
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
int
ndis_libinit()
{
Add support for Windows/x86-64 binaries to Project Evil. Ville-Pertti Keinonen (will at exomi dot comohmygodnospampleasekthx) deserves a big thanks for submitting initial patches to make it work. I have mangled his contributions appropriately. The main gotcha with Windows/x86-64 is that Microsoft uses a different calling convention than everyone else. The standard ABI requires using 6 registers for argument passing, with other arguments on the stack. Microsoft uses only 4 registers, and requires the caller to leave room on the stack for the register arguments incase the callee needs to spill them. Unlike x86, where Microsoft uses a mix of _cdecl, _stdcall and _fastcall, all routines on Windows/x86-64 uses the same convention. This unfortunately means that all the functions we export to the driver require an intermediate translation wrapper. Similarly, we have to wrap all calls back into the driver binary itself. The original patches provided macros to wrap every single routine at compile time, providing a secondary jump table with a customized wrapper for each exported routine. I decided to use a different approach: the call wrapper for each function is created from a template at runtime, and the routine to jump to is patched into the wrapper as it is created. The subr_pe module has been modified to patch in the wrapped function instead of the original. (On x86, the wrapping routine is a no-op.) There are some minor API differences that had to be accounted for: - KeAcquireSpinLock() is a real function on amd64, not a macro wrapper around KfAcquireSpinLock() - NdisFreeBuffer() is actually IoFreeMdl(). I had to change the whole NDIS_BUFFER API a bit to accomodate this. Bugs fixed along the way: - IoAllocateMdl() always returned NULL - kern_windrv.c:windrv_unload() wasn't releasing private driver object extensions correctly (found thanks to memguard) This has only been tested with the driver for the Broadcom 802.11g chipset, which was the only Windows/x86-64 driver I could find.
2005-02-16 05:41:18 +00:00
image_patch_table *patch;
Implement NdisOpenFile()/NdisCloseFile()/NdisMapFile()/NdisUnmapFile(). By default, we search for files in /compat/ndis. This can be changed with a systcl. These routines are used by some drivers which need to download firmware or microcode into their respective devices during initialization. Also, remove extraneous newlines from the 'built-in' sysctl/registry variables.
2004-01-09 03:57:00 +00:00
strcpy(ndis_filepath, "/compat/ndis");
Add support for Windows/x86-64 binaries to Project Evil. Ville-Pertti Keinonen (will at exomi dot comohmygodnospampleasekthx) deserves a big thanks for submitting initial patches to make it work. I have mangled his contributions appropriately. The main gotcha with Windows/x86-64 is that Microsoft uses a different calling convention than everyone else. The standard ABI requires using 6 registers for argument passing, with other arguments on the stack. Microsoft uses only 4 registers, and requires the caller to leave room on the stack for the register arguments incase the callee needs to spill them. Unlike x86, where Microsoft uses a mix of _cdecl, _stdcall and _fastcall, all routines on Windows/x86-64 uses the same convention. This unfortunately means that all the functions we export to the driver require an intermediate translation wrapper. Similarly, we have to wrap all calls back into the driver binary itself. The original patches provided macros to wrap every single routine at compile time, providing a secondary jump table with a customized wrapper for each exported routine. I decided to use a different approach: the call wrapper for each function is created from a template at runtime, and the routine to jump to is patched into the wrapper as it is created. The subr_pe module has been modified to patch in the wrapped function instead of the original. (On x86, the wrapping routine is a no-op.) There are some minor API differences that had to be accounted for: - KeAcquireSpinLock() is a real function on amd64, not a macro wrapper around KfAcquireSpinLock() - NdisFreeBuffer() is actually IoFreeMdl(). I had to change the whole NDIS_BUFFER API a bit to accomodate this. Bugs fixed along the way: - IoAllocateMdl() always returned NULL - kern_windrv.c:windrv_unload() wasn't releasing private driver object extensions correctly (found thanks to memguard) This has only been tested with the driver for the Broadcom 802.11g chipset, which was the only Windows/x86-64 driver I could find.
2005-02-16 05:41:18 +00:00
patch = ndis_functbl;
while (patch->ipt_func != NULL) {
windrv_wrap((funcptr)patch->ipt_func,
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
(funcptr *)&patch->ipt_wrap,
patch->ipt_argcnt, patch->ipt_ftype);
Add support for Windows/x86-64 binaries to Project Evil. Ville-Pertti Keinonen (will at exomi dot comohmygodnospampleasekthx) deserves a big thanks for submitting initial patches to make it work. I have mangled his contributions appropriately. The main gotcha with Windows/x86-64 is that Microsoft uses a different calling convention than everyone else. The standard ABI requires using 6 registers for argument passing, with other arguments on the stack. Microsoft uses only 4 registers, and requires the caller to leave room on the stack for the register arguments incase the callee needs to spill them. Unlike x86, where Microsoft uses a mix of _cdecl, _stdcall and _fastcall, all routines on Windows/x86-64 uses the same convention. This unfortunately means that all the functions we export to the driver require an intermediate translation wrapper. Similarly, we have to wrap all calls back into the driver binary itself. The original patches provided macros to wrap every single routine at compile time, providing a secondary jump table with a customized wrapper for each exported routine. I decided to use a different approach: the call wrapper for each function is created from a template at runtime, and the routine to jump to is patched into the wrapper as it is created. The subr_pe module has been modified to patch in the wrapped function instead of the original. (On x86, the wrapping routine is a no-op.) There are some minor API differences that had to be accounted for: - KeAcquireSpinLock() is a real function on amd64, not a macro wrapper around KfAcquireSpinLock() - NdisFreeBuffer() is actually IoFreeMdl(). I had to change the whole NDIS_BUFFER API a bit to accomodate this. Bugs fixed along the way: - IoAllocateMdl() always returned NULL - kern_windrv.c:windrv_unload() wasn't releasing private driver object extensions correctly (found thanks to memguard) This has only been tested with the driver for the Broadcom 802.11g chipset, which was the only Windows/x86-64 driver I could find.
2005-02-16 05:41:18 +00:00
patch++;
}
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return(0);
}
int
ndis_libfini()
{
Add support for Windows/x86-64 binaries to Project Evil. Ville-Pertti Keinonen (will at exomi dot comohmygodnospampleasekthx) deserves a big thanks for submitting initial patches to make it work. I have mangled his contributions appropriately. The main gotcha with Windows/x86-64 is that Microsoft uses a different calling convention than everyone else. The standard ABI requires using 6 registers for argument passing, with other arguments on the stack. Microsoft uses only 4 registers, and requires the caller to leave room on the stack for the register arguments incase the callee needs to spill them. Unlike x86, where Microsoft uses a mix of _cdecl, _stdcall and _fastcall, all routines on Windows/x86-64 uses the same convention. This unfortunately means that all the functions we export to the driver require an intermediate translation wrapper. Similarly, we have to wrap all calls back into the driver binary itself. The original patches provided macros to wrap every single routine at compile time, providing a secondary jump table with a customized wrapper for each exported routine. I decided to use a different approach: the call wrapper for each function is created from a template at runtime, and the routine to jump to is patched into the wrapper as it is created. The subr_pe module has been modified to patch in the wrapped function instead of the original. (On x86, the wrapping routine is a no-op.) There are some minor API differences that had to be accounted for: - KeAcquireSpinLock() is a real function on amd64, not a macro wrapper around KfAcquireSpinLock() - NdisFreeBuffer() is actually IoFreeMdl(). I had to change the whole NDIS_BUFFER API a bit to accomodate this. Bugs fixed along the way: - IoAllocateMdl() always returned NULL - kern_windrv.c:windrv_unload() wasn't releasing private driver object extensions correctly (found thanks to memguard) This has only been tested with the driver for the Broadcom 802.11g chipset, which was the only Windows/x86-64 driver I could find.
2005-02-16 05:41:18 +00:00
image_patch_table *patch;
patch = ndis_functbl;
while (patch->ipt_func != NULL) {
windrv_unwrap(patch->ipt_wrap);
patch++;
}
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return(0);
}
Fix an amd64 issue I overlooked. When setting up a callout to ndis_timercall() in NdisMInitializeTimer(), we can't use the raw function pointer. This is because ntoskrnl_run_dpc() expects to invoke a function with Microsoft calling conventions. On i386, this works because ndis_timercall() is declared with the __stdcall attribute, but this is a no-op on amd64. To do it correctly, we have to generate a wrapper for ndis_timercall() and us the wrapper instead of of the raw function pointer. Fix this by adding ndis_timercall() to the funcptr table in subr_ndis.c, and create ndis_findwrap() to extract the wrapped function from the table in NdisMInitializeTimer() instead of just passing ndis_timercall() to KeInitializeDpc() directly.
2005-03-31 16:38:48 +00:00
static funcptr
ndis_findwrap(func)
funcptr func;
{
image_patch_table *patch;
Apparently I'm cursed. ndis_findwrap() should be searching ndis_functbl, not ntoskrnl_functbl.
2005-03-31 21:20:19 +00:00
patch = ndis_functbl;
Fix an amd64 issue I overlooked. When setting up a callout to ndis_timercall() in NdisMInitializeTimer(), we can't use the raw function pointer. This is because ntoskrnl_run_dpc() expects to invoke a function with Microsoft calling conventions. On i386, this works because ndis_timercall() is declared with the __stdcall attribute, but this is a no-op on amd64. To do it correctly, we have to generate a wrapper for ndis_timercall() and us the wrapper instead of of the raw function pointer. Fix this by adding ndis_timercall() to the funcptr table in subr_ndis.c, and create ndis_findwrap() to extract the wrapped function from the table in NdisMInitializeTimer() instead of just passing ndis_timercall() to KeInitializeDpc() directly.
2005-03-31 16:38:48 +00:00
while (patch->ipt_func != NULL) {
if ((funcptr)patch->ipt_func == func)
return((funcptr)patch->ipt_wrap);
patch++;
}
return(NULL);
}
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
/*
* This routine does the messy Windows Driver Model device attachment
* stuff on behalf of NDIS drivers. We register our own AddDevice
* routine here
*/
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
NdisInitializeWrapper(wrapper, drv, path, unused)
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
ndis_handle *wrapper;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
driver_object *drv;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
void *path;
void *unused;
{
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
/*
* As of yet, I haven't come up with a compelling
* reason to define a private NDIS wrapper structure,
* so we use a pointer to the driver object as the
* wrapper handle. The driver object has the miniport
* characteristics struct for this driver hung off it
* via IoAllocateDriverObjectExtension(), and that's
* really all the private data we need.
*/
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
*wrapper = drv;
/*
* If this was really Windows, we'd be registering dispatch
* routines for the NDIS miniport module here, but we're
* not Windows so all we really need to do is set up an
* AddDevice function that'll be invoked when a new device
* instance appears.
*/
drv->dro_driverext->dre_adddevicefunc = NdisAddDevice;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisTerminateWrapper(handle, syspec)
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
ndis_handle handle;
void *syspec;
{
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
/* Nothing to see here, move along. */
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMRegisterMiniport(handle, characteristics, len)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_handle handle;
ndis_miniport_characteristics *characteristics;
int len;
{
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
ndis_miniport_characteristics *ch = NULL;
driver_object *drv;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
drv = (driver_object *)handle;
/*
* We need to save the NDIS miniport characteristics
* somewhere. This data is per-driver, not per-device
* (all devices handled by the same driver have the
* same characteristics) so we hook it onto the driver
* object using IoAllocateDriverObjectExtension().
* The extra extension info is automagically deleted when
* the driver is unloaded (see windrv_unload()).
*/
if (IoAllocateDriverObjectExtension(drv, (void *)1,
sizeof(ndis_miniport_characteristics), (void **)&ch) !=
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
STATUS_SUCCESS) {
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
return(NDIS_STATUS_RESOURCES);
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
}
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
bzero((char *)ch, sizeof(ndis_miniport_characteristics));
bcopy((char *)characteristics, (char *)ch, len);
if (ch->nmc_version_major < 5 || ch->nmc_version_minor < 1) {
ch->nmc_shutdown_handler = NULL;
ch->nmc_canceltxpkts_handler = NULL;
ch->nmc_pnpevent_handler = NULL;
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
}
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return(NDIS_STATUS_SUCCESS);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisAllocateMemoryWithTag(vaddr, len, tag)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
void **vaddr;
uint32_t len;
uint32_t tag;
{
void *mem;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
mem = ExAllocatePoolWithTag(NonPagedPool, len, tag);
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
if (mem == NULL) {
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return(NDIS_STATUS_RESOURCES);
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
}
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
*vaddr = mem;
return(NDIS_STATUS_SUCCESS);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisAllocateMemory(vaddr, len, flags, highaddr)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
void **vaddr;
uint32_t len;
uint32_t flags;
ndis_physaddr highaddr;
{
void *mem;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
mem = ExAllocatePoolWithTag(NonPagedPool, len, 0);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
if (mem == NULL)
return(NDIS_STATUS_RESOURCES);
*vaddr = mem;
return(NDIS_STATUS_SUCCESS);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisFreeMemory(vaddr, len, flags)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
void *vaddr;
uint32_t len;
uint32_t flags;
{
if (len == 0)
return;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
ExFreePool(vaddr);
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMSetAttributesEx(adapter_handle, adapter_ctx, hangsecs,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
flags, iftype)
ndis_handle adapter_handle;
ndis_handle adapter_ctx;
uint32_t hangsecs;
uint32_t flags;
ndis_interface_type iftype;
{
ndis_miniport_block *block;
/*
* Save the adapter context, we need it for calling
* the driver's internal functions.
*/
block = (ndis_miniport_block *)adapter_handle;
block->nmb_miniportadapterctx = adapter_ctx;
block->nmb_checkforhangsecs = hangsecs;
Fix the problem with the Cisco Aironet 340 PCMCIA card. Most newer drivers for Windows are deserialized miniports. Such drivers maintain their own queues and do their own locking. This particular driver is not deserialized though, and we need special support to handle it correctly. Typically, in the ndis_rxeof() handler, we pass all incoming packets directly to (*ifp->if_input)(). This in turn may cause another thread to run and preempt us, and the packet may actually be processed and then released before we even exit the ndis_rxeof() routine. The problem with this is that releasing a packet calls the ndis_return_packet() function, which hands the packet and its buffers back to the driver. Calling ndis_return_packet() before ndis_rxeof() returns will screw up the driver's internal queues since, not being deserialized, it does no locking. To avoid this problem, if we detect a serialized driver (by checking the attribute flags passed to NdisSetAttributesEx(), we use an alternate ndis_rxeof() handler, ndis_rxeof_serial(), which puts the call to (*ifp->if_input)() on the NDIS SWI work queue. This guarantees the packet won't be processed until after ndis_rxeof_serial() returns. Note that another approach is to always copy the packet data into another mbuf and just let the driver retain ownership of the ndis_packet structure (ndis_return_packet() never needs to be called in this case). I'm not sure which method is faster.
2004-03-11 09:40:00 +00:00
block->nmb_flags = flags;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return(NDIS_STATUS_SUCCESS);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisOpenConfiguration(status, cfg, wrapctx)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_status *status;
ndis_handle *cfg;
ndis_handle wrapctx;
{
*cfg = wrapctx;
*status = NDIS_STATUS_SUCCESS;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisOpenConfigurationKeyByName(status, cfg, subkey, subhandle)
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
ndis_status *status;
ndis_handle cfg;
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
unicode_string *subkey;
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
ndis_handle *subhandle;
{
*subhandle = cfg;
*status = NDIS_STATUS_SUCCESS;
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisOpenConfigurationKeyByIndex(status, cfg, idx, subkey, subhandle)
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
ndis_status *status;
ndis_handle cfg;
uint32_t idx;
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
unicode_string *subkey;
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
ndis_handle *subhandle;
{
*status = NDIS_STATUS_FAILURE;
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
return;
}
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
static ndis_status
ndis_encode_parm(block, oid, type, parm)
ndis_miniport_block *block;
struct sysctl_oid *oid;
ndis_parm_type type;
ndis_config_parm **parm;
{
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
ndis_config_parm *p;
ndis_parmlist_entry *np;
unicode_string *us;
ansi_string as;
- Rewrite the timer and event API routines in subr_ndis.c so that they are actually layered on top of the KeTimer API in subr_ntoskrnl.c, just as it is in Windows. This reduces code duplication and more closely imitates the way things are done in Windows. - Modify ndis_encode_parm() to deal with the case where we have a registry key expressed as a hex value ("0x1") which is being read via NdisReadConfiguration() as an int. Previously, we tried to decode things like "0x1" with strtol() using a base of 10, which would always yield 0. This is what was causing problems with the Intel 2200BG Centrino 802.11g driver: the .inf file that comes with it has a key called RadioEnable with a value of 0x1. We incorrectly decoded this value to '0' when it was queried, hence the driver thought we wanted the radio turned off. - In if_ndis.c, most drivers don't accept NDIS_80211_AUTHMODE_AUTO, but NDIS_80211_AUTHMODE_SHARED may not be right in some cases, so for now always use NDIS_80211_AUTHMODE_OPEN. NOTE: There is still one problem with the Intel 2200BG driver: it happens that the kernel stack in Windows is larger than the kernel stack in FreeBSD. The 2200BG driver sometimes eats up more than 2 pages of stack space, which can lead to a double fault panic. For the moment, I got things to work by adding the following to my kernel config file: options KSTACK_PAGES=8 I'm pretty sure 8 is too big; I just picked this value out of a hat as a test, and it happened to work, so I left it. 4 pages might be enough. Unfortunately, I don't think you can dynamically give a thread a larger stack, so I'm not sure how to handle this short of putting a note in the man page about it and dealing with the flood of mail from people who never read man pages.
2004-03-20 23:39:43 +00:00
int base = 0;
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
uint32_t val;
char tmp[32];
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
np = ExAllocatePoolWithTag(NonPagedPool,
sizeof(ndis_parmlist_entry), 0);
if (np == NULL)
return(NDIS_STATUS_RESOURCES);
InsertHeadList((&block->nmb_parmlist), (&np->np_list));
*parm = p = &np->np_parm;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
switch(type) {
case ndis_parm_string:
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
/* See if this might be a number. */
val = strtoul((char *)oid->oid_arg1, NULL, 10);
us = &p->ncp_parmdata.ncp_stringdata;
p->ncp_type = ndis_parm_string;
if (val) {
snprintf(tmp, 32, "%x", val);
RtlInitAnsiString(&as, tmp);
} else {
RtlInitAnsiString(&as, (char *)oid->oid_arg1);
}
if (RtlAnsiStringToUnicodeString(us, &as, TRUE)) {
ExFreePool(np);
return(NDIS_STATUS_RESOURCES);
}
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
break;
case ndis_parm_int:
- Rewrite the timer and event API routines in subr_ndis.c so that they are actually layered on top of the KeTimer API in subr_ntoskrnl.c, just as it is in Windows. This reduces code duplication and more closely imitates the way things are done in Windows. - Modify ndis_encode_parm() to deal with the case where we have a registry key expressed as a hex value ("0x1") which is being read via NdisReadConfiguration() as an int. Previously, we tried to decode things like "0x1" with strtol() using a base of 10, which would always yield 0. This is what was causing problems with the Intel 2200BG Centrino 802.11g driver: the .inf file that comes with it has a key called RadioEnable with a value of 0x1. We incorrectly decoded this value to '0' when it was queried, hence the driver thought we wanted the radio turned off. - In if_ndis.c, most drivers don't accept NDIS_80211_AUTHMODE_AUTO, but NDIS_80211_AUTHMODE_SHARED may not be right in some cases, so for now always use NDIS_80211_AUTHMODE_OPEN. NOTE: There is still one problem with the Intel 2200BG driver: it happens that the kernel stack in Windows is larger than the kernel stack in FreeBSD. The 2200BG driver sometimes eats up more than 2 pages of stack space, which can lead to a double fault panic. For the moment, I got things to work by adding the following to my kernel config file: options KSTACK_PAGES=8 I'm pretty sure 8 is too big; I just picked this value out of a hat as a test, and it happened to work, so I left it. 4 pages might be enough. Unfortunately, I don't think you can dynamically give a thread a larger stack, so I'm not sure how to handle this short of putting a note in the man page about it and dealing with the flood of mail from people who never read man pages.
2004-03-20 23:39:43 +00:00
if (strncmp((char *)oid->oid_arg1, "0x", 2) == 0)
base = 16;
else
base = 10;
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
p->ncp_type = ndis_parm_int;
p->ncp_parmdata.ncp_intdata =
- Rewrite the timer and event API routines in subr_ndis.c so that they are actually layered on top of the KeTimer API in subr_ntoskrnl.c, just as it is in Windows. This reduces code duplication and more closely imitates the way things are done in Windows. - Modify ndis_encode_parm() to deal with the case where we have a registry key expressed as a hex value ("0x1") which is being read via NdisReadConfiguration() as an int. Previously, we tried to decode things like "0x1" with strtol() using a base of 10, which would always yield 0. This is what was causing problems with the Intel 2200BG Centrino 802.11g driver: the .inf file that comes with it has a key called RadioEnable with a value of 0x1. We incorrectly decoded this value to '0' when it was queried, hence the driver thought we wanted the radio turned off. - In if_ndis.c, most drivers don't accept NDIS_80211_AUTHMODE_AUTO, but NDIS_80211_AUTHMODE_SHARED may not be right in some cases, so for now always use NDIS_80211_AUTHMODE_OPEN. NOTE: There is still one problem with the Intel 2200BG driver: it happens that the kernel stack in Windows is larger than the kernel stack in FreeBSD. The 2200BG driver sometimes eats up more than 2 pages of stack space, which can lead to a double fault panic. For the moment, I got things to work by adding the following to my kernel config file: options KSTACK_PAGES=8 I'm pretty sure 8 is too big; I just picked this value out of a hat as a test, and it happened to work, so I left it. 4 pages might be enough. Unfortunately, I don't think you can dynamically give a thread a larger stack, so I'm not sure how to handle this short of putting a note in the man page about it and dealing with the flood of mail from people who never read man pages.
2004-03-20 23:39:43 +00:00
strtol((char *)oid->oid_arg1, NULL, base);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
break;
case ndis_parm_hexint:
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
#ifdef notdef
- Rewrite the timer and event API routines in subr_ndis.c so that they are actually layered on top of the KeTimer API in subr_ntoskrnl.c, just as it is in Windows. This reduces code duplication and more closely imitates the way things are done in Windows. - Modify ndis_encode_parm() to deal with the case where we have a registry key expressed as a hex value ("0x1") which is being read via NdisReadConfiguration() as an int. Previously, we tried to decode things like "0x1" with strtol() using a base of 10, which would always yield 0. This is what was causing problems with the Intel 2200BG Centrino 802.11g driver: the .inf file that comes with it has a key called RadioEnable with a value of 0x1. We incorrectly decoded this value to '0' when it was queried, hence the driver thought we wanted the radio turned off. - In if_ndis.c, most drivers don't accept NDIS_80211_AUTHMODE_AUTO, but NDIS_80211_AUTHMODE_SHARED may not be right in some cases, so for now always use NDIS_80211_AUTHMODE_OPEN. NOTE: There is still one problem with the Intel 2200BG driver: it happens that the kernel stack in Windows is larger than the kernel stack in FreeBSD. The 2200BG driver sometimes eats up more than 2 pages of stack space, which can lead to a double fault panic. For the moment, I got things to work by adding the following to my kernel config file: options KSTACK_PAGES=8 I'm pretty sure 8 is too big; I just picked this value out of a hat as a test, and it happened to work, so I left it. 4 pages might be enough. Unfortunately, I don't think you can dynamically give a thread a larger stack, so I'm not sure how to handle this short of putting a note in the man page about it and dealing with the flood of mail from people who never read man pages.
2004-03-20 23:39:43 +00:00
if (strncmp((char *)oid->oid_arg1, "0x", 2) == 0)
base = 16;
else
base = 10;
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
#endif
base = 16;
p->ncp_type = ndis_parm_hexint;
p->ncp_parmdata.ncp_intdata =
- Rewrite the timer and event API routines in subr_ndis.c so that they are actually layered on top of the KeTimer API in subr_ntoskrnl.c, just as it is in Windows. This reduces code duplication and more closely imitates the way things are done in Windows. - Modify ndis_encode_parm() to deal with the case where we have a registry key expressed as a hex value ("0x1") which is being read via NdisReadConfiguration() as an int. Previously, we tried to decode things like "0x1" with strtol() using a base of 10, which would always yield 0. This is what was causing problems with the Intel 2200BG Centrino 802.11g driver: the .inf file that comes with it has a key called RadioEnable with a value of 0x1. We incorrectly decoded this value to '0' when it was queried, hence the driver thought we wanted the radio turned off. - In if_ndis.c, most drivers don't accept NDIS_80211_AUTHMODE_AUTO, but NDIS_80211_AUTHMODE_SHARED may not be right in some cases, so for now always use NDIS_80211_AUTHMODE_OPEN. NOTE: There is still one problem with the Intel 2200BG driver: it happens that the kernel stack in Windows is larger than the kernel stack in FreeBSD. The 2200BG driver sometimes eats up more than 2 pages of stack space, which can lead to a double fault panic. For the moment, I got things to work by adding the following to my kernel config file: options KSTACK_PAGES=8 I'm pretty sure 8 is too big; I just picked this value out of a hat as a test, and it happened to work, so I left it. 4 pages might be enough. Unfortunately, I don't think you can dynamically give a thread a larger stack, so I'm not sure how to handle this short of putting a note in the man page about it and dealing with the flood of mail from people who never read man pages.
2004-03-20 23:39:43 +00:00
strtoul((char *)oid->oid_arg1, NULL, base);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
break;
default:
return(NDIS_STATUS_FAILURE);
break;
}
return(NDIS_STATUS_SUCCESS);
}
Make NdisReadPcmciaAttributeMemory() and NdisWritePcmciaAttributeMemory() actually work. Make the PCI and PCCARD attachments provide a bus_get_resource_list() method so that resource listing for PCCARD works. PCCARD does not have a bus_get_resource_list() method (yet), so I faked up the resource list management in if_ndis_pccard.c, and added bus_get_resource_list() methods to both if_ndis_pccard.c and if_ndis_pci.c. The one in the PCI attechment just hands off to the PCI bus code. The difference is transparent to the NDIS resource handler code. Fixed ndis_open_file() so that opening files which live on NFS filesystems work: pass an actual ucred structure to VOP_GETATTR() (NFS explodes if the ucred structure is NOCRED). Make NdisMMapIoSpace() handle mapping of PCMCIA attribute memory resources correctly. Turn subr_ndis.c:my_strcasecmp() into ndis_strcasecmp() and export it so that if_ndis_pccard.c can use it, and junk the other copy of my_strcasecmp() from if_ndis_pccard.c.
2004-07-11 00:19:30 +00:00
int
ndis_strcasecmp(s1, s2)
In subr_ndis.c, when searching for keys in our make-pretend registry, make the key name matching case-insensitive. There are some drivers and .inf files that have mismatched cases, e.g. the driver will look for "AdhocBand" whereas the .inf file specifies a registry key to be created called "AdHocBand." The mismatch is probably a typo that went undetected (so much for QA), but since Windows seems to be case-insensitive, we should be too. In if_ndis.c, initialize rates and channels correctly so that specify frequences correctly when trying to set channels in the 5Ghz band, and so that 802.11b rates show up for some a/b/g cards (which otherwise appear to have no 802.11b modes). Also, when setting OID_802_11_CONFIGURATION in ndis_80211_setstate(), provide default values for the beacon interval, ATIM window and dwelltime. The Atheros "Aries" driver will crash if you try to select ad-hoc mode and leave the beacon interval set to 0: it blindly uses this value and does a division by 0 in the interrupt handler, causing an integer divide trap.
2004-05-29 06:41:17 +00:00
const char *s1;
const char *s2;
{
Grrr. Really check subr_ndis.c in this time. (fixed my_strcasecmp())
2004-06-04 04:45:38 +00:00
char a, b;
In subr_ndis.c, when searching for keys in our make-pretend registry, make the key name matching case-insensitive. There are some drivers and .inf files that have mismatched cases, e.g. the driver will look for "AdhocBand" whereas the .inf file specifies a registry key to be created called "AdHocBand." The mismatch is probably a typo that went undetected (so much for QA), but since Windows seems to be case-insensitive, we should be too. In if_ndis.c, initialize rates and channels correctly so that specify frequences correctly when trying to set channels in the 5Ghz band, and so that 802.11b rates show up for some a/b/g cards (which otherwise appear to have no 802.11b modes). Also, when setting OID_802_11_CONFIGURATION in ndis_80211_setstate(), provide default values for the beacon interval, ATIM window and dwelltime. The Atheros "Aries" driver will crash if you try to select ad-hoc mode and leave the beacon interval set to 0: it blindly uses this value and does a division by 0 in the interrupt handler, causing an integer divide trap.
2004-05-29 06:41:17 +00:00
Grrr. Really check subr_ndis.c in this time. (fixed my_strcasecmp())
2004-06-04 04:45:38 +00:00
/*
* In the kernel, toupper() is a macro. Have to be careful
* not to use pointer arithmetic when passing it arguments.
*/
while(1) {
a = *s1;
b = *s2++;
if (toupper(a) != toupper(b))
break;
Big mess 'o changes: - Give ndiscvt(8) the ability to process a .SYS file directly into a .o file so that we don't have to emit big messy char arrays into the ndis_driver_data.h file. This behavior is currently optional, but may become the default some day. - Give ndiscvt(8) the ability to turn arbitrary files into .ko files so that they can be pre-loaded or kldloaded. (Both this and the previous change involve using objcopy(1)). - Give NdisOpenFile() the ability to 'read' files out of kernel memory that have been kldloaded or pre-loaded, and disallow the use of the normal vn_open() file opening method during bootstrap (when no filesystems have been mounted yet). Some people have reported that kldloading if_ndis.ko works fine when the system is running multiuser but causes a panic when the modile is pre-loaded by /boot/loader. This happens with drivers that need to use NdisOpenFile() to access external files (i.e. firmware images). NdisOpenFile() won't work during kernel bootstrapping because no filesystems have been mounted. To get around this, you can now do the following: o Say you have a firmware file called firmware.img o Do: ndiscvt -f firmware.img -- this creates firmware.img.ko o Put the firmware.img.ko in /boot/kernel o add firmware.img_load="YES" in /boot/loader.conf o add if_ndis_load="YES" and ndis_load="YES" as well Now the loader will suck the additional file into memory as a .ko. The phony .ko has two symbols in it: filename_start and filename_end, which are generated by objcopy(1). ndis_open_file() will traverse each module in the module list looking for these symbols and, if it finds them, it'll use them to generate the file mapping address and length values that the caller of NdisOpenFile() wants. As a bonus, this will even work if the file has been statically linked into the kernel itself, since the "kernel" module is searched too. (ndiscvt(8) will generate both filename.o and filename.ko for you). - Modify the mechanism used to provide make-pretend FASTCALL support. Rather than using inline assembly to yank the first two arguments out of %ecx and %edx, we now use the __regparm__(3) attribute (and the __stdcall__ attribute) and use some macro magic to re-order the arguments and provide dummy arguments as needed so that the arguments passed in registers end up in the right place. Change taken from DragonflyBSD version of the NDISulator.
2004-08-01 20:04:31 +00:00
if (*s1++ == '\0')
Grrr. Really check subr_ndis.c in this time. (fixed my_strcasecmp())
2004-06-04 04:45:38 +00:00
return(0);
}
In subr_ndis.c, when searching for keys in our make-pretend registry, make the key name matching case-insensitive. There are some drivers and .inf files that have mismatched cases, e.g. the driver will look for "AdhocBand" whereas the .inf file specifies a registry key to be created called "AdHocBand." The mismatch is probably a typo that went undetected (so much for QA), but since Windows seems to be case-insensitive, we should be too. In if_ndis.c, initialize rates and channels correctly so that specify frequences correctly when trying to set channels in the 5Ghz band, and so that 802.11b rates show up for some a/b/g cards (which otherwise appear to have no 802.11b modes). Also, when setting OID_802_11_CONFIGURATION in ndis_80211_setstate(), provide default values for the beacon interval, ATIM window and dwelltime. The Atheros "Aries" driver will crash if you try to select ad-hoc mode and leave the beacon interval set to 0: it blindly uses this value and does a division by 0 in the interrupt handler, causing an integer divide trap.
2004-05-29 06:41:17 +00:00
Grrr. Really check subr_ndis.c in this time. (fixed my_strcasecmp())
2004-06-04 04:45:38 +00:00
return (*(const unsigned char *)s1 - *(const unsigned char *)(s2 - 1));
In subr_ndis.c, when searching for keys in our make-pretend registry, make the key name matching case-insensitive. There are some drivers and .inf files that have mismatched cases, e.g. the driver will look for "AdhocBand" whereas the .inf file specifies a registry key to be created called "AdHocBand." The mismatch is probably a typo that went undetected (so much for QA), but since Windows seems to be case-insensitive, we should be too. In if_ndis.c, initialize rates and channels correctly so that specify frequences correctly when trying to set channels in the 5Ghz band, and so that 802.11b rates show up for some a/b/g cards (which otherwise appear to have no 802.11b modes). Also, when setting OID_802_11_CONFIGURATION in ndis_80211_setstate(), provide default values for the beacon interval, ATIM window and dwelltime. The Atheros "Aries" driver will crash if you try to select ad-hoc mode and leave the beacon interval set to 0: it blindly uses this value and does a division by 0 in the interrupt handler, causing an integer divide trap.
2004-05-29 06:41:17 +00:00
}
Big mess 'o changes: - Give ndiscvt(8) the ability to process a .SYS file directly into a .o file so that we don't have to emit big messy char arrays into the ndis_driver_data.h file. This behavior is currently optional, but may become the default some day. - Give ndiscvt(8) the ability to turn arbitrary files into .ko files so that they can be pre-loaded or kldloaded. (Both this and the previous change involve using objcopy(1)). - Give NdisOpenFile() the ability to 'read' files out of kernel memory that have been kldloaded or pre-loaded, and disallow the use of the normal vn_open() file opening method during bootstrap (when no filesystems have been mounted yet). Some people have reported that kldloading if_ndis.ko works fine when the system is running multiuser but causes a panic when the modile is pre-loaded by /boot/loader. This happens with drivers that need to use NdisOpenFile() to access external files (i.e. firmware images). NdisOpenFile() won't work during kernel bootstrapping because no filesystems have been mounted. To get around this, you can now do the following: o Say you have a firmware file called firmware.img o Do: ndiscvt -f firmware.img -- this creates firmware.img.ko o Put the firmware.img.ko in /boot/kernel o add firmware.img_load="YES" in /boot/loader.conf o add if_ndis_load="YES" and ndis_load="YES" as well Now the loader will suck the additional file into memory as a .ko. The phony .ko has two symbols in it: filename_start and filename_end, which are generated by objcopy(1). ndis_open_file() will traverse each module in the module list looking for these symbols and, if it finds them, it'll use them to generate the file mapping address and length values that the caller of NdisOpenFile() wants. As a bonus, this will even work if the file has been statically linked into the kernel itself, since the "kernel" module is searched too. (ndiscvt(8) will generate both filename.o and filename.ko for you). - Modify the mechanism used to provide make-pretend FASTCALL support. Rather than using inline assembly to yank the first two arguments out of %ecx and %edx, we now use the __regparm__(3) attribute (and the __stdcall__ attribute) and use some macro magic to re-order the arguments and provide dummy arguments as needed so that the arguments passed in registers end up in the right place. Change taken from DragonflyBSD version of the NDISulator.
2004-08-01 20:04:31 +00:00
int
ndis_strncasecmp(s1, s2, n)
const char *s1;
const char *s2;
size_t n;
{
char a, b;
if (n != 0) {
do {
a = *s1;
b = *s2++;
if (toupper(a) != toupper(b))
return (*(const unsigned char *)s1 -
*(const unsigned char *)(s2 - 1));
if (*s1++ == '\0')
break;
} while (--n != 0);
}
return(0);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisReadConfiguration(status, parm, cfg, key, type)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_status *status;
ndis_config_parm **parm;
ndis_handle cfg;
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
unicode_string *key;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_parm_type type;
{
char *keystr = NULL;
ndis_miniport_block *block;
struct ndis_softc *sc;
struct sysctl_oid *oidp;
struct sysctl_ctx_entry *e;
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
ansi_string as;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
block = (ndis_miniport_block *)cfg;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
sc = device_get_softc(block->nmb_physdeviceobj->do_devext);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
if (key->us_len == 0 || key->us_buf == NULL) {
In if_ndis.c:ndis_attach(), temporarily set the IFF_UP flag while calling the haltfunc. If an interrupt is triggered by the init or halt func, the IFF_UP flag must be set in order for us to be able to service it. In kern_ndis.c: implement a handler for NdisMSendResourcesAvailable() (currently does nothing since we don't really need it). In subr_ndis.c: - Correct ndis_init_string() and ndis_unicode_to_ansi(), which were both horribly broken. - Implement NdisImmediateReadPciSlotInformation() and NdisImmediateWritePciSlotInformation(). - Implement NdisBufferLength(). - Work around my first confirmed NDIS driver bug. The SMC 9462 gigE driver (natsemi 83820-based copper) incorrectly creates a spinlock in its DriverEntry() routine and then destroys it in its MiniportHalt() handler. This is wrong: spinlocks should be created in MiniportInit(). In a Windows environment, this is often not a problem because DriverEntry()/MiniportInit() are called once when the system boots and MiniportHalt() or the shutdown handler is called when the system halts. With this stuff in place, this driver now seems to work: ndis0: <SMC EZ Card 1000> port 0xe000-0xe0ff mem 0xda000000-0xda000fff irq 10 at device 9.0 on pci0 ndis0: assign PCI resources... ndis_open_file("FLASH9.hex", 18446744073709551615) ndis0: Ethernet address: 00:04:e2:0e:d3:f0
2004-01-03 13:20:30 +00:00
*status = NDIS_STATUS_FAILURE;
return;
}
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
if (RtlUnicodeStringToAnsiString(&as, key, TRUE)) {
*status = NDIS_STATUS_RESOURCES;
return;
}
keystr = as.as_buf;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
/*
* See if registry key is already in a list of known keys
* included with the driver.
*/
Take advantage of the dev sysctl tree. Approved by: wpaul
2004-06-04 22:24:46 +00:00
#if __FreeBSD_version < 502113
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
TAILQ_FOREACH(e, &sc->ndis_ctx, link) {
Take advantage of the dev sysctl tree. Approved by: wpaul
2004-06-04 22:24:46 +00:00
#else
TAILQ_FOREACH(e, device_get_sysctl_ctx(sc->ndis_dev), link) {
#endif
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
oidp = e->entry;
Make NdisReadPcmciaAttributeMemory() and NdisWritePcmciaAttributeMemory() actually work. Make the PCI and PCCARD attachments provide a bus_get_resource_list() method so that resource listing for PCCARD works. PCCARD does not have a bus_get_resource_list() method (yet), so I faked up the resource list management in if_ndis_pccard.c, and added bus_get_resource_list() methods to both if_ndis_pccard.c and if_ndis_pci.c. The one in the PCI attechment just hands off to the PCI bus code. The difference is transparent to the NDIS resource handler code. Fixed ndis_open_file() so that opening files which live on NFS filesystems work: pass an actual ucred structure to VOP_GETATTR() (NFS explodes if the ucred structure is NOCRED). Make NdisMMapIoSpace() handle mapping of PCMCIA attribute memory resources correctly. Turn subr_ndis.c:my_strcasecmp() into ndis_strcasecmp() and export it so that if_ndis_pccard.c can use it, and junk the other copy of my_strcasecmp() from if_ndis_pccard.c.
2004-07-11 00:19:30 +00:00
if (ndis_strcasecmp(oidp->oid_name, keystr) == 0) {
Correct the behavior of ndis_adjust_buflen(): the NDIS spec says it's an error to set the buffer bytecount to anything larger than the buffer's original allocation size, but anything less than that is ok. Also, in ndis_ptom(), use the same logic: if the bytecount is larger than the allocation size, consider the bytecount invalid and the allocation size as the packet fragment length (m_len) instead of the bytecount. This corrects a consistency problem between the Broadcom wireless driver and some of the ethernet drivers I've tested: the ethernet drivers all report the packet frag sizes in buf->nb_bytecount, but the Broadcom wireless driver reports them in buf->nb_size. This seems like a bug to me, but it clearly must work in Windows, so we have to deal with it here too.
2003-12-12 08:54:48 +00:00
if (strcmp((char *)oidp->oid_arg1, "UNSET") == 0) {
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
RtlFreeAnsiString(&as);
Correct the behavior of ndis_adjust_buflen(): the NDIS spec says it's an error to set the buffer bytecount to anything larger than the buffer's original allocation size, but anything less than that is ok. Also, in ndis_ptom(), use the same logic: if the bytecount is larger than the allocation size, consider the bytecount invalid and the allocation size as the packet fragment length (m_len) instead of the bytecount. This corrects a consistency problem between the Broadcom wireless driver and some of the ethernet drivers I've tested: the ethernet drivers all report the packet frag sizes in buf->nb_bytecount, but the Broadcom wireless driver reports them in buf->nb_size. This seems like a bug to me, but it clearly must work in Windows, so we have to deal with it here too.
2003-12-12 08:54:48 +00:00
*status = NDIS_STATUS_FAILURE;
return;
}
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
*status = ndis_encode_parm(block, oidp, type, parm);
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
RtlFreeAnsiString(&as);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
}
/*
* If the key didn't match, add it to the list of dynamically
* created ones. Sometimes, drivers refer to registry keys
* that aren't documented in their .INF files. These keys
* are supposed to be created by some sort of utility or
* control panel snap-in that comes with the driver software.
* Sometimes it's useful to be able to manipulate these.
* If the driver requests the key in the form of a string,
* make its default value an empty string, otherwise default
* it to "0".
*/
if (type == ndis_parm_int || type == ndis_parm_hexint)
Correct the behavior of ndis_adjust_buflen(): the NDIS spec says it's an error to set the buffer bytecount to anything larger than the buffer's original allocation size, but anything less than that is ok. Also, in ndis_ptom(), use the same logic: if the bytecount is larger than the allocation size, consider the bytecount invalid and the allocation size as the packet fragment length (m_len) instead of the bytecount. This corrects a consistency problem between the Broadcom wireless driver and some of the ethernet drivers I've tested: the ethernet drivers all report the packet frag sizes in buf->nb_bytecount, but the Broadcom wireless driver reports them in buf->nb_size. This seems like a bug to me, but it clearly must work in Windows, so we have to deal with it here too.
2003-12-12 08:54:48 +00:00
ndis_add_sysctl(sc, keystr, "(dynamic integer key)",
"UNSET", CTLFLAG_RW);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
else
Correct the behavior of ndis_adjust_buflen(): the NDIS spec says it's an error to set the buffer bytecount to anything larger than the buffer's original allocation size, but anything less than that is ok. Also, in ndis_ptom(), use the same logic: if the bytecount is larger than the allocation size, consider the bytecount invalid and the allocation size as the packet fragment length (m_len) instead of the bytecount. This corrects a consistency problem between the Broadcom wireless driver and some of the ethernet drivers I've tested: the ethernet drivers all report the packet frag sizes in buf->nb_bytecount, but the Broadcom wireless driver reports them in buf->nb_size. This seems like a bug to me, but it clearly must work in Windows, so we have to deal with it here too.
2003-12-12 08:54:48 +00:00
ndis_add_sysctl(sc, keystr, "(dynamic string key)",
"UNSET", CTLFLAG_RW);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
RtlFreeAnsiString(&as);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
*status = NDIS_STATUS_FAILURE;
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
static ndis_status
ndis_decode_parm(block, parm, val)
ndis_miniport_block *block;
ndis_config_parm *parm;
char *val;
{
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
unicode_string *ustr;
ansi_string as;
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
switch(parm->ncp_type) {
case ndis_parm_string:
ustr = &parm->ncp_parmdata.ncp_stringdata;
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
if (RtlUnicodeStringToAnsiString(&as, ustr, TRUE))
return(NDIS_STATUS_RESOURCES);
bcopy(as.as_buf, val, as.as_len);
RtlFreeAnsiString(&as);
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
break;
case ndis_parm_int:
Convert from using taskqueue_swi to using private kernel threads. The problem with using taskqueue_swi is that some of the things we defer into threads might block for up to several seconds. This is an unfriendly thing to do to taskqueue_swi, since it is assumed the taskqueue threads will execute fairly quickly once a task is submitted. Reorganized the locking in if_ndis.c in the process. Cleaned up ndis_write_cfg() and ndis_decode_parm() a little.
2004-01-18 22:57:11 +00:00
sprintf(val, "%d", parm->ncp_parmdata.ncp_intdata);
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
break;
case ndis_parm_hexint:
sprintf(val, "%xu", parm->ncp_parmdata.ncp_intdata);
break;
default:
return(NDIS_STATUS_FAILURE);
break;
}
return(NDIS_STATUS_SUCCESS);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisWriteConfiguration(status, cfg, key, parm)
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
ndis_status *status;
ndis_handle cfg;
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
unicode_string *key;
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
ndis_config_parm *parm;
{
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
ansi_string as;
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
char *keystr = NULL;
ndis_miniport_block *block;
struct ndis_softc *sc;
struct sysctl_oid *oidp;
struct sysctl_ctx_entry *e;
char val[256];
block = (ndis_miniport_block *)cfg;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
sc = device_get_softc(block->nmb_physdeviceobj->do_devext);
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
if (RtlUnicodeStringToAnsiString(&as, key, TRUE)) {
*status = NDIS_STATUS_RESOURCES;
return;
}
keystr = as.as_buf;
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
/* Decode the parameter into a string. */
Convert from using taskqueue_swi to using private kernel threads. The problem with using taskqueue_swi is that some of the things we defer into threads might block for up to several seconds. This is an unfriendly thing to do to taskqueue_swi, since it is assumed the taskqueue threads will execute fairly quickly once a task is submitted. Reorganized the locking in if_ndis.c in the process. Cleaned up ndis_write_cfg() and ndis_decode_parm() a little.
2004-01-18 22:57:11 +00:00
bzero(val, sizeof(val));
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
*status = ndis_decode_parm(block, parm, val);
if (*status != NDIS_STATUS_SUCCESS) {
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
RtlFreeAnsiString(&as);
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
return;
}
/* See if the key already exists. */
Take advantage of the dev sysctl tree. Approved by: wpaul
2004-06-04 22:24:46 +00:00
#if __FreeBSD_version < 502113
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
TAILQ_FOREACH(e, &sc->ndis_ctx, link) {
Take advantage of the dev sysctl tree. Approved by: wpaul
2004-06-04 22:24:46 +00:00
#else
TAILQ_FOREACH(e, device_get_sysctl_ctx(sc->ndis_dev), link) {
#endif
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
oidp = e->entry;
Make NdisReadPcmciaAttributeMemory() and NdisWritePcmciaAttributeMemory() actually work. Make the PCI and PCCARD attachments provide a bus_get_resource_list() method so that resource listing for PCCARD works. PCCARD does not have a bus_get_resource_list() method (yet), so I faked up the resource list management in if_ndis_pccard.c, and added bus_get_resource_list() methods to both if_ndis_pccard.c and if_ndis_pci.c. The one in the PCI attechment just hands off to the PCI bus code. The difference is transparent to the NDIS resource handler code. Fixed ndis_open_file() so that opening files which live on NFS filesystems work: pass an actual ucred structure to VOP_GETATTR() (NFS explodes if the ucred structure is NOCRED). Make NdisMMapIoSpace() handle mapping of PCMCIA attribute memory resources correctly. Turn subr_ndis.c:my_strcasecmp() into ndis_strcasecmp() and export it so that if_ndis_pccard.c can use it, and junk the other copy of my_strcasecmp() from if_ndis_pccard.c.
2004-07-11 00:19:30 +00:00
if (ndis_strcasecmp(oidp->oid_name, keystr) == 0) {
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
/* Found it, set the value. */
strcpy((char *)oidp->oid_arg1, val);
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
RtlFreeAnsiString(&as);
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
return;
}
}
/* Not found, add a new key with the specified value. */
ndis_add_sysctl(sc, keystr, "(dynamically set key)",
val, CTLFLAG_RW);
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
RtlFreeAnsiString(&as);
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
*status = NDIS_STATUS_SUCCESS;
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisCloseConfiguration(cfg)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_handle cfg;
{
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
list_entry *e;
ndis_parmlist_entry *pe;
ndis_miniport_block *block;
ndis_config_parm *p;
block = (ndis_miniport_block *)cfg;
while (!IsListEmpty(&block->nmb_parmlist)) {
e = RemoveHeadList(&block->nmb_parmlist);
pe = CONTAINING_RECORD(e, ndis_parmlist_entry, np_list);
p = &pe->np_parm;
if (p->ncp_type == ndis_parm_string)
RtlFreeUnicodeString(&p->ncp_parmdata.ncp_stringdata);
ExFreePool(e);
}
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
Continue my efforts to imitate Windows as closely as possible by attempting to duplicate Windows spinlocks. Windows spinlocks differ from FreeBSD spinlocks in the way they block preemption. FreeBSD spinlocks use critical_enter(), which masks off _all_ interrupts. This prevents any other threads from being scheduled, but it also prevents ISRs from running. In Windows, preemption is achieved by raising the processor IRQL to DISPATCH_LEVEL, which prevents other threads from preempting you, but does _not_ prevent device ISRs from running. (This is essentially what Solaris calls dispatcher locks.) The Windows spinlock itself (kspin_lock) is just an integer value which is atomically set when you acquire the lock and atomically cleared when you release it. FreeBSD doesn't have IRQ levels, so we have to cheat a little by using thread priorities: normal thread priority is PASSIVE_LEVEL, lowest interrupt thread priority is DISPATCH_LEVEL, highest thread priority is DEVICE_LEVEL (PI_REALTIME) and critical_enter() is HIGH_LEVEL. In practice, only PASSIVE_LEVEL and DISPATCH_LEVEL matter to us. The immediate benefit of all this is that I no longer have to rely on a mutex pool. Now, I'm sure many people will be seized by the urge to criticize me for doing an end run around our own spinlock implementation, but it makes more sense to do it this way. Well, it does to me anyway. Overview of the changes: - Properly implement hal_lock(), hal_unlock(), hal_irql(), hal_raise_irql() and hal_lower_irql() so that they more closely resemble their Windows counterparts. The IRQL is determined by thread priority. - Make ntoskrnl_lock_dpc() and ntoskrnl_unlock_dpc() do what they do in Windows, which is to atomically set/clear the lock value. These routines are designed to be called from DISPATCH_LEVEL, and are actually half of the work involved in acquiring/releasing spinlocks. - Add FASTCALL1(), FASTCALL2() and FASTCALL3() macros/wrappers that allow us to call a _fastcall function in spite of the fact that our version of gcc doesn't support __attribute__((__fastcall__)) yet. The macros take 1, 2 or 3 arguments, respectively. We need to call hal_lock(), hal_unlock() etc... ourselves, but can't really invoke the function directly. I could have just made the underlying functions native routines and put _fastcall wrappers around them for the benefit of Windows binaries, but that would create needless bloat. - Remove ndis_mtxpool and all references to it. We don't need it anymore. - Re-implement the NdisSpinLock routines so that they use hal_lock() and friends like they do in Windows. - Use the new spinlock methods for handling lookaside lists and linked list updates in place of the mutex locks that were there before. - Remove mutex locking from ndis_isr() and ndis_intrhand() since they're already called with ndis_intrmtx held in if_ndis.c. - Put ndis_destroy_lock() code under explicit #ifdef notdef/#endif. It turns out there are some drivers which stupidly free the memory in which their spinlocks reside before calling ndis_destroy_lock() on them (touch-after-free bug). The ADMtek wireless driver is guilty of this faux pas. (Why this doesn't clobber Windows I have no idea.) - Make NdisDprAcquireSpinLock() and NdisDprReleaseSpinLock() into real functions instead of aliasing them to NdisAcaquireSpinLock() and NdisReleaseSpinLock(). The Dpr routines use KeAcquireSpinLockAtDpcLevel() level and KeReleaseSpinLockFromDpcLevel(), which acquires the lock without twiddling the IRQL. - In ndis_linksts_done(), do _not_ call ndis_80211_getstate(). Some drivers may call the status/status done callbacks as the result of setting an OID: ndis_80211_getstate() gets OIDs, which means we might cause the driver to recursively access some of its internal structures unexpectedly. The ndis_ticktask() routine will call ndis_80211_getstate() for us eventually anyway. - Fix the channel setting code a little in ndis_80211_setstate(), and initialize the channel to IEEE80211_CHAN_ANYC. (The Microsoft spec says you're not supposed to twiddle the channel in BSS mode; I may need to enforce this later.) This fixes the problems I was having with the ADMtek adm8211 driver: we were setting the channel to a non-standard default, which would cause it to fail to associate in BSS mode. - Use hal_raise_irql() to raise our IRQL to DISPATCH_LEVEL when calling certain miniport routines, per the Microsoft documentation. I think that's everything. Hopefully, other than fixing the ADMtek driver, there should be no apparent change in behavior.
2004-04-14 07:48:03 +00:00
/*
* Initialize a Windows spinlock.
*/
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisAllocateSpinLock(lock)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_spin_lock *lock;
{
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
KeInitializeSpinLock(&lock->nsl_spinlock);
Continue my efforts to imitate Windows as closely as possible by attempting to duplicate Windows spinlocks. Windows spinlocks differ from FreeBSD spinlocks in the way they block preemption. FreeBSD spinlocks use critical_enter(), which masks off _all_ interrupts. This prevents any other threads from being scheduled, but it also prevents ISRs from running. In Windows, preemption is achieved by raising the processor IRQL to DISPATCH_LEVEL, which prevents other threads from preempting you, but does _not_ prevent device ISRs from running. (This is essentially what Solaris calls dispatcher locks.) The Windows spinlock itself (kspin_lock) is just an integer value which is atomically set when you acquire the lock and atomically cleared when you release it. FreeBSD doesn't have IRQ levels, so we have to cheat a little by using thread priorities: normal thread priority is PASSIVE_LEVEL, lowest interrupt thread priority is DISPATCH_LEVEL, highest thread priority is DEVICE_LEVEL (PI_REALTIME) and critical_enter() is HIGH_LEVEL. In practice, only PASSIVE_LEVEL and DISPATCH_LEVEL matter to us. The immediate benefit of all this is that I no longer have to rely on a mutex pool. Now, I'm sure many people will be seized by the urge to criticize me for doing an end run around our own spinlock implementation, but it makes more sense to do it this way. Well, it does to me anyway. Overview of the changes: - Properly implement hal_lock(), hal_unlock(), hal_irql(), hal_raise_irql() and hal_lower_irql() so that they more closely resemble their Windows counterparts. The IRQL is determined by thread priority. - Make ntoskrnl_lock_dpc() and ntoskrnl_unlock_dpc() do what they do in Windows, which is to atomically set/clear the lock value. These routines are designed to be called from DISPATCH_LEVEL, and are actually half of the work involved in acquiring/releasing spinlocks. - Add FASTCALL1(), FASTCALL2() and FASTCALL3() macros/wrappers that allow us to call a _fastcall function in spite of the fact that our version of gcc doesn't support __attribute__((__fastcall__)) yet. The macros take 1, 2 or 3 arguments, respectively. We need to call hal_lock(), hal_unlock() etc... ourselves, but can't really invoke the function directly. I could have just made the underlying functions native routines and put _fastcall wrappers around them for the benefit of Windows binaries, but that would create needless bloat. - Remove ndis_mtxpool and all references to it. We don't need it anymore. - Re-implement the NdisSpinLock routines so that they use hal_lock() and friends like they do in Windows. - Use the new spinlock methods for handling lookaside lists and linked list updates in place of the mutex locks that were there before. - Remove mutex locking from ndis_isr() and ndis_intrhand() since they're already called with ndis_intrmtx held in if_ndis.c. - Put ndis_destroy_lock() code under explicit #ifdef notdef/#endif. It turns out there are some drivers which stupidly free the memory in which their spinlocks reside before calling ndis_destroy_lock() on them (touch-after-free bug). The ADMtek wireless driver is guilty of this faux pas. (Why this doesn't clobber Windows I have no idea.) - Make NdisDprAcquireSpinLock() and NdisDprReleaseSpinLock() into real functions instead of aliasing them to NdisAcaquireSpinLock() and NdisReleaseSpinLock(). The Dpr routines use KeAcquireSpinLockAtDpcLevel() level and KeReleaseSpinLockFromDpcLevel(), which acquires the lock without twiddling the IRQL. - In ndis_linksts_done(), do _not_ call ndis_80211_getstate(). Some drivers may call the status/status done callbacks as the result of setting an OID: ndis_80211_getstate() gets OIDs, which means we might cause the driver to recursively access some of its internal structures unexpectedly. The ndis_ticktask() routine will call ndis_80211_getstate() for us eventually anyway. - Fix the channel setting code a little in ndis_80211_setstate(), and initialize the channel to IEEE80211_CHAN_ANYC. (The Microsoft spec says you're not supposed to twiddle the channel in BSS mode; I may need to enforce this later.) This fixes the problems I was having with the ADMtek adm8211 driver: we were setting the channel to a non-standard default, which would cause it to fail to associate in BSS mode. - Use hal_raise_irql() to raise our IRQL to DISPATCH_LEVEL when calling certain miniport routines, per the Microsoft documentation. I think that's everything. Hopefully, other than fixing the ADMtek driver, there should be no apparent change in behavior.
2004-04-14 07:48:03 +00:00
lock->nsl_kirql = 0;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
Continue my efforts to imitate Windows as closely as possible by attempting to duplicate Windows spinlocks. Windows spinlocks differ from FreeBSD spinlocks in the way they block preemption. FreeBSD spinlocks use critical_enter(), which masks off _all_ interrupts. This prevents any other threads from being scheduled, but it also prevents ISRs from running. In Windows, preemption is achieved by raising the processor IRQL to DISPATCH_LEVEL, which prevents other threads from preempting you, but does _not_ prevent device ISRs from running. (This is essentially what Solaris calls dispatcher locks.) The Windows spinlock itself (kspin_lock) is just an integer value which is atomically set when you acquire the lock and atomically cleared when you release it. FreeBSD doesn't have IRQ levels, so we have to cheat a little by using thread priorities: normal thread priority is PASSIVE_LEVEL, lowest interrupt thread priority is DISPATCH_LEVEL, highest thread priority is DEVICE_LEVEL (PI_REALTIME) and critical_enter() is HIGH_LEVEL. In practice, only PASSIVE_LEVEL and DISPATCH_LEVEL matter to us. The immediate benefit of all this is that I no longer have to rely on a mutex pool. Now, I'm sure many people will be seized by the urge to criticize me for doing an end run around our own spinlock implementation, but it makes more sense to do it this way. Well, it does to me anyway. Overview of the changes: - Properly implement hal_lock(), hal_unlock(), hal_irql(), hal_raise_irql() and hal_lower_irql() so that they more closely resemble their Windows counterparts. The IRQL is determined by thread priority. - Make ntoskrnl_lock_dpc() and ntoskrnl_unlock_dpc() do what they do in Windows, which is to atomically set/clear the lock value. These routines are designed to be called from DISPATCH_LEVEL, and are actually half of the work involved in acquiring/releasing spinlocks. - Add FASTCALL1(), FASTCALL2() and FASTCALL3() macros/wrappers that allow us to call a _fastcall function in spite of the fact that our version of gcc doesn't support __attribute__((__fastcall__)) yet. The macros take 1, 2 or 3 arguments, respectively. We need to call hal_lock(), hal_unlock() etc... ourselves, but can't really invoke the function directly. I could have just made the underlying functions native routines and put _fastcall wrappers around them for the benefit of Windows binaries, but that would create needless bloat. - Remove ndis_mtxpool and all references to it. We don't need it anymore. - Re-implement the NdisSpinLock routines so that they use hal_lock() and friends like they do in Windows. - Use the new spinlock methods for handling lookaside lists and linked list updates in place of the mutex locks that were there before. - Remove mutex locking from ndis_isr() and ndis_intrhand() since they're already called with ndis_intrmtx held in if_ndis.c. - Put ndis_destroy_lock() code under explicit #ifdef notdef/#endif. It turns out there are some drivers which stupidly free the memory in which their spinlocks reside before calling ndis_destroy_lock() on them (touch-after-free bug). The ADMtek wireless driver is guilty of this faux pas. (Why this doesn't clobber Windows I have no idea.) - Make NdisDprAcquireSpinLock() and NdisDprReleaseSpinLock() into real functions instead of aliasing them to NdisAcaquireSpinLock() and NdisReleaseSpinLock(). The Dpr routines use KeAcquireSpinLockAtDpcLevel() level and KeReleaseSpinLockFromDpcLevel(), which acquires the lock without twiddling the IRQL. - In ndis_linksts_done(), do _not_ call ndis_80211_getstate(). Some drivers may call the status/status done callbacks as the result of setting an OID: ndis_80211_getstate() gets OIDs, which means we might cause the driver to recursively access some of its internal structures unexpectedly. The ndis_ticktask() routine will call ndis_80211_getstate() for us eventually anyway. - Fix the channel setting code a little in ndis_80211_setstate(), and initialize the channel to IEEE80211_CHAN_ANYC. (The Microsoft spec says you're not supposed to twiddle the channel in BSS mode; I may need to enforce this later.) This fixes the problems I was having with the ADMtek adm8211 driver: we were setting the channel to a non-standard default, which would cause it to fail to associate in BSS mode. - Use hal_raise_irql() to raise our IRQL to DISPATCH_LEVEL when calling certain miniport routines, per the Microsoft documentation. I think that's everything. Hopefully, other than fixing the ADMtek driver, there should be no apparent change in behavior.
2004-04-14 07:48:03 +00:00
/*
* Destroy a Windows spinlock. This is a no-op for now. There are two reasons
* for this. One is that it's sort of superfluous: we don't have to do anything
* special to deallocate the spinlock. The other is that there are some buggy
* drivers which call NdisFreeSpinLock() _after_ calling NdisFreeMemory() on
* the block of memory in which the spinlock resides. (Yes, ADMtek, I'm
* talking to you.)
*/
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisFreeSpinLock(lock)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_spin_lock *lock;
{
Continue my efforts to imitate Windows as closely as possible by attempting to duplicate Windows spinlocks. Windows spinlocks differ from FreeBSD spinlocks in the way they block preemption. FreeBSD spinlocks use critical_enter(), which masks off _all_ interrupts. This prevents any other threads from being scheduled, but it also prevents ISRs from running. In Windows, preemption is achieved by raising the processor IRQL to DISPATCH_LEVEL, which prevents other threads from preempting you, but does _not_ prevent device ISRs from running. (This is essentially what Solaris calls dispatcher locks.) The Windows spinlock itself (kspin_lock) is just an integer value which is atomically set when you acquire the lock and atomically cleared when you release it. FreeBSD doesn't have IRQ levels, so we have to cheat a little by using thread priorities: normal thread priority is PASSIVE_LEVEL, lowest interrupt thread priority is DISPATCH_LEVEL, highest thread priority is DEVICE_LEVEL (PI_REALTIME) and critical_enter() is HIGH_LEVEL. In practice, only PASSIVE_LEVEL and DISPATCH_LEVEL matter to us. The immediate benefit of all this is that I no longer have to rely on a mutex pool. Now, I'm sure many people will be seized by the urge to criticize me for doing an end run around our own spinlock implementation, but it makes more sense to do it this way. Well, it does to me anyway. Overview of the changes: - Properly implement hal_lock(), hal_unlock(), hal_irql(), hal_raise_irql() and hal_lower_irql() so that they more closely resemble their Windows counterparts. The IRQL is determined by thread priority. - Make ntoskrnl_lock_dpc() and ntoskrnl_unlock_dpc() do what they do in Windows, which is to atomically set/clear the lock value. These routines are designed to be called from DISPATCH_LEVEL, and are actually half of the work involved in acquiring/releasing spinlocks. - Add FASTCALL1(), FASTCALL2() and FASTCALL3() macros/wrappers that allow us to call a _fastcall function in spite of the fact that our version of gcc doesn't support __attribute__((__fastcall__)) yet. The macros take 1, 2 or 3 arguments, respectively. We need to call hal_lock(), hal_unlock() etc... ourselves, but can't really invoke the function directly. I could have just made the underlying functions native routines and put _fastcall wrappers around them for the benefit of Windows binaries, but that would create needless bloat. - Remove ndis_mtxpool and all references to it. We don't need it anymore. - Re-implement the NdisSpinLock routines so that they use hal_lock() and friends like they do in Windows. - Use the new spinlock methods for handling lookaside lists and linked list updates in place of the mutex locks that were there before. - Remove mutex locking from ndis_isr() and ndis_intrhand() since they're already called with ndis_intrmtx held in if_ndis.c. - Put ndis_destroy_lock() code under explicit #ifdef notdef/#endif. It turns out there are some drivers which stupidly free the memory in which their spinlocks reside before calling ndis_destroy_lock() on them (touch-after-free bug). The ADMtek wireless driver is guilty of this faux pas. (Why this doesn't clobber Windows I have no idea.) - Make NdisDprAcquireSpinLock() and NdisDprReleaseSpinLock() into real functions instead of aliasing them to NdisAcaquireSpinLock() and NdisReleaseSpinLock(). The Dpr routines use KeAcquireSpinLockAtDpcLevel() level and KeReleaseSpinLockFromDpcLevel(), which acquires the lock without twiddling the IRQL. - In ndis_linksts_done(), do _not_ call ndis_80211_getstate(). Some drivers may call the status/status done callbacks as the result of setting an OID: ndis_80211_getstate() gets OIDs, which means we might cause the driver to recursively access some of its internal structures unexpectedly. The ndis_ticktask() routine will call ndis_80211_getstate() for us eventually anyway. - Fix the channel setting code a little in ndis_80211_setstate(), and initialize the channel to IEEE80211_CHAN_ANYC. (The Microsoft spec says you're not supposed to twiddle the channel in BSS mode; I may need to enforce this later.) This fixes the problems I was having with the ADMtek adm8211 driver: we were setting the channel to a non-standard default, which would cause it to fail to associate in BSS mode. - Use hal_raise_irql() to raise our IRQL to DISPATCH_LEVEL when calling certain miniport routines, per the Microsoft documentation. I think that's everything. Hopefully, other than fixing the ADMtek driver, there should be no apparent change in behavior.
2004-04-14 07:48:03 +00:00
#ifdef notdef
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
KeInitializeSpinLock(&lock->nsl_spinlock);
Continue my efforts to imitate Windows as closely as possible by attempting to duplicate Windows spinlocks. Windows spinlocks differ from FreeBSD spinlocks in the way they block preemption. FreeBSD spinlocks use critical_enter(), which masks off _all_ interrupts. This prevents any other threads from being scheduled, but it also prevents ISRs from running. In Windows, preemption is achieved by raising the processor IRQL to DISPATCH_LEVEL, which prevents other threads from preempting you, but does _not_ prevent device ISRs from running. (This is essentially what Solaris calls dispatcher locks.) The Windows spinlock itself (kspin_lock) is just an integer value which is atomically set when you acquire the lock and atomically cleared when you release it. FreeBSD doesn't have IRQ levels, so we have to cheat a little by using thread priorities: normal thread priority is PASSIVE_LEVEL, lowest interrupt thread priority is DISPATCH_LEVEL, highest thread priority is DEVICE_LEVEL (PI_REALTIME) and critical_enter() is HIGH_LEVEL. In practice, only PASSIVE_LEVEL and DISPATCH_LEVEL matter to us. The immediate benefit of all this is that I no longer have to rely on a mutex pool. Now, I'm sure many people will be seized by the urge to criticize me for doing an end run around our own spinlock implementation, but it makes more sense to do it this way. Well, it does to me anyway. Overview of the changes: - Properly implement hal_lock(), hal_unlock(), hal_irql(), hal_raise_irql() and hal_lower_irql() so that they more closely resemble their Windows counterparts. The IRQL is determined by thread priority. - Make ntoskrnl_lock_dpc() and ntoskrnl_unlock_dpc() do what they do in Windows, which is to atomically set/clear the lock value. These routines are designed to be called from DISPATCH_LEVEL, and are actually half of the work involved in acquiring/releasing spinlocks. - Add FASTCALL1(), FASTCALL2() and FASTCALL3() macros/wrappers that allow us to call a _fastcall function in spite of the fact that our version of gcc doesn't support __attribute__((__fastcall__)) yet. The macros take 1, 2 or 3 arguments, respectively. We need to call hal_lock(), hal_unlock() etc... ourselves, but can't really invoke the function directly. I could have just made the underlying functions native routines and put _fastcall wrappers around them for the benefit of Windows binaries, but that would create needless bloat. - Remove ndis_mtxpool and all references to it. We don't need it anymore. - Re-implement the NdisSpinLock routines so that they use hal_lock() and friends like they do in Windows. - Use the new spinlock methods for handling lookaside lists and linked list updates in place of the mutex locks that were there before. - Remove mutex locking from ndis_isr() and ndis_intrhand() since they're already called with ndis_intrmtx held in if_ndis.c. - Put ndis_destroy_lock() code under explicit #ifdef notdef/#endif. It turns out there are some drivers which stupidly free the memory in which their spinlocks reside before calling ndis_destroy_lock() on them (touch-after-free bug). The ADMtek wireless driver is guilty of this faux pas. (Why this doesn't clobber Windows I have no idea.) - Make NdisDprAcquireSpinLock() and NdisDprReleaseSpinLock() into real functions instead of aliasing them to NdisAcaquireSpinLock() and NdisReleaseSpinLock(). The Dpr routines use KeAcquireSpinLockAtDpcLevel() level and KeReleaseSpinLockFromDpcLevel(), which acquires the lock without twiddling the IRQL. - In ndis_linksts_done(), do _not_ call ndis_80211_getstate(). Some drivers may call the status/status done callbacks as the result of setting an OID: ndis_80211_getstate() gets OIDs, which means we might cause the driver to recursively access some of its internal structures unexpectedly. The ndis_ticktask() routine will call ndis_80211_getstate() for us eventually anyway. - Fix the channel setting code a little in ndis_80211_setstate(), and initialize the channel to IEEE80211_CHAN_ANYC. (The Microsoft spec says you're not supposed to twiddle the channel in BSS mode; I may need to enforce this later.) This fixes the problems I was having with the ADMtek adm8211 driver: we were setting the channel to a non-standard default, which would cause it to fail to associate in BSS mode. - Use hal_raise_irql() to raise our IRQL to DISPATCH_LEVEL when calling certain miniport routines, per the Microsoft documentation. I think that's everything. Hopefully, other than fixing the ADMtek driver, there should be no apparent change in behavior.
2004-04-14 07:48:03 +00:00
lock->nsl_kirql = 0;
#endif
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
Continue my efforts to imitate Windows as closely as possible by attempting to duplicate Windows spinlocks. Windows spinlocks differ from FreeBSD spinlocks in the way they block preemption. FreeBSD spinlocks use critical_enter(), which masks off _all_ interrupts. This prevents any other threads from being scheduled, but it also prevents ISRs from running. In Windows, preemption is achieved by raising the processor IRQL to DISPATCH_LEVEL, which prevents other threads from preempting you, but does _not_ prevent device ISRs from running. (This is essentially what Solaris calls dispatcher locks.) The Windows spinlock itself (kspin_lock) is just an integer value which is atomically set when you acquire the lock and atomically cleared when you release it. FreeBSD doesn't have IRQ levels, so we have to cheat a little by using thread priorities: normal thread priority is PASSIVE_LEVEL, lowest interrupt thread priority is DISPATCH_LEVEL, highest thread priority is DEVICE_LEVEL (PI_REALTIME) and critical_enter() is HIGH_LEVEL. In practice, only PASSIVE_LEVEL and DISPATCH_LEVEL matter to us. The immediate benefit of all this is that I no longer have to rely on a mutex pool. Now, I'm sure many people will be seized by the urge to criticize me for doing an end run around our own spinlock implementation, but it makes more sense to do it this way. Well, it does to me anyway. Overview of the changes: - Properly implement hal_lock(), hal_unlock(), hal_irql(), hal_raise_irql() and hal_lower_irql() so that they more closely resemble their Windows counterparts. The IRQL is determined by thread priority. - Make ntoskrnl_lock_dpc() and ntoskrnl_unlock_dpc() do what they do in Windows, which is to atomically set/clear the lock value. These routines are designed to be called from DISPATCH_LEVEL, and are actually half of the work involved in acquiring/releasing spinlocks. - Add FASTCALL1(), FASTCALL2() and FASTCALL3() macros/wrappers that allow us to call a _fastcall function in spite of the fact that our version of gcc doesn't support __attribute__((__fastcall__)) yet. The macros take 1, 2 or 3 arguments, respectively. We need to call hal_lock(), hal_unlock() etc... ourselves, but can't really invoke the function directly. I could have just made the underlying functions native routines and put _fastcall wrappers around them for the benefit of Windows binaries, but that would create needless bloat. - Remove ndis_mtxpool and all references to it. We don't need it anymore. - Re-implement the NdisSpinLock routines so that they use hal_lock() and friends like they do in Windows. - Use the new spinlock methods for handling lookaside lists and linked list updates in place of the mutex locks that were there before. - Remove mutex locking from ndis_isr() and ndis_intrhand() since they're already called with ndis_intrmtx held in if_ndis.c. - Put ndis_destroy_lock() code under explicit #ifdef notdef/#endif. It turns out there are some drivers which stupidly free the memory in which their spinlocks reside before calling ndis_destroy_lock() on them (touch-after-free bug). The ADMtek wireless driver is guilty of this faux pas. (Why this doesn't clobber Windows I have no idea.) - Make NdisDprAcquireSpinLock() and NdisDprReleaseSpinLock() into real functions instead of aliasing them to NdisAcaquireSpinLock() and NdisReleaseSpinLock(). The Dpr routines use KeAcquireSpinLockAtDpcLevel() level and KeReleaseSpinLockFromDpcLevel(), which acquires the lock without twiddling the IRQL. - In ndis_linksts_done(), do _not_ call ndis_80211_getstate(). Some drivers may call the status/status done callbacks as the result of setting an OID: ndis_80211_getstate() gets OIDs, which means we might cause the driver to recursively access some of its internal structures unexpectedly. The ndis_ticktask() routine will call ndis_80211_getstate() for us eventually anyway. - Fix the channel setting code a little in ndis_80211_setstate(), and initialize the channel to IEEE80211_CHAN_ANYC. (The Microsoft spec says you're not supposed to twiddle the channel in BSS mode; I may need to enforce this later.) This fixes the problems I was having with the ADMtek adm8211 driver: we were setting the channel to a non-standard default, which would cause it to fail to associate in BSS mode. - Use hal_raise_irql() to raise our IRQL to DISPATCH_LEVEL when calling certain miniport routines, per the Microsoft documentation. I think that's everything. Hopefully, other than fixing the ADMtek driver, there should be no apparent change in behavior.
2004-04-14 07:48:03 +00:00
/*
* Acquire a spinlock from IRQL <= DISPATCH_LEVEL.
*/
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisAcquireSpinLock(lock)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_spin_lock *lock;
{
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
KeAcquireSpinLock(&lock->nsl_spinlock, &lock->nsl_kirql);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
Continue my efforts to imitate Windows as closely as possible by attempting to duplicate Windows spinlocks. Windows spinlocks differ from FreeBSD spinlocks in the way they block preemption. FreeBSD spinlocks use critical_enter(), which masks off _all_ interrupts. This prevents any other threads from being scheduled, but it also prevents ISRs from running. In Windows, preemption is achieved by raising the processor IRQL to DISPATCH_LEVEL, which prevents other threads from preempting you, but does _not_ prevent device ISRs from running. (This is essentially what Solaris calls dispatcher locks.) The Windows spinlock itself (kspin_lock) is just an integer value which is atomically set when you acquire the lock and atomically cleared when you release it. FreeBSD doesn't have IRQ levels, so we have to cheat a little by using thread priorities: normal thread priority is PASSIVE_LEVEL, lowest interrupt thread priority is DISPATCH_LEVEL, highest thread priority is DEVICE_LEVEL (PI_REALTIME) and critical_enter() is HIGH_LEVEL. In practice, only PASSIVE_LEVEL and DISPATCH_LEVEL matter to us. The immediate benefit of all this is that I no longer have to rely on a mutex pool. Now, I'm sure many people will be seized by the urge to criticize me for doing an end run around our own spinlock implementation, but it makes more sense to do it this way. Well, it does to me anyway. Overview of the changes: - Properly implement hal_lock(), hal_unlock(), hal_irql(), hal_raise_irql() and hal_lower_irql() so that they more closely resemble their Windows counterparts. The IRQL is determined by thread priority. - Make ntoskrnl_lock_dpc() and ntoskrnl_unlock_dpc() do what they do in Windows, which is to atomically set/clear the lock value. These routines are designed to be called from DISPATCH_LEVEL, and are actually half of the work involved in acquiring/releasing spinlocks. - Add FASTCALL1(), FASTCALL2() and FASTCALL3() macros/wrappers that allow us to call a _fastcall function in spite of the fact that our version of gcc doesn't support __attribute__((__fastcall__)) yet. The macros take 1, 2 or 3 arguments, respectively. We need to call hal_lock(), hal_unlock() etc... ourselves, but can't really invoke the function directly. I could have just made the underlying functions native routines and put _fastcall wrappers around them for the benefit of Windows binaries, but that would create needless bloat. - Remove ndis_mtxpool and all references to it. We don't need it anymore. - Re-implement the NdisSpinLock routines so that they use hal_lock() and friends like they do in Windows. - Use the new spinlock methods for handling lookaside lists and linked list updates in place of the mutex locks that were there before. - Remove mutex locking from ndis_isr() and ndis_intrhand() since they're already called with ndis_intrmtx held in if_ndis.c. - Put ndis_destroy_lock() code under explicit #ifdef notdef/#endif. It turns out there are some drivers which stupidly free the memory in which their spinlocks reside before calling ndis_destroy_lock() on them (touch-after-free bug). The ADMtek wireless driver is guilty of this faux pas. (Why this doesn't clobber Windows I have no idea.) - Make NdisDprAcquireSpinLock() and NdisDprReleaseSpinLock() into real functions instead of aliasing them to NdisAcaquireSpinLock() and NdisReleaseSpinLock(). The Dpr routines use KeAcquireSpinLockAtDpcLevel() level and KeReleaseSpinLockFromDpcLevel(), which acquires the lock without twiddling the IRQL. - In ndis_linksts_done(), do _not_ call ndis_80211_getstate(). Some drivers may call the status/status done callbacks as the result of setting an OID: ndis_80211_getstate() gets OIDs, which means we might cause the driver to recursively access some of its internal structures unexpectedly. The ndis_ticktask() routine will call ndis_80211_getstate() for us eventually anyway. - Fix the channel setting code a little in ndis_80211_setstate(), and initialize the channel to IEEE80211_CHAN_ANYC. (The Microsoft spec says you're not supposed to twiddle the channel in BSS mode; I may need to enforce this later.) This fixes the problems I was having with the ADMtek adm8211 driver: we were setting the channel to a non-standard default, which would cause it to fail to associate in BSS mode. - Use hal_raise_irql() to raise our IRQL to DISPATCH_LEVEL when calling certain miniport routines, per the Microsoft documentation. I think that's everything. Hopefully, other than fixing the ADMtek driver, there should be no apparent change in behavior.
2004-04-14 07:48:03 +00:00
/*
* Release a spinlock from IRQL == DISPATCH_LEVEL.
*/
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisReleaseSpinLock(lock)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_spin_lock *lock;
{
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
KeReleaseSpinLock(&lock->nsl_spinlock, lock->nsl_kirql);
Continue my efforts to imitate Windows as closely as possible by attempting to duplicate Windows spinlocks. Windows spinlocks differ from FreeBSD spinlocks in the way they block preemption. FreeBSD spinlocks use critical_enter(), which masks off _all_ interrupts. This prevents any other threads from being scheduled, but it also prevents ISRs from running. In Windows, preemption is achieved by raising the processor IRQL to DISPATCH_LEVEL, which prevents other threads from preempting you, but does _not_ prevent device ISRs from running. (This is essentially what Solaris calls dispatcher locks.) The Windows spinlock itself (kspin_lock) is just an integer value which is atomically set when you acquire the lock and atomically cleared when you release it. FreeBSD doesn't have IRQ levels, so we have to cheat a little by using thread priorities: normal thread priority is PASSIVE_LEVEL, lowest interrupt thread priority is DISPATCH_LEVEL, highest thread priority is DEVICE_LEVEL (PI_REALTIME) and critical_enter() is HIGH_LEVEL. In practice, only PASSIVE_LEVEL and DISPATCH_LEVEL matter to us. The immediate benefit of all this is that I no longer have to rely on a mutex pool. Now, I'm sure many people will be seized by the urge to criticize me for doing an end run around our own spinlock implementation, but it makes more sense to do it this way. Well, it does to me anyway. Overview of the changes: - Properly implement hal_lock(), hal_unlock(), hal_irql(), hal_raise_irql() and hal_lower_irql() so that they more closely resemble their Windows counterparts. The IRQL is determined by thread priority. - Make ntoskrnl_lock_dpc() and ntoskrnl_unlock_dpc() do what they do in Windows, which is to atomically set/clear the lock value. These routines are designed to be called from DISPATCH_LEVEL, and are actually half of the work involved in acquiring/releasing spinlocks. - Add FASTCALL1(), FASTCALL2() and FASTCALL3() macros/wrappers that allow us to call a _fastcall function in spite of the fact that our version of gcc doesn't support __attribute__((__fastcall__)) yet. The macros take 1, 2 or 3 arguments, respectively. We need to call hal_lock(), hal_unlock() etc... ourselves, but can't really invoke the function directly. I could have just made the underlying functions native routines and put _fastcall wrappers around them for the benefit of Windows binaries, but that would create needless bloat. - Remove ndis_mtxpool and all references to it. We don't need it anymore. - Re-implement the NdisSpinLock routines so that they use hal_lock() and friends like they do in Windows. - Use the new spinlock methods for handling lookaside lists and linked list updates in place of the mutex locks that were there before. - Remove mutex locking from ndis_isr() and ndis_intrhand() since they're already called with ndis_intrmtx held in if_ndis.c. - Put ndis_destroy_lock() code under explicit #ifdef notdef/#endif. It turns out there are some drivers which stupidly free the memory in which their spinlocks reside before calling ndis_destroy_lock() on them (touch-after-free bug). The ADMtek wireless driver is guilty of this faux pas. (Why this doesn't clobber Windows I have no idea.) - Make NdisDprAcquireSpinLock() and NdisDprReleaseSpinLock() into real functions instead of aliasing them to NdisAcaquireSpinLock() and NdisReleaseSpinLock(). The Dpr routines use KeAcquireSpinLockAtDpcLevel() level and KeReleaseSpinLockFromDpcLevel(), which acquires the lock without twiddling the IRQL. - In ndis_linksts_done(), do _not_ call ndis_80211_getstate(). Some drivers may call the status/status done callbacks as the result of setting an OID: ndis_80211_getstate() gets OIDs, which means we might cause the driver to recursively access some of its internal structures unexpectedly. The ndis_ticktask() routine will call ndis_80211_getstate() for us eventually anyway. - Fix the channel setting code a little in ndis_80211_setstate(), and initialize the channel to IEEE80211_CHAN_ANYC. (The Microsoft spec says you're not supposed to twiddle the channel in BSS mode; I may need to enforce this later.) This fixes the problems I was having with the ADMtek adm8211 driver: we were setting the channel to a non-standard default, which would cause it to fail to associate in BSS mode. - Use hal_raise_irql() to raise our IRQL to DISPATCH_LEVEL when calling certain miniport routines, per the Microsoft documentation. I think that's everything. Hopefully, other than fixing the ADMtek driver, there should be no apparent change in behavior.
2004-04-14 07:48:03 +00:00
return;
}
/*
* Acquire a spinlock when already running at IRQL == DISPATCH_LEVEL.
*/
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisDprAcquireSpinLock(lock)
Continue my efforts to imitate Windows as closely as possible by attempting to duplicate Windows spinlocks. Windows spinlocks differ from FreeBSD spinlocks in the way they block preemption. FreeBSD spinlocks use critical_enter(), which masks off _all_ interrupts. This prevents any other threads from being scheduled, but it also prevents ISRs from running. In Windows, preemption is achieved by raising the processor IRQL to DISPATCH_LEVEL, which prevents other threads from preempting you, but does _not_ prevent device ISRs from running. (This is essentially what Solaris calls dispatcher locks.) The Windows spinlock itself (kspin_lock) is just an integer value which is atomically set when you acquire the lock and atomically cleared when you release it. FreeBSD doesn't have IRQ levels, so we have to cheat a little by using thread priorities: normal thread priority is PASSIVE_LEVEL, lowest interrupt thread priority is DISPATCH_LEVEL, highest thread priority is DEVICE_LEVEL (PI_REALTIME) and critical_enter() is HIGH_LEVEL. In practice, only PASSIVE_LEVEL and DISPATCH_LEVEL matter to us. The immediate benefit of all this is that I no longer have to rely on a mutex pool. Now, I'm sure many people will be seized by the urge to criticize me for doing an end run around our own spinlock implementation, but it makes more sense to do it this way. Well, it does to me anyway. Overview of the changes: - Properly implement hal_lock(), hal_unlock(), hal_irql(), hal_raise_irql() and hal_lower_irql() so that they more closely resemble their Windows counterparts. The IRQL is determined by thread priority. - Make ntoskrnl_lock_dpc() and ntoskrnl_unlock_dpc() do what they do in Windows, which is to atomically set/clear the lock value. These routines are designed to be called from DISPATCH_LEVEL, and are actually half of the work involved in acquiring/releasing spinlocks. - Add FASTCALL1(), FASTCALL2() and FASTCALL3() macros/wrappers that allow us to call a _fastcall function in spite of the fact that our version of gcc doesn't support __attribute__((__fastcall__)) yet. The macros take 1, 2 or 3 arguments, respectively. We need to call hal_lock(), hal_unlock() etc... ourselves, but can't really invoke the function directly. I could have just made the underlying functions native routines and put _fastcall wrappers around them for the benefit of Windows binaries, but that would create needless bloat. - Remove ndis_mtxpool and all references to it. We don't need it anymore. - Re-implement the NdisSpinLock routines so that they use hal_lock() and friends like they do in Windows. - Use the new spinlock methods for handling lookaside lists and linked list updates in place of the mutex locks that were there before. - Remove mutex locking from ndis_isr() and ndis_intrhand() since they're already called with ndis_intrmtx held in if_ndis.c. - Put ndis_destroy_lock() code under explicit #ifdef notdef/#endif. It turns out there are some drivers which stupidly free the memory in which their spinlocks reside before calling ndis_destroy_lock() on them (touch-after-free bug). The ADMtek wireless driver is guilty of this faux pas. (Why this doesn't clobber Windows I have no idea.) - Make NdisDprAcquireSpinLock() and NdisDprReleaseSpinLock() into real functions instead of aliasing them to NdisAcaquireSpinLock() and NdisReleaseSpinLock(). The Dpr routines use KeAcquireSpinLockAtDpcLevel() level and KeReleaseSpinLockFromDpcLevel(), which acquires the lock without twiddling the IRQL. - In ndis_linksts_done(), do _not_ call ndis_80211_getstate(). Some drivers may call the status/status done callbacks as the result of setting an OID: ndis_80211_getstate() gets OIDs, which means we might cause the driver to recursively access some of its internal structures unexpectedly. The ndis_ticktask() routine will call ndis_80211_getstate() for us eventually anyway. - Fix the channel setting code a little in ndis_80211_setstate(), and initialize the channel to IEEE80211_CHAN_ANYC. (The Microsoft spec says you're not supposed to twiddle the channel in BSS mode; I may need to enforce this later.) This fixes the problems I was having with the ADMtek adm8211 driver: we were setting the channel to a non-standard default, which would cause it to fail to associate in BSS mode. - Use hal_raise_irql() to raise our IRQL to DISPATCH_LEVEL when calling certain miniport routines, per the Microsoft documentation. I think that's everything. Hopefully, other than fixing the ADMtek driver, there should be no apparent change in behavior.
2004-04-14 07:48:03 +00:00
ndis_spin_lock *lock;
{
Finally bring an end to the great "make the Atheros NDIS driver work on SMP" saga. After several weeks and much gnashing of teeth, I have finally tracked down all the problems, despite their best efforts to confound and annoy me. Problem nunmber one: the Atheros windows driver is _NOT_ a de-serialized miniport! It used to be that NDIS drivers relied on the NDIS library itself for all their locking and serialization needs. Transmit packet queues were all handled internally by NDIS, and all calls to MiniportXXX() routines were guaranteed to be appropriately serialized. This proved to be a performance problem however, and Microsoft introduced de-serialized miniports with the NDIS 5.x spec. Microsoft still supports serialized miniports, but recommends that all new drivers written for Windows XP and later be deserialized. Apparently Atheros wasn't listening when they said this. This means (among other things) that we have to serialize calls to MiniportSendPackets(). We also have to serialize calls to MiniportTimer() that are triggered via the NdisMInitializeTimer() routine. It finally dawned on me why NdisMInitializeTimer() takes a special NDIS_MINIPORT_TIMER structure and a pointer to the miniport block: the timer callback must be serialized, and it's only by saving the miniport block handle that we can get access to the serialization lock during the timer callback. Problem number two: haunted hardware. The thing that was _really_ driving me absolutely bonkers for the longest time is that, for some reason I couldn't understand, my test machine would occasionally freeze or more frustratingly, reset completely. That's reset and in *pow!* back to the BIOS startup. No panic, no crashdump, just a reset. This appeared to happen most often when MiniportReset() was called. (As to why MiniportReset() was being called, see problem three below.) I thought maybe I had created some sort of horrible deadlock condition in the process of adding the serialization, but after three weeks, at least 6 different locking implementations and heroic efforts to debug the spinlock code, the machine still kept resetting. Finally, I started single stepping through the MiniportReset() routine in the driver using the kernel debugger, and this ultimately led me to the source of the problem. One of the last things the Atheros MiniportReset() routine does is call NdisReadPciSlotInformation() several times to inspect a portion of the device's PCI config space. It reads the same chunk of config space repeatedly, in rapid succession. Presumeably, it's polling the hardware for some sort of event. The reset occurs partway through this process. I discovered that when I single-stepped through this portion of the routine, the reset didn't occur. So I inserted a 1 microsecond delay into the read loop in NdisReadPciSlotInformation(). Suddenly, the reset was gone!! I'm still very puzzled by the whole thing. What I suspect is happening is that reading the PCI config space so quickly is causing a severe PCI bus error. My test system is a Sun w2100z dual Opteron system, and the NIC is a miniPCI card mounted in a miniPCI-to-PCI carrier card, plugged into a 100Mhz PCI slot. It's possible that this combination of hardware causes a bus protocol violation in this scenario which leads to a fatal machine check. This is pure speculation though. Really all I know for sure is that inserting the delay makes the problem go away. (To quote Homer Simpson: "I don't know how it works, but fire makes it good!") Problem number three: NdisAllocatePacket() needs to make sure to initialize the npp_validcounts field in the 'private' section of the NDIS_PACKET structure. The reason if_ndis was calling the MiniportReset() routine in the first place is that packet transmits were sometimes hanging. When sending a packet, an NDIS driver will call NdisQueryPacket() to learn how many physical buffers the packet resides in. NdisQueryPacket() is actually a macro, which traverses the NDIS_BUFFER list attached to the NDIS_PACKET and stashes some of the results in the 'private' section of the NDIS_PACKET. It also sets the npp_validcounts field to TRUE To indicate that the results are now valid. The problem is, now that if_ndis creates a pool of transmit packets via NdisAllocatePacketPool(), it's important that each time a new packet is allocated via NdisAllocatePacket() that validcounts be initialized to FALSE. If it isn't, and a previously transmitted NDIS_PACKET is pulled out of the pool, it may contain stale data from a previous transmission which won't get updated by NdisQueryPacket(). This would cause the driver to miscompute the number of fragments for a given packet, and botch the transmission. Fixing these three problems seems to make the Atheros driver happy on SMP, which hopefully means other serialized miniports will be happy too. And there was much rejoicing. Other stuff fixed along the way: - Modified ndis_thsuspend() to take a mutex as an argument. This allows KeWaitForSingleObject() and KeWaitForMultipleObjects() to avoid any possible race conditions with other routines that use the dispatcher lock. - Fixed KeCancelTimer() so that it returns the correct value for 'pending' according to the Microsoft documentation - Modfied NdisGetSystemUpTime() to use ticks and hz rather than calling nanouptime(). Also added comment that this routine wraps after 49.7 days. - Added macros for KeAcquireSpinLock()/KeReleaseSpinLock() to hide all the MSCALL() goop. - For x86, KeAcquireSpinLockRaiseToDpc() needs to be a separate function. This is because it's supposed to be _stdcall on the x86 arch, whereas KeAcquireSpinLock() is supposed to be _fastcall. On amd64, all routines use the same calling convention so we can just map KeAcquireSpinLockRaiseToDpc() directly to KfAcquireSpinLock() and it will work. (The _fastcall attribute is a no-op on amd64.) - Implement and use IoInitializeDpcRequest() and IoRequestDpc() (they're just macros) and use them for interrupt handling. This allows us to move the ndis_intrtask() routine from if_ndis.c to kern_ndis.c. - Fix the MmInitializeMdl() macro so that is uses sizeof(vm_offset_t) when computing mdl_size instead of uint32_t, so that it matches the MmSizeOfMdl() routine. - Change a could of M_WAITOKs to M_NOWAITs in the unicode routines in subr_ndis.c. - Use the dispatcher lock a little more consistently in subr_ntoskrnl.c. - Get rid of the "wait for link event" hack in ndis_init(). Now that I fixed NdisReadPciSlotInformation(), it seems I don't need it anymore. This should fix the witness panic a couple of people have reported. - Use MSCALL1() when calling the MiniportHangCheck() function in ndis_ticktask(). I accidentally missed this one when adding the wrapping for amd64.
2005-03-27 10:14:36 +00:00
KeAcquireSpinLockAtDpcLevel(&lock->nsl_spinlock);
Continue my efforts to imitate Windows as closely as possible by attempting to duplicate Windows spinlocks. Windows spinlocks differ from FreeBSD spinlocks in the way they block preemption. FreeBSD spinlocks use critical_enter(), which masks off _all_ interrupts. This prevents any other threads from being scheduled, but it also prevents ISRs from running. In Windows, preemption is achieved by raising the processor IRQL to DISPATCH_LEVEL, which prevents other threads from preempting you, but does _not_ prevent device ISRs from running. (This is essentially what Solaris calls dispatcher locks.) The Windows spinlock itself (kspin_lock) is just an integer value which is atomically set when you acquire the lock and atomically cleared when you release it. FreeBSD doesn't have IRQ levels, so we have to cheat a little by using thread priorities: normal thread priority is PASSIVE_LEVEL, lowest interrupt thread priority is DISPATCH_LEVEL, highest thread priority is DEVICE_LEVEL (PI_REALTIME) and critical_enter() is HIGH_LEVEL. In practice, only PASSIVE_LEVEL and DISPATCH_LEVEL matter to us. The immediate benefit of all this is that I no longer have to rely on a mutex pool. Now, I'm sure many people will be seized by the urge to criticize me for doing an end run around our own spinlock implementation, but it makes more sense to do it this way. Well, it does to me anyway. Overview of the changes: - Properly implement hal_lock(), hal_unlock(), hal_irql(), hal_raise_irql() and hal_lower_irql() so that they more closely resemble their Windows counterparts. The IRQL is determined by thread priority. - Make ntoskrnl_lock_dpc() and ntoskrnl_unlock_dpc() do what they do in Windows, which is to atomically set/clear the lock value. These routines are designed to be called from DISPATCH_LEVEL, and are actually half of the work involved in acquiring/releasing spinlocks. - Add FASTCALL1(), FASTCALL2() and FASTCALL3() macros/wrappers that allow us to call a _fastcall function in spite of the fact that our version of gcc doesn't support __attribute__((__fastcall__)) yet. The macros take 1, 2 or 3 arguments, respectively. We need to call hal_lock(), hal_unlock() etc... ourselves, but can't really invoke the function directly. I could have just made the underlying functions native routines and put _fastcall wrappers around them for the benefit of Windows binaries, but that would create needless bloat. - Remove ndis_mtxpool and all references to it. We don't need it anymore. - Re-implement the NdisSpinLock routines so that they use hal_lock() and friends like they do in Windows. - Use the new spinlock methods for handling lookaside lists and linked list updates in place of the mutex locks that were there before. - Remove mutex locking from ndis_isr() and ndis_intrhand() since they're already called with ndis_intrmtx held in if_ndis.c. - Put ndis_destroy_lock() code under explicit #ifdef notdef/#endif. It turns out there are some drivers which stupidly free the memory in which their spinlocks reside before calling ndis_destroy_lock() on them (touch-after-free bug). The ADMtek wireless driver is guilty of this faux pas. (Why this doesn't clobber Windows I have no idea.) - Make NdisDprAcquireSpinLock() and NdisDprReleaseSpinLock() into real functions instead of aliasing them to NdisAcaquireSpinLock() and NdisReleaseSpinLock(). The Dpr routines use KeAcquireSpinLockAtDpcLevel() level and KeReleaseSpinLockFromDpcLevel(), which acquires the lock without twiddling the IRQL. - In ndis_linksts_done(), do _not_ call ndis_80211_getstate(). Some drivers may call the status/status done callbacks as the result of setting an OID: ndis_80211_getstate() gets OIDs, which means we might cause the driver to recursively access some of its internal structures unexpectedly. The ndis_ticktask() routine will call ndis_80211_getstate() for us eventually anyway. - Fix the channel setting code a little in ndis_80211_setstate(), and initialize the channel to IEEE80211_CHAN_ANYC. (The Microsoft spec says you're not supposed to twiddle the channel in BSS mode; I may need to enforce this later.) This fixes the problems I was having with the ADMtek adm8211 driver: we were setting the channel to a non-standard default, which would cause it to fail to associate in BSS mode. - Use hal_raise_irql() to raise our IRQL to DISPATCH_LEVEL when calling certain miniport routines, per the Microsoft documentation. I think that's everything. Hopefully, other than fixing the ADMtek driver, there should be no apparent change in behavior.
2004-04-14 07:48:03 +00:00
return;
}
/*
* Release a spinlock without leaving IRQL == DISPATCH_LEVEL.
*/
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisDprReleaseSpinLock(lock)
Continue my efforts to imitate Windows as closely as possible by attempting to duplicate Windows spinlocks. Windows spinlocks differ from FreeBSD spinlocks in the way they block preemption. FreeBSD spinlocks use critical_enter(), which masks off _all_ interrupts. This prevents any other threads from being scheduled, but it also prevents ISRs from running. In Windows, preemption is achieved by raising the processor IRQL to DISPATCH_LEVEL, which prevents other threads from preempting you, but does _not_ prevent device ISRs from running. (This is essentially what Solaris calls dispatcher locks.) The Windows spinlock itself (kspin_lock) is just an integer value which is atomically set when you acquire the lock and atomically cleared when you release it. FreeBSD doesn't have IRQ levels, so we have to cheat a little by using thread priorities: normal thread priority is PASSIVE_LEVEL, lowest interrupt thread priority is DISPATCH_LEVEL, highest thread priority is DEVICE_LEVEL (PI_REALTIME) and critical_enter() is HIGH_LEVEL. In practice, only PASSIVE_LEVEL and DISPATCH_LEVEL matter to us. The immediate benefit of all this is that I no longer have to rely on a mutex pool. Now, I'm sure many people will be seized by the urge to criticize me for doing an end run around our own spinlock implementation, but it makes more sense to do it this way. Well, it does to me anyway. Overview of the changes: - Properly implement hal_lock(), hal_unlock(), hal_irql(), hal_raise_irql() and hal_lower_irql() so that they more closely resemble their Windows counterparts. The IRQL is determined by thread priority. - Make ntoskrnl_lock_dpc() and ntoskrnl_unlock_dpc() do what they do in Windows, which is to atomically set/clear the lock value. These routines are designed to be called from DISPATCH_LEVEL, and are actually half of the work involved in acquiring/releasing spinlocks. - Add FASTCALL1(), FASTCALL2() and FASTCALL3() macros/wrappers that allow us to call a _fastcall function in spite of the fact that our version of gcc doesn't support __attribute__((__fastcall__)) yet. The macros take 1, 2 or 3 arguments, respectively. We need to call hal_lock(), hal_unlock() etc... ourselves, but can't really invoke the function directly. I could have just made the underlying functions native routines and put _fastcall wrappers around them for the benefit of Windows binaries, but that would create needless bloat. - Remove ndis_mtxpool and all references to it. We don't need it anymore. - Re-implement the NdisSpinLock routines so that they use hal_lock() and friends like they do in Windows. - Use the new spinlock methods for handling lookaside lists and linked list updates in place of the mutex locks that were there before. - Remove mutex locking from ndis_isr() and ndis_intrhand() since they're already called with ndis_intrmtx held in if_ndis.c. - Put ndis_destroy_lock() code under explicit #ifdef notdef/#endif. It turns out there are some drivers which stupidly free the memory in which their spinlocks reside before calling ndis_destroy_lock() on them (touch-after-free bug). The ADMtek wireless driver is guilty of this faux pas. (Why this doesn't clobber Windows I have no idea.) - Make NdisDprAcquireSpinLock() and NdisDprReleaseSpinLock() into real functions instead of aliasing them to NdisAcaquireSpinLock() and NdisReleaseSpinLock(). The Dpr routines use KeAcquireSpinLockAtDpcLevel() level and KeReleaseSpinLockFromDpcLevel(), which acquires the lock without twiddling the IRQL. - In ndis_linksts_done(), do _not_ call ndis_80211_getstate(). Some drivers may call the status/status done callbacks as the result of setting an OID: ndis_80211_getstate() gets OIDs, which means we might cause the driver to recursively access some of its internal structures unexpectedly. The ndis_ticktask() routine will call ndis_80211_getstate() for us eventually anyway. - Fix the channel setting code a little in ndis_80211_setstate(), and initialize the channel to IEEE80211_CHAN_ANYC. (The Microsoft spec says you're not supposed to twiddle the channel in BSS mode; I may need to enforce this later.) This fixes the problems I was having with the ADMtek adm8211 driver: we were setting the channel to a non-standard default, which would cause it to fail to associate in BSS mode. - Use hal_raise_irql() to raise our IRQL to DISPATCH_LEVEL when calling certain miniport routines, per the Microsoft documentation. I think that's everything. Hopefully, other than fixing the ADMtek driver, there should be no apparent change in behavior.
2004-04-14 07:48:03 +00:00
ndis_spin_lock *lock;
{
Finally bring an end to the great "make the Atheros NDIS driver work on SMP" saga. After several weeks and much gnashing of teeth, I have finally tracked down all the problems, despite their best efforts to confound and annoy me. Problem nunmber one: the Atheros windows driver is _NOT_ a de-serialized miniport! It used to be that NDIS drivers relied on the NDIS library itself for all their locking and serialization needs. Transmit packet queues were all handled internally by NDIS, and all calls to MiniportXXX() routines were guaranteed to be appropriately serialized. This proved to be a performance problem however, and Microsoft introduced de-serialized miniports with the NDIS 5.x spec. Microsoft still supports serialized miniports, but recommends that all new drivers written for Windows XP and later be deserialized. Apparently Atheros wasn't listening when they said this. This means (among other things) that we have to serialize calls to MiniportSendPackets(). We also have to serialize calls to MiniportTimer() that are triggered via the NdisMInitializeTimer() routine. It finally dawned on me why NdisMInitializeTimer() takes a special NDIS_MINIPORT_TIMER structure and a pointer to the miniport block: the timer callback must be serialized, and it's only by saving the miniport block handle that we can get access to the serialization lock during the timer callback. Problem number two: haunted hardware. The thing that was _really_ driving me absolutely bonkers for the longest time is that, for some reason I couldn't understand, my test machine would occasionally freeze or more frustratingly, reset completely. That's reset and in *pow!* back to the BIOS startup. No panic, no crashdump, just a reset. This appeared to happen most often when MiniportReset() was called. (As to why MiniportReset() was being called, see problem three below.) I thought maybe I had created some sort of horrible deadlock condition in the process of adding the serialization, but after three weeks, at least 6 different locking implementations and heroic efforts to debug the spinlock code, the machine still kept resetting. Finally, I started single stepping through the MiniportReset() routine in the driver using the kernel debugger, and this ultimately led me to the source of the problem. One of the last things the Atheros MiniportReset() routine does is call NdisReadPciSlotInformation() several times to inspect a portion of the device's PCI config space. It reads the same chunk of config space repeatedly, in rapid succession. Presumeably, it's polling the hardware for some sort of event. The reset occurs partway through this process. I discovered that when I single-stepped through this portion of the routine, the reset didn't occur. So I inserted a 1 microsecond delay into the read loop in NdisReadPciSlotInformation(). Suddenly, the reset was gone!! I'm still very puzzled by the whole thing. What I suspect is happening is that reading the PCI config space so quickly is causing a severe PCI bus error. My test system is a Sun w2100z dual Opteron system, and the NIC is a miniPCI card mounted in a miniPCI-to-PCI carrier card, plugged into a 100Mhz PCI slot. It's possible that this combination of hardware causes a bus protocol violation in this scenario which leads to a fatal machine check. This is pure speculation though. Really all I know for sure is that inserting the delay makes the problem go away. (To quote Homer Simpson: "I don't know how it works, but fire makes it good!") Problem number three: NdisAllocatePacket() needs to make sure to initialize the npp_validcounts field in the 'private' section of the NDIS_PACKET structure. The reason if_ndis was calling the MiniportReset() routine in the first place is that packet transmits were sometimes hanging. When sending a packet, an NDIS driver will call NdisQueryPacket() to learn how many physical buffers the packet resides in. NdisQueryPacket() is actually a macro, which traverses the NDIS_BUFFER list attached to the NDIS_PACKET and stashes some of the results in the 'private' section of the NDIS_PACKET. It also sets the npp_validcounts field to TRUE To indicate that the results are now valid. The problem is, now that if_ndis creates a pool of transmit packets via NdisAllocatePacketPool(), it's important that each time a new packet is allocated via NdisAllocatePacket() that validcounts be initialized to FALSE. If it isn't, and a previously transmitted NDIS_PACKET is pulled out of the pool, it may contain stale data from a previous transmission which won't get updated by NdisQueryPacket(). This would cause the driver to miscompute the number of fragments for a given packet, and botch the transmission. Fixing these three problems seems to make the Atheros driver happy on SMP, which hopefully means other serialized miniports will be happy too. And there was much rejoicing. Other stuff fixed along the way: - Modified ndis_thsuspend() to take a mutex as an argument. This allows KeWaitForSingleObject() and KeWaitForMultipleObjects() to avoid any possible race conditions with other routines that use the dispatcher lock. - Fixed KeCancelTimer() so that it returns the correct value for 'pending' according to the Microsoft documentation - Modfied NdisGetSystemUpTime() to use ticks and hz rather than calling nanouptime(). Also added comment that this routine wraps after 49.7 days. - Added macros for KeAcquireSpinLock()/KeReleaseSpinLock() to hide all the MSCALL() goop. - For x86, KeAcquireSpinLockRaiseToDpc() needs to be a separate function. This is because it's supposed to be _stdcall on the x86 arch, whereas KeAcquireSpinLock() is supposed to be _fastcall. On amd64, all routines use the same calling convention so we can just map KeAcquireSpinLockRaiseToDpc() directly to KfAcquireSpinLock() and it will work. (The _fastcall attribute is a no-op on amd64.) - Implement and use IoInitializeDpcRequest() and IoRequestDpc() (they're just macros) and use them for interrupt handling. This allows us to move the ndis_intrtask() routine from if_ndis.c to kern_ndis.c. - Fix the MmInitializeMdl() macro so that is uses sizeof(vm_offset_t) when computing mdl_size instead of uint32_t, so that it matches the MmSizeOfMdl() routine. - Change a could of M_WAITOKs to M_NOWAITs in the unicode routines in subr_ndis.c. - Use the dispatcher lock a little more consistently in subr_ntoskrnl.c. - Get rid of the "wait for link event" hack in ndis_init(). Now that I fixed NdisReadPciSlotInformation(), it seems I don't need it anymore. This should fix the witness panic a couple of people have reported. - Use MSCALL1() when calling the MiniportHangCheck() function in ndis_ticktask(). I accidentally missed this one when adding the wrapping for amd64.
2005-03-27 10:14:36 +00:00
KeReleaseSpinLockFromDpcLevel(&lock->nsl_spinlock);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
static void
NdisInitializeReadWriteLock(lock)
ndis_rw_lock *lock;
{
KeInitializeSpinLock(&lock->nrl_spinlock);
bzero((char *)&lock->nrl_rsvd, sizeof(lock->nrl_rsvd));
return;
}
static void
NdisAcquireReadWriteLock(lock, writeacc, state)
ndis_rw_lock *lock;
uint8_t writeacc;
ndis_lock_state *state;
{
if (writeacc == TRUE) {
KeAcquireSpinLock(&lock->nrl_spinlock, &state->nls_oldirql);
lock->nrl_rsvd[0]++;
} else
lock->nrl_rsvd[1]++;
return;
}
static void
NdisReleaseReadWriteLock(lock, state)
ndis_rw_lock *lock;
ndis_lock_state *state;
{
if (lock->nrl_rsvd[0]) {
lock->nrl_rsvd[0]--;
KeReleaseSpinLock(&lock->nrl_spinlock, state->nls_oldirql);
} else
lock->nrl_rsvd[1]--;
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static uint32_t
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisReadPciSlotInformation(adapter, slot, offset, buf, len)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_handle adapter;
uint32_t slot;
uint32_t offset;
void *buf;
uint32_t len;
{
ndis_miniport_block *block;
int i;
char *dest;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
device_t dev;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
block = (ndis_miniport_block *)adapter;
dest = buf;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
if (block == NULL)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return(0);
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
dev = block->nmb_physdeviceobj->do_devext;
Finally bring an end to the great "make the Atheros NDIS driver work on SMP" saga. After several weeks and much gnashing of teeth, I have finally tracked down all the problems, despite their best efforts to confound and annoy me. Problem nunmber one: the Atheros windows driver is _NOT_ a de-serialized miniport! It used to be that NDIS drivers relied on the NDIS library itself for all their locking and serialization needs. Transmit packet queues were all handled internally by NDIS, and all calls to MiniportXXX() routines were guaranteed to be appropriately serialized. This proved to be a performance problem however, and Microsoft introduced de-serialized miniports with the NDIS 5.x spec. Microsoft still supports serialized miniports, but recommends that all new drivers written for Windows XP and later be deserialized. Apparently Atheros wasn't listening when they said this. This means (among other things) that we have to serialize calls to MiniportSendPackets(). We also have to serialize calls to MiniportTimer() that are triggered via the NdisMInitializeTimer() routine. It finally dawned on me why NdisMInitializeTimer() takes a special NDIS_MINIPORT_TIMER structure and a pointer to the miniport block: the timer callback must be serialized, and it's only by saving the miniport block handle that we can get access to the serialization lock during the timer callback. Problem number two: haunted hardware. The thing that was _really_ driving me absolutely bonkers for the longest time is that, for some reason I couldn't understand, my test machine would occasionally freeze or more frustratingly, reset completely. That's reset and in *pow!* back to the BIOS startup. No panic, no crashdump, just a reset. This appeared to happen most often when MiniportReset() was called. (As to why MiniportReset() was being called, see problem three below.) I thought maybe I had created some sort of horrible deadlock condition in the process of adding the serialization, but after three weeks, at least 6 different locking implementations and heroic efforts to debug the spinlock code, the machine still kept resetting. Finally, I started single stepping through the MiniportReset() routine in the driver using the kernel debugger, and this ultimately led me to the source of the problem. One of the last things the Atheros MiniportReset() routine does is call NdisReadPciSlotInformation() several times to inspect a portion of the device's PCI config space. It reads the same chunk of config space repeatedly, in rapid succession. Presumeably, it's polling the hardware for some sort of event. The reset occurs partway through this process. I discovered that when I single-stepped through this portion of the routine, the reset didn't occur. So I inserted a 1 microsecond delay into the read loop in NdisReadPciSlotInformation(). Suddenly, the reset was gone!! I'm still very puzzled by the whole thing. What I suspect is happening is that reading the PCI config space so quickly is causing a severe PCI bus error. My test system is a Sun w2100z dual Opteron system, and the NIC is a miniPCI card mounted in a miniPCI-to-PCI carrier card, plugged into a 100Mhz PCI slot. It's possible that this combination of hardware causes a bus protocol violation in this scenario which leads to a fatal machine check. This is pure speculation though. Really all I know for sure is that inserting the delay makes the problem go away. (To quote Homer Simpson: "I don't know how it works, but fire makes it good!") Problem number three: NdisAllocatePacket() needs to make sure to initialize the npp_validcounts field in the 'private' section of the NDIS_PACKET structure. The reason if_ndis was calling the MiniportReset() routine in the first place is that packet transmits were sometimes hanging. When sending a packet, an NDIS driver will call NdisQueryPacket() to learn how many physical buffers the packet resides in. NdisQueryPacket() is actually a macro, which traverses the NDIS_BUFFER list attached to the NDIS_PACKET and stashes some of the results in the 'private' section of the NDIS_PACKET. It also sets the npp_validcounts field to TRUE To indicate that the results are now valid. The problem is, now that if_ndis creates a pool of transmit packets via NdisAllocatePacketPool(), it's important that each time a new packet is allocated via NdisAllocatePacket() that validcounts be initialized to FALSE. If it isn't, and a previously transmitted NDIS_PACKET is pulled out of the pool, it may contain stale data from a previous transmission which won't get updated by NdisQueryPacket(). This would cause the driver to miscompute the number of fragments for a given packet, and botch the transmission. Fixing these three problems seems to make the Atheros driver happy on SMP, which hopefully means other serialized miniports will be happy too. And there was much rejoicing. Other stuff fixed along the way: - Modified ndis_thsuspend() to take a mutex as an argument. This allows KeWaitForSingleObject() and KeWaitForMultipleObjects() to avoid any possible race conditions with other routines that use the dispatcher lock. - Fixed KeCancelTimer() so that it returns the correct value for 'pending' according to the Microsoft documentation - Modfied NdisGetSystemUpTime() to use ticks and hz rather than calling nanouptime(). Also added comment that this routine wraps after 49.7 days. - Added macros for KeAcquireSpinLock()/KeReleaseSpinLock() to hide all the MSCALL() goop. - For x86, KeAcquireSpinLockRaiseToDpc() needs to be a separate function. This is because it's supposed to be _stdcall on the x86 arch, whereas KeAcquireSpinLock() is supposed to be _fastcall. On amd64, all routines use the same calling convention so we can just map KeAcquireSpinLockRaiseToDpc() directly to KfAcquireSpinLock() and it will work. (The _fastcall attribute is a no-op on amd64.) - Implement and use IoInitializeDpcRequest() and IoRequestDpc() (they're just macros) and use them for interrupt handling. This allows us to move the ndis_intrtask() routine from if_ndis.c to kern_ndis.c. - Fix the MmInitializeMdl() macro so that is uses sizeof(vm_offset_t) when computing mdl_size instead of uint32_t, so that it matches the MmSizeOfMdl() routine. - Change a could of M_WAITOKs to M_NOWAITs in the unicode routines in subr_ndis.c. - Use the dispatcher lock a little more consistently in subr_ntoskrnl.c. - Get rid of the "wait for link event" hack in ndis_init(). Now that I fixed NdisReadPciSlotInformation(), it seems I don't need it anymore. This should fix the witness panic a couple of people have reported. - Use MSCALL1() when calling the MiniportHangCheck() function in ndis_ticktask(). I accidentally missed this one when adding the wrapping for amd64.
2005-03-27 10:14:36 +00:00
/*
* I have a test system consisting of a Sun w2100z
* dual 2.4Ghz Opteron machine and an Atheros 802.11a/b/g
* "Aries" miniPCI NIC. (The NIC is installed in the
* machine using a miniPCI to PCI bus adapter card.)
* When running in SMP mode, I found that
* performing a large number of consecutive calls to
* NdisReadPciSlotInformation() would result in a
* sudden system reset (or in some cases a freeze).
* My suspicion is that the multiple reads are somehow
* triggering a fatal PCI bus error that leads to a
* machine check. The 1us delay in the loop below
* seems to prevent this problem.
*/
for (i = 0; i < len; i++) {
DELAY(1);
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
dest[i] = pci_read_config(dev, i + offset, 1);
Finally bring an end to the great "make the Atheros NDIS driver work on SMP" saga. After several weeks and much gnashing of teeth, I have finally tracked down all the problems, despite their best efforts to confound and annoy me. Problem nunmber one: the Atheros windows driver is _NOT_ a de-serialized miniport! It used to be that NDIS drivers relied on the NDIS library itself for all their locking and serialization needs. Transmit packet queues were all handled internally by NDIS, and all calls to MiniportXXX() routines were guaranteed to be appropriately serialized. This proved to be a performance problem however, and Microsoft introduced de-serialized miniports with the NDIS 5.x spec. Microsoft still supports serialized miniports, but recommends that all new drivers written for Windows XP and later be deserialized. Apparently Atheros wasn't listening when they said this. This means (among other things) that we have to serialize calls to MiniportSendPackets(). We also have to serialize calls to MiniportTimer() that are triggered via the NdisMInitializeTimer() routine. It finally dawned on me why NdisMInitializeTimer() takes a special NDIS_MINIPORT_TIMER structure and a pointer to the miniport block: the timer callback must be serialized, and it's only by saving the miniport block handle that we can get access to the serialization lock during the timer callback. Problem number two: haunted hardware. The thing that was _really_ driving me absolutely bonkers for the longest time is that, for some reason I couldn't understand, my test machine would occasionally freeze or more frustratingly, reset completely. That's reset and in *pow!* back to the BIOS startup. No panic, no crashdump, just a reset. This appeared to happen most often when MiniportReset() was called. (As to why MiniportReset() was being called, see problem three below.) I thought maybe I had created some sort of horrible deadlock condition in the process of adding the serialization, but after three weeks, at least 6 different locking implementations and heroic efforts to debug the spinlock code, the machine still kept resetting. Finally, I started single stepping through the MiniportReset() routine in the driver using the kernel debugger, and this ultimately led me to the source of the problem. One of the last things the Atheros MiniportReset() routine does is call NdisReadPciSlotInformation() several times to inspect a portion of the device's PCI config space. It reads the same chunk of config space repeatedly, in rapid succession. Presumeably, it's polling the hardware for some sort of event. The reset occurs partway through this process. I discovered that when I single-stepped through this portion of the routine, the reset didn't occur. So I inserted a 1 microsecond delay into the read loop in NdisReadPciSlotInformation(). Suddenly, the reset was gone!! I'm still very puzzled by the whole thing. What I suspect is happening is that reading the PCI config space so quickly is causing a severe PCI bus error. My test system is a Sun w2100z dual Opteron system, and the NIC is a miniPCI card mounted in a miniPCI-to-PCI carrier card, plugged into a 100Mhz PCI slot. It's possible that this combination of hardware causes a bus protocol violation in this scenario which leads to a fatal machine check. This is pure speculation though. Really all I know for sure is that inserting the delay makes the problem go away. (To quote Homer Simpson: "I don't know how it works, but fire makes it good!") Problem number three: NdisAllocatePacket() needs to make sure to initialize the npp_validcounts field in the 'private' section of the NDIS_PACKET structure. The reason if_ndis was calling the MiniportReset() routine in the first place is that packet transmits were sometimes hanging. When sending a packet, an NDIS driver will call NdisQueryPacket() to learn how many physical buffers the packet resides in. NdisQueryPacket() is actually a macro, which traverses the NDIS_BUFFER list attached to the NDIS_PACKET and stashes some of the results in the 'private' section of the NDIS_PACKET. It also sets the npp_validcounts field to TRUE To indicate that the results are now valid. The problem is, now that if_ndis creates a pool of transmit packets via NdisAllocatePacketPool(), it's important that each time a new packet is allocated via NdisAllocatePacket() that validcounts be initialized to FALSE. If it isn't, and a previously transmitted NDIS_PACKET is pulled out of the pool, it may contain stale data from a previous transmission which won't get updated by NdisQueryPacket(). This would cause the driver to miscompute the number of fragments for a given packet, and botch the transmission. Fixing these three problems seems to make the Atheros driver happy on SMP, which hopefully means other serialized miniports will be happy too. And there was much rejoicing. Other stuff fixed along the way: - Modified ndis_thsuspend() to take a mutex as an argument. This allows KeWaitForSingleObject() and KeWaitForMultipleObjects() to avoid any possible race conditions with other routines that use the dispatcher lock. - Fixed KeCancelTimer() so that it returns the correct value for 'pending' according to the Microsoft documentation - Modfied NdisGetSystemUpTime() to use ticks and hz rather than calling nanouptime(). Also added comment that this routine wraps after 49.7 days. - Added macros for KeAcquireSpinLock()/KeReleaseSpinLock() to hide all the MSCALL() goop. - For x86, KeAcquireSpinLockRaiseToDpc() needs to be a separate function. This is because it's supposed to be _stdcall on the x86 arch, whereas KeAcquireSpinLock() is supposed to be _fastcall. On amd64, all routines use the same calling convention so we can just map KeAcquireSpinLockRaiseToDpc() directly to KfAcquireSpinLock() and it will work. (The _fastcall attribute is a no-op on amd64.) - Implement and use IoInitializeDpcRequest() and IoRequestDpc() (they're just macros) and use them for interrupt handling. This allows us to move the ndis_intrtask() routine from if_ndis.c to kern_ndis.c. - Fix the MmInitializeMdl() macro so that is uses sizeof(vm_offset_t) when computing mdl_size instead of uint32_t, so that it matches the MmSizeOfMdl() routine. - Change a could of M_WAITOKs to M_NOWAITs in the unicode routines in subr_ndis.c. - Use the dispatcher lock a little more consistently in subr_ntoskrnl.c. - Get rid of the "wait for link event" hack in ndis_init(). Now that I fixed NdisReadPciSlotInformation(), it seems I don't need it anymore. This should fix the witness panic a couple of people have reported. - Use MSCALL1() when calling the MiniportHangCheck() function in ndis_ticktask(). I accidentally missed this one when adding the wrapping for amd64.
2005-03-27 10:14:36 +00:00
}
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return(len);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static uint32_t
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisWritePciSlotInformation(adapter, slot, offset, buf, len)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_handle adapter;
uint32_t slot;
uint32_t offset;
void *buf;
uint32_t len;
{
ndis_miniport_block *block;
int i;
char *dest;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
device_t dev;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
block = (ndis_miniport_block *)adapter;
dest = buf;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
if (block == NULL)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return(0);
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
dev = block->nmb_physdeviceobj->do_devext;
Finally bring an end to the great "make the Atheros NDIS driver work on SMP" saga. After several weeks and much gnashing of teeth, I have finally tracked down all the problems, despite their best efforts to confound and annoy me. Problem nunmber one: the Atheros windows driver is _NOT_ a de-serialized miniport! It used to be that NDIS drivers relied on the NDIS library itself for all their locking and serialization needs. Transmit packet queues were all handled internally by NDIS, and all calls to MiniportXXX() routines were guaranteed to be appropriately serialized. This proved to be a performance problem however, and Microsoft introduced de-serialized miniports with the NDIS 5.x spec. Microsoft still supports serialized miniports, but recommends that all new drivers written for Windows XP and later be deserialized. Apparently Atheros wasn't listening when they said this. This means (among other things) that we have to serialize calls to MiniportSendPackets(). We also have to serialize calls to MiniportTimer() that are triggered via the NdisMInitializeTimer() routine. It finally dawned on me why NdisMInitializeTimer() takes a special NDIS_MINIPORT_TIMER structure and a pointer to the miniport block: the timer callback must be serialized, and it's only by saving the miniport block handle that we can get access to the serialization lock during the timer callback. Problem number two: haunted hardware. The thing that was _really_ driving me absolutely bonkers for the longest time is that, for some reason I couldn't understand, my test machine would occasionally freeze or more frustratingly, reset completely. That's reset and in *pow!* back to the BIOS startup. No panic, no crashdump, just a reset. This appeared to happen most often when MiniportReset() was called. (As to why MiniportReset() was being called, see problem three below.) I thought maybe I had created some sort of horrible deadlock condition in the process of adding the serialization, but after three weeks, at least 6 different locking implementations and heroic efforts to debug the spinlock code, the machine still kept resetting. Finally, I started single stepping through the MiniportReset() routine in the driver using the kernel debugger, and this ultimately led me to the source of the problem. One of the last things the Atheros MiniportReset() routine does is call NdisReadPciSlotInformation() several times to inspect a portion of the device's PCI config space. It reads the same chunk of config space repeatedly, in rapid succession. Presumeably, it's polling the hardware for some sort of event. The reset occurs partway through this process. I discovered that when I single-stepped through this portion of the routine, the reset didn't occur. So I inserted a 1 microsecond delay into the read loop in NdisReadPciSlotInformation(). Suddenly, the reset was gone!! I'm still very puzzled by the whole thing. What I suspect is happening is that reading the PCI config space so quickly is causing a severe PCI bus error. My test system is a Sun w2100z dual Opteron system, and the NIC is a miniPCI card mounted in a miniPCI-to-PCI carrier card, plugged into a 100Mhz PCI slot. It's possible that this combination of hardware causes a bus protocol violation in this scenario which leads to a fatal machine check. This is pure speculation though. Really all I know for sure is that inserting the delay makes the problem go away. (To quote Homer Simpson: "I don't know how it works, but fire makes it good!") Problem number three: NdisAllocatePacket() needs to make sure to initialize the npp_validcounts field in the 'private' section of the NDIS_PACKET structure. The reason if_ndis was calling the MiniportReset() routine in the first place is that packet transmits were sometimes hanging. When sending a packet, an NDIS driver will call NdisQueryPacket() to learn how many physical buffers the packet resides in. NdisQueryPacket() is actually a macro, which traverses the NDIS_BUFFER list attached to the NDIS_PACKET and stashes some of the results in the 'private' section of the NDIS_PACKET. It also sets the npp_validcounts field to TRUE To indicate that the results are now valid. The problem is, now that if_ndis creates a pool of transmit packets via NdisAllocatePacketPool(), it's important that each time a new packet is allocated via NdisAllocatePacket() that validcounts be initialized to FALSE. If it isn't, and a previously transmitted NDIS_PACKET is pulled out of the pool, it may contain stale data from a previous transmission which won't get updated by NdisQueryPacket(). This would cause the driver to miscompute the number of fragments for a given packet, and botch the transmission. Fixing these three problems seems to make the Atheros driver happy on SMP, which hopefully means other serialized miniports will be happy too. And there was much rejoicing. Other stuff fixed along the way: - Modified ndis_thsuspend() to take a mutex as an argument. This allows KeWaitForSingleObject() and KeWaitForMultipleObjects() to avoid any possible race conditions with other routines that use the dispatcher lock. - Fixed KeCancelTimer() so that it returns the correct value for 'pending' according to the Microsoft documentation - Modfied NdisGetSystemUpTime() to use ticks and hz rather than calling nanouptime(). Also added comment that this routine wraps after 49.7 days. - Added macros for KeAcquireSpinLock()/KeReleaseSpinLock() to hide all the MSCALL() goop. - For x86, KeAcquireSpinLockRaiseToDpc() needs to be a separate function. This is because it's supposed to be _stdcall on the x86 arch, whereas KeAcquireSpinLock() is supposed to be _fastcall. On amd64, all routines use the same calling convention so we can just map KeAcquireSpinLockRaiseToDpc() directly to KfAcquireSpinLock() and it will work. (The _fastcall attribute is a no-op on amd64.) - Implement and use IoInitializeDpcRequest() and IoRequestDpc() (they're just macros) and use them for interrupt handling. This allows us to move the ndis_intrtask() routine from if_ndis.c to kern_ndis.c. - Fix the MmInitializeMdl() macro so that is uses sizeof(vm_offset_t) when computing mdl_size instead of uint32_t, so that it matches the MmSizeOfMdl() routine. - Change a could of M_WAITOKs to M_NOWAITs in the unicode routines in subr_ndis.c. - Use the dispatcher lock a little more consistently in subr_ntoskrnl.c. - Get rid of the "wait for link event" hack in ndis_init(). Now that I fixed NdisReadPciSlotInformation(), it seems I don't need it anymore. This should fix the witness panic a couple of people have reported. - Use MSCALL1() when calling the MiniportHangCheck() function in ndis_ticktask(). I accidentally missed this one when adding the wrapping for amd64.
2005-03-27 10:14:36 +00:00
for (i = 0; i < len; i++) {
DELAY(1);
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
pci_write_config(dev, i + offset, dest[i], 1);
Finally bring an end to the great "make the Atheros NDIS driver work on SMP" saga. After several weeks and much gnashing of teeth, I have finally tracked down all the problems, despite their best efforts to confound and annoy me. Problem nunmber one: the Atheros windows driver is _NOT_ a de-serialized miniport! It used to be that NDIS drivers relied on the NDIS library itself for all their locking and serialization needs. Transmit packet queues were all handled internally by NDIS, and all calls to MiniportXXX() routines were guaranteed to be appropriately serialized. This proved to be a performance problem however, and Microsoft introduced de-serialized miniports with the NDIS 5.x spec. Microsoft still supports serialized miniports, but recommends that all new drivers written for Windows XP and later be deserialized. Apparently Atheros wasn't listening when they said this. This means (among other things) that we have to serialize calls to MiniportSendPackets(). We also have to serialize calls to MiniportTimer() that are triggered via the NdisMInitializeTimer() routine. It finally dawned on me why NdisMInitializeTimer() takes a special NDIS_MINIPORT_TIMER structure and a pointer to the miniport block: the timer callback must be serialized, and it's only by saving the miniport block handle that we can get access to the serialization lock during the timer callback. Problem number two: haunted hardware. The thing that was _really_ driving me absolutely bonkers for the longest time is that, for some reason I couldn't understand, my test machine would occasionally freeze or more frustratingly, reset completely. That's reset and in *pow!* back to the BIOS startup. No panic, no crashdump, just a reset. This appeared to happen most often when MiniportReset() was called. (As to why MiniportReset() was being called, see problem three below.) I thought maybe I had created some sort of horrible deadlock condition in the process of adding the serialization, but after three weeks, at least 6 different locking implementations and heroic efforts to debug the spinlock code, the machine still kept resetting. Finally, I started single stepping through the MiniportReset() routine in the driver using the kernel debugger, and this ultimately led me to the source of the problem. One of the last things the Atheros MiniportReset() routine does is call NdisReadPciSlotInformation() several times to inspect a portion of the device's PCI config space. It reads the same chunk of config space repeatedly, in rapid succession. Presumeably, it's polling the hardware for some sort of event. The reset occurs partway through this process. I discovered that when I single-stepped through this portion of the routine, the reset didn't occur. So I inserted a 1 microsecond delay into the read loop in NdisReadPciSlotInformation(). Suddenly, the reset was gone!! I'm still very puzzled by the whole thing. What I suspect is happening is that reading the PCI config space so quickly is causing a severe PCI bus error. My test system is a Sun w2100z dual Opteron system, and the NIC is a miniPCI card mounted in a miniPCI-to-PCI carrier card, plugged into a 100Mhz PCI slot. It's possible that this combination of hardware causes a bus protocol violation in this scenario which leads to a fatal machine check. This is pure speculation though. Really all I know for sure is that inserting the delay makes the problem go away. (To quote Homer Simpson: "I don't know how it works, but fire makes it good!") Problem number three: NdisAllocatePacket() needs to make sure to initialize the npp_validcounts field in the 'private' section of the NDIS_PACKET structure. The reason if_ndis was calling the MiniportReset() routine in the first place is that packet transmits were sometimes hanging. When sending a packet, an NDIS driver will call NdisQueryPacket() to learn how many physical buffers the packet resides in. NdisQueryPacket() is actually a macro, which traverses the NDIS_BUFFER list attached to the NDIS_PACKET and stashes some of the results in the 'private' section of the NDIS_PACKET. It also sets the npp_validcounts field to TRUE To indicate that the results are now valid. The problem is, now that if_ndis creates a pool of transmit packets via NdisAllocatePacketPool(), it's important that each time a new packet is allocated via NdisAllocatePacket() that validcounts be initialized to FALSE. If it isn't, and a previously transmitted NDIS_PACKET is pulled out of the pool, it may contain stale data from a previous transmission which won't get updated by NdisQueryPacket(). This would cause the driver to miscompute the number of fragments for a given packet, and botch the transmission. Fixing these three problems seems to make the Atheros driver happy on SMP, which hopefully means other serialized miniports will be happy too. And there was much rejoicing. Other stuff fixed along the way: - Modified ndis_thsuspend() to take a mutex as an argument. This allows KeWaitForSingleObject() and KeWaitForMultipleObjects() to avoid any possible race conditions with other routines that use the dispatcher lock. - Fixed KeCancelTimer() so that it returns the correct value for 'pending' according to the Microsoft documentation - Modfied NdisGetSystemUpTime() to use ticks and hz rather than calling nanouptime(). Also added comment that this routine wraps after 49.7 days. - Added macros for KeAcquireSpinLock()/KeReleaseSpinLock() to hide all the MSCALL() goop. - For x86, KeAcquireSpinLockRaiseToDpc() needs to be a separate function. This is because it's supposed to be _stdcall on the x86 arch, whereas KeAcquireSpinLock() is supposed to be _fastcall. On amd64, all routines use the same calling convention so we can just map KeAcquireSpinLockRaiseToDpc() directly to KfAcquireSpinLock() and it will work. (The _fastcall attribute is a no-op on amd64.) - Implement and use IoInitializeDpcRequest() and IoRequestDpc() (they're just macros) and use them for interrupt handling. This allows us to move the ndis_intrtask() routine from if_ndis.c to kern_ndis.c. - Fix the MmInitializeMdl() macro so that is uses sizeof(vm_offset_t) when computing mdl_size instead of uint32_t, so that it matches the MmSizeOfMdl() routine. - Change a could of M_WAITOKs to M_NOWAITs in the unicode routines in subr_ndis.c. - Use the dispatcher lock a little more consistently in subr_ntoskrnl.c. - Get rid of the "wait for link event" hack in ndis_init(). Now that I fixed NdisReadPciSlotInformation(), it seems I don't need it anymore. This should fix the witness panic a couple of people have reported. - Use MSCALL1() when calling the MiniportHangCheck() function in ndis_ticktask(). I accidentally missed this one when adding the wrapping for amd64.
2005-03-27 10:14:36 +00:00
}
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return(len);
}
/*
* The errorlog routine uses a variable argument list, so we
* have to declare it this way.
*/
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
It appears drivers may call NdisWriteErrorLogEntry() with locks held. However, if we need to translate a unicode message table message, ndis_unicode_to_ascii() might malloc() some memory, which causes a warning from witness. Avoid this by using some stack space to hold the translated message. (Also bounds check to make sure we don't overrun the stack buffer.)
2004-01-07 19:26:47 +00:00
#define ERRMSGLEN 512
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisWriteErrorLogEntry(ndis_handle adapter, ndis_error_code code,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
uint32_t numerrors, ...)
{
ndis_miniport_block *block;
va_list ap;
Clean up pe_get_message(). Allow the caller to obtain the resource flag so that it can see if the message string is unicode or not and do the conversion itself rather than doing it in subr_pe.c. This prevents subr_pe.c from being dependent on subr_ndis.c.
2004-01-06 18:06:54 +00:00
int i, error;
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
char *str = NULL;
Clean up pe_get_message(). Allow the caller to obtain the resource flag so that it can see if the message string is unicode or not and do the conversion itself rather than doing it in subr_pe.c. This prevents subr_pe.c from being dependent on subr_ndis.c.
2004-01-06 18:06:54 +00:00
uint16_t flags;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
device_t dev;
Add support for Windows/x86-64 binaries to Project Evil. Ville-Pertti Keinonen (will at exomi dot comohmygodnospampleasekthx) deserves a big thanks for submitting initial patches to make it work. I have mangled his contributions appropriately. The main gotcha with Windows/x86-64 is that Microsoft uses a different calling convention than everyone else. The standard ABI requires using 6 registers for argument passing, with other arguments on the stack. Microsoft uses only 4 registers, and requires the caller to leave room on the stack for the register arguments incase the callee needs to spill them. Unlike x86, where Microsoft uses a mix of _cdecl, _stdcall and _fastcall, all routines on Windows/x86-64 uses the same convention. This unfortunately means that all the functions we export to the driver require an intermediate translation wrapper. Similarly, we have to wrap all calls back into the driver binary itself. The original patches provided macros to wrap every single routine at compile time, providing a secondary jump table with a customized wrapper for each exported routine. I decided to use a different approach: the call wrapper for each function is created from a template at runtime, and the routine to jump to is patched into the wrapper as it is created. The subr_pe module has been modified to patch in the wrapped function instead of the original. (On x86, the wrapping routine is a no-op.) There are some minor API differences that had to be accounted for: - KeAcquireSpinLock() is a real function on amd64, not a macro wrapper around KfAcquireSpinLock() - NdisFreeBuffer() is actually IoFreeMdl(). I had to change the whole NDIS_BUFFER API a bit to accomodate this. Bugs fixed along the way: - IoAllocateMdl() always returned NULL - kern_windrv.c:windrv_unload() wasn't releasing private driver object extensions correctly (found thanks to memguard) This has only been tested with the driver for the Broadcom 802.11g chipset, which was the only Windows/x86-64 driver I could find.
2005-02-16 05:41:18 +00:00
driver_object *drv;
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
struct ndis_softc *sc;
struct ifnet *ifp;
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
unicode_string us;
Fix handling of message table messages that got broken when I converted NdisWriteErrorLogEntry() to use the RtlXXX unicode/ansi conversion routines.
2005-10-24 05:05:09 +00:00
ansi_string as = { 0, 0, NULL };
Make NdisReadPcmciaAttributeMemory() and NdisWritePcmciaAttributeMemory() actually work. Make the PCI and PCCARD attachments provide a bus_get_resource_list() method so that resource listing for PCCARD works. PCCARD does not have a bus_get_resource_list() method (yet), so I faked up the resource list management in if_ndis_pccard.c, and added bus_get_resource_list() methods to both if_ndis_pccard.c and if_ndis_pci.c. The one in the PCI attechment just hands off to the PCI bus code. The difference is transparent to the NDIS resource handler code. Fixed ndis_open_file() so that opening files which live on NFS filesystems work: pass an actual ucred structure to VOP_GETATTR() (NFS explodes if the ucred structure is NOCRED). Make NdisMMapIoSpace() handle mapping of PCMCIA attribute memory resources correctly. Turn subr_ndis.c:my_strcasecmp() into ndis_strcasecmp() and export it so that if_ndis_pccard.c can use it, and junk the other copy of my_strcasecmp() from if_ndis_pccard.c.
2004-07-11 00:19:30 +00:00
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
block = (ndis_miniport_block *)adapter;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
dev = block->nmb_physdeviceobj->do_devext;
Throw the switch on the new driver generation/loading mechanism. From here on in, if_ndis.ko will be pre-built as a module, and can be built into a static kernel (though it's not part of GENERIC). Drivers are created using the new ndisgen(8) script, which uses ndiscvt(8) under the covers, along with a few other tools. The result is a driver module that can be kldloaded into the kernel. A driver with foo.inf and foo.sys files will be converted into foo_sys.ko (and foo_sys.o, for those who want/need to make static kernels). This module contains all of the necessary info from the .INF file and the driver binary image, converted into an ELF module. You can kldload this module (or add it to /boot/loader.conf) to have it loaded automatically. Any required firmware files can be bundled into the module as well (or converted/loaded separately). Also, add a workaround for a problem in NdisMSleep(). During system bootstrap (cold == 1), msleep() always returns 0 without actually sleeping. The Intel 2200BG driver uses NdisMSleep() to wait for the NIC's firmware to come to life, and fails to load if NdisMSleep() doesn't actually delay. As a workaround, if msleep() (and hence ndis_thsuspend()) returns 0, use a hard DELAY() to sleep instead). This is not really the right thing to do, but we can't really do much else. At the very least, this makes the Intel driver happy. There are probably other drivers that fail in this way during bootstrap. Unfortunately, the only workaround for those is to avoid pre-loading them and kldload them once the system is running instead.
2005-04-24 20:21:22 +00:00
drv = block->nmb_deviceobj->do_drvobj;
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
sc = device_get_softc(dev);
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
ifp = sc->ifp;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
Add support for Windows/x86-64 binaries to Project Evil. Ville-Pertti Keinonen (will at exomi dot comohmygodnospampleasekthx) deserves a big thanks for submitting initial patches to make it work. I have mangled his contributions appropriately. The main gotcha with Windows/x86-64 is that Microsoft uses a different calling convention than everyone else. The standard ABI requires using 6 registers for argument passing, with other arguments on the stack. Microsoft uses only 4 registers, and requires the caller to leave room on the stack for the register arguments incase the callee needs to spill them. Unlike x86, where Microsoft uses a mix of _cdecl, _stdcall and _fastcall, all routines on Windows/x86-64 uses the same convention. This unfortunately means that all the functions we export to the driver require an intermediate translation wrapper. Similarly, we have to wrap all calls back into the driver binary itself. The original patches provided macros to wrap every single routine at compile time, providing a secondary jump table with a customized wrapper for each exported routine. I decided to use a different approach: the call wrapper for each function is created from a template at runtime, and the routine to jump to is patched into the wrapper as it is created. The subr_pe module has been modified to patch in the wrapped function instead of the original. (On x86, the wrapping routine is a no-op.) There are some minor API differences that had to be accounted for: - KeAcquireSpinLock() is a real function on amd64, not a macro wrapper around KfAcquireSpinLock() - NdisFreeBuffer() is actually IoFreeMdl(). I had to change the whole NDIS_BUFFER API a bit to accomodate this. Bugs fixed along the way: - IoAllocateMdl() always returned NULL - kern_windrv.c:windrv_unload() wasn't releasing private driver object extensions correctly (found thanks to memguard) This has only been tested with the driver for the Broadcom 802.11g chipset, which was the only Windows/x86-64 driver I could find.
2005-02-16 05:41:18 +00:00
error = pe_get_message((vm_offset_t)drv->dro_driverstart,
code, &str, &i, &flags);
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
if (error == 0 && flags & MESSAGE_RESOURCE_UNICODE &&
ifp->if_flags & IFF_DEBUG) {
Fix handling of message table messages that got broken when I converted NdisWriteErrorLogEntry() to use the RtlXXX unicode/ansi conversion routines.
2005-10-24 05:05:09 +00:00
RtlInitUnicodeString(&us, (uint16_t *)str);
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
if (RtlUnicodeStringToAnsiString(&as, &us, TRUE))
return;
str = as.as_buf;
Fix handling of message table messages that got broken when I converted NdisWriteErrorLogEntry() to use the RtlXXX unicode/ansi conversion routines.
2005-10-24 05:05:09 +00:00
} else
str = NULL;
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
device_printf (dev, "NDIS ERROR: %x (%s)\n", code,
- Add pe_get_message() and pe_get_messagetable() for processing the RT_MESSAGETABLE resources that some driver binaries have. This allows us to print error messages in ndis_syslog(). - Correct the implementation of InterlockedIncrement() and InterlockedDecrement() -- they return uint32_t, not void. - Correct the declarations of the 64-bit arithmetic shift routines in subr_ntoskrnl.c (_allshr, allshl, etc...). These do not follow the _stdcall convention: instead, they appear to be __attribute__((regparm(3)). - Change the implementation of KeInitializeSpinLock(). There is no complementary KeFreeSpinLock() function, so creating a new mutex on each call to KeInitializeSpinLock() leaks resources when a driver is unloaded. For now, KeInitializeSpinLock() returns a handle to the ntoskrnl interlock mutex. - Use a driver's MiniportDisableInterrupt() and MiniportEnableInterrupt() routines if they exist. I'm not sure if I'm doing this right yet, but at the very least this shouldn't break any currently working drivers, and it makes the Intel PRO/1000 driver work. - In ndis_register_intr(), save some state that might be needed later, and save a pointer to the driver's interrupt structure in the ndis_miniport_block. - Save a pointer to the driver image for use by ndis_syslog() when it calls pe_get_message().
2004-01-06 07:09:26 +00:00
str == NULL ? "unknown error" : str);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
if (ifp->if_flags & IFF_DEBUG) {
device_printf (dev, "NDIS NUMERRORS: %x\n", numerrors);
va_start(ap, numerrors);
for (i = 0; i < numerrors; i++)
device_printf (dev, "argptr: %p\n",
va_arg(ap, void *));
va_end(ap);
}
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
Fix handling of message table messages that got broken when I converted NdisWriteErrorLogEntry() to use the RtlXXX unicode/ansi conversion routines.
2005-10-24 05:05:09 +00:00
if (as.as_len)
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
RtlFreeAnsiString(&as);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
static void
ndis_map_cb(arg, segs, nseg, error)
void *arg;
bus_dma_segment_t *segs;
int nseg;
int error;
{
struct ndis_map_arg *ctx;
int i;
if (error)
return;
ctx = arg;
for (i = 0; i < nseg; i++) {
ctx->nma_fraglist[i].npu_physaddr.np_quad = segs[i].ds_addr;
ctx->nma_fraglist[i].npu_len = segs[i].ds_len;
}
ctx->nma_cnt = nseg;
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMStartBufferPhysicalMapping(adapter, buf, mapreg, writedev, addrarray, arraysize)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_handle adapter;
ndis_buffer *buf;
uint32_t mapreg;
uint8_t writedev;
ndis_paddr_unit *addrarray;
uint32_t *arraysize;
{
ndis_miniport_block *block;
struct ndis_softc *sc;
struct ndis_map_arg nma;
bus_dmamap_t map;
int error;
if (adapter == NULL)
return;
block = (ndis_miniport_block *)adapter;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
sc = device_get_softc(block->nmb_physdeviceobj->do_devext);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
if (mapreg > sc->ndis_mmapcnt)
return;
map = sc->ndis_mmaps[mapreg];
nma.nma_fraglist = addrarray;
error = bus_dmamap_load(sc->ndis_mtag, map,
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
MmGetMdlVirtualAddress(buf), MmGetMdlByteCount(buf), ndis_map_cb,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
(void *)&nma, BUS_DMA_NOWAIT);
if (error)
return;
bus_dmamap_sync(sc->ndis_mtag, map,
writedev ? BUS_DMASYNC_PREWRITE : BUS_DMASYNC_PREREAD);
*arraysize = nma.nma_cnt;
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMCompleteBufferPhysicalMapping(adapter, buf, mapreg)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_handle adapter;
ndis_buffer *buf;
uint32_t mapreg;
{
ndis_miniport_block *block;
struct ndis_softc *sc;
bus_dmamap_t map;
if (adapter == NULL)
return;
block = (ndis_miniport_block *)adapter;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
sc = device_get_softc(block->nmb_physdeviceobj->do_devext);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
if (mapreg > sc->ndis_mmapcnt)
return;
map = sc->ndis_mmaps[mapreg];
bus_dmamap_sync(sc->ndis_mtag, map,
BUS_DMASYNC_POSTREAD|BUS_DMASYNC_POSTWRITE);
bus_dmamap_unload(sc->ndis_mtag, map);
return;
}
Reorganize the timer code a little and implement NdisInitializeTimer() and NdisCancelTimer(). NdisInitializeTimer() doesn't accept an NDIS miniport context argument, so we have to derive it from the timer function context (which is supposed to be the adapter private context). NdisCancelTimer is now an alias for NdisMCancelTimer(). Also add stubs for NdisMRegisterDevice() and NdisMDeregisterDevice(). These are no-ops for now, but will likely get fleshed in once I start working on the Am1771/Am1772 wireless driver.
2004-01-26 21:21:53 +00:00
/*
Finally bring an end to the great "make the Atheros NDIS driver work on SMP" saga. After several weeks and much gnashing of teeth, I have finally tracked down all the problems, despite their best efforts to confound and annoy me. Problem nunmber one: the Atheros windows driver is _NOT_ a de-serialized miniport! It used to be that NDIS drivers relied on the NDIS library itself for all their locking and serialization needs. Transmit packet queues were all handled internally by NDIS, and all calls to MiniportXXX() routines were guaranteed to be appropriately serialized. This proved to be a performance problem however, and Microsoft introduced de-serialized miniports with the NDIS 5.x spec. Microsoft still supports serialized miniports, but recommends that all new drivers written for Windows XP and later be deserialized. Apparently Atheros wasn't listening when they said this. This means (among other things) that we have to serialize calls to MiniportSendPackets(). We also have to serialize calls to MiniportTimer() that are triggered via the NdisMInitializeTimer() routine. It finally dawned on me why NdisMInitializeTimer() takes a special NDIS_MINIPORT_TIMER structure and a pointer to the miniport block: the timer callback must be serialized, and it's only by saving the miniport block handle that we can get access to the serialization lock during the timer callback. Problem number two: haunted hardware. The thing that was _really_ driving me absolutely bonkers for the longest time is that, for some reason I couldn't understand, my test machine would occasionally freeze or more frustratingly, reset completely. That's reset and in *pow!* back to the BIOS startup. No panic, no crashdump, just a reset. This appeared to happen most often when MiniportReset() was called. (As to why MiniportReset() was being called, see problem three below.) I thought maybe I had created some sort of horrible deadlock condition in the process of adding the serialization, but after three weeks, at least 6 different locking implementations and heroic efforts to debug the spinlock code, the machine still kept resetting. Finally, I started single stepping through the MiniportReset() routine in the driver using the kernel debugger, and this ultimately led me to the source of the problem. One of the last things the Atheros MiniportReset() routine does is call NdisReadPciSlotInformation() several times to inspect a portion of the device's PCI config space. It reads the same chunk of config space repeatedly, in rapid succession. Presumeably, it's polling the hardware for some sort of event. The reset occurs partway through this process. I discovered that when I single-stepped through this portion of the routine, the reset didn't occur. So I inserted a 1 microsecond delay into the read loop in NdisReadPciSlotInformation(). Suddenly, the reset was gone!! I'm still very puzzled by the whole thing. What I suspect is happening is that reading the PCI config space so quickly is causing a severe PCI bus error. My test system is a Sun w2100z dual Opteron system, and the NIC is a miniPCI card mounted in a miniPCI-to-PCI carrier card, plugged into a 100Mhz PCI slot. It's possible that this combination of hardware causes a bus protocol violation in this scenario which leads to a fatal machine check. This is pure speculation though. Really all I know for sure is that inserting the delay makes the problem go away. (To quote Homer Simpson: "I don't know how it works, but fire makes it good!") Problem number three: NdisAllocatePacket() needs to make sure to initialize the npp_validcounts field in the 'private' section of the NDIS_PACKET structure. The reason if_ndis was calling the MiniportReset() routine in the first place is that packet transmits were sometimes hanging. When sending a packet, an NDIS driver will call NdisQueryPacket() to learn how many physical buffers the packet resides in. NdisQueryPacket() is actually a macro, which traverses the NDIS_BUFFER list attached to the NDIS_PACKET and stashes some of the results in the 'private' section of the NDIS_PACKET. It also sets the npp_validcounts field to TRUE To indicate that the results are now valid. The problem is, now that if_ndis creates a pool of transmit packets via NdisAllocatePacketPool(), it's important that each time a new packet is allocated via NdisAllocatePacket() that validcounts be initialized to FALSE. If it isn't, and a previously transmitted NDIS_PACKET is pulled out of the pool, it may contain stale data from a previous transmission which won't get updated by NdisQueryPacket(). This would cause the driver to miscompute the number of fragments for a given packet, and botch the transmission. Fixing these three problems seems to make the Atheros driver happy on SMP, which hopefully means other serialized miniports will be happy too. And there was much rejoicing. Other stuff fixed along the way: - Modified ndis_thsuspend() to take a mutex as an argument. This allows KeWaitForSingleObject() and KeWaitForMultipleObjects() to avoid any possible race conditions with other routines that use the dispatcher lock. - Fixed KeCancelTimer() so that it returns the correct value for 'pending' according to the Microsoft documentation - Modfied NdisGetSystemUpTime() to use ticks and hz rather than calling nanouptime(). Also added comment that this routine wraps after 49.7 days. - Added macros for KeAcquireSpinLock()/KeReleaseSpinLock() to hide all the MSCALL() goop. - For x86, KeAcquireSpinLockRaiseToDpc() needs to be a separate function. This is because it's supposed to be _stdcall on the x86 arch, whereas KeAcquireSpinLock() is supposed to be _fastcall. On amd64, all routines use the same calling convention so we can just map KeAcquireSpinLockRaiseToDpc() directly to KfAcquireSpinLock() and it will work. (The _fastcall attribute is a no-op on amd64.) - Implement and use IoInitializeDpcRequest() and IoRequestDpc() (they're just macros) and use them for interrupt handling. This allows us to move the ndis_intrtask() routine from if_ndis.c to kern_ndis.c. - Fix the MmInitializeMdl() macro so that is uses sizeof(vm_offset_t) when computing mdl_size instead of uint32_t, so that it matches the MmSizeOfMdl() routine. - Change a could of M_WAITOKs to M_NOWAITs in the unicode routines in subr_ndis.c. - Use the dispatcher lock a little more consistently in subr_ntoskrnl.c. - Get rid of the "wait for link event" hack in ndis_init(). Now that I fixed NdisReadPciSlotInformation(), it seems I don't need it anymore. This should fix the witness panic a couple of people have reported. - Use MSCALL1() when calling the MiniportHangCheck() function in ndis_ticktask(). I accidentally missed this one when adding the wrapping for amd64.
2005-03-27 10:14:36 +00:00
* This is an older (?) timer init routine which doesn't
* accept a miniport context handle. Serialized miniports should
* never call this function.
Reorganize the timer code a little and implement NdisInitializeTimer() and NdisCancelTimer(). NdisInitializeTimer() doesn't accept an NDIS miniport context argument, so we have to derive it from the timer function context (which is supposed to be the adapter private context). NdisCancelTimer is now an alias for NdisMCancelTimer(). Also add stubs for NdisMRegisterDevice() and NdisMDeregisterDevice(). These are no-ops for now, but will likely get fleshed in once I start working on the Am1771/Am1772 wireless driver.
2004-01-26 21:21:53 +00:00
*/
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisInitializeTimer(timer, func, ctx)
Reorganize the timer code a little and implement NdisInitializeTimer() and NdisCancelTimer(). NdisInitializeTimer() doesn't accept an NDIS miniport context argument, so we have to derive it from the timer function context (which is supposed to be the adapter private context). NdisCancelTimer is now an alias for NdisMCancelTimer(). Also add stubs for NdisMRegisterDevice() and NdisMDeregisterDevice(). These are no-ops for now, but will likely get fleshed in once I start working on the Am1771/Am1772 wireless driver.
2004-01-26 21:21:53 +00:00
ndis_timer *timer;
ndis_timer_function func;
void *ctx;
{
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
KeInitializeTimer(&timer->nt_ktimer);
KeInitializeDpc(&timer->nt_kdpc, func, ctx);
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
KeSetImportanceDpc(&timer->nt_kdpc, KDPC_IMPORTANCE_LOW);
Reorganize the timer code a little and implement NdisInitializeTimer() and NdisCancelTimer(). NdisInitializeTimer() doesn't accept an NDIS miniport context argument, so we have to derive it from the timer function context (which is supposed to be the adapter private context). NdisCancelTimer is now an alias for NdisMCancelTimer(). Also add stubs for NdisMRegisterDevice() and NdisMDeregisterDevice(). These are no-ops for now, but will likely get fleshed in once I start working on the Am1771/Am1772 wireless driver.
2004-01-26 21:21:53 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Finally bring an end to the great "make the Atheros NDIS driver work on SMP" saga. After several weeks and much gnashing of teeth, I have finally tracked down all the problems, despite their best efforts to confound and annoy me. Problem nunmber one: the Atheros windows driver is _NOT_ a de-serialized miniport! It used to be that NDIS drivers relied on the NDIS library itself for all their locking and serialization needs. Transmit packet queues were all handled internally by NDIS, and all calls to MiniportXXX() routines were guaranteed to be appropriately serialized. This proved to be a performance problem however, and Microsoft introduced de-serialized miniports with the NDIS 5.x spec. Microsoft still supports serialized miniports, but recommends that all new drivers written for Windows XP and later be deserialized. Apparently Atheros wasn't listening when they said this. This means (among other things) that we have to serialize calls to MiniportSendPackets(). We also have to serialize calls to MiniportTimer() that are triggered via the NdisMInitializeTimer() routine. It finally dawned on me why NdisMInitializeTimer() takes a special NDIS_MINIPORT_TIMER structure and a pointer to the miniport block: the timer callback must be serialized, and it's only by saving the miniport block handle that we can get access to the serialization lock during the timer callback. Problem number two: haunted hardware. The thing that was _really_ driving me absolutely bonkers for the longest time is that, for some reason I couldn't understand, my test machine would occasionally freeze or more frustratingly, reset completely. That's reset and in *pow!* back to the BIOS startup. No panic, no crashdump, just a reset. This appeared to happen most often when MiniportReset() was called. (As to why MiniportReset() was being called, see problem three below.) I thought maybe I had created some sort of horrible deadlock condition in the process of adding the serialization, but after three weeks, at least 6 different locking implementations and heroic efforts to debug the spinlock code, the machine still kept resetting. Finally, I started single stepping through the MiniportReset() routine in the driver using the kernel debugger, and this ultimately led me to the source of the problem. One of the last things the Atheros MiniportReset() routine does is call NdisReadPciSlotInformation() several times to inspect a portion of the device's PCI config space. It reads the same chunk of config space repeatedly, in rapid succession. Presumeably, it's polling the hardware for some sort of event. The reset occurs partway through this process. I discovered that when I single-stepped through this portion of the routine, the reset didn't occur. So I inserted a 1 microsecond delay into the read loop in NdisReadPciSlotInformation(). Suddenly, the reset was gone!! I'm still very puzzled by the whole thing. What I suspect is happening is that reading the PCI config space so quickly is causing a severe PCI bus error. My test system is a Sun w2100z dual Opteron system, and the NIC is a miniPCI card mounted in a miniPCI-to-PCI carrier card, plugged into a 100Mhz PCI slot. It's possible that this combination of hardware causes a bus protocol violation in this scenario which leads to a fatal machine check. This is pure speculation though. Really all I know for sure is that inserting the delay makes the problem go away. (To quote Homer Simpson: "I don't know how it works, but fire makes it good!") Problem number three: NdisAllocatePacket() needs to make sure to initialize the npp_validcounts field in the 'private' section of the NDIS_PACKET structure. The reason if_ndis was calling the MiniportReset() routine in the first place is that packet transmits were sometimes hanging. When sending a packet, an NDIS driver will call NdisQueryPacket() to learn how many physical buffers the packet resides in. NdisQueryPacket() is actually a macro, which traverses the NDIS_BUFFER list attached to the NDIS_PACKET and stashes some of the results in the 'private' section of the NDIS_PACKET. It also sets the npp_validcounts field to TRUE To indicate that the results are now valid. The problem is, now that if_ndis creates a pool of transmit packets via NdisAllocatePacketPool(), it's important that each time a new packet is allocated via NdisAllocatePacket() that validcounts be initialized to FALSE. If it isn't, and a previously transmitted NDIS_PACKET is pulled out of the pool, it may contain stale data from a previous transmission which won't get updated by NdisQueryPacket(). This would cause the driver to miscompute the number of fragments for a given packet, and botch the transmission. Fixing these three problems seems to make the Atheros driver happy on SMP, which hopefully means other serialized miniports will be happy too. And there was much rejoicing. Other stuff fixed along the way: - Modified ndis_thsuspend() to take a mutex as an argument. This allows KeWaitForSingleObject() and KeWaitForMultipleObjects() to avoid any possible race conditions with other routines that use the dispatcher lock. - Fixed KeCancelTimer() so that it returns the correct value for 'pending' according to the Microsoft documentation - Modfied NdisGetSystemUpTime() to use ticks and hz rather than calling nanouptime(). Also added comment that this routine wraps after 49.7 days. - Added macros for KeAcquireSpinLock()/KeReleaseSpinLock() to hide all the MSCALL() goop. - For x86, KeAcquireSpinLockRaiseToDpc() needs to be a separate function. This is because it's supposed to be _stdcall on the x86 arch, whereas KeAcquireSpinLock() is supposed to be _fastcall. On amd64, all routines use the same calling convention so we can just map KeAcquireSpinLockRaiseToDpc() directly to KfAcquireSpinLock() and it will work. (The _fastcall attribute is a no-op on amd64.) - Implement and use IoInitializeDpcRequest() and IoRequestDpc() (they're just macros) and use them for interrupt handling. This allows us to move the ndis_intrtask() routine from if_ndis.c to kern_ndis.c. - Fix the MmInitializeMdl() macro so that is uses sizeof(vm_offset_t) when computing mdl_size instead of uint32_t, so that it matches the MmSizeOfMdl() routine. - Change a could of M_WAITOKs to M_NOWAITs in the unicode routines in subr_ndis.c. - Use the dispatcher lock a little more consistently in subr_ntoskrnl.c. - Get rid of the "wait for link event" hack in ndis_init(). Now that I fixed NdisReadPciSlotInformation(), it seems I don't need it anymore. This should fix the witness panic a couple of people have reported. - Use MSCALL1() when calling the MiniportHangCheck() function in ndis_ticktask(). I accidentally missed this one when adding the wrapping for amd64.
2005-03-27 10:14:36 +00:00
ndis_timercall(dpc, timer, sysarg1, sysarg2)
kdpc *dpc;
ndis_miniport_timer *timer;
void *sysarg1;
void *sysarg2;
{
/*
* Since we're called as a DPC, we should be running
* at DISPATCH_LEVEL here. This means to acquire the
* spinlock, we can use KeAcquireSpinLockAtDpcLevel()
* rather than KeAcquireSpinLock().
*/
if (NDIS_SERIALIZED(timer->nmt_block))
KeAcquireSpinLockAtDpcLevel(&timer->nmt_block->nmb_lock);
MSCALL4(timer->nmt_timerfunc, dpc, timer->nmt_timerctx,
sysarg1, sysarg2);
if (NDIS_SERIALIZED(timer->nmt_block))
KeReleaseSpinLockFromDpcLevel(&timer->nmt_block->nmb_lock);
return;
}
/*
* For a long time I wondered why there were two NDIS timer initialization
* routines, and why this one needed an NDIS_MINIPORT_TIMER and the
* MiniportAdapterHandle. The NDIS_MINIPORT_TIMER has its own callout
* function and context pointers separate from those in the DPC, which
* allows for another level of indirection: when the timer fires, we
* can have our own timer function invoked, and from there we can call
* the driver's function. But why go to all that trouble? Then it hit
* me: for serialized miniports, the timer callouts are not re-entrant.
* By trapping the callouts and having access to the MiniportAdapterHandle,
* we can protect the driver callouts by acquiring the NDIS serialization
* lock. This is essential for allowing serialized miniports to work
* correctly on SMP systems. On UP hosts, setting IRQL to DISPATCH_LEVEL
* is enough to prevent other threads from pre-empting you, but with
* SMP, you must acquire a lock as well, otherwise the other CPU is
* free to clobber you.
*/
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMInitializeTimer(timer, handle, func, ctx)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_miniport_timer *timer;
- Rewrite the timer and event API routines in subr_ndis.c so that they are actually layered on top of the KeTimer API in subr_ntoskrnl.c, just as it is in Windows. This reduces code duplication and more closely imitates the way things are done in Windows. - Modify ndis_encode_parm() to deal with the case where we have a registry key expressed as a hex value ("0x1") which is being read via NdisReadConfiguration() as an int. Previously, we tried to decode things like "0x1" with strtol() using a base of 10, which would always yield 0. This is what was causing problems with the Intel 2200BG Centrino 802.11g driver: the .inf file that comes with it has a key called RadioEnable with a value of 0x1. We incorrectly decoded this value to '0' when it was queried, hence the driver thought we wanted the radio turned off. - In if_ndis.c, most drivers don't accept NDIS_80211_AUTHMODE_AUTO, but NDIS_80211_AUTHMODE_SHARED may not be right in some cases, so for now always use NDIS_80211_AUTHMODE_OPEN. NOTE: There is still one problem with the Intel 2200BG driver: it happens that the kernel stack in Windows is larger than the kernel stack in FreeBSD. The 2200BG driver sometimes eats up more than 2 pages of stack space, which can lead to a double fault panic. For the moment, I got things to work by adding the following to my kernel config file: options KSTACK_PAGES=8 I'm pretty sure 8 is too big; I just picked this value out of a hat as a test, and it happened to work, so I left it. 4 pages might be enough. Unfortunately, I don't think you can dynamically give a thread a larger stack, so I'm not sure how to handle this short of putting a note in the man page about it and dealing with the flood of mail from people who never read man pages.
2004-03-20 23:39:43 +00:00
ndis_handle handle;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_timer_function func;
void *ctx;
{
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
uint8_t irql;
Finally bring an end to the great "make the Atheros NDIS driver work on SMP" saga. After several weeks and much gnashing of teeth, I have finally tracked down all the problems, despite their best efforts to confound and annoy me. Problem nunmber one: the Atheros windows driver is _NOT_ a de-serialized miniport! It used to be that NDIS drivers relied on the NDIS library itself for all their locking and serialization needs. Transmit packet queues were all handled internally by NDIS, and all calls to MiniportXXX() routines were guaranteed to be appropriately serialized. This proved to be a performance problem however, and Microsoft introduced de-serialized miniports with the NDIS 5.x spec. Microsoft still supports serialized miniports, but recommends that all new drivers written for Windows XP and later be deserialized. Apparently Atheros wasn't listening when they said this. This means (among other things) that we have to serialize calls to MiniportSendPackets(). We also have to serialize calls to MiniportTimer() that are triggered via the NdisMInitializeTimer() routine. It finally dawned on me why NdisMInitializeTimer() takes a special NDIS_MINIPORT_TIMER structure and a pointer to the miniport block: the timer callback must be serialized, and it's only by saving the miniport block handle that we can get access to the serialization lock during the timer callback. Problem number two: haunted hardware. The thing that was _really_ driving me absolutely bonkers for the longest time is that, for some reason I couldn't understand, my test machine would occasionally freeze or more frustratingly, reset completely. That's reset and in *pow!* back to the BIOS startup. No panic, no crashdump, just a reset. This appeared to happen most often when MiniportReset() was called. (As to why MiniportReset() was being called, see problem three below.) I thought maybe I had created some sort of horrible deadlock condition in the process of adding the serialization, but after three weeks, at least 6 different locking implementations and heroic efforts to debug the spinlock code, the machine still kept resetting. Finally, I started single stepping through the MiniportReset() routine in the driver using the kernel debugger, and this ultimately led me to the source of the problem. One of the last things the Atheros MiniportReset() routine does is call NdisReadPciSlotInformation() several times to inspect a portion of the device's PCI config space. It reads the same chunk of config space repeatedly, in rapid succession. Presumeably, it's polling the hardware for some sort of event. The reset occurs partway through this process. I discovered that when I single-stepped through this portion of the routine, the reset didn't occur. So I inserted a 1 microsecond delay into the read loop in NdisReadPciSlotInformation(). Suddenly, the reset was gone!! I'm still very puzzled by the whole thing. What I suspect is happening is that reading the PCI config space so quickly is causing a severe PCI bus error. My test system is a Sun w2100z dual Opteron system, and the NIC is a miniPCI card mounted in a miniPCI-to-PCI carrier card, plugged into a 100Mhz PCI slot. It's possible that this combination of hardware causes a bus protocol violation in this scenario which leads to a fatal machine check. This is pure speculation though. Really all I know for sure is that inserting the delay makes the problem go away. (To quote Homer Simpson: "I don't know how it works, but fire makes it good!") Problem number three: NdisAllocatePacket() needs to make sure to initialize the npp_validcounts field in the 'private' section of the NDIS_PACKET structure. The reason if_ndis was calling the MiniportReset() routine in the first place is that packet transmits were sometimes hanging. When sending a packet, an NDIS driver will call NdisQueryPacket() to learn how many physical buffers the packet resides in. NdisQueryPacket() is actually a macro, which traverses the NDIS_BUFFER list attached to the NDIS_PACKET and stashes some of the results in the 'private' section of the NDIS_PACKET. It also sets the npp_validcounts field to TRUE To indicate that the results are now valid. The problem is, now that if_ndis creates a pool of transmit packets via NdisAllocatePacketPool(), it's important that each time a new packet is allocated via NdisAllocatePacket() that validcounts be initialized to FALSE. If it isn't, and a previously transmitted NDIS_PACKET is pulled out of the pool, it may contain stale data from a previous transmission which won't get updated by NdisQueryPacket(). This would cause the driver to miscompute the number of fragments for a given packet, and botch the transmission. Fixing these three problems seems to make the Atheros driver happy on SMP, which hopefully means other serialized miniports will be happy too. And there was much rejoicing. Other stuff fixed along the way: - Modified ndis_thsuspend() to take a mutex as an argument. This allows KeWaitForSingleObject() and KeWaitForMultipleObjects() to avoid any possible race conditions with other routines that use the dispatcher lock. - Fixed KeCancelTimer() so that it returns the correct value for 'pending' according to the Microsoft documentation - Modfied NdisGetSystemUpTime() to use ticks and hz rather than calling nanouptime(). Also added comment that this routine wraps after 49.7 days. - Added macros for KeAcquireSpinLock()/KeReleaseSpinLock() to hide all the MSCALL() goop. - For x86, KeAcquireSpinLockRaiseToDpc() needs to be a separate function. This is because it's supposed to be _stdcall on the x86 arch, whereas KeAcquireSpinLock() is supposed to be _fastcall. On amd64, all routines use the same calling convention so we can just map KeAcquireSpinLockRaiseToDpc() directly to KfAcquireSpinLock() and it will work. (The _fastcall attribute is a no-op on amd64.) - Implement and use IoInitializeDpcRequest() and IoRequestDpc() (they're just macros) and use them for interrupt handling. This allows us to move the ndis_intrtask() routine from if_ndis.c to kern_ndis.c. - Fix the MmInitializeMdl() macro so that is uses sizeof(vm_offset_t) when computing mdl_size instead of uint32_t, so that it matches the MmSizeOfMdl() routine. - Change a could of M_WAITOKs to M_NOWAITs in the unicode routines in subr_ndis.c. - Use the dispatcher lock a little more consistently in subr_ntoskrnl.c. - Get rid of the "wait for link event" hack in ndis_init(). Now that I fixed NdisReadPciSlotInformation(), it seems I don't need it anymore. This should fix the witness panic a couple of people have reported. - Use MSCALL1() when calling the MiniportHangCheck() function in ndis_ticktask(). I accidentally missed this one when adding the wrapping for amd64.
2005-03-27 10:14:36 +00:00
/* Save the driver's funcptr and context */
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
- Rewrite the timer and event API routines in subr_ndis.c so that they are actually layered on top of the KeTimer API in subr_ntoskrnl.c, just as it is in Windows. This reduces code duplication and more closely imitates the way things are done in Windows. - Modify ndis_encode_parm() to deal with the case where we have a registry key expressed as a hex value ("0x1") which is being read via NdisReadConfiguration() as an int. Previously, we tried to decode things like "0x1" with strtol() using a base of 10, which would always yield 0. This is what was causing problems with the Intel 2200BG Centrino 802.11g driver: the .inf file that comes with it has a key called RadioEnable with a value of 0x1. We incorrectly decoded this value to '0' when it was queried, hence the driver thought we wanted the radio turned off. - In if_ndis.c, most drivers don't accept NDIS_80211_AUTHMODE_AUTO, but NDIS_80211_AUTHMODE_SHARED may not be right in some cases, so for now always use NDIS_80211_AUTHMODE_OPEN. NOTE: There is still one problem with the Intel 2200BG driver: it happens that the kernel stack in Windows is larger than the kernel stack in FreeBSD. The 2200BG driver sometimes eats up more than 2 pages of stack space, which can lead to a double fault panic. For the moment, I got things to work by adding the following to my kernel config file: options KSTACK_PAGES=8 I'm pretty sure 8 is too big; I just picked this value out of a hat as a test, and it happened to work, so I left it. 4 pages might be enough. Unfortunately, I don't think you can dynamically give a thread a larger stack, so I'm not sure how to handle this short of putting a note in the man page about it and dealing with the flood of mail from people who never read man pages.
2004-03-20 23:39:43 +00:00
timer->nmt_timerfunc = func;
timer->nmt_timerctx = ctx;
timer->nmt_block = handle;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
Fix an amd64 issue I overlooked. When setting up a callout to ndis_timercall() in NdisMInitializeTimer(), we can't use the raw function pointer. This is because ntoskrnl_run_dpc() expects to invoke a function with Microsoft calling conventions. On i386, this works because ndis_timercall() is declared with the __stdcall attribute, but this is a no-op on amd64. To do it correctly, we have to generate a wrapper for ndis_timercall() and us the wrapper instead of of the raw function pointer. Fix this by adding ndis_timercall() to the funcptr table in subr_ndis.c, and create ndis_findwrap() to extract the wrapped function from the table in NdisMInitializeTimer() instead of just passing ndis_timercall() to KeInitializeDpc() directly.
2005-03-31 16:38:48 +00:00
/*
* Set up the timer so it will call our intermediate DPC.
* Be sure to use the wrapped entry point, since
* ntoskrnl_run_dpc() expects to invoke a function with
* Microsoft calling conventions.
*/
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
KeInitializeTimer(&timer->nmt_ktimer);
Fix an amd64 issue I overlooked. When setting up a callout to ndis_timercall() in NdisMInitializeTimer(), we can't use the raw function pointer. This is because ntoskrnl_run_dpc() expects to invoke a function with Microsoft calling conventions. On i386, this works because ndis_timercall() is declared with the __stdcall attribute, but this is a no-op on amd64. To do it correctly, we have to generate a wrapper for ndis_timercall() and us the wrapper instead of of the raw function pointer. Fix this by adding ndis_timercall() to the funcptr table in subr_ndis.c, and create ndis_findwrap() to extract the wrapped function from the table in NdisMInitializeTimer() instead of just passing ndis_timercall() to KeInitializeDpc() directly.
2005-03-31 16:38:48 +00:00
KeInitializeDpc(&timer->nmt_kdpc,
ndis_findwrap((funcptr)ndis_timercall), timer);
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
timer->nmt_ktimer.k_dpc = &timer->nmt_kdpc;
KeAcquireSpinLock(&timer->nmt_block->nmb_lock, &irql);
timer->nmt_nexttimer = timer->nmt_block->nmb_timerlist;
timer->nmt_block->nmb_timerlist = timer;
KeReleaseSpinLock(&timer->nmt_block->nmb_lock, irql);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
/*
- Rewrite the timer and event API routines in subr_ndis.c so that they are actually layered on top of the KeTimer API in subr_ntoskrnl.c, just as it is in Windows. This reduces code duplication and more closely imitates the way things are done in Windows. - Modify ndis_encode_parm() to deal with the case where we have a registry key expressed as a hex value ("0x1") which is being read via NdisReadConfiguration() as an int. Previously, we tried to decode things like "0x1" with strtol() using a base of 10, which would always yield 0. This is what was causing problems with the Intel 2200BG Centrino 802.11g driver: the .inf file that comes with it has a key called RadioEnable with a value of 0x1. We incorrectly decoded this value to '0' when it was queried, hence the driver thought we wanted the radio turned off. - In if_ndis.c, most drivers don't accept NDIS_80211_AUTHMODE_AUTO, but NDIS_80211_AUTHMODE_SHARED may not be right in some cases, so for now always use NDIS_80211_AUTHMODE_OPEN. NOTE: There is still one problem with the Intel 2200BG driver: it happens that the kernel stack in Windows is larger than the kernel stack in FreeBSD. The 2200BG driver sometimes eats up more than 2 pages of stack space, which can lead to a double fault panic. For the moment, I got things to work by adding the following to my kernel config file: options KSTACK_PAGES=8 I'm pretty sure 8 is too big; I just picked this value out of a hat as a test, and it happened to work, so I left it. 4 pages might be enough. Unfortunately, I don't think you can dynamically give a thread a larger stack, so I'm not sure how to handle this short of putting a note in the man page about it and dealing with the flood of mail from people who never read man pages.
2004-03-20 23:39:43 +00:00
* In Windows, there's both an NdisMSetTimer() and an NdisSetTimer(),
* but the former is just a macro wrapper around the latter.
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
*/
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisSetTimer(timer, msecs)
- Rewrite the timer and event API routines in subr_ndis.c so that they are actually layered on top of the KeTimer API in subr_ntoskrnl.c, just as it is in Windows. This reduces code duplication and more closely imitates the way things are done in Windows. - Modify ndis_encode_parm() to deal with the case where we have a registry key expressed as a hex value ("0x1") which is being read via NdisReadConfiguration() as an int. Previously, we tried to decode things like "0x1" with strtol() using a base of 10, which would always yield 0. This is what was causing problems with the Intel 2200BG Centrino 802.11g driver: the .inf file that comes with it has a key called RadioEnable with a value of 0x1. We incorrectly decoded this value to '0' when it was queried, hence the driver thought we wanted the radio turned off. - In if_ndis.c, most drivers don't accept NDIS_80211_AUTHMODE_AUTO, but NDIS_80211_AUTHMODE_SHARED may not be right in some cases, so for now always use NDIS_80211_AUTHMODE_OPEN. NOTE: There is still one problem with the Intel 2200BG driver: it happens that the kernel stack in Windows is larger than the kernel stack in FreeBSD. The 2200BG driver sometimes eats up more than 2 pages of stack space, which can lead to a double fault panic. For the moment, I got things to work by adding the following to my kernel config file: options KSTACK_PAGES=8 I'm pretty sure 8 is too big; I just picked this value out of a hat as a test, and it happened to work, so I left it. 4 pages might be enough. Unfortunately, I don't think you can dynamically give a thread a larger stack, so I'm not sure how to handle this short of putting a note in the man page about it and dealing with the flood of mail from people who never read man pages.
2004-03-20 23:39:43 +00:00
ndis_timer *timer;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
uint32_t msecs;
{
- Rewrite the timer and event API routines in subr_ndis.c so that they are actually layered on top of the KeTimer API in subr_ntoskrnl.c, just as it is in Windows. This reduces code duplication and more closely imitates the way things are done in Windows. - Modify ndis_encode_parm() to deal with the case where we have a registry key expressed as a hex value ("0x1") which is being read via NdisReadConfiguration() as an int. Previously, we tried to decode things like "0x1" with strtol() using a base of 10, which would always yield 0. This is what was causing problems with the Intel 2200BG Centrino 802.11g driver: the .inf file that comes with it has a key called RadioEnable with a value of 0x1. We incorrectly decoded this value to '0' when it was queried, hence the driver thought we wanted the radio turned off. - In if_ndis.c, most drivers don't accept NDIS_80211_AUTHMODE_AUTO, but NDIS_80211_AUTHMODE_SHARED may not be right in some cases, so for now always use NDIS_80211_AUTHMODE_OPEN. NOTE: There is still one problem with the Intel 2200BG driver: it happens that the kernel stack in Windows is larger than the kernel stack in FreeBSD. The 2200BG driver sometimes eats up more than 2 pages of stack space, which can lead to a double fault panic. For the moment, I got things to work by adding the following to my kernel config file: options KSTACK_PAGES=8 I'm pretty sure 8 is too big; I just picked this value out of a hat as a test, and it happened to work, so I left it. 4 pages might be enough. Unfortunately, I don't think you can dynamically give a thread a larger stack, so I'm not sure how to handle this short of putting a note in the man page about it and dealing with the flood of mail from people who never read man pages.
2004-03-20 23:39:43 +00:00
/*
* KeSetTimer() wants the period in
* hundred nanosecond intervals.
*/
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
KeSetTimer(&timer->nt_ktimer,
- Rewrite the timer and event API routines in subr_ndis.c so that they are actually layered on top of the KeTimer API in subr_ntoskrnl.c, just as it is in Windows. This reduces code duplication and more closely imitates the way things are done in Windows. - Modify ndis_encode_parm() to deal with the case where we have a registry key expressed as a hex value ("0x1") which is being read via NdisReadConfiguration() as an int. Previously, we tried to decode things like "0x1" with strtol() using a base of 10, which would always yield 0. This is what was causing problems with the Intel 2200BG Centrino 802.11g driver: the .inf file that comes with it has a key called RadioEnable with a value of 0x1. We incorrectly decoded this value to '0' when it was queried, hence the driver thought we wanted the radio turned off. - In if_ndis.c, most drivers don't accept NDIS_80211_AUTHMODE_AUTO, but NDIS_80211_AUTHMODE_SHARED may not be right in some cases, so for now always use NDIS_80211_AUTHMODE_OPEN. NOTE: There is still one problem with the Intel 2200BG driver: it happens that the kernel stack in Windows is larger than the kernel stack in FreeBSD. The 2200BG driver sometimes eats up more than 2 pages of stack space, which can lead to a double fault panic. For the moment, I got things to work by adding the following to my kernel config file: options KSTACK_PAGES=8 I'm pretty sure 8 is too big; I just picked this value out of a hat as a test, and it happened to work, so I left it. 4 pages might be enough. Unfortunately, I don't think you can dynamically give a thread a larger stack, so I'm not sure how to handle this short of putting a note in the man page about it and dealing with the flood of mail from people who never read man pages.
2004-03-20 23:39:43 +00:00
((int64_t)msecs * -10000), &timer->nt_kdpc);
Fix several issues related to the KeInitializeTimer() etc... API stuff that I added recently: - When a periodic timer fires, it's automatically re-armed. We must make sure to re-arm the timer _before_ invoking any caller-supplied defered procedure call: the DPC may choose to call KeCancelTimer(), and re-arming the timer after the DPC un-does the effect of the cancel. - Fix similar issue with periodic timers in subr_ndis.c. - When calling KeSetTimer() or KeSetTimerEx(), if the timer is already pending, untimeout() it first before timeout()ing it again. - The old Atheros driver for the 5211 seems to use KeSetTimerEx() incorrectly, or at the very least in a very strange way that doesn't quite follow the Microsoft documentation. In one case, it calls KeSetTimerEx() with a duetime of 0 and a period of 5000. The Microsoft documentation says that negative duetime values are relative to the current time and positive values are absolute. But it doesn't say what's supposed to happen with positive values that less than the current time, i.e. absolute values that are in the past. Lacking any further information, I have decided that timers with positive duetimes that are in the past should fire right away (or in our case, after only 1 tick). This also takes care of the other strange usage in the Atheros driver, where the duetime is specified as 500000 and the period is 50. I think someone may have meant to use -500000 and misinterpreted the documentation. - Also modified KeWaitForSingleObject() and KeWaitForMultipleObjects() to make the same duetime adjustment, since they have the same rules regarding timeout values. - Cosmetic: change name of 'timeout' variable in KeWaitForSingleObject() and KeWaitForMultipleObjects() to 'duetime' to avoid senseless (though harmless) overlap with timeout() function name. With these fixes, I can get the 5211 card to associate properly with my adhoc net using driver AR5211.SYS version 2.4.1.6.
2004-03-10 07:43:11 +00:00
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMSetPeriodicTimer(timer, msecs)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_miniport_timer *timer;
uint32_t msecs;
{
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
KeSetTimerEx(&timer->nmt_ktimer,
- Rewrite the timer and event API routines in subr_ndis.c so that they are actually layered on top of the KeTimer API in subr_ntoskrnl.c, just as it is in Windows. This reduces code duplication and more closely imitates the way things are done in Windows. - Modify ndis_encode_parm() to deal with the case where we have a registry key expressed as a hex value ("0x1") which is being read via NdisReadConfiguration() as an int. Previously, we tried to decode things like "0x1" with strtol() using a base of 10, which would always yield 0. This is what was causing problems with the Intel 2200BG Centrino 802.11g driver: the .inf file that comes with it has a key called RadioEnable with a value of 0x1. We incorrectly decoded this value to '0' when it was queried, hence the driver thought we wanted the radio turned off. - In if_ndis.c, most drivers don't accept NDIS_80211_AUTHMODE_AUTO, but NDIS_80211_AUTHMODE_SHARED may not be right in some cases, so for now always use NDIS_80211_AUTHMODE_OPEN. NOTE: There is still one problem with the Intel 2200BG driver: it happens that the kernel stack in Windows is larger than the kernel stack in FreeBSD. The 2200BG driver sometimes eats up more than 2 pages of stack space, which can lead to a double fault panic. For the moment, I got things to work by adding the following to my kernel config file: options KSTACK_PAGES=8 I'm pretty sure 8 is too big; I just picked this value out of a hat as a test, and it happened to work, so I left it. 4 pages might be enough. Unfortunately, I don't think you can dynamically give a thread a larger stack, so I'm not sure how to handle this short of putting a note in the man page about it and dealing with the flood of mail from people who never read man pages.
2004-03-20 23:39:43 +00:00
((int64_t)msecs * -10000), msecs, &timer->nmt_kdpc);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
- Rewrite the timer and event API routines in subr_ndis.c so that they are actually layered on top of the KeTimer API in subr_ntoskrnl.c, just as it is in Windows. This reduces code duplication and more closely imitates the way things are done in Windows. - Modify ndis_encode_parm() to deal with the case where we have a registry key expressed as a hex value ("0x1") which is being read via NdisReadConfiguration() as an int. Previously, we tried to decode things like "0x1" with strtol() using a base of 10, which would always yield 0. This is what was causing problems with the Intel 2200BG Centrino 802.11g driver: the .inf file that comes with it has a key called RadioEnable with a value of 0x1. We incorrectly decoded this value to '0' when it was queried, hence the driver thought we wanted the radio turned off. - In if_ndis.c, most drivers don't accept NDIS_80211_AUTHMODE_AUTO, but NDIS_80211_AUTHMODE_SHARED may not be right in some cases, so for now always use NDIS_80211_AUTHMODE_OPEN. NOTE: There is still one problem with the Intel 2200BG driver: it happens that the kernel stack in Windows is larger than the kernel stack in FreeBSD. The 2200BG driver sometimes eats up more than 2 pages of stack space, which can lead to a double fault panic. For the moment, I got things to work by adding the following to my kernel config file: options KSTACK_PAGES=8 I'm pretty sure 8 is too big; I just picked this value out of a hat as a test, and it happened to work, so I left it. 4 pages might be enough. Unfortunately, I don't think you can dynamically give a thread a larger stack, so I'm not sure how to handle this short of putting a note in the man page about it and dealing with the flood of mail from people who never read man pages.
2004-03-20 23:39:43 +00:00
/*
* Technically, this is really NdisCancelTimer(), but we also
* (ab)use it for NdisMCancelTimer(), since in our implementation
* we don't need the extra info in the ndis_miniport_timer
Finally bring an end to the great "make the Atheros NDIS driver work on SMP" saga. After several weeks and much gnashing of teeth, I have finally tracked down all the problems, despite their best efforts to confound and annoy me. Problem nunmber one: the Atheros windows driver is _NOT_ a de-serialized miniport! It used to be that NDIS drivers relied on the NDIS library itself for all their locking and serialization needs. Transmit packet queues were all handled internally by NDIS, and all calls to MiniportXXX() routines were guaranteed to be appropriately serialized. This proved to be a performance problem however, and Microsoft introduced de-serialized miniports with the NDIS 5.x spec. Microsoft still supports serialized miniports, but recommends that all new drivers written for Windows XP and later be deserialized. Apparently Atheros wasn't listening when they said this. This means (among other things) that we have to serialize calls to MiniportSendPackets(). We also have to serialize calls to MiniportTimer() that are triggered via the NdisMInitializeTimer() routine. It finally dawned on me why NdisMInitializeTimer() takes a special NDIS_MINIPORT_TIMER structure and a pointer to the miniport block: the timer callback must be serialized, and it's only by saving the miniport block handle that we can get access to the serialization lock during the timer callback. Problem number two: haunted hardware. The thing that was _really_ driving me absolutely bonkers for the longest time is that, for some reason I couldn't understand, my test machine would occasionally freeze or more frustratingly, reset completely. That's reset and in *pow!* back to the BIOS startup. No panic, no crashdump, just a reset. This appeared to happen most often when MiniportReset() was called. (As to why MiniportReset() was being called, see problem three below.) I thought maybe I had created some sort of horrible deadlock condition in the process of adding the serialization, but after three weeks, at least 6 different locking implementations and heroic efforts to debug the spinlock code, the machine still kept resetting. Finally, I started single stepping through the MiniportReset() routine in the driver using the kernel debugger, and this ultimately led me to the source of the problem. One of the last things the Atheros MiniportReset() routine does is call NdisReadPciSlotInformation() several times to inspect a portion of the device's PCI config space. It reads the same chunk of config space repeatedly, in rapid succession. Presumeably, it's polling the hardware for some sort of event. The reset occurs partway through this process. I discovered that when I single-stepped through this portion of the routine, the reset didn't occur. So I inserted a 1 microsecond delay into the read loop in NdisReadPciSlotInformation(). Suddenly, the reset was gone!! I'm still very puzzled by the whole thing. What I suspect is happening is that reading the PCI config space so quickly is causing a severe PCI bus error. My test system is a Sun w2100z dual Opteron system, and the NIC is a miniPCI card mounted in a miniPCI-to-PCI carrier card, plugged into a 100Mhz PCI slot. It's possible that this combination of hardware causes a bus protocol violation in this scenario which leads to a fatal machine check. This is pure speculation though. Really all I know for sure is that inserting the delay makes the problem go away. (To quote Homer Simpson: "I don't know how it works, but fire makes it good!") Problem number three: NdisAllocatePacket() needs to make sure to initialize the npp_validcounts field in the 'private' section of the NDIS_PACKET structure. The reason if_ndis was calling the MiniportReset() routine in the first place is that packet transmits were sometimes hanging. When sending a packet, an NDIS driver will call NdisQueryPacket() to learn how many physical buffers the packet resides in. NdisQueryPacket() is actually a macro, which traverses the NDIS_BUFFER list attached to the NDIS_PACKET and stashes some of the results in the 'private' section of the NDIS_PACKET. It also sets the npp_validcounts field to TRUE To indicate that the results are now valid. The problem is, now that if_ndis creates a pool of transmit packets via NdisAllocatePacketPool(), it's important that each time a new packet is allocated via NdisAllocatePacket() that validcounts be initialized to FALSE. If it isn't, and a previously transmitted NDIS_PACKET is pulled out of the pool, it may contain stale data from a previous transmission which won't get updated by NdisQueryPacket(). This would cause the driver to miscompute the number of fragments for a given packet, and botch the transmission. Fixing these three problems seems to make the Atheros driver happy on SMP, which hopefully means other serialized miniports will be happy too. And there was much rejoicing. Other stuff fixed along the way: - Modified ndis_thsuspend() to take a mutex as an argument. This allows KeWaitForSingleObject() and KeWaitForMultipleObjects() to avoid any possible race conditions with other routines that use the dispatcher lock. - Fixed KeCancelTimer() so that it returns the correct value for 'pending' according to the Microsoft documentation - Modfied NdisGetSystemUpTime() to use ticks and hz rather than calling nanouptime(). Also added comment that this routine wraps after 49.7 days. - Added macros for KeAcquireSpinLock()/KeReleaseSpinLock() to hide all the MSCALL() goop. - For x86, KeAcquireSpinLockRaiseToDpc() needs to be a separate function. This is because it's supposed to be _stdcall on the x86 arch, whereas KeAcquireSpinLock() is supposed to be _fastcall. On amd64, all routines use the same calling convention so we can just map KeAcquireSpinLockRaiseToDpc() directly to KfAcquireSpinLock() and it will work. (The _fastcall attribute is a no-op on amd64.) - Implement and use IoInitializeDpcRequest() and IoRequestDpc() (they're just macros) and use them for interrupt handling. This allows us to move the ndis_intrtask() routine from if_ndis.c to kern_ndis.c. - Fix the MmInitializeMdl() macro so that is uses sizeof(vm_offset_t) when computing mdl_size instead of uint32_t, so that it matches the MmSizeOfMdl() routine. - Change a could of M_WAITOKs to M_NOWAITs in the unicode routines in subr_ndis.c. - Use the dispatcher lock a little more consistently in subr_ntoskrnl.c. - Get rid of the "wait for link event" hack in ndis_init(). Now that I fixed NdisReadPciSlotInformation(), it seems I don't need it anymore. This should fix the witness panic a couple of people have reported. - Use MSCALL1() when calling the MiniportHangCheck() function in ndis_ticktask(). I accidentally missed this one when adding the wrapping for amd64.
2005-03-27 10:14:36 +00:00
* structure just to cancel a timer.
- Rewrite the timer and event API routines in subr_ndis.c so that they are actually layered on top of the KeTimer API in subr_ntoskrnl.c, just as it is in Windows. This reduces code duplication and more closely imitates the way things are done in Windows. - Modify ndis_encode_parm() to deal with the case where we have a registry key expressed as a hex value ("0x1") which is being read via NdisReadConfiguration() as an int. Previously, we tried to decode things like "0x1" with strtol() using a base of 10, which would always yield 0. This is what was causing problems with the Intel 2200BG Centrino 802.11g driver: the .inf file that comes with it has a key called RadioEnable with a value of 0x1. We incorrectly decoded this value to '0' when it was queried, hence the driver thought we wanted the radio turned off. - In if_ndis.c, most drivers don't accept NDIS_80211_AUTHMODE_AUTO, but NDIS_80211_AUTHMODE_SHARED may not be right in some cases, so for now always use NDIS_80211_AUTHMODE_OPEN. NOTE: There is still one problem with the Intel 2200BG driver: it happens that the kernel stack in Windows is larger than the kernel stack in FreeBSD. The 2200BG driver sometimes eats up more than 2 pages of stack space, which can lead to a double fault panic. For the moment, I got things to work by adding the following to my kernel config file: options KSTACK_PAGES=8 I'm pretty sure 8 is too big; I just picked this value out of a hat as a test, and it happened to work, so I left it. 4 pages might be enough. Unfortunately, I don't think you can dynamically give a thread a larger stack, so I'm not sure how to handle this short of putting a note in the man page about it and dealing with the flood of mail from people who never read man pages.
2004-03-20 23:39:43 +00:00
*/
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMCancelTimer(timer, cancelled)
- Rewrite the timer and event API routines in subr_ndis.c so that they are actually layered on top of the KeTimer API in subr_ntoskrnl.c, just as it is in Windows. This reduces code duplication and more closely imitates the way things are done in Windows. - Modify ndis_encode_parm() to deal with the case where we have a registry key expressed as a hex value ("0x1") which is being read via NdisReadConfiguration() as an int. Previously, we tried to decode things like "0x1" with strtol() using a base of 10, which would always yield 0. This is what was causing problems with the Intel 2200BG Centrino 802.11g driver: the .inf file that comes with it has a key called RadioEnable with a value of 0x1. We incorrectly decoded this value to '0' when it was queried, hence the driver thought we wanted the radio turned off. - In if_ndis.c, most drivers don't accept NDIS_80211_AUTHMODE_AUTO, but NDIS_80211_AUTHMODE_SHARED may not be right in some cases, so for now always use NDIS_80211_AUTHMODE_OPEN. NOTE: There is still one problem with the Intel 2200BG driver: it happens that the kernel stack in Windows is larger than the kernel stack in FreeBSD. The 2200BG driver sometimes eats up more than 2 pages of stack space, which can lead to a double fault panic. For the moment, I got things to work by adding the following to my kernel config file: options KSTACK_PAGES=8 I'm pretty sure 8 is too big; I just picked this value out of a hat as a test, and it happened to work, so I left it. 4 pages might be enough. Unfortunately, I don't think you can dynamically give a thread a larger stack, so I'm not sure how to handle this short of putting a note in the man page about it and dealing with the flood of mail from people who never read man pages.
2004-03-20 23:39:43 +00:00
ndis_timer *timer;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
uint8_t *cancelled;
{
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
*cancelled = KeCancelTimer(&timer->nt_ktimer);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMQueryAdapterResources(status, adapter, list, buflen)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_status *status;
ndis_handle adapter;
ndis_resource_list *list;
uint32_t *buflen;
{
ndis_miniport_block *block;
struct ndis_softc *sc;
Tweak ndiscvt to support yet another flavor of .INF files (look for the NTx86 section decoration). subr_ndis.c: correct the behavior of ndis_query_resources(): if the caller doesn't provide enough space to return the resources, tell it how much it needs to provide and return an error. subr_hal.c & subr_ntoskrnl.c: implement/stub a bunch of new routines; ntoskrnl: KefAcquireSpinLockAtDpcLevel KefReleaseSpinLockFromDpcLevel MmMapLockedPages InterlockedDecrement InterlockedIncrement IoFreeMdl KeInitializeSpinLock HAL: KfReleaseSpinLock KeGetCurrentIrql KfAcquireSpinLock Lastly, correct spelling of "_aullshr" in the ntoskrnl functable.
2004-01-03 02:25:21 +00:00
int rsclen;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
block = (ndis_miniport_block *)adapter;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
sc = device_get_softc(block->nmb_physdeviceobj->do_devext);
Tweak ndiscvt to support yet another flavor of .INF files (look for the NTx86 section decoration). subr_ndis.c: correct the behavior of ndis_query_resources(): if the caller doesn't provide enough space to return the resources, tell it how much it needs to provide and return an error. subr_hal.c & subr_ntoskrnl.c: implement/stub a bunch of new routines; ntoskrnl: KefAcquireSpinLockAtDpcLevel KefReleaseSpinLockFromDpcLevel MmMapLockedPages InterlockedDecrement InterlockedIncrement IoFreeMdl KeInitializeSpinLock HAL: KfReleaseSpinLock KeGetCurrentIrql KfAcquireSpinLock Lastly, correct spelling of "_aullshr" in the ntoskrnl functable.
2004-01-03 02:25:21 +00:00
rsclen = sizeof(ndis_resource_list) +
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
(sizeof(cm_partial_resource_desc) * (sc->ndis_rescnt - 1));
Tweak ndiscvt to support yet another flavor of .INF files (look for the NTx86 section decoration). subr_ndis.c: correct the behavior of ndis_query_resources(): if the caller doesn't provide enough space to return the resources, tell it how much it needs to provide and return an error. subr_hal.c & subr_ntoskrnl.c: implement/stub a bunch of new routines; ntoskrnl: KefAcquireSpinLockAtDpcLevel KefReleaseSpinLockFromDpcLevel MmMapLockedPages InterlockedDecrement InterlockedIncrement IoFreeMdl KeInitializeSpinLock HAL: KfReleaseSpinLock KeGetCurrentIrql KfAcquireSpinLock Lastly, correct spelling of "_aullshr" in the ntoskrnl functable.
2004-01-03 02:25:21 +00:00
if (*buflen < rsclen) {
*buflen = rsclen;
*status = NDIS_STATUS_INVALID_LENGTH;
return;
}
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
- The MiniportReset() function can return NDIS_STATUS_PENDING, in which case we should wait for the resetdone handler to be called before returning. - When providing resources via ndis_query_resources(), uses the computed rsclen when using bcopy() to copy out the resource data rather than the caller-supplied buffer length. - Avoid using ndis_reset_nic() in if_ndis.c unless we really need to reset the NIC because of a problem. - Allow interrupts to be fielded during ndis_attach(), at least as far as allowing ndis_isr() and ndis_intrhand() to run. - Use ndis_80211_rates_ex when probing for supported rates. Technically, this isn't supposed to work since, although Microsoft added the extended rate structure with the NDIS 5.1 update, the spec still says that the OID_802_11_SUPPORTED_RATES OID uses ndis_80211_rates. In spite of this, it appears some drivers use it anyway. - When adding in our guessed rates, check to see if they already exist so that we avoid any duplicates. - Add a printf() to ndis_open_file() that alerts the user when a driver attempts to open a file under /compat/ndis. With these changes, I can get the driver for the SMC 2802W 54g PCI card to load and run. This board uses a Prism54G chip. Note that in order for this driver to work, you must place the supplied smc2802w.arm firmware image under /compat/ndis. (The firmware is not resident on the device.) Note that this should also allow the 3Com 3CRWE154G72 card to work as well; as far as I can tell, these cards also use a Prism54G chip.
2004-04-05 08:26:52 +00:00
bcopy((char *)block->nmb_rlist, (char *)list, rsclen);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
*status = NDIS_STATUS_SUCCESS;
Add support for Windows/x86-64 binaries to Project Evil. Ville-Pertti Keinonen (will at exomi dot comohmygodnospampleasekthx) deserves a big thanks for submitting initial patches to make it work. I have mangled his contributions appropriately. The main gotcha with Windows/x86-64 is that Microsoft uses a different calling convention than everyone else. The standard ABI requires using 6 registers for argument passing, with other arguments on the stack. Microsoft uses only 4 registers, and requires the caller to leave room on the stack for the register arguments incase the callee needs to spill them. Unlike x86, where Microsoft uses a mix of _cdecl, _stdcall and _fastcall, all routines on Windows/x86-64 uses the same convention. This unfortunately means that all the functions we export to the driver require an intermediate translation wrapper. Similarly, we have to wrap all calls back into the driver binary itself. The original patches provided macros to wrap every single routine at compile time, providing a secondary jump table with a customized wrapper for each exported routine. I decided to use a different approach: the call wrapper for each function is created from a template at runtime, and the routine to jump to is patched into the wrapper as it is created. The subr_pe module has been modified to patch in the wrapped function instead of the original. (On x86, the wrapping routine is a no-op.) There are some minor API differences that had to be accounted for: - KeAcquireSpinLock() is a real function on amd64, not a macro wrapper around KfAcquireSpinLock() - NdisFreeBuffer() is actually IoFreeMdl(). I had to change the whole NDIS_BUFFER API a bit to accomodate this. Bugs fixed along the way: - IoAllocateMdl() always returned NULL - kern_windrv.c:windrv_unload() wasn't releasing private driver object extensions correctly (found thanks to memguard) This has only been tested with the driver for the Broadcom 802.11g chipset, which was the only Windows/x86-64 driver I could find.
2005-02-16 05:41:18 +00:00
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMRegisterIoPortRange(offset, adapter, port, numports)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
void **offset;
ndis_handle adapter;
uint32_t port;
uint32_t numports;
{
struct ndis_miniport_block *block;
struct ndis_softc *sc;
if (adapter == NULL)
return(NDIS_STATUS_FAILURE);
block = (ndis_miniport_block *)adapter;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
sc = device_get_softc(block->nmb_physdeviceobj->do_devext);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
if (sc->ndis_res_io == NULL)
return(NDIS_STATUS_FAILURE);
Loosen up the range test in ndis_register_ioport(). Allow drivers to map ranges that are smaller than what our resource manager code knows is available, rather than requiring that they match exactly. This fixes a problem with the Intel PRO/1000 gigE driver: it wants to map a range of 32 I/O ports, even though some chips appear set up to decode a range of 64. With this fix, it loads and runs correctly.
2004-01-13 01:12:28 +00:00
/* Don't let the device map more ports than we have. */
if (rman_get_size(sc->ndis_res_io) < numports)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return(NDIS_STATUS_INVALID_LENGTH);
*offset = (void *)rman_get_start(sc->ndis_res_io);
return(NDIS_STATUS_SUCCESS);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMDeregisterIoPortRange(adapter, port, numports, offset)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_handle adapter;
uint32_t port;
uint32_t numports;
void *offset;
{
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisReadNetworkAddress(status, addr, addrlen, adapter)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_status *status;
void **addr;
uint32_t *addrlen;
ndis_handle adapter;
{
struct ndis_softc *sc;
ndis_miniport_block *block;
uint8_t empty[] = { 0, 0, 0, 0, 0, 0 };
block = (ndis_miniport_block *)adapter;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
sc = device_get_softc(block->nmb_physdeviceobj->do_devext);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
#ifdef IFP2ENADDR
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
if (bcmp(IFP2ENADDR(sc->ifp), empty, ETHER_ADDR_LEN) == 0)
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
#else
if (bcmp(sc->arpcom.ac_enaddr, empty, ETHER_ADDR_LEN) == 0)
#endif
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
*status = NDIS_STATUS_FAILURE;
else {
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
#ifdef IFP2ENADDR
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
*addr = IFP2ENADDR(sc->ifp);
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
#else
*addr = sc->arpcom.ac_enaddr;
#endif
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
*addrlen = ETHER_ADDR_LEN;
*status = NDIS_STATUS_SUCCESS;
}
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisQueryMapRegisterCount(bustype, cnt)
Back out the last batch of changes until I have a chance to properly evaluate them. Whatever they're meant to do, they're doing it wrong. Also: - Clean up last bits of NULL fallout in subr_pe - Don't let ndis_ifmedia_sts() do anything if the IFF_UP flag isn't set - Implement NdisSystemProcessorCount() and NdisQueryMapRegisterCount().
2003-12-26 03:31:34 +00:00
uint32_t bustype;
uint32_t *cnt;
{
subr_hal.c: implement WRITE_PORT_BUFFER_xxx() and READ_PORT_BUFFER_xxx() subr_ndis.c: implement NdisDprAllocatePacket() and NdisDprFreePacket() (which are aliased to NdisAllocatePacket() and NdisFreePacket()), and bump the value we return in ndis_mapreg_cnt() to something ridiculously large, since some drivers apparently expect to be able to allocate way more than just 64. These changes allow the Level 1 1000baseSX driver to work for the following card: ndis0: <SMC TigerCard 1000 Adapter> port 0xe000-0xe0ff mem 0xda004000-0xda0043ff irq 10 at device 9.0 on pci0 ndis0: Ethernet address: 00:e0:29:6f:cc:04 This is already supported by the lge(4) driver, but I decided to take a try at making the Windows driver that came with it work too, since I still had the floppy diskette for it lying around.
2004-01-03 09:20:48 +00:00
*cnt = 8192;
Back out the last batch of changes until I have a chance to properly evaluate them. Whatever they're meant to do, they're doing it wrong. Also: - Clean up last bits of NULL fallout in subr_pe - Don't let ndis_ifmedia_sts() do anything if the IFF_UP flag isn't set - Implement NdisSystemProcessorCount() and NdisQueryMapRegisterCount().
2003-12-26 03:31:34 +00:00
return(NDIS_STATUS_SUCCESS);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMAllocateMapRegisters(adapter, dmachannel, dmasize, physmapneeded, maxmap)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_handle adapter;
uint32_t dmachannel;
uint8_t dmasize;
uint32_t physmapneeded;
uint32_t maxmap;
{
struct ndis_softc *sc;
ndis_miniport_block *block;
int error, i, nseg = NDIS_MAXSEG;
block = (ndis_miniport_block *)adapter;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
sc = device_get_softc(block->nmb_physdeviceobj->do_devext);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
sc->ndis_mmaps = malloc(sizeof(bus_dmamap_t) * physmapneeded,
M_DEVBUF, M_NOWAIT|M_ZERO);
if (sc->ndis_mmaps == NULL)
return(NDIS_STATUS_RESOURCES);
error = bus_dma_tag_create(sc->ndis_parent_tag, ETHER_ALIGN, 0,
BUS_SPACE_MAXADDR_32BIT, BUS_SPACE_MAXADDR, NULL,
NULL, maxmap * nseg, nseg, maxmap, BUS_DMA_ALLOCNOW,
NULL, NULL, &sc->ndis_mtag);
if (error) {
free(sc->ndis_mmaps, M_DEVBUF);
return(NDIS_STATUS_RESOURCES);
}
for (i = 0; i < physmapneeded; i++)
bus_dmamap_create(sc->ndis_mtag, 0, &sc->ndis_mmaps[i]);
sc->ndis_mmapcnt = physmapneeded;
return(NDIS_STATUS_SUCCESS);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMFreeMapRegisters(adapter)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_handle adapter;
{
struct ndis_softc *sc;
ndis_miniport_block *block;
int i;
block = (ndis_miniport_block *)adapter;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
sc = device_get_softc(block->nmb_physdeviceobj->do_devext);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
for (i = 0; i < sc->ndis_mmapcnt; i++)
bus_dmamap_destroy(sc->ndis_mtag, sc->ndis_mmaps[i]);
free(sc->ndis_mmaps, M_DEVBUF);
bus_dma_tag_destroy(sc->ndis_mtag);
return;
}
static void
ndis_mapshared_cb(arg, segs, nseg, error)
void *arg;
bus_dma_segment_t *segs;
int nseg;
int error;
{
ndis_physaddr *p;
if (error || nseg > 1)
return;
p = arg;
p->np_quad = segs[0].ds_addr;
return;
}
/*
* This maps to bus_dmamem_alloc().
*/
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMAllocateSharedMemory(adapter, len, cached, vaddr, paddr)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_handle adapter;
uint32_t len;
uint8_t cached;
void **vaddr;
ndis_physaddr *paddr;
{
ndis_miniport_block *block;
struct ndis_softc *sc;
struct ndis_shmem *sh;
int error;
if (adapter == NULL)
return;
block = (ndis_miniport_block *)adapter;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
sc = device_get_softc(block->nmb_physdeviceobj->do_devext);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
sh = malloc(sizeof(struct ndis_shmem), M_DEVBUF, M_NOWAIT|M_ZERO);
if (sh == NULL)
return;
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
InitializeListHead(&sh->ndis_list);
Fix two problems: - In subr_ndis.c:ndis_allocate_sharemem(), create the busdma tags used for shared memory allocations with a lowaddr of 0x3E7FFFFF. This forces the buffers to be mapped to physical/bus addresses within the first 1GB of physical memory. It seems that at least one card (Linksys Instant Wireless PCI V2.7) depends on this behavior. I don't know if this is a hardware restriction, or if the NDIS driver for this card is truncating the addresses itself, but using physical/bus addresses beyong the 1GB limit causes initialization failures. - Create am NDIS_INITIALIZED() macro in if_ndisvar.h and use it in if_ndis.c to test whether the device has been initialized rather than checking for the presence of the IFF_UP flag in if_flags. While debugging the previous problem, I noticed that bringing up the device would always produce failures from ndis_setmulti(). It turns out that the following steps now occur during device initialization: - IFF_UP flag is set in if_flags - ifp->if_ioctl() called with SIOCSIFADDR (which we don't handle) - ifp->if_ioctl() called with SIOCADDMULTI - ifp->if_ioctl() called with SIOCADDMULTI (again) - ifp->if_ioctl() called with SIOCADDMULTI (yet again) - ifp->if_ioctl() called with SIOCSIFFLAGS Setting the receive filter and multicast filters can only be done when the underlying NDIS driver has been initialized, which is done by ifp->if_init(). However, we don't call ifp->if_init() until ifp->if_ioctl() is called with SIOCSIFFLAGS and IFF_UP has been set. It appears that now, the network stack tries to add multicast addresses to interface's filter before those steps occur. Normally, ndis_setmulti() would trap this condition by checking for the IFF_UP flag, but the network code has in fact set this flag already, so ndis_setmulti() is fooled into thinking the interface has been initialized when it really hasn't. It turns out this is usually harmless because the ifp->if_init() routine (in this case ndis_init()) will set up the multicast filter when it initializes the hardware anyway, and the underlying routines (ndis_get_info()/ndis_set_info()) know that the driver/NIC haven't been initialized yet, but you end up spurious error messages on the console all the time. Something tells me this new behavior isn't really correct. I think the intention was to fix it so that ifp->if_init() is only called once when we ifconfig an interface up, but the end result seems a little bogus: the change of the IFF_UP flag should be propagated down to the driver before calling any other ioctl() that might actually require the hardware to be up and running.
2004-07-07 17:46:30 +00:00
/*
* When performing shared memory allocations, create a tag
* with a lowaddr limit that restricts physical memory mappings
* so that they all fall within the first 1GB of memory.
* At least one device/driver combination (Linksys Instant
* Wireless PCI Card V2.7, Broadcom 802.11b) seems to have
* problems with performing DMA operations with physical
Add support for Windows/x86-64 binaries to Project Evil. Ville-Pertti Keinonen (will at exomi dot comohmygodnospampleasekthx) deserves a big thanks for submitting initial patches to make it work. I have mangled his contributions appropriately. The main gotcha with Windows/x86-64 is that Microsoft uses a different calling convention than everyone else. The standard ABI requires using 6 registers for argument passing, with other arguments on the stack. Microsoft uses only 4 registers, and requires the caller to leave room on the stack for the register arguments incase the callee needs to spill them. Unlike x86, where Microsoft uses a mix of _cdecl, _stdcall and _fastcall, all routines on Windows/x86-64 uses the same convention. This unfortunately means that all the functions we export to the driver require an intermediate translation wrapper. Similarly, we have to wrap all calls back into the driver binary itself. The original patches provided macros to wrap every single routine at compile time, providing a secondary jump table with a customized wrapper for each exported routine. I decided to use a different approach: the call wrapper for each function is created from a template at runtime, and the routine to jump to is patched into the wrapper as it is created. The subr_pe module has been modified to patch in the wrapped function instead of the original. (On x86, the wrapping routine is a no-op.) There are some minor API differences that had to be accounted for: - KeAcquireSpinLock() is a real function on amd64, not a macro wrapper around KfAcquireSpinLock() - NdisFreeBuffer() is actually IoFreeMdl(). I had to change the whole NDIS_BUFFER API a bit to accomodate this. Bugs fixed along the way: - IoAllocateMdl() always returned NULL - kern_windrv.c:windrv_unload() wasn't releasing private driver object extensions correctly (found thanks to memguard) This has only been tested with the driver for the Broadcom 802.11g chipset, which was the only Windows/x86-64 driver I could find.
2005-02-16 05:41:18 +00:00
* addresses that lie above the 1GB mark. I don't know if this
* is a hardware limitation or if the addresses are being
* truncated within the driver, but this seems to be the only
* way to make these cards work reliably in systems with more
* than 1GB of physical memory.
Fix two problems: - In subr_ndis.c:ndis_allocate_sharemem(), create the busdma tags used for shared memory allocations with a lowaddr of 0x3E7FFFFF. This forces the buffers to be mapped to physical/bus addresses within the first 1GB of physical memory. It seems that at least one card (Linksys Instant Wireless PCI V2.7) depends on this behavior. I don't know if this is a hardware restriction, or if the NDIS driver for this card is truncating the addresses itself, but using physical/bus addresses beyong the 1GB limit causes initialization failures. - Create am NDIS_INITIALIZED() macro in if_ndisvar.h and use it in if_ndis.c to test whether the device has been initialized rather than checking for the presence of the IFF_UP flag in if_flags. While debugging the previous problem, I noticed that bringing up the device would always produce failures from ndis_setmulti(). It turns out that the following steps now occur during device initialization: - IFF_UP flag is set in if_flags - ifp->if_ioctl() called with SIOCSIFADDR (which we don't handle) - ifp->if_ioctl() called with SIOCADDMULTI - ifp->if_ioctl() called with SIOCADDMULTI (again) - ifp->if_ioctl() called with SIOCADDMULTI (yet again) - ifp->if_ioctl() called with SIOCSIFFLAGS Setting the receive filter and multicast filters can only be done when the underlying NDIS driver has been initialized, which is done by ifp->if_init(). However, we don't call ifp->if_init() until ifp->if_ioctl() is called with SIOCSIFFLAGS and IFF_UP has been set. It appears that now, the network stack tries to add multicast addresses to interface's filter before those steps occur. Normally, ndis_setmulti() would trap this condition by checking for the IFF_UP flag, but the network code has in fact set this flag already, so ndis_setmulti() is fooled into thinking the interface has been initialized when it really hasn't. It turns out this is usually harmless because the ifp->if_init() routine (in this case ndis_init()) will set up the multicast filter when it initializes the hardware anyway, and the underlying routines (ndis_get_info()/ndis_set_info()) know that the driver/NIC haven't been initialized yet, but you end up spurious error messages on the console all the time. Something tells me this new behavior isn't really correct. I think the intention was to fix it so that ifp->if_init() is only called once when we ifconfig an interface up, but the end result seems a little bogus: the change of the IFF_UP flag should be propagated down to the driver before calling any other ioctl() that might actually require the hardware to be up and running.
2004-07-07 17:46:30 +00:00
*/
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
error = bus_dma_tag_create(sc->ndis_parent_tag, 64,
Fix two problems: - In subr_ndis.c:ndis_allocate_sharemem(), create the busdma tags used for shared memory allocations with a lowaddr of 0x3E7FFFFF. This forces the buffers to be mapped to physical/bus addresses within the first 1GB of physical memory. It seems that at least one card (Linksys Instant Wireless PCI V2.7) depends on this behavior. I don't know if this is a hardware restriction, or if the NDIS driver for this card is truncating the addresses itself, but using physical/bus addresses beyong the 1GB limit causes initialization failures. - Create am NDIS_INITIALIZED() macro in if_ndisvar.h and use it in if_ndis.c to test whether the device has been initialized rather than checking for the presence of the IFF_UP flag in if_flags. While debugging the previous problem, I noticed that bringing up the device would always produce failures from ndis_setmulti(). It turns out that the following steps now occur during device initialization: - IFF_UP flag is set in if_flags - ifp->if_ioctl() called with SIOCSIFADDR (which we don't handle) - ifp->if_ioctl() called with SIOCADDMULTI - ifp->if_ioctl() called with SIOCADDMULTI (again) - ifp->if_ioctl() called with SIOCADDMULTI (yet again) - ifp->if_ioctl() called with SIOCSIFFLAGS Setting the receive filter and multicast filters can only be done when the underlying NDIS driver has been initialized, which is done by ifp->if_init(). However, we don't call ifp->if_init() until ifp->if_ioctl() is called with SIOCSIFFLAGS and IFF_UP has been set. It appears that now, the network stack tries to add multicast addresses to interface's filter before those steps occur. Normally, ndis_setmulti() would trap this condition by checking for the IFF_UP flag, but the network code has in fact set this flag already, so ndis_setmulti() is fooled into thinking the interface has been initialized when it really hasn't. It turns out this is usually harmless because the ifp->if_init() routine (in this case ndis_init()) will set up the multicast filter when it initializes the hardware anyway, and the underlying routines (ndis_get_info()/ndis_set_info()) know that the driver/NIC haven't been initialized yet, but you end up spurious error messages on the console all the time. Something tells me this new behavior isn't really correct. I think the intention was to fix it so that ifp->if_init() is only called once when we ifconfig an interface up, but the end result seems a little bogus: the change of the IFF_UP flag should be propagated down to the driver before calling any other ioctl() that might actually require the hardware to be up and running.
2004-07-07 17:46:30 +00:00
0, NDIS_BUS_SPACE_SHARED_MAXADDR, BUS_SPACE_MAXADDR, NULL,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
NULL, len, 1, len, BUS_DMA_ALLOCNOW, NULL, NULL,
&sh->ndis_stag);
if (error) {
free(sh, M_DEVBUF);
return;
}
error = bus_dmamem_alloc(sh->ndis_stag, vaddr,
BUS_DMA_NOWAIT | BUS_DMA_ZERO, &sh->ndis_smap);
if (error) {
bus_dma_tag_destroy(sh->ndis_stag);
free(sh, M_DEVBUF);
return;
}
error = bus_dmamap_load(sh->ndis_stag, sh->ndis_smap, *vaddr,
len, ndis_mapshared_cb, (void *)paddr, BUS_DMA_NOWAIT);
if (error) {
bus_dmamem_free(sh->ndis_stag, *vaddr, sh->ndis_smap);
bus_dma_tag_destroy(sh->ndis_stag);
free(sh, M_DEVBUF);
return;
}
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
/*
* Save the physical address along with the source address.
* The AirGo MIMO driver will call NdisMFreeSharedMemory()
* with a bogus virtual address sometimes, but with a valid
* physical address. To keep this from causing trouble, we
* use the physical address to as a sanity check in case
* searching based on the virtual address fails.
*/
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
NDIS_LOCK(sc);
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
sh->ndis_paddr.np_quad = paddr->np_quad;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
sh->ndis_saddr = *vaddr;
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
InsertHeadList((&sc->ndis_shlist), (&sh->ndis_list));
NDIS_UNLOCK(sc);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
Correct/improve the implementation of NdisMAllocateSharedMemoryAsync(). Since we have a worker thread now, we can actually do the allocation asynchronously in that thread's context. Also, we need to return a status value: if we're unable to queue up the async allocation, we return NDIS_STATUS_FAILURE, otherwise we return NDIS_STATUS_PENDING to indicate the allocation has been queued and will occur later. This replaces the kludge where we just invoked the callback routine right away in the current context.
2004-02-04 04:44:16 +00:00
struct ndis_allocwork {
uint32_t na_len;
uint8_t na_cached;
void *na_ctx;
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
io_workitem *na_iw;
Correct/improve the implementation of NdisMAllocateSharedMemoryAsync(). Since we have a worker thread now, we can actually do the allocation asynchronously in that thread's context. Also, we need to return a status value: if we're unable to queue up the async allocation, we return NDIS_STATUS_FAILURE, otherwise we return NDIS_STATUS_PENDING to indicate the allocation has been queued and will occur later. This replaces the kludge where we just invoked the callback routine right away in the current context.
2004-02-04 04:44:16 +00:00
};
static void
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
ndis_asyncmem_complete(dobj, arg)
device_object *dobj;
Correct/improve the implementation of NdisMAllocateSharedMemoryAsync(). Since we have a worker thread now, we can actually do the allocation asynchronously in that thread's context. Also, we need to return a status value: if we're unable to queue up the async allocation, we return NDIS_STATUS_FAILURE, otherwise we return NDIS_STATUS_PENDING to indicate the allocation has been queued and will occur later. This replaces the kludge where we just invoked the callback routine right away in the current context.
2004-02-04 04:44:16 +00:00
void *arg;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
{
ndis_miniport_block *block;
struct ndis_softc *sc;
Correct/improve the implementation of NdisMAllocateSharedMemoryAsync(). Since we have a worker thread now, we can actually do the allocation asynchronously in that thread's context. Also, we need to return a status value: if we're unable to queue up the async allocation, we return NDIS_STATUS_FAILURE, otherwise we return NDIS_STATUS_PENDING to indicate the allocation has been queued and will occur later. This replaces the kludge where we just invoked the callback routine right away in the current context.
2004-02-04 04:44:16 +00:00
struct ndis_allocwork *w;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
void *vaddr;
ndis_physaddr paddr;
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
ndis_allocdone_handler donefunc;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
Correct/improve the implementation of NdisMAllocateSharedMemoryAsync(). Since we have a worker thread now, we can actually do the allocation asynchronously in that thread's context. Also, we need to return a status value: if we're unable to queue up the async allocation, we return NDIS_STATUS_FAILURE, otherwise we return NDIS_STATUS_PENDING to indicate the allocation has been queued and will occur later. This replaces the kludge where we just invoked the callback routine right away in the current context.
2004-02-04 04:44:16 +00:00
w = arg;
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
block = (ndis_miniport_block *)dobj->do_devext;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
sc = device_get_softc(block->nmb_physdeviceobj->do_devext);
Correct/improve the implementation of NdisMAllocateSharedMemoryAsync(). Since we have a worker thread now, we can actually do the allocation asynchronously in that thread's context. Also, we need to return a status value: if we're unable to queue up the async allocation, we return NDIS_STATUS_FAILURE, otherwise we return NDIS_STATUS_PENDING to indicate the allocation has been queued and will occur later. This replaces the kludge where we just invoked the callback routine right away in the current context.
2004-02-04 04:44:16 +00:00
vaddr = NULL;
paddr.np_quad = 0;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
donefunc = sc->ndis_chars->nmc_allocate_complete_func;
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
NdisMAllocateSharedMemory(block, w->na_len,
Correct/improve the implementation of NdisMAllocateSharedMemoryAsync(). Since we have a worker thread now, we can actually do the allocation asynchronously in that thread's context. Also, we need to return a status value: if we're unable to queue up the async allocation, we return NDIS_STATUS_FAILURE, otherwise we return NDIS_STATUS_PENDING to indicate the allocation has been queued and will occur later. This replaces the kludge where we just invoked the callback routine right away in the current context.
2004-02-04 04:44:16 +00:00
w->na_cached, &vaddr, &paddr);
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
MSCALL5(donefunc, block, vaddr, &paddr, w->na_len, w->na_ctx);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
IoFreeWorkItem(w->na_iw);
free(w, M_DEVBUF);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMAllocateSharedMemoryAsync(adapter, len, cached, ctx)
Correct/improve the implementation of NdisMAllocateSharedMemoryAsync(). Since we have a worker thread now, we can actually do the allocation asynchronously in that thread's context. Also, we need to return a status value: if we're unable to queue up the async allocation, we return NDIS_STATUS_FAILURE, otherwise we return NDIS_STATUS_PENDING to indicate the allocation has been queued and will occur later. This replaces the kludge where we just invoked the callback routine right away in the current context.
2004-02-04 04:44:16 +00:00
ndis_handle adapter;
uint32_t len;
uint8_t cached;
void *ctx;
{
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
ndis_miniport_block *block;
Correct/improve the implementation of NdisMAllocateSharedMemoryAsync(). Since we have a worker thread now, we can actually do the allocation asynchronously in that thread's context. Also, we need to return a status value: if we're unable to queue up the async allocation, we return NDIS_STATUS_FAILURE, otherwise we return NDIS_STATUS_PENDING to indicate the allocation has been queued and will occur later. This replaces the kludge where we just invoked the callback routine right away in the current context.
2004-02-04 04:44:16 +00:00
struct ndis_allocwork *w;
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
io_workitem *iw;
io_workitem_func ifw;
Correct/improve the implementation of NdisMAllocateSharedMemoryAsync(). Since we have a worker thread now, we can actually do the allocation asynchronously in that thread's context. Also, we need to return a status value: if we're unable to queue up the async allocation, we return NDIS_STATUS_FAILURE, otherwise we return NDIS_STATUS_PENDING to indicate the allocation has been queued and will occur later. This replaces the kludge where we just invoked the callback routine right away in the current context.
2004-02-04 04:44:16 +00:00
if (adapter == NULL)
return(NDIS_STATUS_FAILURE);
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
block = adapter;
iw = IoAllocateWorkItem(block->nmb_deviceobj);
if (iw == NULL)
return(NDIS_STATUS_FAILURE);
Correct/improve the implementation of NdisMAllocateSharedMemoryAsync(). Since we have a worker thread now, we can actually do the allocation asynchronously in that thread's context. Also, we need to return a status value: if we're unable to queue up the async allocation, we return NDIS_STATUS_FAILURE, otherwise we return NDIS_STATUS_PENDING to indicate the allocation has been queued and will occur later. This replaces the kludge where we just invoked the callback routine right away in the current context.
2004-02-04 04:44:16 +00:00
w = malloc(sizeof(struct ndis_allocwork), M_TEMP, M_NOWAIT);
if (w == NULL)
return(NDIS_STATUS_FAILURE);
w->na_cached = cached;
w->na_len = len;
w->na_ctx = ctx;
Correct some problems with workitem usage. NdisScheduleWorkItem() does not use exactly the same workitem sturcture as ExQueueWorkItem() like I originally thought it did.
2005-05-16 15:29:21 +00:00
w->na_iw = iw;
Correct/improve the implementation of NdisMAllocateSharedMemoryAsync(). Since we have a worker thread now, we can actually do the allocation asynchronously in that thread's context. Also, we need to return a status value: if we're unable to queue up the async allocation, we return NDIS_STATUS_FAILURE, otherwise we return NDIS_STATUS_PENDING to indicate the allocation has been queued and will occur later. This replaces the kludge where we just invoked the callback routine right away in the current context.
2004-02-04 04:44:16 +00:00
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
ifw = (io_workitem_func)ndis_findwrap((funcptr)ndis_asyncmem_complete);
IoQueueWorkItem(iw, ifw, WORKQUEUE_DELAYED, w);
Correct/improve the implementation of NdisMAllocateSharedMemoryAsync(). Since we have a worker thread now, we can actually do the allocation asynchronously in that thread's context. Also, we need to return a status value: if we're unable to queue up the async allocation, we return NDIS_STATUS_FAILURE, otherwise we return NDIS_STATUS_PENDING to indicate the allocation has been queued and will occur later. This replaces the kludge where we just invoked the callback routine right away in the current context.
2004-02-04 04:44:16 +00:00
return(NDIS_STATUS_PENDING);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMFreeSharedMemory(adapter, len, cached, vaddr, paddr)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_handle adapter;
uint32_t len;
uint8_t cached;
void *vaddr;
ndis_physaddr paddr;
{
ndis_miniport_block *block;
struct ndis_softc *sc;
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
struct ndis_shmem *sh = NULL;
list_entry *l;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
if (vaddr == NULL || adapter == NULL)
return;
block = (ndis_miniport_block *)adapter;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
sc = device_get_softc(block->nmb_physdeviceobj->do_devext);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
Another Coverity fix from Sam: add NULL pointer test in NdisMFreeSharedMemory() (if the list is already empty, just bail).
2005-03-28 21:09:00 +00:00
/* Sanity check: is list empty? */
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
if (IsListEmpty(&sc->ndis_shlist))
Another Coverity fix from Sam: add NULL pointer test in NdisMFreeSharedMemory() (if the list is already empty, just bail).
2005-03-28 21:09:00 +00:00
return;
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
NDIS_LOCK(sc);
l = sc->ndis_shlist.nle_flink;
while (l != &sc->ndis_shlist) {
sh = CONTAINING_RECORD(l, struct ndis_shmem, ndis_list);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
if (sh->ndis_saddr == vaddr)
break;
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
/*
* Check the physaddr too, just in case the driver lied
* about the virtual address.
*/
if (sh->ndis_paddr.np_quad == paddr.np_quad)
break;
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
l = l->nle_flink;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
}
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
if (sh == NULL) {
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
NDIS_UNLOCK(sc);
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
printf("NDIS: buggy driver tried to free "
Use %jx instead of %qx to silence compiler warning on amd64.
2005-05-05 15:56:41 +00:00
"invalid shared memory: vaddr: %p paddr: 0x%jx\n",
Cast 64 bit quantity to uintmax_t to print it with %jx. This is technically a no-op since uintmax_t is uint64_t on all currently supported architectures, but we should use an explicit cast instead of depending on this obscure coincidence.
2005-05-05 22:33:06 +00:00
vaddr, (uintmax_t)paddr.np_quad);
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
return;
}
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
RemoveEntryList(&sh->ndis_list);
NDIS_UNLOCK(sc);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
bus_dmamap_unload(sh->ndis_stag, sh->ndis_smap);
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
bus_dmamem_free(sh->ndis_stag, sh->ndis_saddr, sh->ndis_smap);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
bus_dma_tag_destroy(sh->ndis_stag);
free(sh, M_DEVBUF);
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMMapIoSpace(vaddr, adapter, paddr, len)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
void **vaddr;
ndis_handle adapter;
ndis_physaddr paddr;
uint32_t len;
{
ndis_miniport_block *block;
struct ndis_softc *sc;
if (adapter == NULL)
return(NDIS_STATUS_FAILURE);
block = (ndis_miniport_block *)adapter;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
sc = device_get_softc(block->nmb_physdeviceobj->do_devext);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
Make NdisMMapIoSpace() guard against NULL/uninitialized resource pointers too.
2004-02-08 20:39:35 +00:00
if (sc->ndis_res_mem != NULL &&
paddr.np_quad == rman_get_start(sc->ndis_res_mem))
Make NdisMMapIoSpace() handle the case where a device has both mem and altmem ranges mapped.
2004-02-08 20:32:41 +00:00
*vaddr = (void *)rman_get_virtual(sc->ndis_res_mem);
Make NdisMMapIoSpace() guard against NULL/uninitialized resource pointers too.
2004-02-08 20:39:35 +00:00
else if (sc->ndis_res_altmem != NULL &&
paddr.np_quad == rman_get_start(sc->ndis_res_altmem))
Make NdisMMapIoSpace() handle the case where a device has both mem and altmem ranges mapped.
2004-02-08 20:32:41 +00:00
*vaddr = (void *)rman_get_virtual(sc->ndis_res_altmem);
Make NdisReadPcmciaAttributeMemory() and NdisWritePcmciaAttributeMemory() actually work. Make the PCI and PCCARD attachments provide a bus_get_resource_list() method so that resource listing for PCCARD works. PCCARD does not have a bus_get_resource_list() method (yet), so I faked up the resource list management in if_ndis_pccard.c, and added bus_get_resource_list() methods to both if_ndis_pccard.c and if_ndis_pci.c. The one in the PCI attechment just hands off to the PCI bus code. The difference is transparent to the NDIS resource handler code. Fixed ndis_open_file() so that opening files which live on NFS filesystems work: pass an actual ucred structure to VOP_GETATTR() (NFS explodes if the ucred structure is NOCRED). Make NdisMMapIoSpace() handle mapping of PCMCIA attribute memory resources correctly. Turn subr_ndis.c:my_strcasecmp() into ndis_strcasecmp() and export it so that if_ndis_pccard.c can use it, and junk the other copy of my_strcasecmp() from if_ndis_pccard.c.
2004-07-11 00:19:30 +00:00
else if (sc->ndis_res_am != NULL &&
paddr.np_quad == rman_get_start(sc->ndis_res_am))
*vaddr = (void *)rman_get_virtual(sc->ndis_res_am);
Make NdisMMapIoSpace() handle the case where a device has both mem and altmem ranges mapped.
2004-02-08 20:32:41 +00:00
else
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return(NDIS_STATUS_FAILURE);
return(NDIS_STATUS_SUCCESS);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMUnmapIoSpace(adapter, vaddr, len)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_handle adapter;
void *vaddr;
uint32_t len;
{
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static uint32_t
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisGetCacheFillSize(void)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
{
return(128);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static uint32_t
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMGetDmaAlignment(handle)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_handle handle;
{
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
return(16);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
}
/*
* NDIS has two methods for dealing with NICs that support DMA.
* One is to just pass packets to the driver and let it call
* NdisMStartBufferPhysicalMapping() to map each buffer in the packet
* all by itself, and the other is to let the NDIS library handle the
* buffer mapping internally, and hand the driver an already populated
* scatter/gather fragment list. If the driver calls
* NdisMInitializeScatterGatherDma(), it wants to use the latter
* method.
*/
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMInitializeScatterGatherDma(adapter, is64, maxphysmap)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_handle adapter;
uint8_t is64;
uint32_t maxphysmap;
{
struct ndis_softc *sc;
ndis_miniport_block *block;
int error;
if (adapter == NULL)
return(NDIS_STATUS_FAILURE);
block = (ndis_miniport_block *)adapter;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
sc = device_get_softc(block->nmb_physdeviceobj->do_devext);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
/* Don't do this twice. */
if (sc->ndis_sc == 1)
return(NDIS_STATUS_SUCCESS);
error = bus_dma_tag_create(sc->ndis_parent_tag, ETHER_ALIGN, 0,
BUS_SPACE_MAXADDR_32BIT, BUS_SPACE_MAXADDR, NULL, NULL,
MCLBYTES * NDIS_MAXSEG, NDIS_MAXSEG, MCLBYTES, BUS_DMA_ALLOCNOW,
NULL, NULL, &sc->ndis_ttag);
sc->ndis_sc = 1;
return(NDIS_STATUS_SUCCESS);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisAllocatePacketPool(status, pool, descnum, protrsvdlen)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_status *status;
ndis_handle *pool;
uint32_t descnum;
uint32_t protrsvdlen;
{
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
ndis_packet_pool *p;
ndis_packet *packets;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
int i;
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
p = ExAllocatePoolWithTag(NonPagedPool, sizeof(ndis_packet_pool), 0);
if (p == NULL) {
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
*status = NDIS_STATUS_RESOURCES;
return;
}
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
p->np_cnt = descnum + NDIS_POOL_EXTRA;
p->np_protrsvd = protrsvdlen;
p->np_len = sizeof(ndis_packet) + protrsvdlen;
packets = ExAllocatePoolWithTag(NonPagedPool, p->np_cnt *
p->np_len, 0);
if (packets == NULL) {
ExFreePool(p);
*status = NDIS_STATUS_RESOURCES;
return;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
}
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
p->np_pktmem = packets;
for (i = 0; i < p->np_cnt; i++)
InterlockedPushEntrySList(&p->np_head,
(struct slist_entry *)&packets[i]);
#ifdef NDIS_DEBUG_PACKETS
p->np_dead = 0;
KeInitializeSpinLock(&p->np_lock);
KeInitializeEvent(&p->np_event, EVENT_TYPE_NOTIFY, TRUE);
#endif
*pool = p;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
*status = NDIS_STATUS_SUCCESS;
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisAllocatePacketPoolEx(status, pool, descnum, oflowdescnum, protrsvdlen)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_status *status;
ndis_handle *pool;
uint32_t descnum;
uint32_t oflowdescnum;
uint32_t protrsvdlen;
{
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
return(NdisAllocatePacketPool(status, pool,
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
descnum + oflowdescnum, protrsvdlen));
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
uint32_t
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisPacketPoolUsage(pool)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_handle pool;
{
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
ndis_packet_pool *p;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
p = (ndis_packet_pool *)pool;
return(p->np_cnt - ExQueryDepthSList(&p->np_head));
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisFreePacketPool(pool)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_handle pool;
{
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
ndis_packet_pool *p;
int usage;
#ifdef NDIS_DEBUG_PACKETS
MAXPATHLEN is 1024, which means NdisOpenFile() and ndis_find_sym() were both consuming 1K of stack space. This is unfriendly. Allocate the buffers off the heap instead. It's a little slower, but these aren't performance critical routines. Also, add a spinlock to NdisAllocatePacketPool(), NdisAllocatePacket(), NdisFreePacketPool() and NdisFreePacket(). The pool is maintained as a linked list. I don't know for a fact that it can be corrupted, but why take chances.
2005-03-03 03:51:02 +00:00
uint8_t irql;
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
#endif
Add sanity checks to the ndis_packet and ndis_buffer pool handling routines to guard against problems caused by (possibly) buggy drivers. The RealTek 8180 wireless driver calls NdisFreeBuffer() to release some of its buffers _after_ it's already called NdisFreeBufferPool() to destroy the pool to which the buffers belong. In our implementation, this error causes NdisFreeBuffer() to touch stale heap memory. If you are running a release kernel, and hence have INVARIANTS et al turned off, it turns out nothing happens. But if you're using a development kernel config with INVARIANTS on, the malloc()/free() sanity checks will scribble over the pool memory with 0xdeadc0de once it's released so that any attempts to touch it will cause a trap, and indeed this is what happens. It happens that I run 5.2-RELEASE on my laptop, so when I tested the rtl8180.sys driver, it worked fine for me, but people trying to run it with development systems checked out or cvsupped from -current would get a page fault on driver load. I can't find any reason why the NDISulator would cause the RealTek driver to do the NdisFreeBufferPool() prematurely, and the same driver obviously works with Windows -- or at least, it doesn't cause a crash: the Microsoft documentation for NdisFreeBufferPool() says that failing to return all buffers to the pool before calling NdisFreeBufferPool() causes a memory leak. I've written to my contacts at RealTek asking them to check if this is indeed a bug in their driver. In the meantime, these new sanity checks will catch this problem and issue a warning rather than causing a trap. The trick is to keep a count of outstanding buffers for each buffer pool, and if the driver tries to call NdisFreeBufferPool() while there are still buffers outstanding, we mark the pool for deletion and then defer destroying it until after the last buffer has been reclaimed.
2004-03-04 00:17:14 +00:00
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
p = (ndis_packet_pool *)pool;
Add sanity checks to the ndis_packet and ndis_buffer pool handling routines to guard against problems caused by (possibly) buggy drivers. The RealTek 8180 wireless driver calls NdisFreeBuffer() to release some of its buffers _after_ it's already called NdisFreeBufferPool() to destroy the pool to which the buffers belong. In our implementation, this error causes NdisFreeBuffer() to touch stale heap memory. If you are running a release kernel, and hence have INVARIANTS et al turned off, it turns out nothing happens. But if you're using a development kernel config with INVARIANTS on, the malloc()/free() sanity checks will scribble over the pool memory with 0xdeadc0de once it's released so that any attempts to touch it will cause a trap, and indeed this is what happens. It happens that I run 5.2-RELEASE on my laptop, so when I tested the rtl8180.sys driver, it worked fine for me, but people trying to run it with development systems checked out or cvsupped from -current would get a page fault on driver load. I can't find any reason why the NDISulator would cause the RealTek driver to do the NdisFreeBufferPool() prematurely, and the same driver obviously works with Windows -- or at least, it doesn't cause a crash: the Microsoft documentation for NdisFreeBufferPool() says that failing to return all buffers to the pool before calling NdisFreeBufferPool() causes a memory leak. I've written to my contacts at RealTek asking them to check if this is indeed a bug in their driver. In the meantime, these new sanity checks will catch this problem and issue a warning rather than causing a trap. The trick is to keep a count of outstanding buffers for each buffer pool, and if the driver tries to call NdisFreeBufferPool() while there are still buffers outstanding, we mark the pool for deletion and then defer destroying it until after the last buffer has been reclaimed.
2004-03-04 00:17:14 +00:00
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
#ifdef NDIS_DEBUG_PACKETS
KeAcquireSpinLock(&p->np_lock, &irql);
#endif
Add sanity checks to the ndis_packet and ndis_buffer pool handling routines to guard against problems caused by (possibly) buggy drivers. The RealTek 8180 wireless driver calls NdisFreeBuffer() to release some of its buffers _after_ it's already called NdisFreeBufferPool() to destroy the pool to which the buffers belong. In our implementation, this error causes NdisFreeBuffer() to touch stale heap memory. If you are running a release kernel, and hence have INVARIANTS et al turned off, it turns out nothing happens. But if you're using a development kernel config with INVARIANTS on, the malloc()/free() sanity checks will scribble over the pool memory with 0xdeadc0de once it's released so that any attempts to touch it will cause a trap, and indeed this is what happens. It happens that I run 5.2-RELEASE on my laptop, so when I tested the rtl8180.sys driver, it worked fine for me, but people trying to run it with development systems checked out or cvsupped from -current would get a page fault on driver load. I can't find any reason why the NDISulator would cause the RealTek driver to do the NdisFreeBufferPool() prematurely, and the same driver obviously works with Windows -- or at least, it doesn't cause a crash: the Microsoft documentation for NdisFreeBufferPool() says that failing to return all buffers to the pool before calling NdisFreeBufferPool() causes a memory leak. I've written to my contacts at RealTek asking them to check if this is indeed a bug in their driver. In the meantime, these new sanity checks will catch this problem and issue a warning rather than causing a trap. The trick is to keep a count of outstanding buffers for each buffer pool, and if the driver tries to call NdisFreeBufferPool() while there are still buffers outstanding, we mark the pool for deletion and then defer destroying it until after the last buffer has been reclaimed.
2004-03-04 00:17:14 +00:00
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
usage = NdisPacketPoolUsage(pool);
Add sanity checks to the ndis_packet and ndis_buffer pool handling routines to guard against problems caused by (possibly) buggy drivers. The RealTek 8180 wireless driver calls NdisFreeBuffer() to release some of its buffers _after_ it's already called NdisFreeBufferPool() to destroy the pool to which the buffers belong. In our implementation, this error causes NdisFreeBuffer() to touch stale heap memory. If you are running a release kernel, and hence have INVARIANTS et al turned off, it turns out nothing happens. But if you're using a development kernel config with INVARIANTS on, the malloc()/free() sanity checks will scribble over the pool memory with 0xdeadc0de once it's released so that any attempts to touch it will cause a trap, and indeed this is what happens. It happens that I run 5.2-RELEASE on my laptop, so when I tested the rtl8180.sys driver, it worked fine for me, but people trying to run it with development systems checked out or cvsupped from -current would get a page fault on driver load. I can't find any reason why the NDISulator would cause the RealTek driver to do the NdisFreeBufferPool() prematurely, and the same driver obviously works with Windows -- or at least, it doesn't cause a crash: the Microsoft documentation for NdisFreeBufferPool() says that failing to return all buffers to the pool before calling NdisFreeBufferPool() causes a memory leak. I've written to my contacts at RealTek asking them to check if this is indeed a bug in their driver. In the meantime, these new sanity checks will catch this problem and issue a warning rather than causing a trap. The trick is to keep a count of outstanding buffers for each buffer pool, and if the driver tries to call NdisFreeBufferPool() while there are still buffers outstanding, we mark the pool for deletion and then defer destroying it until after the last buffer has been reclaimed.
2004-03-04 00:17:14 +00:00
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
#ifdef NDIS_DEBUG_PACKETS
if (usage) {
p->np_dead = 1;
KeResetEvent(&p->np_event);
KeReleaseSpinLock(&p->np_lock, irql);
KeWaitForSingleObject(&p->np_event, 0, 0, FALSE, NULL);
} else
KeReleaseSpinLock(&p->np_lock, irql);
#endif
Add sanity checks to the ndis_packet and ndis_buffer pool handling routines to guard against problems caused by (possibly) buggy drivers. The RealTek 8180 wireless driver calls NdisFreeBuffer() to release some of its buffers _after_ it's already called NdisFreeBufferPool() to destroy the pool to which the buffers belong. In our implementation, this error causes NdisFreeBuffer() to touch stale heap memory. If you are running a release kernel, and hence have INVARIANTS et al turned off, it turns out nothing happens. But if you're using a development kernel config with INVARIANTS on, the malloc()/free() sanity checks will scribble over the pool memory with 0xdeadc0de once it's released so that any attempts to touch it will cause a trap, and indeed this is what happens. It happens that I run 5.2-RELEASE on my laptop, so when I tested the rtl8180.sys driver, it worked fine for me, but people trying to run it with development systems checked out or cvsupped from -current would get a page fault on driver load. I can't find any reason why the NDISulator would cause the RealTek driver to do the NdisFreeBufferPool() prematurely, and the same driver obviously works with Windows -- or at least, it doesn't cause a crash: the Microsoft documentation for NdisFreeBufferPool() says that failing to return all buffers to the pool before calling NdisFreeBufferPool() causes a memory leak. I've written to my contacts at RealTek asking them to check if this is indeed a bug in their driver. In the meantime, these new sanity checks will catch this problem and issue a warning rather than causing a trap. The trick is to keep a count of outstanding buffers for each buffer pool, and if the driver tries to call NdisFreeBufferPool() while there are still buffers outstanding, we mark the pool for deletion and then defer destroying it until after the last buffer has been reclaimed.
2004-03-04 00:17:14 +00:00
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
ExFreePool(p->np_pktmem);
ExFreePool(p);
Add sanity checks to the ndis_packet and ndis_buffer pool handling routines to guard against problems caused by (possibly) buggy drivers. The RealTek 8180 wireless driver calls NdisFreeBuffer() to release some of its buffers _after_ it's already called NdisFreeBufferPool() to destroy the pool to which the buffers belong. In our implementation, this error causes NdisFreeBuffer() to touch stale heap memory. If you are running a release kernel, and hence have INVARIANTS et al turned off, it turns out nothing happens. But if you're using a development kernel config with INVARIANTS on, the malloc()/free() sanity checks will scribble over the pool memory with 0xdeadc0de once it's released so that any attempts to touch it will cause a trap, and indeed this is what happens. It happens that I run 5.2-RELEASE on my laptop, so when I tested the rtl8180.sys driver, it worked fine for me, but people trying to run it with development systems checked out or cvsupped from -current would get a page fault on driver load. I can't find any reason why the NDISulator would cause the RealTek driver to do the NdisFreeBufferPool() prematurely, and the same driver obviously works with Windows -- or at least, it doesn't cause a crash: the Microsoft documentation for NdisFreeBufferPool() says that failing to return all buffers to the pool before calling NdisFreeBufferPool() causes a memory leak. I've written to my contacts at RealTek asking them to check if this is indeed a bug in their driver. In the meantime, these new sanity checks will catch this problem and issue a warning rather than causing a trap. The trick is to keep a count of outstanding buffers for each buffer pool, and if the driver tries to call NdisFreeBufferPool() while there are still buffers outstanding, we mark the pool for deletion and then defer destroying it until after the last buffer has been reclaimed.
2004-03-04 00:17:14 +00:00
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisAllocatePacket(status, packet, pool)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_status *status;
ndis_packet **packet;
ndis_handle pool;
{
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
ndis_packet_pool *p;
ndis_packet *pkt;
#ifdef NDIS_DEBUG_PACKETS
MAXPATHLEN is 1024, which means NdisOpenFile() and ndis_find_sym() were both consuming 1K of stack space. This is unfriendly. Allocate the buffers off the heap instead. It's a little slower, but these aren't performance critical routines. Also, add a spinlock to NdisAllocatePacketPool(), NdisAllocatePacket(), NdisFreePacketPool() and NdisFreePacket(). The pool is maintained as a linked list. I don't know for a fact that it can be corrupted, but why take chances.
2005-03-03 03:51:02 +00:00
uint8_t irql;
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
#endif
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
p = (ndis_packet_pool *)pool;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
#ifdef NDIS_DEBUG_PACKETS
KeAcquireSpinLock(&p->np_lock, &irql);
if (p->np_dead) {
KeReleaseSpinLock(&p->np_lock, irql);
printf("NDIS: tried to allocate packet from dead pool %p\n",
pool);
*status = NDIS_STATUS_RESOURCES;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
#endif
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
pkt = (ndis_packet *)InterlockedPopEntrySList(&p->np_head);
Add sanity checks to the ndis_packet and ndis_buffer pool handling routines to guard against problems caused by (possibly) buggy drivers. The RealTek 8180 wireless driver calls NdisFreeBuffer() to release some of its buffers _after_ it's already called NdisFreeBufferPool() to destroy the pool to which the buffers belong. In our implementation, this error causes NdisFreeBuffer() to touch stale heap memory. If you are running a release kernel, and hence have INVARIANTS et al turned off, it turns out nothing happens. But if you're using a development kernel config with INVARIANTS on, the malloc()/free() sanity checks will scribble over the pool memory with 0xdeadc0de once it's released so that any attempts to touch it will cause a trap, and indeed this is what happens. It happens that I run 5.2-RELEASE on my laptop, so when I tested the rtl8180.sys driver, it worked fine for me, but people trying to run it with development systems checked out or cvsupped from -current would get a page fault on driver load. I can't find any reason why the NDISulator would cause the RealTek driver to do the NdisFreeBufferPool() prematurely, and the same driver obviously works with Windows -- or at least, it doesn't cause a crash: the Microsoft documentation for NdisFreeBufferPool() says that failing to return all buffers to the pool before calling NdisFreeBufferPool() causes a memory leak. I've written to my contacts at RealTek asking them to check if this is indeed a bug in their driver. In the meantime, these new sanity checks will catch this problem and issue a warning rather than causing a trap. The trick is to keep a count of outstanding buffers for each buffer pool, and if the driver tries to call NdisFreeBufferPool() while there are still buffers outstanding, we mark the pool for deletion and then defer destroying it until after the last buffer has been reclaimed.
2004-03-04 00:17:14 +00:00
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
#ifdef NDIS_DEBUG_PACKETS
KeReleaseSpinLock(&p->np_lock, irql);
#endif
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
if (pkt == NULL) {
*status = NDIS_STATUS_RESOURCES;
return;
}
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
bzero((char *)pkt, sizeof(ndis_packet));
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
/* Save pointer to the pool. */
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
pkt->np_private.npp_pool = pool;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
/* Set the oob offset pointer. Lots of things expect this. */
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
pkt->np_private.npp_packetooboffset = offsetof(ndis_packet, np_oob);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
The private data section of ndis_packets has a 'packet flags' byte which has two important flags in it: the 'allocated by NDIS' flag and the 'media specific info present' flag. There are two Windows macros for getting/setting media specific info fields within the ndis_packet structure which can behave improperly if these flags are not initialized correctly when a packet is allocated. It seems the correct thing to do is always set the NDIS_PACKET_ALLOCATED_BY_NDIS flag on all newly allocated packets. This fixes the crashes with the Intel Centrino wireless driver. My sample card now seems to work correctly. Also, fix a potential LOR involving ndis_txeof() in if_ndis.c.
2004-01-09 06:53:49 +00:00
/*
* We must initialize the packet flags correctly in order
* for the NDIS_SET_PACKET_MEDIA_SPECIFIC_INFO() and
Add support for Windows/x86-64 binaries to Project Evil. Ville-Pertti Keinonen (will at exomi dot comohmygodnospampleasekthx) deserves a big thanks for submitting initial patches to make it work. I have mangled his contributions appropriately. The main gotcha with Windows/x86-64 is that Microsoft uses a different calling convention than everyone else. The standard ABI requires using 6 registers for argument passing, with other arguments on the stack. Microsoft uses only 4 registers, and requires the caller to leave room on the stack for the register arguments incase the callee needs to spill them. Unlike x86, where Microsoft uses a mix of _cdecl, _stdcall and _fastcall, all routines on Windows/x86-64 uses the same convention. This unfortunately means that all the functions we export to the driver require an intermediate translation wrapper. Similarly, we have to wrap all calls back into the driver binary itself. The original patches provided macros to wrap every single routine at compile time, providing a secondary jump table with a customized wrapper for each exported routine. I decided to use a different approach: the call wrapper for each function is created from a template at runtime, and the routine to jump to is patched into the wrapper as it is created. The subr_pe module has been modified to patch in the wrapped function instead of the original. (On x86, the wrapping routine is a no-op.) There are some minor API differences that had to be accounted for: - KeAcquireSpinLock() is a real function on amd64, not a macro wrapper around KfAcquireSpinLock() - NdisFreeBuffer() is actually IoFreeMdl(). I had to change the whole NDIS_BUFFER API a bit to accomodate this. Bugs fixed along the way: - IoAllocateMdl() always returned NULL - kern_windrv.c:windrv_unload() wasn't releasing private driver object extensions correctly (found thanks to memguard) This has only been tested with the driver for the Broadcom 802.11g chipset, which was the only Windows/x86-64 driver I could find.
2005-02-16 05:41:18 +00:00
* NDIS_GET_PACKET_MEDIA_SPECIFIC_INFO() macros to work
* correctly.
The private data section of ndis_packets has a 'packet flags' byte which has two important flags in it: the 'allocated by NDIS' flag and the 'media specific info present' flag. There are two Windows macros for getting/setting media specific info fields within the ndis_packet structure which can behave improperly if these flags are not initialized correctly when a packet is allocated. It seems the correct thing to do is always set the NDIS_PACKET_ALLOCATED_BY_NDIS flag on all newly allocated packets. This fixes the crashes with the Intel Centrino wireless driver. My sample card now seems to work correctly. Also, fix a potential LOR involving ndis_txeof() in if_ndis.c.
2004-01-09 06:53:49 +00:00
*/
pkt->np_private.npp_ndispktflags = NDIS_PACKET_ALLOCATED_BY_NDIS;
Spell '0' as 'FALSE' when initializing npp_validcounts. (Doesn't change the code, but emphasises that this field is used as a boolean.)
2005-03-28 17:06:47 +00:00
pkt->np_private.npp_validcounts = FALSE;
The private data section of ndis_packets has a 'packet flags' byte which has two important flags in it: the 'allocated by NDIS' flag and the 'media specific info present' flag. There are two Windows macros for getting/setting media specific info fields within the ndis_packet structure which can behave improperly if these flags are not initialized correctly when a packet is allocated. It seems the correct thing to do is always set the NDIS_PACKET_ALLOCATED_BY_NDIS flag on all newly allocated packets. This fixes the crashes with the Intel Centrino wireless driver. My sample card now seems to work correctly. Also, fix a potential LOR involving ndis_txeof() in if_ndis.c.
2004-01-09 06:53:49 +00:00
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
*packet = pkt;
*status = NDIS_STATUS_SUCCESS;
Add support for Windows/x86-64 binaries to Project Evil. Ville-Pertti Keinonen (will at exomi dot comohmygodnospampleasekthx) deserves a big thanks for submitting initial patches to make it work. I have mangled his contributions appropriately. The main gotcha with Windows/x86-64 is that Microsoft uses a different calling convention than everyone else. The standard ABI requires using 6 registers for argument passing, with other arguments on the stack. Microsoft uses only 4 registers, and requires the caller to leave room on the stack for the register arguments incase the callee needs to spill them. Unlike x86, where Microsoft uses a mix of _cdecl, _stdcall and _fastcall, all routines on Windows/x86-64 uses the same convention. This unfortunately means that all the functions we export to the driver require an intermediate translation wrapper. Similarly, we have to wrap all calls back into the driver binary itself. The original patches provided macros to wrap every single routine at compile time, providing a secondary jump table with a customized wrapper for each exported routine. I decided to use a different approach: the call wrapper for each function is created from a template at runtime, and the routine to jump to is patched into the wrapper as it is created. The subr_pe module has been modified to patch in the wrapped function instead of the original. (On x86, the wrapping routine is a no-op.) There are some minor API differences that had to be accounted for: - KeAcquireSpinLock() is a real function on amd64, not a macro wrapper around KfAcquireSpinLock() - NdisFreeBuffer() is actually IoFreeMdl(). I had to change the whole NDIS_BUFFER API a bit to accomodate this. Bugs fixed along the way: - IoAllocateMdl() always returned NULL - kern_windrv.c:windrv_unload() wasn't releasing private driver object extensions correctly (found thanks to memguard) This has only been tested with the driver for the Broadcom 802.11g chipset, which was the only Windows/x86-64 driver I could find.
2005-02-16 05:41:18 +00:00
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisFreePacket(packet)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_packet *packet;
{
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
ndis_packet_pool *p;
#ifdef NDIS_DEBUG_PACKETS
MAXPATHLEN is 1024, which means NdisOpenFile() and ndis_find_sym() were both consuming 1K of stack space. This is unfriendly. Allocate the buffers off the heap instead. It's a little slower, but these aren't performance critical routines. Also, add a spinlock to NdisAllocatePacketPool(), NdisAllocatePacket(), NdisFreePacketPool() and NdisFreePacket(). The pool is maintained as a linked list. I don't know for a fact that it can be corrupted, but why take chances.
2005-03-03 03:51:02 +00:00
uint8_t irql;
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
#endif
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
p = (ndis_packet_pool *)packet->np_private.npp_pool;
MAXPATHLEN is 1024, which means NdisOpenFile() and ndis_find_sym() were both consuming 1K of stack space. This is unfriendly. Allocate the buffers off the heap instead. It's a little slower, but these aren't performance critical routines. Also, add a spinlock to NdisAllocatePacketPool(), NdisAllocatePacket(), NdisFreePacketPool() and NdisFreePacket(). The pool is maintained as a linked list. I don't know for a fact that it can be corrupted, but why take chances.
2005-03-03 03:51:02 +00:00
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
#ifdef NDIS_DEBUG_PACKETS
KeAcquireSpinLock(&p->np_lock, &irql);
#endif
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
InterlockedPushEntrySList(&p->np_head, (slist_entry *)packet);
Add sanity checks to the ndis_packet and ndis_buffer pool handling routines to guard against problems caused by (possibly) buggy drivers. The RealTek 8180 wireless driver calls NdisFreeBuffer() to release some of its buffers _after_ it's already called NdisFreeBufferPool() to destroy the pool to which the buffers belong. In our implementation, this error causes NdisFreeBuffer() to touch stale heap memory. If you are running a release kernel, and hence have INVARIANTS et al turned off, it turns out nothing happens. But if you're using a development kernel config with INVARIANTS on, the malloc()/free() sanity checks will scribble over the pool memory with 0xdeadc0de once it's released so that any attempts to touch it will cause a trap, and indeed this is what happens. It happens that I run 5.2-RELEASE on my laptop, so when I tested the rtl8180.sys driver, it worked fine for me, but people trying to run it with development systems checked out or cvsupped from -current would get a page fault on driver load. I can't find any reason why the NDISulator would cause the RealTek driver to do the NdisFreeBufferPool() prematurely, and the same driver obviously works with Windows -- or at least, it doesn't cause a crash: the Microsoft documentation for NdisFreeBufferPool() says that failing to return all buffers to the pool before calling NdisFreeBufferPool() causes a memory leak. I've written to my contacts at RealTek asking them to check if this is indeed a bug in their driver. In the meantime, these new sanity checks will catch this problem and issue a warning rather than causing a trap. The trick is to keep a count of outstanding buffers for each buffer pool, and if the driver tries to call NdisFreeBufferPool() while there are still buffers outstanding, we mark the pool for deletion and then defer destroying it until after the last buffer has been reclaimed.
2004-03-04 00:17:14 +00:00
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
#ifdef NDIS_DEBUG_PACKETS
if (p->np_dead) {
if (ExQueryDepthSList(&p->np_head) == p->np_cnt)
KeSetEvent(&p->np_event, IO_NO_INCREMENT, FALSE);
}
KeReleaseSpinLock(&p->np_lock, irql);
#endif
Add sanity checks to the ndis_packet and ndis_buffer pool handling routines to guard against problems caused by (possibly) buggy drivers. The RealTek 8180 wireless driver calls NdisFreeBuffer() to release some of its buffers _after_ it's already called NdisFreeBufferPool() to destroy the pool to which the buffers belong. In our implementation, this error causes NdisFreeBuffer() to touch stale heap memory. If you are running a release kernel, and hence have INVARIANTS et al turned off, it turns out nothing happens. But if you're using a development kernel config with INVARIANTS on, the malloc()/free() sanity checks will scribble over the pool memory with 0xdeadc0de once it's released so that any attempts to touch it will cause a trap, and indeed this is what happens. It happens that I run 5.2-RELEASE on my laptop, so when I tested the rtl8180.sys driver, it worked fine for me, but people trying to run it with development systems checked out or cvsupped from -current would get a page fault on driver load. I can't find any reason why the NDISulator would cause the RealTek driver to do the NdisFreeBufferPool() prematurely, and the same driver obviously works with Windows -- or at least, it doesn't cause a crash: the Microsoft documentation for NdisFreeBufferPool() says that failing to return all buffers to the pool before calling NdisFreeBufferPool() causes a memory leak. I've written to my contacts at RealTek asking them to check if this is indeed a bug in their driver. In the meantime, these new sanity checks will catch this problem and issue a warning rather than causing a trap. The trick is to keep a count of outstanding buffers for each buffer pool, and if the driver tries to call NdisFreeBufferPool() while there are still buffers outstanding, we mark the pool for deletion and then defer destroying it until after the last buffer has been reclaimed.
2004-03-04 00:17:14 +00:00
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisUnchainBufferAtFront(packet, buf)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_packet *packet;
ndis_buffer **buf;
{
ndis_packet_private *priv;
if (packet == NULL || buf == NULL)
return;
priv = &packet->np_private;
priv->npp_validcounts = FALSE;
if (priv->npp_head == priv->npp_tail) {
*buf = priv->npp_head;
priv->npp_head = priv->npp_tail = NULL;
} else {
*buf = priv->npp_head;
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
priv->npp_head = (*buf)->mdl_next;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
}
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisUnchainBufferAtBack(packet, buf)
Now that I finally have power back, implement a couple more NDIS API routines: NdisUnchainBufferAtBack(), NdisGetFirstBufferFromPacketSafe() and NdisGetFirstBufferFromPacket(). This should bring us a little closer to getting the Intel centrino wireless NIC to work. Note: I have not actually tested these additions since I don't have a driver that calls them, however they're pretty simple, and one of them is taken pretty much directly from the Windows ndis.h header file, so I'm fairly confident they work, but disclaimers apply.
2003-12-22 08:24:32 +00:00
ndis_packet *packet;
ndis_buffer **buf;
{
ndis_packet_private *priv;
ndis_buffer *tmp;
if (packet == NULL || buf == NULL)
return;
priv = &packet->np_private;
priv->npp_validcounts = FALSE;
if (priv->npp_head == priv->npp_tail) {
*buf = priv->npp_head;
priv->npp_head = priv->npp_tail = NULL;
} else {
*buf = priv->npp_tail;
tmp = priv->npp_head;
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
while (tmp->mdl_next != priv->npp_tail)
tmp = tmp->mdl_next;
Now that I finally have power back, implement a couple more NDIS API routines: NdisUnchainBufferAtBack(), NdisGetFirstBufferFromPacketSafe() and NdisGetFirstBufferFromPacket(). This should bring us a little closer to getting the Intel centrino wireless NIC to work. Note: I have not actually tested these additions since I don't have a driver that calls them, however they're pretty simple, and one of them is taken pretty much directly from the Windows ndis.h header file, so I'm fairly confident they work, but disclaimers apply.
2003-12-22 08:24:32 +00:00
priv->npp_tail = tmp;
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
tmp->mdl_next = NULL;
Now that I finally have power back, implement a couple more NDIS API routines: NdisUnchainBufferAtBack(), NdisGetFirstBufferFromPacketSafe() and NdisGetFirstBufferFromPacket(). This should bring us a little closer to getting the Intel centrino wireless NIC to work. Note: I have not actually tested these additions since I don't have a driver that calls them, however they're pretty simple, and one of them is taken pretty much directly from the Windows ndis.h header file, so I'm fairly confident they work, but disclaimers apply.
2003-12-22 08:24:32 +00:00
}
return;
}
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
/*
Add support for Windows/x86-64 binaries to Project Evil. Ville-Pertti Keinonen (will at exomi dot comohmygodnospampleasekthx) deserves a big thanks for submitting initial patches to make it work. I have mangled his contributions appropriately. The main gotcha with Windows/x86-64 is that Microsoft uses a different calling convention than everyone else. The standard ABI requires using 6 registers for argument passing, with other arguments on the stack. Microsoft uses only 4 registers, and requires the caller to leave room on the stack for the register arguments incase the callee needs to spill them. Unlike x86, where Microsoft uses a mix of _cdecl, _stdcall and _fastcall, all routines on Windows/x86-64 uses the same convention. This unfortunately means that all the functions we export to the driver require an intermediate translation wrapper. Similarly, we have to wrap all calls back into the driver binary itself. The original patches provided macros to wrap every single routine at compile time, providing a secondary jump table with a customized wrapper for each exported routine. I decided to use a different approach: the call wrapper for each function is created from a template at runtime, and the routine to jump to is patched into the wrapper as it is created. The subr_pe module has been modified to patch in the wrapped function instead of the original. (On x86, the wrapping routine is a no-op.) There are some minor API differences that had to be accounted for: - KeAcquireSpinLock() is a real function on amd64, not a macro wrapper around KfAcquireSpinLock() - NdisFreeBuffer() is actually IoFreeMdl(). I had to change the whole NDIS_BUFFER API a bit to accomodate this. Bugs fixed along the way: - IoAllocateMdl() always returned NULL - kern_windrv.c:windrv_unload() wasn't releasing private driver object extensions correctly (found thanks to memguard) This has only been tested with the driver for the Broadcom 802.11g chipset, which was the only Windows/x86-64 driver I could find.
2005-02-16 05:41:18 +00:00
* The NDIS "buffer" is really an MDL (memory descriptor list)
* which is used to describe a buffer in a way that allows it
* to mapped into different contexts. We have to be careful how
* we handle them: in some versions of Windows, the NdisFreeBuffer()
* routine is an actual function in the NDIS API, but in others
* it's just a macro wrapper around IoFreeMdl(). There's really
* no way to use the 'descnum' parameter to count how many
* "buffers" are allocated since in order to use IoFreeMdl() to
* dispose of a buffer, we have to use IoAllocateMdl() to allocate
* them, and IoAllocateMdl() just grabs them out of the heap.
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
*/
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisAllocateBufferPool(status, pool, descnum)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_status *status;
ndis_handle *pool;
uint32_t descnum;
{
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
Add support for Windows/x86-64 binaries to Project Evil. Ville-Pertti Keinonen (will at exomi dot comohmygodnospampleasekthx) deserves a big thanks for submitting initial patches to make it work. I have mangled his contributions appropriately. The main gotcha with Windows/x86-64 is that Microsoft uses a different calling convention than everyone else. The standard ABI requires using 6 registers for argument passing, with other arguments on the stack. Microsoft uses only 4 registers, and requires the caller to leave room on the stack for the register arguments incase the callee needs to spill them. Unlike x86, where Microsoft uses a mix of _cdecl, _stdcall and _fastcall, all routines on Windows/x86-64 uses the same convention. This unfortunately means that all the functions we export to the driver require an intermediate translation wrapper. Similarly, we have to wrap all calls back into the driver binary itself. The original patches provided macros to wrap every single routine at compile time, providing a secondary jump table with a customized wrapper for each exported routine. I decided to use a different approach: the call wrapper for each function is created from a template at runtime, and the routine to jump to is patched into the wrapper as it is created. The subr_pe module has been modified to patch in the wrapped function instead of the original. (On x86, the wrapping routine is a no-op.) There are some minor API differences that had to be accounted for: - KeAcquireSpinLock() is a real function on amd64, not a macro wrapper around KfAcquireSpinLock() - NdisFreeBuffer() is actually IoFreeMdl(). I had to change the whole NDIS_BUFFER API a bit to accomodate this. Bugs fixed along the way: - IoAllocateMdl() always returned NULL - kern_windrv.c:windrv_unload() wasn't releasing private driver object extensions correctly (found thanks to memguard) This has only been tested with the driver for the Broadcom 802.11g chipset, which was the only Windows/x86-64 driver I could find.
2005-02-16 05:41:18 +00:00
/*
* The only thing we can really do here is verify that descnum
* is a reasonable value, but I really don't know what to check
* it against.
*/
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
Add support for Windows/x86-64 binaries to Project Evil. Ville-Pertti Keinonen (will at exomi dot comohmygodnospampleasekthx) deserves a big thanks for submitting initial patches to make it work. I have mangled his contributions appropriately. The main gotcha with Windows/x86-64 is that Microsoft uses a different calling convention than everyone else. The standard ABI requires using 6 registers for argument passing, with other arguments on the stack. Microsoft uses only 4 registers, and requires the caller to leave room on the stack for the register arguments incase the callee needs to spill them. Unlike x86, where Microsoft uses a mix of _cdecl, _stdcall and _fastcall, all routines on Windows/x86-64 uses the same convention. This unfortunately means that all the functions we export to the driver require an intermediate translation wrapper. Similarly, we have to wrap all calls back into the driver binary itself. The original patches provided macros to wrap every single routine at compile time, providing a secondary jump table with a customized wrapper for each exported routine. I decided to use a different approach: the call wrapper for each function is created from a template at runtime, and the routine to jump to is patched into the wrapper as it is created. The subr_pe module has been modified to patch in the wrapped function instead of the original. (On x86, the wrapping routine is a no-op.) There are some minor API differences that had to be accounted for: - KeAcquireSpinLock() is a real function on amd64, not a macro wrapper around KfAcquireSpinLock() - NdisFreeBuffer() is actually IoFreeMdl(). I had to change the whole NDIS_BUFFER API a bit to accomodate this. Bugs fixed along the way: - IoAllocateMdl() always returned NULL - kern_windrv.c:windrv_unload() wasn't releasing private driver object extensions correctly (found thanks to memguard) This has only been tested with the driver for the Broadcom 802.11g chipset, which was the only Windows/x86-64 driver I could find.
2005-02-16 05:41:18 +00:00
*pool = NonPagedPool;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
*status = NDIS_STATUS_SUCCESS;
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisFreeBufferPool(pool)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_handle pool;
{
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisAllocateBuffer(status, buffer, pool, vaddr, len)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_status *status;
ndis_buffer **buffer;
ndis_handle pool;
void *vaddr;
uint32_t len;
{
Add support for Windows/x86-64 binaries to Project Evil. Ville-Pertti Keinonen (will at exomi dot comohmygodnospampleasekthx) deserves a big thanks for submitting initial patches to make it work. I have mangled his contributions appropriately. The main gotcha with Windows/x86-64 is that Microsoft uses a different calling convention than everyone else. The standard ABI requires using 6 registers for argument passing, with other arguments on the stack. Microsoft uses only 4 registers, and requires the caller to leave room on the stack for the register arguments incase the callee needs to spill them. Unlike x86, where Microsoft uses a mix of _cdecl, _stdcall and _fastcall, all routines on Windows/x86-64 uses the same convention. This unfortunately means that all the functions we export to the driver require an intermediate translation wrapper. Similarly, we have to wrap all calls back into the driver binary itself. The original patches provided macros to wrap every single routine at compile time, providing a secondary jump table with a customized wrapper for each exported routine. I decided to use a different approach: the call wrapper for each function is created from a template at runtime, and the routine to jump to is patched into the wrapper as it is created. The subr_pe module has been modified to patch in the wrapped function instead of the original. (On x86, the wrapping routine is a no-op.) There are some minor API differences that had to be accounted for: - KeAcquireSpinLock() is a real function on amd64, not a macro wrapper around KfAcquireSpinLock() - NdisFreeBuffer() is actually IoFreeMdl(). I had to change the whole NDIS_BUFFER API a bit to accomodate this. Bugs fixed along the way: - IoAllocateMdl() always returned NULL - kern_windrv.c:windrv_unload() wasn't releasing private driver object extensions correctly (found thanks to memguard) This has only been tested with the driver for the Broadcom 802.11g chipset, which was the only Windows/x86-64 driver I could find.
2005-02-16 05:41:18 +00:00
ndis_buffer *buf;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
Add support for Windows/x86-64 binaries to Project Evil. Ville-Pertti Keinonen (will at exomi dot comohmygodnospampleasekthx) deserves a big thanks for submitting initial patches to make it work. I have mangled his contributions appropriately. The main gotcha with Windows/x86-64 is that Microsoft uses a different calling convention than everyone else. The standard ABI requires using 6 registers for argument passing, with other arguments on the stack. Microsoft uses only 4 registers, and requires the caller to leave room on the stack for the register arguments incase the callee needs to spill them. Unlike x86, where Microsoft uses a mix of _cdecl, _stdcall and _fastcall, all routines on Windows/x86-64 uses the same convention. This unfortunately means that all the functions we export to the driver require an intermediate translation wrapper. Similarly, we have to wrap all calls back into the driver binary itself. The original patches provided macros to wrap every single routine at compile time, providing a secondary jump table with a customized wrapper for each exported routine. I decided to use a different approach: the call wrapper for each function is created from a template at runtime, and the routine to jump to is patched into the wrapper as it is created. The subr_pe module has been modified to patch in the wrapped function instead of the original. (On x86, the wrapping routine is a no-op.) There are some minor API differences that had to be accounted for: - KeAcquireSpinLock() is a real function on amd64, not a macro wrapper around KfAcquireSpinLock() - NdisFreeBuffer() is actually IoFreeMdl(). I had to change the whole NDIS_BUFFER API a bit to accomodate this. Bugs fixed along the way: - IoAllocateMdl() always returned NULL - kern_windrv.c:windrv_unload() wasn't releasing private driver object extensions correctly (found thanks to memguard) This has only been tested with the driver for the Broadcom 802.11g chipset, which was the only Windows/x86-64 driver I could find.
2005-02-16 05:41:18 +00:00
buf = IoAllocateMdl(vaddr, len, FALSE, FALSE, NULL);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
if (buf == NULL) {
*status = NDIS_STATUS_RESOURCES;
return;
}
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
MmBuildMdlForNonPagedPool(buf);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
*buffer = buf;
*status = NDIS_STATUS_SUCCESS;
Add support for Windows/x86-64 binaries to Project Evil. Ville-Pertti Keinonen (will at exomi dot comohmygodnospampleasekthx) deserves a big thanks for submitting initial patches to make it work. I have mangled his contributions appropriately. The main gotcha with Windows/x86-64 is that Microsoft uses a different calling convention than everyone else. The standard ABI requires using 6 registers for argument passing, with other arguments on the stack. Microsoft uses only 4 registers, and requires the caller to leave room on the stack for the register arguments incase the callee needs to spill them. Unlike x86, where Microsoft uses a mix of _cdecl, _stdcall and _fastcall, all routines on Windows/x86-64 uses the same convention. This unfortunately means that all the functions we export to the driver require an intermediate translation wrapper. Similarly, we have to wrap all calls back into the driver binary itself. The original patches provided macros to wrap every single routine at compile time, providing a secondary jump table with a customized wrapper for each exported routine. I decided to use a different approach: the call wrapper for each function is created from a template at runtime, and the routine to jump to is patched into the wrapper as it is created. The subr_pe module has been modified to patch in the wrapped function instead of the original. (On x86, the wrapping routine is a no-op.) There are some minor API differences that had to be accounted for: - KeAcquireSpinLock() is a real function on amd64, not a macro wrapper around KfAcquireSpinLock() - NdisFreeBuffer() is actually IoFreeMdl(). I had to change the whole NDIS_BUFFER API a bit to accomodate this. Bugs fixed along the way: - IoAllocateMdl() always returned NULL - kern_windrv.c:windrv_unload() wasn't releasing private driver object extensions correctly (found thanks to memguard) This has only been tested with the driver for the Broadcom 802.11g chipset, which was the only Windows/x86-64 driver I could find.
2005-02-16 05:41:18 +00:00
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisFreeBuffer(buf)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_buffer *buf;
{
Add support for Windows/x86-64 binaries to Project Evil. Ville-Pertti Keinonen (will at exomi dot comohmygodnospampleasekthx) deserves a big thanks for submitting initial patches to make it work. I have mangled his contributions appropriately. The main gotcha with Windows/x86-64 is that Microsoft uses a different calling convention than everyone else. The standard ABI requires using 6 registers for argument passing, with other arguments on the stack. Microsoft uses only 4 registers, and requires the caller to leave room on the stack for the register arguments incase the callee needs to spill them. Unlike x86, where Microsoft uses a mix of _cdecl, _stdcall and _fastcall, all routines on Windows/x86-64 uses the same convention. This unfortunately means that all the functions we export to the driver require an intermediate translation wrapper. Similarly, we have to wrap all calls back into the driver binary itself. The original patches provided macros to wrap every single routine at compile time, providing a secondary jump table with a customized wrapper for each exported routine. I decided to use a different approach: the call wrapper for each function is created from a template at runtime, and the routine to jump to is patched into the wrapper as it is created. The subr_pe module has been modified to patch in the wrapped function instead of the original. (On x86, the wrapping routine is a no-op.) There are some minor API differences that had to be accounted for: - KeAcquireSpinLock() is a real function on amd64, not a macro wrapper around KfAcquireSpinLock() - NdisFreeBuffer() is actually IoFreeMdl(). I had to change the whole NDIS_BUFFER API a bit to accomodate this. Bugs fixed along the way: - IoAllocateMdl() always returned NULL - kern_windrv.c:windrv_unload() wasn't releasing private driver object extensions correctly (found thanks to memguard) This has only been tested with the driver for the Broadcom 802.11g chipset, which was the only Windows/x86-64 driver I could find.
2005-02-16 05:41:18 +00:00
IoFreeMdl(buf);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
In if_ndis.c:ndis_attach(), temporarily set the IFF_UP flag while calling the haltfunc. If an interrupt is triggered by the init or halt func, the IFF_UP flag must be set in order for us to be able to service it. In kern_ndis.c: implement a handler for NdisMSendResourcesAvailable() (currently does nothing since we don't really need it). In subr_ndis.c: - Correct ndis_init_string() and ndis_unicode_to_ansi(), which were both horribly broken. - Implement NdisImmediateReadPciSlotInformation() and NdisImmediateWritePciSlotInformation(). - Implement NdisBufferLength(). - Work around my first confirmed NDIS driver bug. The SMC 9462 gigE driver (natsemi 83820-based copper) incorrectly creates a spinlock in its DriverEntry() routine and then destroys it in its MiniportHalt() handler. This is wrong: spinlocks should be created in MiniportInit(). In a Windows environment, this is often not a problem because DriverEntry()/MiniportInit() are called once when the system boots and MiniportHalt() or the shutdown handler is called when the system halts. With this stuff in place, this driver now seems to work: ndis0: <SMC EZ Card 1000> port 0xe000-0xe0ff mem 0xda000000-0xda000fff irq 10 at device 9.0 on pci0 ndis0: assign PCI resources... ndis_open_file("FLASH9.hex", 18446744073709551615) ndis0: Ethernet address: 00:04:e2:0e:d3:f0
2004-01-03 13:20:30 +00:00
/* Aw c'mon. */
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static uint32_t
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisBufferLength(buf)
In if_ndis.c:ndis_attach(), temporarily set the IFF_UP flag while calling the haltfunc. If an interrupt is triggered by the init or halt func, the IFF_UP flag must be set in order for us to be able to service it. In kern_ndis.c: implement a handler for NdisMSendResourcesAvailable() (currently does nothing since we don't really need it). In subr_ndis.c: - Correct ndis_init_string() and ndis_unicode_to_ansi(), which were both horribly broken. - Implement NdisImmediateReadPciSlotInformation() and NdisImmediateWritePciSlotInformation(). - Implement NdisBufferLength(). - Work around my first confirmed NDIS driver bug. The SMC 9462 gigE driver (natsemi 83820-based copper) incorrectly creates a spinlock in its DriverEntry() routine and then destroys it in its MiniportHalt() handler. This is wrong: spinlocks should be created in MiniportInit(). In a Windows environment, this is often not a problem because DriverEntry()/MiniportInit() are called once when the system boots and MiniportHalt() or the shutdown handler is called when the system halts. With this stuff in place, this driver now seems to work: ndis0: <SMC EZ Card 1000> port 0xe000-0xe0ff mem 0xda000000-0xda000fff irq 10 at device 9.0 on pci0 ndis0: assign PCI resources... ndis_open_file("FLASH9.hex", 18446744073709551615) ndis0: Ethernet address: 00:04:e2:0e:d3:f0
2004-01-03 13:20:30 +00:00
ndis_buffer *buf;
{
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
return(MmGetMdlByteCount(buf));
In if_ndis.c:ndis_attach(), temporarily set the IFF_UP flag while calling the haltfunc. If an interrupt is triggered by the init or halt func, the IFF_UP flag must be set in order for us to be able to service it. In kern_ndis.c: implement a handler for NdisMSendResourcesAvailable() (currently does nothing since we don't really need it). In subr_ndis.c: - Correct ndis_init_string() and ndis_unicode_to_ansi(), which were both horribly broken. - Implement NdisImmediateReadPciSlotInformation() and NdisImmediateWritePciSlotInformation(). - Implement NdisBufferLength(). - Work around my first confirmed NDIS driver bug. The SMC 9462 gigE driver (natsemi 83820-based copper) incorrectly creates a spinlock in its DriverEntry() routine and then destroys it in its MiniportHalt() handler. This is wrong: spinlocks should be created in MiniportInit(). In a Windows environment, this is often not a problem because DriverEntry()/MiniportInit() are called once when the system boots and MiniportHalt() or the shutdown handler is called when the system halts. With this stuff in place, this driver now seems to work: ndis0: <SMC EZ Card 1000> port 0xe000-0xe0ff mem 0xda000000-0xda000fff irq 10 at device 9.0 on pci0 ndis0: assign PCI resources... ndis_open_file("FLASH9.hex", 18446744073709551615) ndis0: Ethernet address: 00:04:e2:0e:d3:f0
2004-01-03 13:20:30 +00:00
}
Some minor touchups: In NdisQueryBuffer() and NdisQueryBufferSafe(), the vaddr argument is optional, so test it before trying to dereference it. Also correct NdisGetFirstBufferFromPacket()/NdisGetFirstBufferFromPacketSafe(): we need to use nb_mappedsystemva from the buffer, not nb_systemva.
2003-12-22 10:22:57 +00:00
/*
* Get the virtual address and length of a buffer.
* Note: the vaddr argument is optional.
*/
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisQueryBuffer(buf, vaddr, len)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_buffer *buf;
void **vaddr;
uint32_t *len;
{
Some minor touchups: In NdisQueryBuffer() and NdisQueryBufferSafe(), the vaddr argument is optional, so test it before trying to dereference it. Also correct NdisGetFirstBufferFromPacket()/NdisGetFirstBufferFromPacketSafe(): we need to use nb_mappedsystemva from the buffer, not nb_systemva.
2003-12-22 10:22:57 +00:00
if (vaddr != NULL)
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
*vaddr = MmGetMdlVirtualAddress(buf);
*len = MmGetMdlByteCount(buf);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
/* Same as above -- we don't care about the priority. */
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisQueryBufferSafe(buf, vaddr, len, prio)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_buffer *buf;
void **vaddr;
uint32_t *len;
uint32_t prio;
{
Some minor touchups: In NdisQueryBuffer() and NdisQueryBufferSafe(), the vaddr argument is optional, so test it before trying to dereference it. Also correct NdisGetFirstBufferFromPacket()/NdisGetFirstBufferFromPacketSafe(): we need to use nb_mappedsystemva from the buffer, not nb_systemva.
2003-12-22 10:22:57 +00:00
if (vaddr != NULL)
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
*vaddr = MmGetMdlVirtualAddress(buf);
*len = MmGetMdlByteCount(buf);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
Implement NdisVirtualBufferAddress() and NdisVirtualBufferAddressSafe(). The RealTek 8180 driver seems to need this.
2004-01-27 08:10:34 +00:00
/* Damnit Microsoft!! How many ways can you do the same thing?! */
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void *
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisBufferVirtualAddress(buf)
Implement NdisVirtualBufferAddress() and NdisVirtualBufferAddressSafe(). The RealTek 8180 driver seems to need this.
2004-01-27 08:10:34 +00:00
ndis_buffer *buf;
{
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
return(MmGetMdlVirtualAddress(buf));
Implement NdisVirtualBufferAddress() and NdisVirtualBufferAddressSafe(). The RealTek 8180 driver seems to need this.
2004-01-27 08:10:34 +00:00
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void *
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisBufferVirtualAddressSafe(buf, prio)
Implement NdisVirtualBufferAddress() and NdisVirtualBufferAddressSafe(). The RealTek 8180 driver seems to need this.
2004-01-27 08:10:34 +00:00
ndis_buffer *buf;
uint32_t prio;
{
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
return(MmGetMdlVirtualAddress(buf));
Implement NdisVirtualBufferAddress() and NdisVirtualBufferAddressSafe(). The RealTek 8180 driver seems to need this.
2004-01-27 08:10:34 +00:00
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisAdjustBufferLength(buf, len)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_buffer *buf;
int len;
{
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
MmGetMdlByteCount(buf) = len;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static uint32_t
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisInterlockedIncrement(addend)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
uint32_t *addend;
{
Use atomic ops for the interlocked increment and decrement routines in subr_ndis and subr_ntoskrnl. This is faster and avoids potential LOR whinage from witness (an LOR couldn't happen with the old code since the interlocked inc/dec routines could not sleep with a lock held, but this will keep witness happy and it's more efficient anyway. I think.)
2004-01-07 07:29:27 +00:00
atomic_add_long((u_long *)addend, 1);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return(*addend);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static uint32_t
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisInterlockedDecrement(addend)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
uint32_t *addend;
{
Use atomic ops for the interlocked increment and decrement routines in subr_ndis and subr_ntoskrnl. This is faster and avoids potential LOR whinage from witness (an LOR couldn't happen with the old code since the interlocked inc/dec routines could not sleep with a lock held, but this will keep witness happy and it's more efficient anyway. I think.)
2004-01-07 07:29:27 +00:00
atomic_subtract_long((u_long *)addend, 1);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return(*addend);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisInitializeEvent(event)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_event *event;
{
- Rewrite the timer and event API routines in subr_ndis.c so that they are actually layered on top of the KeTimer API in subr_ntoskrnl.c, just as it is in Windows. This reduces code duplication and more closely imitates the way things are done in Windows. - Modify ndis_encode_parm() to deal with the case where we have a registry key expressed as a hex value ("0x1") which is being read via NdisReadConfiguration() as an int. Previously, we tried to decode things like "0x1" with strtol() using a base of 10, which would always yield 0. This is what was causing problems with the Intel 2200BG Centrino 802.11g driver: the .inf file that comes with it has a key called RadioEnable with a value of 0x1. We incorrectly decoded this value to '0' when it was queried, hence the driver thought we wanted the radio turned off. - In if_ndis.c, most drivers don't accept NDIS_80211_AUTHMODE_AUTO, but NDIS_80211_AUTHMODE_SHARED may not be right in some cases, so for now always use NDIS_80211_AUTHMODE_OPEN. NOTE: There is still one problem with the Intel 2200BG driver: it happens that the kernel stack in Windows is larger than the kernel stack in FreeBSD. The 2200BG driver sometimes eats up more than 2 pages of stack space, which can lead to a double fault panic. For the moment, I got things to work by adding the following to my kernel config file: options KSTACK_PAGES=8 I'm pretty sure 8 is too big; I just picked this value out of a hat as a test, and it happened to work, so I left it. 4 pages might be enough. Unfortunately, I don't think you can dynamically give a thread a larger stack, so I'm not sure how to handle this short of putting a note in the man page about it and dealing with the flood of mail from people who never read man pages.
2004-03-20 23:39:43 +00:00
/*
- In subr_ndis.c:ndis_init_event(), initialize events as notification objects rather than synchronization objects. When a sync object is signaled, only the first thread waiting on it is woken up, and then it's automatically reset to the not-signaled state. When a notification object is signaled, all threads waiting on it will be woken up, and it remains in the signaled state until someone resets it manually. We want the latter behavior for NDIS events. - In kern_ndis.c:ndis_convert_res(), we have to create a temporary copy of the list returned by BUS_GET_RESOURCE_LIST(). When the PCI bus code probes resources for a given device, it enters them into a singly linked list, head first. The result is that traversing this list gives you the resources in reverse order. This means when we create the Windows resource list, it will be in reverse order too. Unfortunately, this can hose drivers for devices with multiple I/O ranges of the same type, like, say, two memory mapped I/O regions (one for registers, one to map the NVRAM/bootrom/whatever). Some drivers test the range size to figure out which region is which, but others just assume that the resources will be listed in ascending order from lowest numbered BAR to highest. Reversing the order means such drivers will choose the wrong resource as their I/O register range. Since we can't traverse the resource SLIST backwards, we have to make a temporary copy of the list in the right order and then build the Windows resource list from that. I suppose we could just fix the PCI bus code to use a TAILQ instead, but then I'd have to track down all the consumers of the BUS_GET_RESOURCE_LIST() and fix them too.
2004-03-25 18:31:52 +00:00
* NDIS events are always notification
- Rewrite the timer and event API routines in subr_ndis.c so that they are actually layered on top of the KeTimer API in subr_ntoskrnl.c, just as it is in Windows. This reduces code duplication and more closely imitates the way things are done in Windows. - Modify ndis_encode_parm() to deal with the case where we have a registry key expressed as a hex value ("0x1") which is being read via NdisReadConfiguration() as an int. Previously, we tried to decode things like "0x1" with strtol() using a base of 10, which would always yield 0. This is what was causing problems with the Intel 2200BG Centrino 802.11g driver: the .inf file that comes with it has a key called RadioEnable with a value of 0x1. We incorrectly decoded this value to '0' when it was queried, hence the driver thought we wanted the radio turned off. - In if_ndis.c, most drivers don't accept NDIS_80211_AUTHMODE_AUTO, but NDIS_80211_AUTHMODE_SHARED may not be right in some cases, so for now always use NDIS_80211_AUTHMODE_OPEN. NOTE: There is still one problem with the Intel 2200BG driver: it happens that the kernel stack in Windows is larger than the kernel stack in FreeBSD. The 2200BG driver sometimes eats up more than 2 pages of stack space, which can lead to a double fault panic. For the moment, I got things to work by adding the following to my kernel config file: options KSTACK_PAGES=8 I'm pretty sure 8 is too big; I just picked this value out of a hat as a test, and it happened to work, so I left it. 4 pages might be enough. Unfortunately, I don't think you can dynamically give a thread a larger stack, so I'm not sure how to handle this short of putting a note in the man page about it and dealing with the flood of mail from people who never read man pages.
2004-03-20 23:39:43 +00:00
* events, and should be initialized to the
* not signaled state.
*/
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
KeInitializeEvent(&event->ne_event, EVENT_TYPE_NOTIFY, FALSE);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisSetEvent(event)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_event *event;
{
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
KeSetEvent(&event->ne_event, IO_NO_INCREMENT, FALSE);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisResetEvent(event)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_event *event;
{
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
KeResetEvent(&event->ne_event);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static uint8_t
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisWaitEvent(event, msecs)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_event *event;
uint32_t msecs;
{
- Rewrite the timer and event API routines in subr_ndis.c so that they are actually layered on top of the KeTimer API in subr_ntoskrnl.c, just as it is in Windows. This reduces code duplication and more closely imitates the way things are done in Windows. - Modify ndis_encode_parm() to deal with the case where we have a registry key expressed as a hex value ("0x1") which is being read via NdisReadConfiguration() as an int. Previously, we tried to decode things like "0x1" with strtol() using a base of 10, which would always yield 0. This is what was causing problems with the Intel 2200BG Centrino 802.11g driver: the .inf file that comes with it has a key called RadioEnable with a value of 0x1. We incorrectly decoded this value to '0' when it was queried, hence the driver thought we wanted the radio turned off. - In if_ndis.c, most drivers don't accept NDIS_80211_AUTHMODE_AUTO, but NDIS_80211_AUTHMODE_SHARED may not be right in some cases, so for now always use NDIS_80211_AUTHMODE_OPEN. NOTE: There is still one problem with the Intel 2200BG driver: it happens that the kernel stack in Windows is larger than the kernel stack in FreeBSD. The 2200BG driver sometimes eats up more than 2 pages of stack space, which can lead to a double fault panic. For the moment, I got things to work by adding the following to my kernel config file: options KSTACK_PAGES=8 I'm pretty sure 8 is too big; I just picked this value out of a hat as a test, and it happened to work, so I left it. 4 pages might be enough. Unfortunately, I don't think you can dynamically give a thread a larger stack, so I'm not sure how to handle this short of putting a note in the man page about it and dealing with the flood of mail from people who never read man pages.
2004-03-20 23:39:43 +00:00
int64_t duetime;
uint32_t rval;
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
- Rewrite the timer and event API routines in subr_ndis.c so that they are actually layered on top of the KeTimer API in subr_ntoskrnl.c, just as it is in Windows. This reduces code duplication and more closely imitates the way things are done in Windows. - Modify ndis_encode_parm() to deal with the case where we have a registry key expressed as a hex value ("0x1") which is being read via NdisReadConfiguration() as an int. Previously, we tried to decode things like "0x1" with strtol() using a base of 10, which would always yield 0. This is what was causing problems with the Intel 2200BG Centrino 802.11g driver: the .inf file that comes with it has a key called RadioEnable with a value of 0x1. We incorrectly decoded this value to '0' when it was queried, hence the driver thought we wanted the radio turned off. - In if_ndis.c, most drivers don't accept NDIS_80211_AUTHMODE_AUTO, but NDIS_80211_AUTHMODE_SHARED may not be right in some cases, so for now always use NDIS_80211_AUTHMODE_OPEN. NOTE: There is still one problem with the Intel 2200BG driver: it happens that the kernel stack in Windows is larger than the kernel stack in FreeBSD. The 2200BG driver sometimes eats up more than 2 pages of stack space, which can lead to a double fault panic. For the moment, I got things to work by adding the following to my kernel config file: options KSTACK_PAGES=8 I'm pretty sure 8 is too big; I just picked this value out of a hat as a test, and it happened to work, so I left it. 4 pages might be enough. Unfortunately, I don't think you can dynamically give a thread a larger stack, so I'm not sure how to handle this short of putting a note in the man page about it and dealing with the flood of mail from people who never read man pages.
2004-03-20 23:39:43 +00:00
duetime = ((int64_t)msecs * -10000);
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
rval = KeWaitForSingleObject(event,
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
0, 0, TRUE, msecs ? & duetime : NULL);
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
- Rewrite the timer and event API routines in subr_ndis.c so that they are actually layered on top of the KeTimer API in subr_ntoskrnl.c, just as it is in Windows. This reduces code duplication and more closely imitates the way things are done in Windows. - Modify ndis_encode_parm() to deal with the case where we have a registry key expressed as a hex value ("0x1") which is being read via NdisReadConfiguration() as an int. Previously, we tried to decode things like "0x1" with strtol() using a base of 10, which would always yield 0. This is what was causing problems with the Intel 2200BG Centrino 802.11g driver: the .inf file that comes with it has a key called RadioEnable with a value of 0x1. We incorrectly decoded this value to '0' when it was queried, hence the driver thought we wanted the radio turned off. - In if_ndis.c, most drivers don't accept NDIS_80211_AUTHMODE_AUTO, but NDIS_80211_AUTHMODE_SHARED may not be right in some cases, so for now always use NDIS_80211_AUTHMODE_OPEN. NOTE: There is still one problem with the Intel 2200BG driver: it happens that the kernel stack in Windows is larger than the kernel stack in FreeBSD. The 2200BG driver sometimes eats up more than 2 pages of stack space, which can lead to a double fault panic. For the moment, I got things to work by adding the following to my kernel config file: options KSTACK_PAGES=8 I'm pretty sure 8 is too big; I just picked this value out of a hat as a test, and it happened to work, so I left it. 4 pages might be enough. Unfortunately, I don't think you can dynamically give a thread a larger stack, so I'm not sure how to handle this short of putting a note in the man page about it and dealing with the flood of mail from people who never read man pages.
2004-03-20 23:39:43 +00:00
if (rval == STATUS_TIMEOUT)
return(FALSE);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
- Rewrite the timer and event API routines in subr_ndis.c so that they are actually layered on top of the KeTimer API in subr_ntoskrnl.c, just as it is in Windows. This reduces code duplication and more closely imitates the way things are done in Windows. - Modify ndis_encode_parm() to deal with the case where we have a registry key expressed as a hex value ("0x1") which is being read via NdisReadConfiguration() as an int. Previously, we tried to decode things like "0x1" with strtol() using a base of 10, which would always yield 0. This is what was causing problems with the Intel 2200BG Centrino 802.11g driver: the .inf file that comes with it has a key called RadioEnable with a value of 0x1. We incorrectly decoded this value to '0' when it was queried, hence the driver thought we wanted the radio turned off. - In if_ndis.c, most drivers don't accept NDIS_80211_AUTHMODE_AUTO, but NDIS_80211_AUTHMODE_SHARED may not be right in some cases, so for now always use NDIS_80211_AUTHMODE_OPEN. NOTE: There is still one problem with the Intel 2200BG driver: it happens that the kernel stack in Windows is larger than the kernel stack in FreeBSD. The 2200BG driver sometimes eats up more than 2 pages of stack space, which can lead to a double fault panic. For the moment, I got things to work by adding the following to my kernel config file: options KSTACK_PAGES=8 I'm pretty sure 8 is too big; I just picked this value out of a hat as a test, and it happened to work, so I left it. 4 pages might be enough. Unfortunately, I don't think you can dynamically give a thread a larger stack, so I'm not sure how to handle this short of putting a note in the man page about it and dealing with the flood of mail from people who never read man pages.
2004-03-20 23:39:43 +00:00
return(TRUE);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisUnicodeStringToAnsiString(dstr, sstr)
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
ansi_string *dstr;
unicode_string *sstr;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
{
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
uint32_t rval;
rval = RtlUnicodeStringToAnsiString(dstr, sstr, FALSE);
if (rval == STATUS_INSUFFICIENT_RESOURCES)
return(NDIS_STATUS_RESOURCES);
if (rval)
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
return(NDIS_STATUS_FAILURE);
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
return (NDIS_STATUS_SUCCESS);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisAnsiStringToUnicodeString(dstr, sstr)
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
unicode_string *dstr;
ansi_string *sstr;
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
{
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
uint32_t rval;
rval = RtlAnsiStringToUnicodeString(dstr, sstr, FALSE);
if (rval == STATUS_INSUFFICIENT_RESOURCES)
return(NDIS_STATUS_RESOURCES);
if (rval)
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
return(NDIS_STATUS_FAILURE);
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return (NDIS_STATUS_SUCCESS);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMPciAssignResources(adapter, slot, list)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_handle adapter;
uint32_t slot;
ndis_resource_list **list;
{
ndis_miniport_block *block;
if (adapter == NULL || list == NULL)
return (NDIS_STATUS_FAILURE);
block = (ndis_miniport_block *)adapter;
*list = block->nmb_rlist;
return (NDIS_STATUS_SUCCESS);
}
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
static void
ndis_intr(arg)
void *arg;
{
struct ndis_softc *sc;
struct ifnet *ifp;
int is_our_intr = 0;
int call_isr = 0;
ndis_miniport_interrupt *intr;
sc = arg;
ifp = sc->ifp;
intr = sc->ndis_block->nmb_interrupt;
if (intr == NULL || sc->ndis_block->nmb_miniportadapterctx == NULL)
return;
if (sc->ndis_block->nmb_interrupt->ni_isrreq == TRUE)
ndis_isr(sc, &is_our_intr, &call_isr);
else {
ndis_disable_intr(sc);
call_isr = 1;
}
if ((is_our_intr || call_isr))
IoRequestDpc(sc->ndis_block->nmb_deviceobj, NULL, sc);
return;
}
static void
ndis_intrhand(dpc, intr, sysarg1, sysarg2)
kdpc *dpc;
ndis_miniport_interrupt *intr;
void *sysarg1;
void *sysarg2;
{
struct ndis_softc *sc;
ndis_miniport_block *block;
ndis_handle adapter;
block = intr->ni_block;
adapter = block->nmb_miniportadapterctx;
sc = device_get_softc(block->nmb_physdeviceobj->do_devext);
if (NDIS_SERIALIZED(sc->ndis_block))
KeAcquireSpinLockAtDpcLevel(&block->nmb_lock);
MSCALL1(intr->ni_isrfunc, adapter);
/* If there's a MiniportEnableInterrupt() routine, call it. */
ndis_enable_intr(sc);
if (NDIS_SERIALIZED(sc->ndis_block))
KeReleaseSpinLockFromDpcLevel(&block->nmb_lock);
/*
* Set the completion event if we've drained all
* pending interrupts.
*/
KeAcquireSpinLockAtDpcLevel(&intr->ni_dpccountlock);
intr->ni_dpccnt--;
if (intr->ni_dpccnt == 0)
KeSetEvent(&intr->ni_dpcevt, IO_NO_INCREMENT, FALSE);
KeReleaseSpinLockFromDpcLevel(&intr->ni_dpccountlock);
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMRegisterInterrupt(intr, adapter, ivec, ilevel, reqisr, shared, imode)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_miniport_interrupt *intr;
ndis_handle adapter;
uint32_t ivec;
uint32_t ilevel;
uint8_t reqisr;
uint8_t shared;
ndis_interrupt_mode imode;
{
- Add pe_get_message() and pe_get_messagetable() for processing the RT_MESSAGETABLE resources that some driver binaries have. This allows us to print error messages in ndis_syslog(). - Correct the implementation of InterlockedIncrement() and InterlockedDecrement() -- they return uint32_t, not void. - Correct the declarations of the 64-bit arithmetic shift routines in subr_ntoskrnl.c (_allshr, allshl, etc...). These do not follow the _stdcall convention: instead, they appear to be __attribute__((regparm(3)). - Change the implementation of KeInitializeSpinLock(). There is no complementary KeFreeSpinLock() function, so creating a new mutex on each call to KeInitializeSpinLock() leaks resources when a driver is unloaded. For now, KeInitializeSpinLock() returns a handle to the ntoskrnl interlock mutex. - Use a driver's MiniportDisableInterrupt() and MiniportEnableInterrupt() routines if they exist. I'm not sure if I'm doing this right yet, but at the very least this shouldn't break any currently working drivers, and it makes the Intel PRO/1000 driver work. - In ndis_register_intr(), save some state that might be needed later, and save a pointer to the driver's interrupt structure in the ndis_miniport_block. - Save a pointer to the driver image for use by ndis_syslog() when it calls pe_get_message().
2004-01-06 07:09:26 +00:00
ndis_miniport_block *block;
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
ndis_miniport_characteristics *ch;
struct ndis_softc *sc;
int error;
- Add pe_get_message() and pe_get_messagetable() for processing the RT_MESSAGETABLE resources that some driver binaries have. This allows us to print error messages in ndis_syslog(). - Correct the implementation of InterlockedIncrement() and InterlockedDecrement() -- they return uint32_t, not void. - Correct the declarations of the 64-bit arithmetic shift routines in subr_ntoskrnl.c (_allshr, allshl, etc...). These do not follow the _stdcall convention: instead, they appear to be __attribute__((regparm(3)). - Change the implementation of KeInitializeSpinLock(). There is no complementary KeFreeSpinLock() function, so creating a new mutex on each call to KeInitializeSpinLock() leaks resources when a driver is unloaded. For now, KeInitializeSpinLock() returns a handle to the ntoskrnl interlock mutex. - Use a driver's MiniportDisableInterrupt() and MiniportEnableInterrupt() routines if they exist. I'm not sure if I'm doing this right yet, but at the very least this shouldn't break any currently working drivers, and it makes the Intel PRO/1000 driver work. - In ndis_register_intr(), save some state that might be needed later, and save a pointer to the driver's interrupt structure in the ndis_miniport_block. - Save a pointer to the driver image for use by ndis_syslog() when it calls pe_get_message().
2004-01-06 07:09:26 +00:00
block = adapter;
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
sc = device_get_softc(block->nmb_physdeviceobj->do_devext);
ch = IoGetDriverObjectExtension(block->nmb_deviceobj->do_drvobj,
(void *)1);
intr->ni_rsvd = ExAllocatePoolWithTag(NonPagedPool,
sizeof(struct mtx), 0);
if (intr->ni_rsvd == NULL)
return(NDIS_STATUS_RESOURCES);
- Add pe_get_message() and pe_get_messagetable() for processing the RT_MESSAGETABLE resources that some driver binaries have. This allows us to print error messages in ndis_syslog(). - Correct the implementation of InterlockedIncrement() and InterlockedDecrement() -- they return uint32_t, not void. - Correct the declarations of the 64-bit arithmetic shift routines in subr_ntoskrnl.c (_allshr, allshl, etc...). These do not follow the _stdcall convention: instead, they appear to be __attribute__((regparm(3)). - Change the implementation of KeInitializeSpinLock(). There is no complementary KeFreeSpinLock() function, so creating a new mutex on each call to KeInitializeSpinLock() leaks resources when a driver is unloaded. For now, KeInitializeSpinLock() returns a handle to the ntoskrnl interlock mutex. - Use a driver's MiniportDisableInterrupt() and MiniportEnableInterrupt() routines if they exist. I'm not sure if I'm doing this right yet, but at the very least this shouldn't break any currently working drivers, and it makes the Intel PRO/1000 driver work. - In ndis_register_intr(), save some state that might be needed later, and save a pointer to the driver's interrupt structure in the ndis_miniport_block. - Save a pointer to the driver image for use by ndis_syslog() when it calls pe_get_message().
2004-01-06 07:09:26 +00:00
Modify if_ndis.c so that the MiniportISR function runs in ndis_intr() and MiniportHandleInterrupt() is fired off later via a task queue in ndis_intrtask(). This more accurately follows the NDIS interrupt handling model, where the ISR does a minimal amount of work in interrupt context and the handler is defered and run at a lower priority. Create a separate ndis_intrmtx mutex just for the guarding the ISR. Modify NdisSynchronizeWithInterrupt() to aquire the ndis_intrmtx mutex before invoking the synchronized procedure. (The purpose of this function is to provide mutual exclusion for code that shares variables with the ISR.) Modify NdisMRegisterInterrupt() to save a pointer to the miniport block in the ndis_miniport_interrupt structure so that NdisSynchronizeWithInterrupt() can grab it later and derive ndis_intrmtx from it.
2004-01-04 21:22:25 +00:00
intr->ni_block = adapter;
- Add pe_get_message() and pe_get_messagetable() for processing the RT_MESSAGETABLE resources that some driver binaries have. This allows us to print error messages in ndis_syslog(). - Correct the implementation of InterlockedIncrement() and InterlockedDecrement() -- they return uint32_t, not void. - Correct the declarations of the 64-bit arithmetic shift routines in subr_ntoskrnl.c (_allshr, allshl, etc...). These do not follow the _stdcall convention: instead, they appear to be __attribute__((regparm(3)). - Change the implementation of KeInitializeSpinLock(). There is no complementary KeFreeSpinLock() function, so creating a new mutex on each call to KeInitializeSpinLock() leaks resources when a driver is unloaded. For now, KeInitializeSpinLock() returns a handle to the ntoskrnl interlock mutex. - Use a driver's MiniportDisableInterrupt() and MiniportEnableInterrupt() routines if they exist. I'm not sure if I'm doing this right yet, but at the very least this shouldn't break any currently working drivers, and it makes the Intel PRO/1000 driver work. - In ndis_register_intr(), save some state that might be needed later, and save a pointer to the driver's interrupt structure in the ndis_miniport_block. - Save a pointer to the driver image for use by ndis_syslog() when it calls pe_get_message().
2004-01-06 07:09:26 +00:00
intr->ni_isrreq = reqisr;
intr->ni_shared = shared;
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
intr->ni_dpccnt = 0;
intr->ni_isrfunc = ch->nmc_interrupt_func;
intr->ni_dpcfunc = ch->nmc_isr_func;
KeInitializeEvent(&intr->ni_dpcevt, EVENT_TYPE_NOTIFY, TRUE);
KeInitializeDpc(&intr->ni_dpc,
ndis_findwrap((funcptr)ndis_intrhand), intr);
KeSetImportanceDpc(&intr->ni_dpc, KDPC_IMPORTANCE_LOW);
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
error = IoConnectInterrupt(&intr->ni_introbj,
ndis_findwrap((funcptr)ndis_intr), sc, NULL,
ivec, ilevel, 0, imode, shared, 0, FALSE);
if (error != STATUS_SUCCESS)
return(NDIS_STATUS_FAILURE);
block->nmb_interrupt = intr;
Add support for Windows/x86-64 binaries to Project Evil. Ville-Pertti Keinonen (will at exomi dot comohmygodnospampleasekthx) deserves a big thanks for submitting initial patches to make it work. I have mangled his contributions appropriately. The main gotcha with Windows/x86-64 is that Microsoft uses a different calling convention than everyone else. The standard ABI requires using 6 registers for argument passing, with other arguments on the stack. Microsoft uses only 4 registers, and requires the caller to leave room on the stack for the register arguments incase the callee needs to spill them. Unlike x86, where Microsoft uses a mix of _cdecl, _stdcall and _fastcall, all routines on Windows/x86-64 uses the same convention. This unfortunately means that all the functions we export to the driver require an intermediate translation wrapper. Similarly, we have to wrap all calls back into the driver binary itself. The original patches provided macros to wrap every single routine at compile time, providing a secondary jump table with a customized wrapper for each exported routine. I decided to use a different approach: the call wrapper for each function is created from a template at runtime, and the routine to jump to is patched into the wrapper as it is created. The subr_pe module has been modified to patch in the wrapped function instead of the original. (On x86, the wrapping routine is a no-op.) There are some minor API differences that had to be accounted for: - KeAcquireSpinLock() is a real function on amd64, not a macro wrapper around KfAcquireSpinLock() - NdisFreeBuffer() is actually IoFreeMdl(). I had to change the whole NDIS_BUFFER API a bit to accomodate this. Bugs fixed along the way: - IoAllocateMdl() always returned NULL - kern_windrv.c:windrv_unload() wasn't releasing private driver object extensions correctly (found thanks to memguard) This has only been tested with the driver for the Broadcom 802.11g chipset, which was the only Windows/x86-64 driver I could find.
2005-02-16 05:41:18 +00:00
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return(NDIS_STATUS_SUCCESS);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMDeregisterInterrupt(intr)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_miniport_interrupt *intr;
{
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
ndis_miniport_block *block;
struct ndis_softc *sc;
uint8_t irql;
block = intr->ni_block;
sc = device_get_softc(block->nmb_physdeviceobj->do_devext);
/* Should really be KeSynchronizeExecution() */
KeAcquireSpinLock(intr->ni_introbj->ki_lock, &irql);
block->nmb_interrupt = NULL;
KeReleaseSpinLock(intr->ni_introbj->ki_lock, irql);
/*
KeFlushQueuedDpcs();
*/
/* Disconnect our ISR */
IoDisconnectInterrupt(intr->ni_introbj);
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
KeWaitForSingleObject(&intr->ni_dpcevt, 0, 0, FALSE, NULL);
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
KeResetEvent(&intr->ni_dpcevt);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMRegisterAdapterShutdownHandler(adapter, shutdownctx, shutdownfunc)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_handle adapter;
void *shutdownctx;
ndis_shutdown_handler shutdownfunc;
{
ndis_miniport_block *block;
ndis_miniport_characteristics *chars;
struct ndis_softc *sc;
if (adapter == NULL)
return;
block = (ndis_miniport_block *)adapter;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
sc = device_get_softc(block->nmb_physdeviceobj->do_devext);
chars = sc->ndis_chars;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
chars->nmc_shutdown_handler = shutdownfunc;
chars->nmc_rsvd0 = shutdownctx;
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMDeregisterAdapterShutdownHandler(adapter)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_handle adapter;
{
ndis_miniport_block *block;
ndis_miniport_characteristics *chars;
struct ndis_softc *sc;
if (adapter == NULL)
return;
block = (ndis_miniport_block *)adapter;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
sc = device_get_softc(block->nmb_physdeviceobj->do_devext);
chars = sc->ndis_chars;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
chars->nmc_shutdown_handler = NULL;
chars->nmc_rsvd0 = NULL;
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static uint32_t
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NDIS_BUFFER_TO_SPAN_PAGES(buf)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_buffer *buf;
{
Re-do the handling of ndis_buffers. The NDIS_BUFFER structure is supposed to be opaque to the driver, however it is exposed through several macros which expect certain behavior. In my original implementation, I used the mappedsystemva member of the structure to hold a pointer to the buffer and bytecount to hold the length. It turns out you must use the startva pointer to point to the page containing the start of the buffer and set byteoffset to the offset within the page where the buffer starts. So, for a buffer with address 'baseva,' startva is baseva & ~(PAGE_SIZE -1) and byteoffset is baseva & (PAGE_SIZE -1). We have to maintain this convention everywhere that ndis_buffers are used. Fortunately, Microsoft defines some macros for initializing and manipulating NDIS_BUFFER structures in ntddk.h. I adapted some of them for use here and used them where appropriate. This fixes the discrepancy I observed between how RX'ed packet sizes were being reported in the Broadcom wireless driver and the sample ethernet drivers that I've tested. This should also help the Intel Centrino wireless driver work. Also try to properly initialize the 802.11 BSS and IBSS channels. (Sadly, the channel value is meaningless since there's no way in the existing NDIS API to get/set the channel, but this should take care of any 'invalid channel (NULL)' messages printed on the console.
2003-12-23 04:08:22 +00:00
if (buf == NULL)
return(0);
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
if (MmGetMdlByteCount(buf) == 0)
Correct the implementation of NDIS_BUFFER_TO_SPAN_PAGES().
2003-12-13 09:07:35 +00:00
return(1);
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
return(SPAN_PAGES(MmGetMdlVirtualAddress(buf),
MmGetMdlByteCount(buf)));
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisGetBufferPhysicalArraySize(buf, pages)
Implement NdisGetBufferPhysicalArraySize(), which apparently is a synonym for NDIS_BUFFER_TO_SPAN_PAGES().
2003-12-16 18:56:33 +00:00
ndis_buffer *buf;
uint32_t *pages;
{
Re-do the handling of ndis_buffers. The NDIS_BUFFER structure is supposed to be opaque to the driver, however it is exposed through several macros which expect certain behavior. In my original implementation, I used the mappedsystemva member of the structure to hold a pointer to the buffer and bytecount to hold the length. It turns out you must use the startva pointer to point to the page containing the start of the buffer and set byteoffset to the offset within the page where the buffer starts. So, for a buffer with address 'baseva,' startva is baseva & ~(PAGE_SIZE -1) and byteoffset is baseva & (PAGE_SIZE -1). We have to maintain this convention everywhere that ndis_buffers are used. Fortunately, Microsoft defines some macros for initializing and manipulating NDIS_BUFFER structures in ntddk.h. I adapted some of them for use here and used them where appropriate. This fixes the discrepancy I observed between how RX'ed packet sizes were being reported in the Broadcom wireless driver and the sample ethernet drivers that I've tested. This should also help the Intel Centrino wireless driver work. Also try to properly initialize the 802.11 BSS and IBSS channels. (Sadly, the channel value is meaningless since there's no way in the existing NDIS API to get/set the channel, but this should take care of any 'invalid channel (NULL)' messages printed on the console.
2003-12-23 04:08:22 +00:00
if (buf == NULL)
return;
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
*pages = NDIS_BUFFER_TO_SPAN_PAGES(buf);
Implement NdisGetBufferPhysicalArraySize(), which apparently is a synonym for NDIS_BUFFER_TO_SPAN_PAGES().
2003-12-16 18:56:33 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisQueryBufferOffset(buf, off, len)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_buffer *buf;
uint32_t *off;
uint32_t *len;
{
Re-do the handling of ndis_buffers. The NDIS_BUFFER structure is supposed to be opaque to the driver, however it is exposed through several macros which expect certain behavior. In my original implementation, I used the mappedsystemva member of the structure to hold a pointer to the buffer and bytecount to hold the length. It turns out you must use the startva pointer to point to the page containing the start of the buffer and set byteoffset to the offset within the page where the buffer starts. So, for a buffer with address 'baseva,' startva is baseva & ~(PAGE_SIZE -1) and byteoffset is baseva & (PAGE_SIZE -1). We have to maintain this convention everywhere that ndis_buffers are used. Fortunately, Microsoft defines some macros for initializing and manipulating NDIS_BUFFER structures in ntddk.h. I adapted some of them for use here and used them where appropriate. This fixes the discrepancy I observed between how RX'ed packet sizes were being reported in the Broadcom wireless driver and the sample ethernet drivers that I've tested. This should also help the Intel Centrino wireless driver work. Also try to properly initialize the 802.11 BSS and IBSS channels. (Sadly, the channel value is meaningless since there's no way in the existing NDIS API to get/set the channel, but this should take care of any 'invalid channel (NULL)' messages printed on the console.
2003-12-23 04:08:22 +00:00
if (buf == NULL)
return;
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
*off = MmGetMdlByteOffset(buf);
*len = MmGetMdlByteCount(buf);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMSleep(usecs)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
uint32_t usecs;
{
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
ktimer timer;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
Throw the switch on the new driver generation/loading mechanism. From here on in, if_ndis.ko will be pre-built as a module, and can be built into a static kernel (though it's not part of GENERIC). Drivers are created using the new ndisgen(8) script, which uses ndiscvt(8) under the covers, along with a few other tools. The result is a driver module that can be kldloaded into the kernel. A driver with foo.inf and foo.sys files will be converted into foo_sys.ko (and foo_sys.o, for those who want/need to make static kernels). This module contains all of the necessary info from the .INF file and the driver binary image, converted into an ELF module. You can kldload this module (or add it to /boot/loader.conf) to have it loaded automatically. Any required firmware files can be bundled into the module as well (or converted/loaded separately). Also, add a workaround for a problem in NdisMSleep(). During system bootstrap (cold == 1), msleep() always returns 0 without actually sleeping. The Intel 2200BG driver uses NdisMSleep() to wait for the NIC's firmware to come to life, and fails to load if NdisMSleep() doesn't actually delay. As a workaround, if msleep() (and hence ndis_thsuspend()) returns 0, use a hard DELAY() to sleep instead). This is not really the right thing to do, but we can't really do much else. At the very least, this makes the Intel driver happy. There are probably other drivers that fail in this way during bootstrap. Unfortunately, the only workaround for those is to avoid pre-loading them and kldload them once the system is running instead.
2005-04-24 20:21:22 +00:00
/*
* During system bootstrap, (i.e. cold == 1), we aren't
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
* allowed to sleep, so we have to do a hard DELAY()
* instead.
Throw the switch on the new driver generation/loading mechanism. From here on in, if_ndis.ko will be pre-built as a module, and can be built into a static kernel (though it's not part of GENERIC). Drivers are created using the new ndisgen(8) script, which uses ndiscvt(8) under the covers, along with a few other tools. The result is a driver module that can be kldloaded into the kernel. A driver with foo.inf and foo.sys files will be converted into foo_sys.ko (and foo_sys.o, for those who want/need to make static kernels). This module contains all of the necessary info from the .INF file and the driver binary image, converted into an ELF module. You can kldload this module (or add it to /boot/loader.conf) to have it loaded automatically. Any required firmware files can be bundled into the module as well (or converted/loaded separately). Also, add a workaround for a problem in NdisMSleep(). During system bootstrap (cold == 1), msleep() always returns 0 without actually sleeping. The Intel 2200BG driver uses NdisMSleep() to wait for the NIC's firmware to come to life, and fails to load if NdisMSleep() doesn't actually delay. As a workaround, if msleep() (and hence ndis_thsuspend()) returns 0, use a hard DELAY() to sleep instead). This is not really the right thing to do, but we can't really do much else. At the very least, this makes the Intel driver happy. There are probably other drivers that fail in this way during bootstrap. Unfortunately, the only workaround for those is to avoid pre-loading them and kldload them once the system is running instead.
2005-04-24 20:21:22 +00:00
*/
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
if (cold)
Throw the switch on the new driver generation/loading mechanism. From here on in, if_ndis.ko will be pre-built as a module, and can be built into a static kernel (though it's not part of GENERIC). Drivers are created using the new ndisgen(8) script, which uses ndiscvt(8) under the covers, along with a few other tools. The result is a driver module that can be kldloaded into the kernel. A driver with foo.inf and foo.sys files will be converted into foo_sys.ko (and foo_sys.o, for those who want/need to make static kernels). This module contains all of the necessary info from the .INF file and the driver binary image, converted into an ELF module. You can kldload this module (or add it to /boot/loader.conf) to have it loaded automatically. Any required firmware files can be bundled into the module as well (or converted/loaded separately). Also, add a workaround for a problem in NdisMSleep(). During system bootstrap (cold == 1), msleep() always returns 0 without actually sleeping. The Intel 2200BG driver uses NdisMSleep() to wait for the NIC's firmware to come to life, and fails to load if NdisMSleep() doesn't actually delay. As a workaround, if msleep() (and hence ndis_thsuspend()) returns 0, use a hard DELAY() to sleep instead). This is not really the right thing to do, but we can't really do much else. At the very least, this makes the Intel driver happy. There are probably other drivers that fail in this way during bootstrap. Unfortunately, the only workaround for those is to avoid pre-loading them and kldload them once the system is running instead.
2005-04-24 20:21:22 +00:00
DELAY(usecs);
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
else {
KeInitializeTimer(&timer);
KeSetTimer(&timer, ((int64_t)usecs * -10), NULL);
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
KeWaitForSingleObject(&timer, 0, 0, FALSE, NULL);
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
}
Another round of cleanups and fixes: - Change ndis_return() from a DPC to a workitem so that it doesn't run at DISPATCH_LEVEL (with the dispatcher lock held). - In if_ndis.c, submit packets to the stack via (*ifp->if_input)() in a workitem instead of doing it directly in ndis_rxeof(), because ndis_rxeof() runs in a DPC, and hence at DISPATCH_LEVEL. This implies that the 'dispatch level' mutex for the current CPU is being held, and we don't want to call if_input while holding any locks. - Reimplement IoConnectInterrupt()/IoDisconnectInterrupt(). The original approach I used to track down the interrupt resource (by scanning the device tree starting at the nexus) is prone to problems when two devices share an interrupt. (E.g removing ndis1 might disable interrupts for ndis0.) The new approach is to multiplex all the NDIS interrupts through a common internal dispatcher (ntoskrnl_intr()) and allow IoConnectInterrupt()/IoDisconnectInterrupt() to add or remove interrupts from the dispatch list. - Implement KeAcquireInterruptSpinLock() and KeReleaseInterruptSpinLock(). - Change the DPC and workitem threads to use the KeXXXSpinLock API instead of mtx_lock_spin()/mtx_unlock_spin(). - Simplify the NdisXXXPacket routines by creating an actual packet pool structure and using the InterlockedSList routines to manage the packet queue. - Only honor the value returned by OID_GEN_MAXIMUM_SEND_PACKETS for serialized drivers. For deserialized drivers, we now create a packet array of 64 entries. (The Microsoft DDK documentation says that for deserialized miniports, OID_GEN_MAXIMUM_SEND_PACKETS is ignored, and the driver for the Marvell 8335 chip, which is a deserialized miniport, returns 1 when queried.) - Clean up timer handling in subr_ntoskrnl. - Add the following conditional debugging code: NTOSKRNL_DEBUG_TIMERS - add debugging and stats for timers NDIS_DEBUG_PACKETS - add extra sanity checking for NdisXXXPacket API NTOSKRNL_DEBUG_SPINLOCKS - add test for spinning too long - In kern_ndis.c, always start the HAL first and shut it down last, since Windows spinlocks depend on it. Ntoskrnl should similarly be started second and shut down next to last.
2005-10-18 19:52:15 +00:00
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static uint32_t
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisReadPcmciaAttributeMemory(handle, offset, buf, len)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_handle handle;
uint32_t offset;
void *buf;
uint32_t len;
{
struct ndis_softc *sc;
ndis_miniport_block *block;
bus_space_handle_t bh;
bus_space_tag_t bt;
char *dest;
int i;
if (handle == NULL)
return(0);
block = (ndis_miniport_block *)handle;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
sc = device_get_softc(block->nmb_physdeviceobj->do_devext);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
dest = buf;
bh = rman_get_bushandle(sc->ndis_res_am);
bt = rman_get_bustag(sc->ndis_res_am);
for (i = 0; i < len; i++)
Make NdisReadPcmciaAttributeMemory() and NdisWritePcmciaAttributeMemory() actually work. Make the PCI and PCCARD attachments provide a bus_get_resource_list() method so that resource listing for PCCARD works. PCCARD does not have a bus_get_resource_list() method (yet), so I faked up the resource list management in if_ndis_pccard.c, and added bus_get_resource_list() methods to both if_ndis_pccard.c and if_ndis_pci.c. The one in the PCI attechment just hands off to the PCI bus code. The difference is transparent to the NDIS resource handler code. Fixed ndis_open_file() so that opening files which live on NFS filesystems work: pass an actual ucred structure to VOP_GETATTR() (NFS explodes if the ucred structure is NOCRED). Make NdisMMapIoSpace() handle mapping of PCMCIA attribute memory resources correctly. Turn subr_ndis.c:my_strcasecmp() into ndis_strcasecmp() and export it so that if_ndis_pccard.c can use it, and junk the other copy of my_strcasecmp() from if_ndis_pccard.c.
2004-07-11 00:19:30 +00:00
dest[i] = bus_space_read_1(bt, bh, (offset + i) * 2);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return(i);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static uint32_t
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisWritePcmciaAttributeMemory(handle, offset, buf, len)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_handle handle;
uint32_t offset;
void *buf;
uint32_t len;
{
struct ndis_softc *sc;
ndis_miniport_block *block;
bus_space_handle_t bh;
bus_space_tag_t bt;
char *src;
int i;
if (handle == NULL)
return(0);
block = (ndis_miniport_block *)handle;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
sc = device_get_softc(block->nmb_physdeviceobj->do_devext);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
src = buf;
bh = rman_get_bushandle(sc->ndis_res_am);
bt = rman_get_bustag(sc->ndis_res_am);
for (i = 0; i < len; i++)
Make NdisReadPcmciaAttributeMemory() and NdisWritePcmciaAttributeMemory() actually work. Make the PCI and PCCARD attachments provide a bus_get_resource_list() method so that resource listing for PCCARD works. PCCARD does not have a bus_get_resource_list() method (yet), so I faked up the resource list management in if_ndis_pccard.c, and added bus_get_resource_list() methods to both if_ndis_pccard.c and if_ndis_pci.c. The one in the PCI attechment just hands off to the PCI bus code. The difference is transparent to the NDIS resource handler code. Fixed ndis_open_file() so that opening files which live on NFS filesystems work: pass an actual ucred structure to VOP_GETATTR() (NFS explodes if the ucred structure is NOCRED). Make NdisMMapIoSpace() handle mapping of PCMCIA attribute memory resources correctly. Turn subr_ndis.c:my_strcasecmp() into ndis_strcasecmp() and export it so that if_ndis_pccard.c can use it, and junk the other copy of my_strcasecmp() from if_ndis_pccard.c.
2004-07-11 00:19:30 +00:00
bus_space_write_1(bt, bh, (offset + i) * 2, src[i]);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return(i);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static list_entry *
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisInterlockedInsertHeadList(head, entry, lock)
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
list_entry *head;
list_entry *entry;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_spin_lock *lock;
{
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
list_entry *flink;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
KeAcquireSpinLock(&lock->nsl_spinlock, &lock->nsl_kirql);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
flink = head->nle_flink;
entry->nle_flink = flink;
entry->nle_blink = head;
flink->nle_blink = entry;
head->nle_flink = entry;
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
KeReleaseSpinLock(&lock->nsl_spinlock, lock->nsl_kirql);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return(flink);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static list_entry *
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisInterlockedRemoveHeadList(head, lock)
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
list_entry *head;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_spin_lock *lock;
{
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
list_entry *flink;
list_entry *entry;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
KeAcquireSpinLock(&lock->nsl_spinlock, &lock->nsl_kirql);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
entry = head->nle_flink;
flink = entry->nle_flink;
head->nle_flink = flink;
flink->nle_blink = head;
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
KeReleaseSpinLock(&lock->nsl_spinlock, lock->nsl_kirql);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return(entry);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static list_entry *
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisInterlockedInsertTailList(head, entry, lock)
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
list_entry *head;
list_entry *entry;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_spin_lock *lock;
{
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
list_entry *blink;
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
KeAcquireSpinLock(&lock->nsl_spinlock, &lock->nsl_kirql);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
blink = head->nle_blink;
entry->nle_flink = head;
entry->nle_blink = blink;
blink->nle_flink = entry;
head->nle_blink = entry;
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
KeReleaseSpinLock(&lock->nsl_spinlock, lock->nsl_kirql);
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
return(blink);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static uint8_t
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMSynchronizeWithInterrupt(intr, syncfunc, syncctx)
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
ndis_miniport_interrupt *intr;
void *syncfunc;
void *syncctx;
{
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
return(KeSynchronizeExecution(intr->ni_introbj, syncfunc, syncctx));
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
}
Implement some more NDIS and ntoskrnl API calls: subr_ndis.c: NdisGetCurrentSystemTime() which, according to the Microsoft documentation returns "the number of 100 nanosecond intervals since January 1, 1601." I have no idea what's so special about that epoch or why they chose 100 nanosecond ticks. I don't know the proper offset to convert nanotime() from the UNIX epoch to January 1, 1601, so for now I'm just doing the unit convertion to 100s of nanoseconds. subr_ntoskrnl.c: memcpy(), memset(), ExInterlockedPopEntrySList(), ExInterlockedPushEntrySList(). The latter two are different from InterlockedPopEntrySList() and InterlockedPushEntrySList() in that they accept a spinlock to hold while executing, whereas the non-Ex routines use a lock internal to ntoskrnl. I also modified ExInitializePagedLookasideList() and ExInitializeNPagedLookasideList() to initialize mutex locks within the lookaside structures. It seems that in NDIS 5.0, the lookaside allocate/free routines ExInterlockedPopEntrySList() and ExInterlockedPushEntrySList(), which require the use of the per-lookaside spinlock, whereas in NDIS 5.1, the per-lookaside spinlock is deprecated. We need to support both cases. Note that I appear to be doing something wrong with ExInterlockedPopEntrySList() and ExInterlockedPushEntrySList(): they don't appear to obtain proper pointers to their arguments, so I'm probably doing something wrong in terms of their calling convention (they're declared to be FASTCALL in Widnows, and I'm not sure what that means for gcc). It happens that in my stub lookaside implementation, they don't need to do any work anyway, so for now I've hacked them to always return NULL, which avoids corrupting the stack. I need to do this right though.
2003-12-12 22:35:13 +00:00
/*
* Return the number of 100 nanosecond intervals since
* January 1, 1601. (?!?!)
*/
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisGetCurrentSystemTime(tval)
Implement some more NDIS and ntoskrnl API calls: subr_ndis.c: NdisGetCurrentSystemTime() which, according to the Microsoft documentation returns "the number of 100 nanosecond intervals since January 1, 1601." I have no idea what's so special about that epoch or why they chose 100 nanosecond ticks. I don't know the proper offset to convert nanotime() from the UNIX epoch to January 1, 1601, so for now I'm just doing the unit convertion to 100s of nanoseconds. subr_ntoskrnl.c: memcpy(), memset(), ExInterlockedPopEntrySList(), ExInterlockedPushEntrySList(). The latter two are different from InterlockedPopEntrySList() and InterlockedPushEntrySList() in that they accept a spinlock to hold while executing, whereas the non-Ex routines use a lock internal to ntoskrnl. I also modified ExInitializePagedLookasideList() and ExInitializeNPagedLookasideList() to initialize mutex locks within the lookaside structures. It seems that in NDIS 5.0, the lookaside allocate/free routines ExInterlockedPopEntrySList() and ExInterlockedPushEntrySList(), which require the use of the per-lookaside spinlock, whereas in NDIS 5.1, the per-lookaside spinlock is deprecated. We need to support both cases. Note that I appear to be doing something wrong with ExInterlockedPopEntrySList() and ExInterlockedPushEntrySList(): they don't appear to obtain proper pointers to their arguments, so I'm probably doing something wrong in terms of their calling convention (they're declared to be FASTCALL in Widnows, and I'm not sure what that means for gcc). It happens that in my stub lookaside implementation, they don't need to do any work anyway, so for now I've hacked them to always return NULL, which avoids corrupting the stack. I need to do this right though.
2003-12-12 22:35:13 +00:00
uint64_t *tval;
{
struct timespec ts;
- Add stubs for Ndis*File() functions - Fix ndis_time(). - Implement NdisGetSystemUpTime(). - Implement RtlCopyUnicodeString() and RtlUnicodeStringToAnsiString(). - In ndis_getstate_80211(), use sc->ndis_link to determine connect status. Submitted by: Brian Feldman <green@freebsd.org>
2003-12-25 00:40:02 +00:00
Implement some more NDIS and ntoskrnl API calls: subr_ndis.c: NdisGetCurrentSystemTime() which, according to the Microsoft documentation returns "the number of 100 nanosecond intervals since January 1, 1601." I have no idea what's so special about that epoch or why they chose 100 nanosecond ticks. I don't know the proper offset to convert nanotime() from the UNIX epoch to January 1, 1601, so for now I'm just doing the unit convertion to 100s of nanoseconds. subr_ntoskrnl.c: memcpy(), memset(), ExInterlockedPopEntrySList(), ExInterlockedPushEntrySList(). The latter two are different from InterlockedPopEntrySList() and InterlockedPushEntrySList() in that they accept a spinlock to hold while executing, whereas the non-Ex routines use a lock internal to ntoskrnl. I also modified ExInitializePagedLookasideList() and ExInitializeNPagedLookasideList() to initialize mutex locks within the lookaside structures. It seems that in NDIS 5.0, the lookaside allocate/free routines ExInterlockedPopEntrySList() and ExInterlockedPushEntrySList(), which require the use of the per-lookaside spinlock, whereas in NDIS 5.1, the per-lookaside spinlock is deprecated. We need to support both cases. Note that I appear to be doing something wrong with ExInterlockedPopEntrySList() and ExInterlockedPushEntrySList(): they don't appear to obtain proper pointers to their arguments, so I'm probably doing something wrong in terms of their calling convention (they're declared to be FASTCALL in Widnows, and I'm not sure what that means for gcc). It happens that in my stub lookaside implementation, they don't need to do any work anyway, so for now I've hacked them to always return NULL, which avoids corrupting the stack. I need to do this right though.
2003-12-12 22:35:13 +00:00
nanotime(&ts);
- Add stubs for Ndis*File() functions - Fix ndis_time(). - Implement NdisGetSystemUpTime(). - Implement RtlCopyUnicodeString() and RtlUnicodeStringToAnsiString(). - In ndis_getstate_80211(), use sc->ndis_link to determine connect status. Submitted by: Brian Feldman <green@freebsd.org>
2003-12-25 00:40:02 +00:00
*tval = (uint64_t)ts.tv_nsec / 100 + (uint64_t)ts.tv_sec * 10000000 +
11644473600;
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
return;
- Add stubs for Ndis*File() functions - Fix ndis_time(). - Implement NdisGetSystemUpTime(). - Implement RtlCopyUnicodeString() and RtlUnicodeStringToAnsiString(). - In ndis_getstate_80211(), use sc->ndis_link to determine connect status. Submitted by: Brian Feldman <green@freebsd.org>
2003-12-25 00:40:02 +00:00
}
/*
* Return the number of milliseconds since the system booted.
*/
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisGetSystemUpTime(tval)
- Add stubs for Ndis*File() functions - Fix ndis_time(). - Implement NdisGetSystemUpTime(). - Implement RtlCopyUnicodeString() and RtlUnicodeStringToAnsiString(). - In ndis_getstate_80211(), use sc->ndis_link to determine connect status. Submitted by: Brian Feldman <green@freebsd.org>
2003-12-25 00:40:02 +00:00
uint32_t *tval;
{
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
struct timespec ts;
nanouptime(&ts);
*tval = ts.tv_nsec / 1000000 + ts.tv_sec * 1000;
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
return;
Implement some more NDIS and ntoskrnl API calls: subr_ndis.c: NdisGetCurrentSystemTime() which, according to the Microsoft documentation returns "the number of 100 nanosecond intervals since January 1, 1601." I have no idea what's so special about that epoch or why they chose 100 nanosecond ticks. I don't know the proper offset to convert nanotime() from the UNIX epoch to January 1, 1601, so for now I'm just doing the unit convertion to 100s of nanoseconds. subr_ntoskrnl.c: memcpy(), memset(), ExInterlockedPopEntrySList(), ExInterlockedPushEntrySList(). The latter two are different from InterlockedPopEntrySList() and InterlockedPushEntrySList() in that they accept a spinlock to hold while executing, whereas the non-Ex routines use a lock internal to ntoskrnl. I also modified ExInitializePagedLookasideList() and ExInitializeNPagedLookasideList() to initialize mutex locks within the lookaside structures. It seems that in NDIS 5.0, the lookaside allocate/free routines ExInterlockedPopEntrySList() and ExInterlockedPushEntrySList(), which require the use of the per-lookaside spinlock, whereas in NDIS 5.1, the per-lookaside spinlock is deprecated. We need to support both cases. Note that I appear to be doing something wrong with ExInterlockedPopEntrySList() and ExInterlockedPushEntrySList(): they don't appear to obtain proper pointers to their arguments, so I'm probably doing something wrong in terms of their calling convention (they're declared to be FASTCALL in Widnows, and I'm not sure what that means for gcc). It happens that in my stub lookaside implementation, they don't need to do any work anyway, so for now I've hacked them to always return NULL, which avoids corrupting the stack. I need to do this right though.
2003-12-12 22:35:13 +00:00
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisInitializeString(dst, src)
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
unicode_string *dst;
subr_ndis.c: - fix ndis_time() so that it returns a time based on the proper epoch (wacky though it may be) - implement NdisInitializeString() and NdisFreeString(), and add stub for NdisMRemoveMiniport() ntoskrnl_var.h: - add missing member to the general_lookaside struct (gl_listentry) subr_ntoskrnl.c: - Fix arguments to the interlocked push/pop routines: 'head' is an slist_header *, not an slist_entry * - Kludge up _fastcall support for the push/pop routines. The _fastcall convention is similar to _stdcall, except the first two available DWORD-sized arguments are passed in %ecx and %edx, respectively. One kludge for this __attribute__ ((regparm(3))), however this isn't entirely right, as it assumes %eax, %ecx and %edx will be used (regparm(2) assumes %eax and %edx). Another kludge is to declare the two fastcall-ed args as local register variables and explicitly assign them to %ecx and %edx, but experimentation showed that gcc would not guard %ecx and %edx against being clobbered. Thus, I came up with a 3rd kludge, which is to use some inline assembly of the form: void *arg1; void *arg2; __asm__("movl %%ecx, %%ecx" : "=c" (arg1)); __asm__("movl %%edx, %%edx" : "=d" (arg2)); This lets gcc know that we're going to reference %ecx and %edx and that it should make an effort not to let it get trampled. This wastes an instruction (movl %reg, %reg is a no-op) but insures proper behavior. It's possible there's a better way to do this though: this is the first time I've used inline assembler in this fashion. The above fixes to ntoskrnl_var.h an subr_ntoskrnl.c make lookaside lists work for the two drivers I have that use them, one of which is an NDIS 5.0 miniport and another which is 5.1.
2003-12-13 07:41:12 +00:00
char *src;
{
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
ansi_string as;
RtlInitAnsiString(&as, src);
RtlAnsiStringToUnicodeString(dst, &as, TRUE);
subr_ndis.c: - fix ndis_time() so that it returns a time based on the proper epoch (wacky though it may be) - implement NdisInitializeString() and NdisFreeString(), and add stub for NdisMRemoveMiniport() ntoskrnl_var.h: - add missing member to the general_lookaside struct (gl_listentry) subr_ntoskrnl.c: - Fix arguments to the interlocked push/pop routines: 'head' is an slist_header *, not an slist_entry * - Kludge up _fastcall support for the push/pop routines. The _fastcall convention is similar to _stdcall, except the first two available DWORD-sized arguments are passed in %ecx and %edx, respectively. One kludge for this __attribute__ ((regparm(3))), however this isn't entirely right, as it assumes %eax, %ecx and %edx will be used (regparm(2) assumes %eax and %edx). Another kludge is to declare the two fastcall-ed args as local register variables and explicitly assign them to %ecx and %edx, but experimentation showed that gcc would not guard %ecx and %edx against being clobbered. Thus, I came up with a 3rd kludge, which is to use some inline assembly of the form: void *arg1; void *arg2; __asm__("movl %%ecx, %%ecx" : "=c" (arg1)); __asm__("movl %%edx, %%edx" : "=d" (arg2)); This lets gcc know that we're going to reference %ecx and %edx and that it should make an effort not to let it get trampled. This wastes an instruction (movl %reg, %reg is a no-op) but insures proper behavior. It's possible there's a better way to do this though: this is the first time I've used inline assembler in this fashion. The above fixes to ntoskrnl_var.h an subr_ntoskrnl.c make lookaside lists work for the two drivers I have that use them, one of which is an NDIS 5.0 miniport and another which is 5.1.
2003-12-13 07:41:12 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisFreeString(str)
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
unicode_string *str;
subr_ndis.c: - fix ndis_time() so that it returns a time based on the proper epoch (wacky though it may be) - implement NdisInitializeString() and NdisFreeString(), and add stub for NdisMRemoveMiniport() ntoskrnl_var.h: - add missing member to the general_lookaside struct (gl_listentry) subr_ntoskrnl.c: - Fix arguments to the interlocked push/pop routines: 'head' is an slist_header *, not an slist_entry * - Kludge up _fastcall support for the push/pop routines. The _fastcall convention is similar to _stdcall, except the first two available DWORD-sized arguments are passed in %ecx and %edx, respectively. One kludge for this __attribute__ ((regparm(3))), however this isn't entirely right, as it assumes %eax, %ecx and %edx will be used (regparm(2) assumes %eax and %edx). Another kludge is to declare the two fastcall-ed args as local register variables and explicitly assign them to %ecx and %edx, but experimentation showed that gcc would not guard %ecx and %edx against being clobbered. Thus, I came up with a 3rd kludge, which is to use some inline assembly of the form: void *arg1; void *arg2; __asm__("movl %%ecx, %%ecx" : "=c" (arg1)); __asm__("movl %%edx, %%edx" : "=d" (arg2)); This lets gcc know that we're going to reference %ecx and %edx and that it should make an effort not to let it get trampled. This wastes an instruction (movl %reg, %reg is a no-op) but insures proper behavior. It's possible there's a better way to do this though: this is the first time I've used inline assembler in this fashion. The above fixes to ntoskrnl_var.h an subr_ntoskrnl.c make lookaside lists work for the two drivers I have that use them, one of which is an NDIS 5.0 miniport and another which is 5.1.
2003-12-13 07:41:12 +00:00
{
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
RtlFreeUnicodeString(str);
subr_ndis.c: - fix ndis_time() so that it returns a time based on the proper epoch (wacky though it may be) - implement NdisInitializeString() and NdisFreeString(), and add stub for NdisMRemoveMiniport() ntoskrnl_var.h: - add missing member to the general_lookaside struct (gl_listentry) subr_ntoskrnl.c: - Fix arguments to the interlocked push/pop routines: 'head' is an slist_header *, not an slist_entry * - Kludge up _fastcall support for the push/pop routines. The _fastcall convention is similar to _stdcall, except the first two available DWORD-sized arguments are passed in %ecx and %edx, respectively. One kludge for this __attribute__ ((regparm(3))), however this isn't entirely right, as it assumes %eax, %ecx and %edx will be used (regparm(2) assumes %eax and %edx). Another kludge is to declare the two fastcall-ed args as local register variables and explicitly assign them to %ecx and %edx, but experimentation showed that gcc would not guard %ecx and %edx against being clobbered. Thus, I came up with a 3rd kludge, which is to use some inline assembly of the form: void *arg1; void *arg2; __asm__("movl %%ecx, %%ecx" : "=c" (arg1)); __asm__("movl %%edx, %%edx" : "=d" (arg2)); This lets gcc know that we're going to reference %ecx and %edx and that it should make an effort not to let it get trampled. This wastes an instruction (movl %reg, %reg is a no-op) but insures proper behavior. It's possible there's a better way to do this though: this is the first time I've used inline assembler in this fashion. The above fixes to ntoskrnl_var.h an subr_ntoskrnl.c make lookaside lists work for the two drivers I have that use them, one of which is an NDIS 5.0 miniport and another which is 5.1.
2003-12-13 07:41:12 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMRemoveMiniport(adapter)
subr_ndis.c: - fix ndis_time() so that it returns a time based on the proper epoch (wacky though it may be) - implement NdisInitializeString() and NdisFreeString(), and add stub for NdisMRemoveMiniport() ntoskrnl_var.h: - add missing member to the general_lookaside struct (gl_listentry) subr_ntoskrnl.c: - Fix arguments to the interlocked push/pop routines: 'head' is an slist_header *, not an slist_entry * - Kludge up _fastcall support for the push/pop routines. The _fastcall convention is similar to _stdcall, except the first two available DWORD-sized arguments are passed in %ecx and %edx, respectively. One kludge for this __attribute__ ((regparm(3))), however this isn't entirely right, as it assumes %eax, %ecx and %edx will be used (regparm(2) assumes %eax and %edx). Another kludge is to declare the two fastcall-ed args as local register variables and explicitly assign them to %ecx and %edx, but experimentation showed that gcc would not guard %ecx and %edx against being clobbered. Thus, I came up with a 3rd kludge, which is to use some inline assembly of the form: void *arg1; void *arg2; __asm__("movl %%ecx, %%ecx" : "=c" (arg1)); __asm__("movl %%edx, %%edx" : "=d" (arg2)); This lets gcc know that we're going to reference %ecx and %edx and that it should make an effort not to let it get trampled. This wastes an instruction (movl %reg, %reg is a no-op) but insures proper behavior. It's possible there's a better way to do this though: this is the first time I've used inline assembler in this fashion. The above fixes to ntoskrnl_var.h an subr_ntoskrnl.c make lookaside lists work for the two drivers I have that use them, one of which is an NDIS 5.0 miniport and another which is 5.1.
2003-12-13 07:41:12 +00:00
ndis_handle *adapter;
{
return(NDIS_STATUS_SUCCESS);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisInitAnsiString(dst, src)
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
ansi_string *dst;
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
char *src;
{
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
RtlInitAnsiString(dst, src);
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisInitUnicodeString(dst, src)
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
unicode_string *dst;
Implement NdisInitUnicodeString().
2003-12-28 21:36:03 +00:00
uint16_t *src;
{
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
RtlInitUnicodeString(dst, src);
Implement NdisInitUnicodeString().
2003-12-28 21:36:03 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void NdisMGetDeviceProperty(adapter, phydevobj,
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
funcdevobj, nextdevobj, resources, transresources)
ndis_handle adapter;
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
device_object **phydevobj;
device_object **funcdevobj;
device_object **nextdevobj;
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
cm_resource_list *resources;
cm_resource_list *transresources;
{
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
ndis_miniport_block *block;
block = (ndis_miniport_block *)adapter;
if (phydevobj != NULL)
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
*phydevobj = block->nmb_physdeviceobj;
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
if (funcdevobj != NULL)
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
*funcdevobj = block->nmb_deviceobj;
if (nextdevobj != NULL)
*nextdevobj = block->nmb_nextdeviceobj;
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
Implement a few new NDIS routines: NdisInitAnsiString(), NdisAnsiStringToUnicodeString(), NdisWriteConfiguration(). Also add stubs for NdisMGetDeviceProperty(), NdisTerminateWrapper(), NdisOpenConfigurationKeyByName(), NdisOpenConfigurationKeyByIndex() and NdisMGetDeviceProperty().
2003-12-14 00:43:40 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisGetFirstBufferFromPacket(packet, buf, firstva, firstlen, totlen)
Now that I finally have power back, implement a couple more NDIS API routines: NdisUnchainBufferAtBack(), NdisGetFirstBufferFromPacketSafe() and NdisGetFirstBufferFromPacket(). This should bring us a little closer to getting the Intel centrino wireless NIC to work. Note: I have not actually tested these additions since I don't have a driver that calls them, however they're pretty simple, and one of them is taken pretty much directly from the Windows ndis.h header file, so I'm fairly confident they work, but disclaimers apply.
2003-12-22 08:24:32 +00:00
ndis_packet *packet;
ndis_buffer **buf;
void **firstva;
uint32_t *firstlen;
uint32_t *totlen;
{
ndis_buffer *tmp;
tmp = packet->np_private.npp_head;
*buf = tmp;
if (tmp == NULL) {
*firstva = NULL;
*firstlen = *totlen = 0;
} else {
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
*firstva = MmGetMdlVirtualAddress(tmp);
*firstlen = *totlen = MmGetMdlByteCount(tmp);
for (tmp = tmp->mdl_next; tmp != NULL; tmp = tmp->mdl_next)
*totlen += MmGetMdlByteCount(tmp);
Now that I finally have power back, implement a couple more NDIS API routines: NdisUnchainBufferAtBack(), NdisGetFirstBufferFromPacketSafe() and NdisGetFirstBufferFromPacket(). This should bring us a little closer to getting the Intel centrino wireless NIC to work. Note: I have not actually tested these additions since I don't have a driver that calls them, however they're pretty simple, and one of them is taken pretty much directly from the Windows ndis.h header file, so I'm fairly confident they work, but disclaimers apply.
2003-12-22 08:24:32 +00:00
}
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisGetFirstBufferFromPacketSafe(packet, buf, firstva, firstlen, totlen, prio)
Now that I finally have power back, implement a couple more NDIS API routines: NdisUnchainBufferAtBack(), NdisGetFirstBufferFromPacketSafe() and NdisGetFirstBufferFromPacket(). This should bring us a little closer to getting the Intel centrino wireless NIC to work. Note: I have not actually tested these additions since I don't have a driver that calls them, however they're pretty simple, and one of them is taken pretty much directly from the Windows ndis.h header file, so I'm fairly confident they work, but disclaimers apply.
2003-12-22 08:24:32 +00:00
ndis_packet *packet;
ndis_buffer **buf;
void **firstva;
uint32_t *firstlen;
uint32_t *totlen;
uint32_t prio;
{
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisGetFirstBufferFromPacket(packet, buf, firstva, firstlen, totlen);
Now that I finally have power back, implement a couple more NDIS API routines: NdisUnchainBufferAtBack(), NdisGetFirstBufferFromPacketSafe() and NdisGetFirstBufferFromPacket(). This should bring us a little closer to getting the Intel centrino wireless NIC to work. Note: I have not actually tested these additions since I don't have a driver that calls them, however they're pretty simple, and one of them is taken pretty much directly from the Windows ndis.h header file, so I'm fairly confident they work, but disclaimers apply.
2003-12-22 08:24:32 +00:00
}
Big mess 'o changes: - Give ndiscvt(8) the ability to process a .SYS file directly into a .o file so that we don't have to emit big messy char arrays into the ndis_driver_data.h file. This behavior is currently optional, but may become the default some day. - Give ndiscvt(8) the ability to turn arbitrary files into .ko files so that they can be pre-loaded or kldloaded. (Both this and the previous change involve using objcopy(1)). - Give NdisOpenFile() the ability to 'read' files out of kernel memory that have been kldloaded or pre-loaded, and disallow the use of the normal vn_open() file opening method during bootstrap (when no filesystems have been mounted yet). Some people have reported that kldloading if_ndis.ko works fine when the system is running multiuser but causes a panic when the modile is pre-loaded by /boot/loader. This happens with drivers that need to use NdisOpenFile() to access external files (i.e. firmware images). NdisOpenFile() won't work during kernel bootstrapping because no filesystems have been mounted. To get around this, you can now do the following: o Say you have a firmware file called firmware.img o Do: ndiscvt -f firmware.img -- this creates firmware.img.ko o Put the firmware.img.ko in /boot/kernel o add firmware.img_load="YES" in /boot/loader.conf o add if_ndis_load="YES" and ndis_load="YES" as well Now the loader will suck the additional file into memory as a .ko. The phony .ko has two symbols in it: filename_start and filename_end, which are generated by objcopy(1). ndis_open_file() will traverse each module in the module list looking for these symbols and, if it finds them, it'll use them to generate the file mapping address and length values that the caller of NdisOpenFile() wants. As a bonus, this will even work if the file has been statically linked into the kernel itself, since the "kernel" module is searched too. (ndiscvt(8) will generate both filename.o and filename.ko for you). - Modify the mechanism used to provide make-pretend FASTCALL support. Rather than using inline assembly to yank the first two arguments out of %ecx and %edx, we now use the __regparm__(3) attribute (and the __stdcall__ attribute) and use some macro magic to re-order the arguments and provide dummy arguments as needed so that the arguments passed in registers end up in the right place. Change taken from DragonflyBSD version of the NDISulator.
2004-08-01 20:04:31 +00:00
static int
ndis_find_sym(lf, filename, suffix, sym)
linker_file_t lf;
char *filename;
char *suffix;
caddr_t *sym;
{
MAXPATHLEN is 1024, which means NdisOpenFile() and ndis_find_sym() were both consuming 1K of stack space. This is unfriendly. Allocate the buffers off the heap instead. It's a little slower, but these aren't performance critical routines. Also, add a spinlock to NdisAllocatePacketPool(), NdisAllocatePacket(), NdisFreePacketPool() and NdisFreePacket(). The pool is maintained as a linked list. I don't know for a fact that it can be corrupted, but why take chances.
2005-03-03 03:51:02 +00:00
char *fullsym;
Make the Texas Instruments 802.11g chipset work with the NDISulator. This was tested with a Netgear WG311v2 802.11b/g PCI card. Things that were fixed: - This chip has two memory mapped regions, one at PCIR_BAR(0) and the other at PCIR_BAR(1). This is a little different from the other chips I've seen with two PCI shared memory regions, since they tend to have the second BAR ad PCIR_BAR(2). if_ndis_pci.c tests explicitly for PCIR_BAR(2). This has been changed to simply fill in ndis_res_mem first and ndis_res_altmem second, if a second shared memory range exists. Given that NDIS drivers seem to scan for BARs in ascending order, I think this should be ok. - Fixed the code that tries to process firmware images that have been loaded as .ko files. To save a step, I was setting up the address mapping in ndis_open_file(), but ndis_map_file() flags pre-existing mappings as an error (to avoid duplicate mappings). Changed this so that the mapping is now donw in ndis_map_file() as expected. - Made the typedef for 'driver_entry' explicitly include __stdcall to silence gcc warning in ndis_load_driver(). NOTE: the Texas Instruments ACX111 driver needs firmware. With my card, there were 3 .bin files shipped with the driver. You must either put these files in /compat/ndis or convert them with ndiscvt -f and kldload them so the driver can use them. Without the firmware image, the NIC won't work.
2004-08-16 18:50:20 +00:00
char *suf;
Big mess 'o changes: - Give ndiscvt(8) the ability to process a .SYS file directly into a .o file so that we don't have to emit big messy char arrays into the ndis_driver_data.h file. This behavior is currently optional, but may become the default some day. - Give ndiscvt(8) the ability to turn arbitrary files into .ko files so that they can be pre-loaded or kldloaded. (Both this and the previous change involve using objcopy(1)). - Give NdisOpenFile() the ability to 'read' files out of kernel memory that have been kldloaded or pre-loaded, and disallow the use of the normal vn_open() file opening method during bootstrap (when no filesystems have been mounted yet). Some people have reported that kldloading if_ndis.ko works fine when the system is running multiuser but causes a panic when the modile is pre-loaded by /boot/loader. This happens with drivers that need to use NdisOpenFile() to access external files (i.e. firmware images). NdisOpenFile() won't work during kernel bootstrapping because no filesystems have been mounted. To get around this, you can now do the following: o Say you have a firmware file called firmware.img o Do: ndiscvt -f firmware.img -- this creates firmware.img.ko o Put the firmware.img.ko in /boot/kernel o add firmware.img_load="YES" in /boot/loader.conf o add if_ndis_load="YES" and ndis_load="YES" as well Now the loader will suck the additional file into memory as a .ko. The phony .ko has two symbols in it: filename_start and filename_end, which are generated by objcopy(1). ndis_open_file() will traverse each module in the module list looking for these symbols and, if it finds them, it'll use them to generate the file mapping address and length values that the caller of NdisOpenFile() wants. As a bonus, this will even work if the file has been statically linked into the kernel itself, since the "kernel" module is searched too. (ndiscvt(8) will generate both filename.o and filename.ko for you). - Modify the mechanism used to provide make-pretend FASTCALL support. Rather than using inline assembly to yank the first two arguments out of %ecx and %edx, we now use the __regparm__(3) attribute (and the __stdcall__ attribute) and use some macro magic to re-order the arguments and provide dummy arguments as needed so that the arguments passed in registers end up in the right place. Change taken from DragonflyBSD version of the NDISulator.
2004-08-01 20:04:31 +00:00
int i;
MAXPATHLEN is 1024, which means NdisOpenFile() and ndis_find_sym() were both consuming 1K of stack space. This is unfriendly. Allocate the buffers off the heap instead. It's a little slower, but these aren't performance critical routines. Also, add a spinlock to NdisAllocatePacketPool(), NdisAllocatePacket(), NdisFreePacketPool() and NdisFreePacket(). The pool is maintained as a linked list. I don't know for a fact that it can be corrupted, but why take chances.
2005-03-03 03:51:02 +00:00
fullsym = ExAllocatePoolWithTag(NonPagedPool, MAXPATHLEN, 0);
if (fullsym == NULL)
return(ENOMEM);
bzero(fullsym, MAXPATHLEN);
strncpy(fullsym, filename, MAXPATHLEN);
if (strlen(filename) < 4) {
ExFreePool(fullsym);
Make the Texas Instruments 802.11g chipset work with the NDISulator. This was tested with a Netgear WG311v2 802.11b/g PCI card. Things that were fixed: - This chip has two memory mapped regions, one at PCIR_BAR(0) and the other at PCIR_BAR(1). This is a little different from the other chips I've seen with two PCI shared memory regions, since they tend to have the second BAR ad PCIR_BAR(2). if_ndis_pci.c tests explicitly for PCIR_BAR(2). This has been changed to simply fill in ndis_res_mem first and ndis_res_altmem second, if a second shared memory range exists. Given that NDIS drivers seem to scan for BARs in ascending order, I think this should be ok. - Fixed the code that tries to process firmware images that have been loaded as .ko files. To save a step, I was setting up the address mapping in ndis_open_file(), but ndis_map_file() flags pre-existing mappings as an error (to avoid duplicate mappings). Changed this so that the mapping is now donw in ndis_map_file() as expected. - Made the typedef for 'driver_entry' explicitly include __stdcall to silence gcc warning in ndis_load_driver(). NOTE: the Texas Instruments ACX111 driver needs firmware. With my card, there were 3 .bin files shipped with the driver. You must either put these files in /compat/ndis or convert them with ndiscvt -f and kldload them so the driver can use them. Without the firmware image, the NIC won't work.
2004-08-16 18:50:20 +00:00
return(EINVAL);
MAXPATHLEN is 1024, which means NdisOpenFile() and ndis_find_sym() were both consuming 1K of stack space. This is unfriendly. Allocate the buffers off the heap instead. It's a little slower, but these aren't performance critical routines. Also, add a spinlock to NdisAllocatePacketPool(), NdisAllocatePacket(), NdisFreePacketPool() and NdisFreePacket(). The pool is maintained as a linked list. I don't know for a fact that it can be corrupted, but why take chances.
2005-03-03 03:51:02 +00:00
}
Make the Texas Instruments 802.11g chipset work with the NDISulator. This was tested with a Netgear WG311v2 802.11b/g PCI card. Things that were fixed: - This chip has two memory mapped regions, one at PCIR_BAR(0) and the other at PCIR_BAR(1). This is a little different from the other chips I've seen with two PCI shared memory regions, since they tend to have the second BAR ad PCIR_BAR(2). if_ndis_pci.c tests explicitly for PCIR_BAR(2). This has been changed to simply fill in ndis_res_mem first and ndis_res_altmem second, if a second shared memory range exists. Given that NDIS drivers seem to scan for BARs in ascending order, I think this should be ok. - Fixed the code that tries to process firmware images that have been loaded as .ko files. To save a step, I was setting up the address mapping in ndis_open_file(), but ndis_map_file() flags pre-existing mappings as an error (to avoid duplicate mappings). Changed this so that the mapping is now donw in ndis_map_file() as expected. - Made the typedef for 'driver_entry' explicitly include __stdcall to silence gcc warning in ndis_load_driver(). NOTE: the Texas Instruments ACX111 driver needs firmware. With my card, there were 3 .bin files shipped with the driver. You must either put these files in /compat/ndis or convert them with ndiscvt -f and kldload them so the driver can use them. Without the firmware image, the NIC won't work.
2004-08-16 18:50:20 +00:00
/* If the filename has a .ko suffix, strip if off. */
suf = fullsym + (strlen(filename) - 3);
if (strcmp(suf, ".ko") == 0)
*suf = '\0';
Big mess 'o changes: - Give ndiscvt(8) the ability to process a .SYS file directly into a .o file so that we don't have to emit big messy char arrays into the ndis_driver_data.h file. This behavior is currently optional, but may become the default some day. - Give ndiscvt(8) the ability to turn arbitrary files into .ko files so that they can be pre-loaded or kldloaded. (Both this and the previous change involve using objcopy(1)). - Give NdisOpenFile() the ability to 'read' files out of kernel memory that have been kldloaded or pre-loaded, and disallow the use of the normal vn_open() file opening method during bootstrap (when no filesystems have been mounted yet). Some people have reported that kldloading if_ndis.ko works fine when the system is running multiuser but causes a panic when the modile is pre-loaded by /boot/loader. This happens with drivers that need to use NdisOpenFile() to access external files (i.e. firmware images). NdisOpenFile() won't work during kernel bootstrapping because no filesystems have been mounted. To get around this, you can now do the following: o Say you have a firmware file called firmware.img o Do: ndiscvt -f firmware.img -- this creates firmware.img.ko o Put the firmware.img.ko in /boot/kernel o add firmware.img_load="YES" in /boot/loader.conf o add if_ndis_load="YES" and ndis_load="YES" as well Now the loader will suck the additional file into memory as a .ko. The phony .ko has two symbols in it: filename_start and filename_end, which are generated by objcopy(1). ndis_open_file() will traverse each module in the module list looking for these symbols and, if it finds them, it'll use them to generate the file mapping address and length values that the caller of NdisOpenFile() wants. As a bonus, this will even work if the file has been statically linked into the kernel itself, since the "kernel" module is searched too. (ndiscvt(8) will generate both filename.o and filename.ko for you). - Modify the mechanism used to provide make-pretend FASTCALL support. Rather than using inline assembly to yank the first two arguments out of %ecx and %edx, we now use the __regparm__(3) attribute (and the __stdcall__ attribute) and use some macro magic to re-order the arguments and provide dummy arguments as needed so that the arguments passed in registers end up in the right place. Change taken from DragonflyBSD version of the NDISulator.
2004-08-01 20:04:31 +00:00
for (i = 0; i < strlen(fullsym); i++) {
if (fullsym[i] == '.')
fullsym[i] = '_';
else
fullsym[i] = tolower(fullsym[i]);
}
strcat(fullsym, suffix);
*sym = linker_file_lookup_symbol(lf, fullsym, 0);
MAXPATHLEN is 1024, which means NdisOpenFile() and ndis_find_sym() were both consuming 1K of stack space. This is unfriendly. Allocate the buffers off the heap instead. It's a little slower, but these aren't performance critical routines. Also, add a spinlock to NdisAllocatePacketPool(), NdisAllocatePacket(), NdisFreePacketPool() and NdisFreePacket(). The pool is maintained as a linked list. I don't know for a fact that it can be corrupted, but why take chances.
2005-03-03 03:51:02 +00:00
ExFreePool(fullsym);
Big mess 'o changes: - Give ndiscvt(8) the ability to process a .SYS file directly into a .o file so that we don't have to emit big messy char arrays into the ndis_driver_data.h file. This behavior is currently optional, but may become the default some day. - Give ndiscvt(8) the ability to turn arbitrary files into .ko files so that they can be pre-loaded or kldloaded. (Both this and the previous change involve using objcopy(1)). - Give NdisOpenFile() the ability to 'read' files out of kernel memory that have been kldloaded or pre-loaded, and disallow the use of the normal vn_open() file opening method during bootstrap (when no filesystems have been mounted yet). Some people have reported that kldloading if_ndis.ko works fine when the system is running multiuser but causes a panic when the modile is pre-loaded by /boot/loader. This happens with drivers that need to use NdisOpenFile() to access external files (i.e. firmware images). NdisOpenFile() won't work during kernel bootstrapping because no filesystems have been mounted. To get around this, you can now do the following: o Say you have a firmware file called firmware.img o Do: ndiscvt -f firmware.img -- this creates firmware.img.ko o Put the firmware.img.ko in /boot/kernel o add firmware.img_load="YES" in /boot/loader.conf o add if_ndis_load="YES" and ndis_load="YES" as well Now the loader will suck the additional file into memory as a .ko. The phony .ko has two symbols in it: filename_start and filename_end, which are generated by objcopy(1). ndis_open_file() will traverse each module in the module list looking for these symbols and, if it finds them, it'll use them to generate the file mapping address and length values that the caller of NdisOpenFile() wants. As a bonus, this will even work if the file has been statically linked into the kernel itself, since the "kernel" module is searched too. (ndiscvt(8) will generate both filename.o and filename.ko for you). - Modify the mechanism used to provide make-pretend FASTCALL support. Rather than using inline assembly to yank the first two arguments out of %ecx and %edx, we now use the __regparm__(3) attribute (and the __stdcall__ attribute) and use some macro magic to re-order the arguments and provide dummy arguments as needed so that the arguments passed in registers end up in the right place. Change taken from DragonflyBSD version of the NDISulator.
2004-08-01 20:04:31 +00:00
if (*sym == 0)
return(ENOENT);
return(0);
}
- Add stubs for Ndis*File() functions - Fix ndis_time(). - Implement NdisGetSystemUpTime(). - Implement RtlCopyUnicodeString() and RtlUnicodeStringToAnsiString(). - In ndis_getstate_80211(), use sc->ndis_link to determine connect status. Submitted by: Brian Feldman <green@freebsd.org>
2003-12-25 00:40:02 +00:00
/* can also return NDIS_STATUS_RESOURCES/NDIS_STATUS_ERROR_READING_FILE */
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisOpenFile(status, filehandle, filelength, filename, highestaddr)
- Add stubs for Ndis*File() functions - Fix ndis_time(). - Implement NdisGetSystemUpTime(). - Implement RtlCopyUnicodeString() and RtlUnicodeStringToAnsiString(). - In ndis_getstate_80211(), use sc->ndis_link to determine connect status. Submitted by: Brian Feldman <green@freebsd.org>
2003-12-25 00:40:02 +00:00
ndis_status *status;
ndis_handle *filehandle;
uint32_t *filelength;
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
unicode_string *filename;
- Add stubs for Ndis*File() functions - Fix ndis_time(). - Implement NdisGetSystemUpTime(). - Implement RtlCopyUnicodeString() and RtlUnicodeStringToAnsiString(). - In ndis_getstate_80211(), use sc->ndis_link to determine connect status. Submitted by: Brian Feldman <green@freebsd.org>
2003-12-25 00:40:02 +00:00
ndis_physaddr highestaddr;
{
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
ansi_string as;
- Add stubs for Ndis*File() functions - Fix ndis_time(). - Implement NdisGetSystemUpTime(). - Implement RtlCopyUnicodeString() and RtlUnicodeStringToAnsiString(). - In ndis_getstate_80211(), use sc->ndis_link to determine connect status. Submitted by: Brian Feldman <green@freebsd.org>
2003-12-25 00:40:02 +00:00
char *afilename = NULL;
Implement NdisOpenFile()/NdisCloseFile()/NdisMapFile()/NdisUnmapFile(). By default, we search for files in /compat/ndis. This can be changed with a systcl. These routines are used by some drivers which need to download firmware or microcode into their respective devices during initialization. Also, remove extraneous newlines from the 'built-in' sysctl/registry variables.
2004-01-09 03:57:00 +00:00
struct thread *td = curthread;
struct nameidata nd;
int flags, error;
struct vattr vat;
struct vattr *vap = &vat;
ndis_fh *fh;
MAXPATHLEN is 1024, which means NdisOpenFile() and ndis_find_sym() were both consuming 1K of stack space. This is unfriendly. Allocate the buffers off the heap instead. It's a little slower, but these aren't performance critical routines. Also, add a spinlock to NdisAllocatePacketPool(), NdisAllocatePacket(), NdisFreePacketPool() and NdisFreePacket(). The pool is maintained as a linked list. I don't know for a fact that it can be corrupted, but why take chances.
2005-03-03 03:51:02 +00:00
char *path;
Big mess 'o changes: - Give ndiscvt(8) the ability to process a .SYS file directly into a .o file so that we don't have to emit big messy char arrays into the ndis_driver_data.h file. This behavior is currently optional, but may become the default some day. - Give ndiscvt(8) the ability to turn arbitrary files into .ko files so that they can be pre-loaded or kldloaded. (Both this and the previous change involve using objcopy(1)). - Give NdisOpenFile() the ability to 'read' files out of kernel memory that have been kldloaded or pre-loaded, and disallow the use of the normal vn_open() file opening method during bootstrap (when no filesystems have been mounted yet). Some people have reported that kldloading if_ndis.ko works fine when the system is running multiuser but causes a panic when the modile is pre-loaded by /boot/loader. This happens with drivers that need to use NdisOpenFile() to access external files (i.e. firmware images). NdisOpenFile() won't work during kernel bootstrapping because no filesystems have been mounted. To get around this, you can now do the following: o Say you have a firmware file called firmware.img o Do: ndiscvt -f firmware.img -- this creates firmware.img.ko o Put the firmware.img.ko in /boot/kernel o add firmware.img_load="YES" in /boot/loader.conf o add if_ndis_load="YES" and ndis_load="YES" as well Now the loader will suck the additional file into memory as a .ko. The phony .ko has two symbols in it: filename_start and filename_end, which are generated by objcopy(1). ndis_open_file() will traverse each module in the module list looking for these symbols and, if it finds them, it'll use them to generate the file mapping address and length values that the caller of NdisOpenFile() wants. As a bonus, this will even work if the file has been statically linked into the kernel itself, since the "kernel" module is searched too. (ndiscvt(8) will generate both filename.o and filename.ko for you). - Modify the mechanism used to provide make-pretend FASTCALL support. Rather than using inline assembly to yank the first two arguments out of %ecx and %edx, we now use the __regparm__(3) attribute (and the __stdcall__ attribute) and use some macro magic to re-order the arguments and provide dummy arguments as needed so that the arguments passed in registers end up in the right place. Change taken from DragonflyBSD version of the NDISulator.
2004-08-01 20:04:31 +00:00
linker_file_t head, lf;
caddr_t kldstart, kldend;
Implement NdisOpenFile()/NdisCloseFile()/NdisMapFile()/NdisUnmapFile(). By default, we search for files in /compat/ndis. This can be changed with a systcl. These routines are used by some drivers which need to download firmware or microcode into their respective devices during initialization. Also, remove extraneous newlines from the 'built-in' sysctl/registry variables.
2004-01-09 03:57:00 +00:00
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
if (RtlUnicodeStringToAnsiString(&as, filename, TRUE)) {
*status = NDIS_STATUS_RESOURCES;
return;
}
afilename = strdup(as.as_buf, M_DEVBUF);
RtlFreeAnsiString(&as);
Implement NdisOpenFile()/NdisCloseFile()/NdisMapFile()/NdisUnmapFile(). By default, we search for files in /compat/ndis. This can be changed with a systcl. These routines are used by some drivers which need to download firmware or microcode into their respective devices during initialization. Also, remove extraneous newlines from the 'built-in' sysctl/registry variables.
2004-01-09 03:57:00 +00:00
MAXPATHLEN is 1024, which means NdisOpenFile() and ndis_find_sym() were both consuming 1K of stack space. This is unfriendly. Allocate the buffers off the heap instead. It's a little slower, but these aren't performance critical routines. Also, add a spinlock to NdisAllocatePacketPool(), NdisAllocatePacket(), NdisFreePacketPool() and NdisFreePacket(). The pool is maintained as a linked list. I don't know for a fact that it can be corrupted, but why take chances.
2005-03-03 03:51:02 +00:00
fh = ExAllocatePoolWithTag(NonPagedPool, sizeof(ndis_fh), 0);
Implement NdisOpenFile()/NdisCloseFile()/NdisMapFile()/NdisUnmapFile(). By default, we search for files in /compat/ndis. This can be changed with a systcl. These routines are used by some drivers which need to download firmware or microcode into their respective devices during initialization. Also, remove extraneous newlines from the 'built-in' sysctl/registry variables.
2004-01-09 03:57:00 +00:00
if (fh == NULL) {
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
free(afilename, M_DEVBUF);
Implement NdisOpenFile()/NdisCloseFile()/NdisMapFile()/NdisUnmapFile(). By default, we search for files in /compat/ndis. This can be changed with a systcl. These routines are used by some drivers which need to download firmware or microcode into their respective devices during initialization. Also, remove extraneous newlines from the 'built-in' sysctl/registry variables.
2004-01-09 03:57:00 +00:00
*status = NDIS_STATUS_RESOURCES;
return;
}
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
fh->nf_name = afilename;
Big mess 'o changes: - Give ndiscvt(8) the ability to process a .SYS file directly into a .o file so that we don't have to emit big messy char arrays into the ndis_driver_data.h file. This behavior is currently optional, but may become the default some day. - Give ndiscvt(8) the ability to turn arbitrary files into .ko files so that they can be pre-loaded or kldloaded. (Both this and the previous change involve using objcopy(1)). - Give NdisOpenFile() the ability to 'read' files out of kernel memory that have been kldloaded or pre-loaded, and disallow the use of the normal vn_open() file opening method during bootstrap (when no filesystems have been mounted yet). Some people have reported that kldloading if_ndis.ko works fine when the system is running multiuser but causes a panic when the modile is pre-loaded by /boot/loader. This happens with drivers that need to use NdisOpenFile() to access external files (i.e. firmware images). NdisOpenFile() won't work during kernel bootstrapping because no filesystems have been mounted. To get around this, you can now do the following: o Say you have a firmware file called firmware.img o Do: ndiscvt -f firmware.img -- this creates firmware.img.ko o Put the firmware.img.ko in /boot/kernel o add firmware.img_load="YES" in /boot/loader.conf o add if_ndis_load="YES" and ndis_load="YES" as well Now the loader will suck the additional file into memory as a .ko. The phony .ko has two symbols in it: filename_start and filename_end, which are generated by objcopy(1). ndis_open_file() will traverse each module in the module list looking for these symbols and, if it finds them, it'll use them to generate the file mapping address and length values that the caller of NdisOpenFile() wants. As a bonus, this will even work if the file has been statically linked into the kernel itself, since the "kernel" module is searched too. (ndiscvt(8) will generate both filename.o and filename.ko for you). - Modify the mechanism used to provide make-pretend FASTCALL support. Rather than using inline assembly to yank the first two arguments out of %ecx and %edx, we now use the __regparm__(3) attribute (and the __stdcall__ attribute) and use some macro magic to re-order the arguments and provide dummy arguments as needed so that the arguments passed in registers end up in the right place. Change taken from DragonflyBSD version of the NDISulator.
2004-08-01 20:04:31 +00:00
/*
* During system bootstrap, it's impossible to load files
* from the rootfs since it's not mounted yet. We therefore
* offer the possibility of opening files that have been
* preloaded as modules instead. Both choices will work
* when kldloading a module from multiuser, but only the
* module option will work during bootstrap. The module
* loading option works by using the ndiscvt(8) utility
* to convert the arbitrary file into a .ko using objcopy(1).
* This file will contain two special symbols: filename_start
* and filename_end. All we have to do is traverse the KLD
* list in search of those symbols and we've found the file
* data. As an added bonus, ndiscvt(8) will also generate
* a normal .o file which can be linked statically with
* the kernel. This means that the symbols will actual reside
* in the kernel's symbol table, but that doesn't matter to
* us since the kernel appears to us as just another module.
*/
/*
* This is an evil trick for getting the head of the linked
* file list, which is not exported from kern_linker.o. It
* happens that linker file #1 is always the kernel, and is
* always the first element in the list.
*/
head = linker_find_file_by_id(1);
for (lf = head; lf != NULL; lf = TAILQ_NEXT(lf, link)) {
if (ndis_find_sym(lf, afilename, "_start", &kldstart))
continue;
if (ndis_find_sym(lf, afilename, "_end", &kldend))
continue;
fh->nf_vp = lf;
I'm a dumbass: remember to initialize fh->nf_map to NULL in ndis_open_file() in the module loading case.
2004-08-16 19:25:27 +00:00
fh->nf_map = NULL;
Big mess 'o changes: - Give ndiscvt(8) the ability to process a .SYS file directly into a .o file so that we don't have to emit big messy char arrays into the ndis_driver_data.h file. This behavior is currently optional, but may become the default some day. - Give ndiscvt(8) the ability to turn arbitrary files into .ko files so that they can be pre-loaded or kldloaded. (Both this and the previous change involve using objcopy(1)). - Give NdisOpenFile() the ability to 'read' files out of kernel memory that have been kldloaded or pre-loaded, and disallow the use of the normal vn_open() file opening method during bootstrap (when no filesystems have been mounted yet). Some people have reported that kldloading if_ndis.ko works fine when the system is running multiuser but causes a panic when the modile is pre-loaded by /boot/loader. This happens with drivers that need to use NdisOpenFile() to access external files (i.e. firmware images). NdisOpenFile() won't work during kernel bootstrapping because no filesystems have been mounted. To get around this, you can now do the following: o Say you have a firmware file called firmware.img o Do: ndiscvt -f firmware.img -- this creates firmware.img.ko o Put the firmware.img.ko in /boot/kernel o add firmware.img_load="YES" in /boot/loader.conf o add if_ndis_load="YES" and ndis_load="YES" as well Now the loader will suck the additional file into memory as a .ko. The phony .ko has two symbols in it: filename_start and filename_end, which are generated by objcopy(1). ndis_open_file() will traverse each module in the module list looking for these symbols and, if it finds them, it'll use them to generate the file mapping address and length values that the caller of NdisOpenFile() wants. As a bonus, this will even work if the file has been statically linked into the kernel itself, since the "kernel" module is searched too. (ndiscvt(8) will generate both filename.o and filename.ko for you). - Modify the mechanism used to provide make-pretend FASTCALL support. Rather than using inline assembly to yank the first two arguments out of %ecx and %edx, we now use the __regparm__(3) attribute (and the __stdcall__ attribute) and use some macro magic to re-order the arguments and provide dummy arguments as needed so that the arguments passed in registers end up in the right place. Change taken from DragonflyBSD version of the NDISulator.
2004-08-01 20:04:31 +00:00
fh->nf_type = NDIS_FH_TYPE_MODULE;
*filelength = fh->nf_maplen = (kldend - kldstart) & 0xFFFFFFFF;
*filehandle = fh;
*status = NDIS_STATUS_SUCCESS;
return;
}
if (TAILQ_EMPTY(&mountlist)) {
The filehandle allocated in NdisOpenFile() is allocated using ExAllocatePoolWithTag(), not malloc(), so it should be released with ExFreePool(), not free(). Fix a couple if instances of free(fh, ...) that got overlooked.
2005-03-28 22:03:47 +00:00
ExFreePool(fh);
Big mess 'o changes: - Give ndiscvt(8) the ability to process a .SYS file directly into a .o file so that we don't have to emit big messy char arrays into the ndis_driver_data.h file. This behavior is currently optional, but may become the default some day. - Give ndiscvt(8) the ability to turn arbitrary files into .ko files so that they can be pre-loaded or kldloaded. (Both this and the previous change involve using objcopy(1)). - Give NdisOpenFile() the ability to 'read' files out of kernel memory that have been kldloaded or pre-loaded, and disallow the use of the normal vn_open() file opening method during bootstrap (when no filesystems have been mounted yet). Some people have reported that kldloading if_ndis.ko works fine when the system is running multiuser but causes a panic when the modile is pre-loaded by /boot/loader. This happens with drivers that need to use NdisOpenFile() to access external files (i.e. firmware images). NdisOpenFile() won't work during kernel bootstrapping because no filesystems have been mounted. To get around this, you can now do the following: o Say you have a firmware file called firmware.img o Do: ndiscvt -f firmware.img -- this creates firmware.img.ko o Put the firmware.img.ko in /boot/kernel o add firmware.img_load="YES" in /boot/loader.conf o add if_ndis_load="YES" and ndis_load="YES" as well Now the loader will suck the additional file into memory as a .ko. The phony .ko has two symbols in it: filename_start and filename_end, which are generated by objcopy(1). ndis_open_file() will traverse each module in the module list looking for these symbols and, if it finds them, it'll use them to generate the file mapping address and length values that the caller of NdisOpenFile() wants. As a bonus, this will even work if the file has been statically linked into the kernel itself, since the "kernel" module is searched too. (ndiscvt(8) will generate both filename.o and filename.ko for you). - Modify the mechanism used to provide make-pretend FASTCALL support. Rather than using inline assembly to yank the first two arguments out of %ecx and %edx, we now use the __regparm__(3) attribute (and the __stdcall__ attribute) and use some macro magic to re-order the arguments and provide dummy arguments as needed so that the arguments passed in registers end up in the right place. Change taken from DragonflyBSD version of the NDISulator.
2004-08-01 20:04:31 +00:00
*status = NDIS_STATUS_FILE_NOT_FOUND;
printf("NDIS: could not find file %s in linker list\n",
afilename);
printf("NDIS: and no filesystems mounted yet, "
"aborting NdisOpenFile()\n");
free(afilename, M_DEVBUF);
return;
}
MAXPATHLEN is 1024, which means NdisOpenFile() and ndis_find_sym() were both consuming 1K of stack space. This is unfriendly. Allocate the buffers off the heap instead. It's a little slower, but these aren't performance critical routines. Also, add a spinlock to NdisAllocatePacketPool(), NdisAllocatePacket(), NdisFreePacketPool() and NdisFreePacket(). The pool is maintained as a linked list. I don't know for a fact that it can be corrupted, but why take chances.
2005-03-03 03:51:02 +00:00
path = ExAllocatePoolWithTag(NonPagedPool, MAXPATHLEN, 0);
if (path == NULL) {
The filehandle allocated in NdisOpenFile() is allocated using ExAllocatePoolWithTag(), not malloc(), so it should be released with ExFreePool(), not free(). Fix a couple if instances of free(fh, ...) that got overlooked.
2005-03-28 22:03:47 +00:00
ExFreePool(fh);
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
free(afilename, M_DEVBUF);
MAXPATHLEN is 1024, which means NdisOpenFile() and ndis_find_sym() were both consuming 1K of stack space. This is unfriendly. Allocate the buffers off the heap instead. It's a little slower, but these aren't performance critical routines. Also, add a spinlock to NdisAllocatePacketPool(), NdisAllocatePacket(), NdisFreePacketPool() and NdisFreePacket(). The pool is maintained as a linked list. I don't know for a fact that it can be corrupted, but why take chances.
2005-03-03 03:51:02 +00:00
*status = NDIS_STATUS_RESOURCES;
return;
}
snprintf(path, MAXPATHLEN, "%s/%s", ndis_filepath, afilename);
Big mess 'o changes: - Give ndiscvt(8) the ability to process a .SYS file directly into a .o file so that we don't have to emit big messy char arrays into the ndis_driver_data.h file. This behavior is currently optional, but may become the default some day. - Give ndiscvt(8) the ability to turn arbitrary files into .ko files so that they can be pre-loaded or kldloaded. (Both this and the previous change involve using objcopy(1)). - Give NdisOpenFile() the ability to 'read' files out of kernel memory that have been kldloaded or pre-loaded, and disallow the use of the normal vn_open() file opening method during bootstrap (when no filesystems have been mounted yet). Some people have reported that kldloading if_ndis.ko works fine when the system is running multiuser but causes a panic when the modile is pre-loaded by /boot/loader. This happens with drivers that need to use NdisOpenFile() to access external files (i.e. firmware images). NdisOpenFile() won't work during kernel bootstrapping because no filesystems have been mounted. To get around this, you can now do the following: o Say you have a firmware file called firmware.img o Do: ndiscvt -f firmware.img -- this creates firmware.img.ko o Put the firmware.img.ko in /boot/kernel o add firmware.img_load="YES" in /boot/loader.conf o add if_ndis_load="YES" and ndis_load="YES" as well Now the loader will suck the additional file into memory as a .ko. The phony .ko has two symbols in it: filename_start and filename_end, which are generated by objcopy(1). ndis_open_file() will traverse each module in the module list looking for these symbols and, if it finds them, it'll use them to generate the file mapping address and length values that the caller of NdisOpenFile() wants. As a bonus, this will even work if the file has been statically linked into the kernel itself, since the "kernel" module is searched too. (ndiscvt(8) will generate both filename.o and filename.ko for you). - Modify the mechanism used to provide make-pretend FASTCALL support. Rather than using inline assembly to yank the first two arguments out of %ecx and %edx, we now use the __regparm__(3) attribute (and the __stdcall__ attribute) and use some macro magic to re-order the arguments and provide dummy arguments as needed so that the arguments passed in registers end up in the right place. Change taken from DragonflyBSD version of the NDISulator.
2004-08-01 20:04:31 +00:00
Implement NdisOpenFile()/NdisCloseFile()/NdisMapFile()/NdisUnmapFile(). By default, we search for files in /compat/ndis. This can be changed with a systcl. These routines are used by some drivers which need to download firmware or microcode into their respective devices during initialization. Also, remove extraneous newlines from the 'built-in' sysctl/registry variables.
2004-01-09 03:57:00 +00:00
mtx_lock(&Giant);
Implement support for single packet sends. The Intel Centrino driver that Asus provides on its CDs has both a MiniportSend() routine and a MiniportSendPackets() function. The Microsoft NDIS docs say that if a driver has both, only the MiniportSendPackets() routine will be used. Although I think I implemented the support correctly, calling the MiniportSend() routine seems to result in no packets going out on the air, even though no error status is returned. The MiniportSendPackets() function does work though, so at least in this case it doesn't matter. In if_ndis.c:ndis_getstate_80211(), if ndis_get_assoc() returns an error, don't bother trying to obtain any other state since the calls may fail, or worse cause the underlying driver to crash. (The above two changes make the Asus-supplied Centrino work.) Also, when calling the OID_802_11_CONFIGURATION OID, remember to initialize the structure lengths correctly. In subr_ndis.c:ndis_open_file(), set the current working directory to rootvnode if we're in a thread that doesn't have a current working directory set.
2004-02-03 07:39:23 +00:00
/* Some threads don't have a current working directory. */
if (td->td_proc->p_fd->fd_rdir == NULL)
td->td_proc->p_fd->fd_rdir = rootvnode;
if (td->td_proc->p_fd->fd_cdir == NULL)
td->td_proc->p_fd->fd_cdir = rootvnode;
Implement NdisOpenFile()/NdisCloseFile()/NdisMapFile()/NdisUnmapFile(). By default, we search for files in /compat/ndis. This can be changed with a systcl. These routines are used by some drivers which need to download firmware or microcode into their respective devices during initialization. Also, remove extraneous newlines from the 'built-in' sysctl/registry variables.
2004-01-09 03:57:00 +00:00
NDINIT(&nd, LOOKUP, FOLLOW, UIO_SYSSPACE, path, td);
flags = FREAD;
error = vn_open(&nd, &flags, 0, -1);
if (error) {
mtx_unlock(&Giant);
*status = NDIS_STATUS_FILE_NOT_FOUND;
MAXPATHLEN is 1024, which means NdisOpenFile() and ndis_find_sym() were both consuming 1K of stack space. This is unfriendly. Allocate the buffers off the heap instead. It's a little slower, but these aren't performance critical routines. Also, add a spinlock to NdisAllocatePacketPool(), NdisAllocatePacket(), NdisFreePacketPool() and NdisFreePacket(). The pool is maintained as a linked list. I don't know for a fact that it can be corrupted, but why take chances.
2005-03-03 03:51:02 +00:00
ExFreePool(fh);
- The MiniportReset() function can return NDIS_STATUS_PENDING, in which case we should wait for the resetdone handler to be called before returning. - When providing resources via ndis_query_resources(), uses the computed rsclen when using bcopy() to copy out the resource data rather than the caller-supplied buffer length. - Avoid using ndis_reset_nic() in if_ndis.c unless we really need to reset the NIC because of a problem. - Allow interrupts to be fielded during ndis_attach(), at least as far as allowing ndis_isr() and ndis_intrhand() to run. - Use ndis_80211_rates_ex when probing for supported rates. Technically, this isn't supposed to work since, although Microsoft added the extended rate structure with the NDIS 5.1 update, the spec still says that the OID_802_11_SUPPORTED_RATES OID uses ndis_80211_rates. In spite of this, it appears some drivers use it anyway. - When adding in our guessed rates, check to see if they already exist so that we avoid any duplicates. - Add a printf() to ndis_open_file() that alerts the user when a driver attempts to open a file under /compat/ndis. With these changes, I can get the driver for the SMC 2802W 54g PCI card to load and run. This board uses a Prism54G chip. Note that in order for this driver to work, you must place the supplied smc2802w.arm firmware image under /compat/ndis. (The firmware is not resident on the device.) Note that this should also allow the 3Com 3CRWE154G72 card to work as well; as far as I can tell, these cards also use a Prism54G chip.
2004-04-05 08:26:52 +00:00
printf("NDIS: open file %s failed: %d\n", path, error);
MAXPATHLEN is 1024, which means NdisOpenFile() and ndis_find_sym() were both consuming 1K of stack space. This is unfriendly. Allocate the buffers off the heap instead. It's a little slower, but these aren't performance critical routines. Also, add a spinlock to NdisAllocatePacketPool(), NdisAllocatePacket(), NdisFreePacketPool() and NdisFreePacket(). The pool is maintained as a linked list. I don't know for a fact that it can be corrupted, but why take chances.
2005-03-03 03:51:02 +00:00
ExFreePool(path);
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
free(afilename, M_DEVBUF);
Implement NdisOpenFile()/NdisCloseFile()/NdisMapFile()/NdisUnmapFile(). By default, we search for files in /compat/ndis. This can be changed with a systcl. These routines are used by some drivers which need to download firmware or microcode into their respective devices during initialization. Also, remove extraneous newlines from the 'built-in' sysctl/registry variables.
2004-01-09 03:57:00 +00:00
return;
}
MAXPATHLEN is 1024, which means NdisOpenFile() and ndis_find_sym() were both consuming 1K of stack space. This is unfriendly. Allocate the buffers off the heap instead. It's a little slower, but these aren't performance critical routines. Also, add a spinlock to NdisAllocatePacketPool(), NdisAllocatePacket(), NdisFreePacketPool() and NdisFreePacket(). The pool is maintained as a linked list. I don't know for a fact that it can be corrupted, but why take chances.
2005-03-03 03:51:02 +00:00
ExFreePool(path);
Implement NdisOpenFile()/NdisCloseFile()/NdisMapFile()/NdisUnmapFile(). By default, we search for files in /compat/ndis. This can be changed with a systcl. These routines are used by some drivers which need to download firmware or microcode into their respective devices during initialization. Also, remove extraneous newlines from the 'built-in' sysctl/registry variables.
2004-01-09 03:57:00 +00:00
NDFREE(&nd, NDF_ONLY_PNBUF);
/* Get the file size. */
Make NdisReadPcmciaAttributeMemory() and NdisWritePcmciaAttributeMemory() actually work. Make the PCI and PCCARD attachments provide a bus_get_resource_list() method so that resource listing for PCCARD works. PCCARD does not have a bus_get_resource_list() method (yet), so I faked up the resource list management in if_ndis_pccard.c, and added bus_get_resource_list() methods to both if_ndis_pccard.c and if_ndis_pci.c. The one in the PCI attechment just hands off to the PCI bus code. The difference is transparent to the NDIS resource handler code. Fixed ndis_open_file() so that opening files which live on NFS filesystems work: pass an actual ucred structure to VOP_GETATTR() (NFS explodes if the ucred structure is NOCRED). Make NdisMMapIoSpace() handle mapping of PCMCIA attribute memory resources correctly. Turn subr_ndis.c:my_strcasecmp() into ndis_strcasecmp() and export it so that if_ndis_pccard.c can use it, and junk the other copy of my_strcasecmp() from if_ndis_pccard.c.
2004-07-11 00:19:30 +00:00
VOP_GETATTR(nd.ni_vp, vap, td->td_ucred, td);
Implement NdisOpenFile()/NdisCloseFile()/NdisMapFile()/NdisUnmapFile(). By default, we search for files in /compat/ndis. This can be changed with a systcl. These routines are used by some drivers which need to download firmware or microcode into their respective devices during initialization. Also, remove extraneous newlines from the 'built-in' sysctl/registry variables.
2004-01-09 03:57:00 +00:00
VOP_UNLOCK(nd.ni_vp, 0, td);
mtx_unlock(&Giant);
fh->nf_vp = nd.ni_vp;
fh->nf_map = NULL;
Big mess 'o changes: - Give ndiscvt(8) the ability to process a .SYS file directly into a .o file so that we don't have to emit big messy char arrays into the ndis_driver_data.h file. This behavior is currently optional, but may become the default some day. - Give ndiscvt(8) the ability to turn arbitrary files into .ko files so that they can be pre-loaded or kldloaded. (Both this and the previous change involve using objcopy(1)). - Give NdisOpenFile() the ability to 'read' files out of kernel memory that have been kldloaded or pre-loaded, and disallow the use of the normal vn_open() file opening method during bootstrap (when no filesystems have been mounted yet). Some people have reported that kldloading if_ndis.ko works fine when the system is running multiuser but causes a panic when the modile is pre-loaded by /boot/loader. This happens with drivers that need to use NdisOpenFile() to access external files (i.e. firmware images). NdisOpenFile() won't work during kernel bootstrapping because no filesystems have been mounted. To get around this, you can now do the following: o Say you have a firmware file called firmware.img o Do: ndiscvt -f firmware.img -- this creates firmware.img.ko o Put the firmware.img.ko in /boot/kernel o add firmware.img_load="YES" in /boot/loader.conf o add if_ndis_load="YES" and ndis_load="YES" as well Now the loader will suck the additional file into memory as a .ko. The phony .ko has two symbols in it: filename_start and filename_end, which are generated by objcopy(1). ndis_open_file() will traverse each module in the module list looking for these symbols and, if it finds them, it'll use them to generate the file mapping address and length values that the caller of NdisOpenFile() wants. As a bonus, this will even work if the file has been statically linked into the kernel itself, since the "kernel" module is searched too. (ndiscvt(8) will generate both filename.o and filename.ko for you). - Modify the mechanism used to provide make-pretend FASTCALL support. Rather than using inline assembly to yank the first two arguments out of %ecx and %edx, we now use the __regparm__(3) attribute (and the __stdcall__ attribute) and use some macro magic to re-order the arguments and provide dummy arguments as needed so that the arguments passed in registers end up in the right place. Change taken from DragonflyBSD version of the NDISulator.
2004-08-01 20:04:31 +00:00
fh->nf_type = NDIS_FH_TYPE_VFS;
Implement NdisOpenFile()/NdisCloseFile()/NdisMapFile()/NdisUnmapFile(). By default, we search for files in /compat/ndis. This can be changed with a systcl. These routines are used by some drivers which need to download firmware or microcode into their respective devices during initialization. Also, remove extraneous newlines from the 'built-in' sysctl/registry variables.
2004-01-09 03:57:00 +00:00
*filehandle = fh;
*filelength = fh->nf_maplen = vap->va_size & 0xFFFFFFFF;
*status = NDIS_STATUS_SUCCESS;
Make NdisReadPcmciaAttributeMemory() and NdisWritePcmciaAttributeMemory() actually work. Make the PCI and PCCARD attachments provide a bus_get_resource_list() method so that resource listing for PCCARD works. PCCARD does not have a bus_get_resource_list() method (yet), so I faked up the resource list management in if_ndis_pccard.c, and added bus_get_resource_list() methods to both if_ndis_pccard.c and if_ndis_pci.c. The one in the PCI attechment just hands off to the PCI bus code. The difference is transparent to the NDIS resource handler code. Fixed ndis_open_file() so that opening files which live on NFS filesystems work: pass an actual ucred structure to VOP_GETATTR() (NFS explodes if the ucred structure is NOCRED). Make NdisMMapIoSpace() handle mapping of PCMCIA attribute memory resources correctly. Turn subr_ndis.c:my_strcasecmp() into ndis_strcasecmp() and export it so that if_ndis_pccard.c can use it, and junk the other copy of my_strcasecmp() from if_ndis_pccard.c.
2004-07-11 00:19:30 +00:00
- Add stubs for Ndis*File() functions - Fix ndis_time(). - Implement NdisGetSystemUpTime(). - Implement RtlCopyUnicodeString() and RtlUnicodeStringToAnsiString(). - In ndis_getstate_80211(), use sc->ndis_link to determine connect status. Submitted by: Brian Feldman <green@freebsd.org>
2003-12-25 00:40:02 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMapFile(status, mappedbuffer, filehandle)
- Add stubs for Ndis*File() functions - Fix ndis_time(). - Implement NdisGetSystemUpTime(). - Implement RtlCopyUnicodeString() and RtlUnicodeStringToAnsiString(). - In ndis_getstate_80211(), use sc->ndis_link to determine connect status. Submitted by: Brian Feldman <green@freebsd.org>
2003-12-25 00:40:02 +00:00
ndis_status *status;
void **mappedbuffer;
ndis_handle filehandle;
{
Implement NdisOpenFile()/NdisCloseFile()/NdisMapFile()/NdisUnmapFile(). By default, we search for files in /compat/ndis. This can be changed with a systcl. These routines are used by some drivers which need to download firmware or microcode into their respective devices during initialization. Also, remove extraneous newlines from the 'built-in' sysctl/registry variables.
2004-01-09 03:57:00 +00:00
ndis_fh *fh;
struct thread *td = curthread;
Make the Texas Instruments 802.11g chipset work with the NDISulator. This was tested with a Netgear WG311v2 802.11b/g PCI card. Things that were fixed: - This chip has two memory mapped regions, one at PCIR_BAR(0) and the other at PCIR_BAR(1). This is a little different from the other chips I've seen with two PCI shared memory regions, since they tend to have the second BAR ad PCIR_BAR(2). if_ndis_pci.c tests explicitly for PCIR_BAR(2). This has been changed to simply fill in ndis_res_mem first and ndis_res_altmem second, if a second shared memory range exists. Given that NDIS drivers seem to scan for BARs in ascending order, I think this should be ok. - Fixed the code that tries to process firmware images that have been loaded as .ko files. To save a step, I was setting up the address mapping in ndis_open_file(), but ndis_map_file() flags pre-existing mappings as an error (to avoid duplicate mappings). Changed this so that the mapping is now donw in ndis_map_file() as expected. - Made the typedef for 'driver_entry' explicitly include __stdcall to silence gcc warning in ndis_load_driver(). NOTE: the Texas Instruments ACX111 driver needs firmware. With my card, there were 3 .bin files shipped with the driver. You must either put these files in /compat/ndis or convert them with ndiscvt -f and kldload them so the driver can use them. Without the firmware image, the NIC won't work.
2004-08-16 18:50:20 +00:00
linker_file_t lf;
caddr_t kldstart;
Implement NdisOpenFile()/NdisCloseFile()/NdisMapFile()/NdisUnmapFile(). By default, we search for files in /compat/ndis. This can be changed with a systcl. These routines are used by some drivers which need to download firmware or microcode into their respective devices during initialization. Also, remove extraneous newlines from the 'built-in' sysctl/registry variables.
2004-01-09 03:57:00 +00:00
int error, resid;
if (filehandle == NULL) {
*status = NDIS_STATUS_FAILURE;
return;
}
fh = (ndis_fh *)filehandle;
if (fh->nf_vp == NULL) {
*status = NDIS_STATUS_FAILURE;
return;
}
if (fh->nf_map != NULL) {
*status = NDIS_STATUS_ALREADY_MAPPED;
return;
}
Big mess 'o changes: - Give ndiscvt(8) the ability to process a .SYS file directly into a .o file so that we don't have to emit big messy char arrays into the ndis_driver_data.h file. This behavior is currently optional, but may become the default some day. - Give ndiscvt(8) the ability to turn arbitrary files into .ko files so that they can be pre-loaded or kldloaded. (Both this and the previous change involve using objcopy(1)). - Give NdisOpenFile() the ability to 'read' files out of kernel memory that have been kldloaded or pre-loaded, and disallow the use of the normal vn_open() file opening method during bootstrap (when no filesystems have been mounted yet). Some people have reported that kldloading if_ndis.ko works fine when the system is running multiuser but causes a panic when the modile is pre-loaded by /boot/loader. This happens with drivers that need to use NdisOpenFile() to access external files (i.e. firmware images). NdisOpenFile() won't work during kernel bootstrapping because no filesystems have been mounted. To get around this, you can now do the following: o Say you have a firmware file called firmware.img o Do: ndiscvt -f firmware.img -- this creates firmware.img.ko o Put the firmware.img.ko in /boot/kernel o add firmware.img_load="YES" in /boot/loader.conf o add if_ndis_load="YES" and ndis_load="YES" as well Now the loader will suck the additional file into memory as a .ko. The phony .ko has two symbols in it: filename_start and filename_end, which are generated by objcopy(1). ndis_open_file() will traverse each module in the module list looking for these symbols and, if it finds them, it'll use them to generate the file mapping address and length values that the caller of NdisOpenFile() wants. As a bonus, this will even work if the file has been statically linked into the kernel itself, since the "kernel" module is searched too. (ndiscvt(8) will generate both filename.o and filename.ko for you). - Modify the mechanism used to provide make-pretend FASTCALL support. Rather than using inline assembly to yank the first two arguments out of %ecx and %edx, we now use the __regparm__(3) attribute (and the __stdcall__ attribute) and use some macro magic to re-order the arguments and provide dummy arguments as needed so that the arguments passed in registers end up in the right place. Change taken from DragonflyBSD version of the NDISulator.
2004-08-01 20:04:31 +00:00
if (fh->nf_type == NDIS_FH_TYPE_MODULE) {
Make the Texas Instruments 802.11g chipset work with the NDISulator. This was tested with a Netgear WG311v2 802.11b/g PCI card. Things that were fixed: - This chip has two memory mapped regions, one at PCIR_BAR(0) and the other at PCIR_BAR(1). This is a little different from the other chips I've seen with two PCI shared memory regions, since they tend to have the second BAR ad PCIR_BAR(2). if_ndis_pci.c tests explicitly for PCIR_BAR(2). This has been changed to simply fill in ndis_res_mem first and ndis_res_altmem second, if a second shared memory range exists. Given that NDIS drivers seem to scan for BARs in ascending order, I think this should be ok. - Fixed the code that tries to process firmware images that have been loaded as .ko files. To save a step, I was setting up the address mapping in ndis_open_file(), but ndis_map_file() flags pre-existing mappings as an error (to avoid duplicate mappings). Changed this so that the mapping is now donw in ndis_map_file() as expected. - Made the typedef for 'driver_entry' explicitly include __stdcall to silence gcc warning in ndis_load_driver(). NOTE: the Texas Instruments ACX111 driver needs firmware. With my card, there were 3 .bin files shipped with the driver. You must either put these files in /compat/ndis or convert them with ndiscvt -f and kldload them so the driver can use them. Without the firmware image, the NIC won't work.
2004-08-16 18:50:20 +00:00
lf = fh->nf_vp;
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
if (ndis_find_sym(lf, fh->nf_name, "_start", &kldstart)) {
Make the Texas Instruments 802.11g chipset work with the NDISulator. This was tested with a Netgear WG311v2 802.11b/g PCI card. Things that were fixed: - This chip has two memory mapped regions, one at PCIR_BAR(0) and the other at PCIR_BAR(1). This is a little different from the other chips I've seen with two PCI shared memory regions, since they tend to have the second BAR ad PCIR_BAR(2). if_ndis_pci.c tests explicitly for PCIR_BAR(2). This has been changed to simply fill in ndis_res_mem first and ndis_res_altmem second, if a second shared memory range exists. Given that NDIS drivers seem to scan for BARs in ascending order, I think this should be ok. - Fixed the code that tries to process firmware images that have been loaded as .ko files. To save a step, I was setting up the address mapping in ndis_open_file(), but ndis_map_file() flags pre-existing mappings as an error (to avoid duplicate mappings). Changed this so that the mapping is now donw in ndis_map_file() as expected. - Made the typedef for 'driver_entry' explicitly include __stdcall to silence gcc warning in ndis_load_driver(). NOTE: the Texas Instruments ACX111 driver needs firmware. With my card, there were 3 .bin files shipped with the driver. You must either put these files in /compat/ndis or convert them with ndiscvt -f and kldload them so the driver can use them. Without the firmware image, the NIC won't work.
2004-08-16 18:50:20 +00:00
*status = NDIS_STATUS_FAILURE;
return;
}
fh->nf_map = kldstart;
Big mess 'o changes: - Give ndiscvt(8) the ability to process a .SYS file directly into a .o file so that we don't have to emit big messy char arrays into the ndis_driver_data.h file. This behavior is currently optional, but may become the default some day. - Give ndiscvt(8) the ability to turn arbitrary files into .ko files so that they can be pre-loaded or kldloaded. (Both this and the previous change involve using objcopy(1)). - Give NdisOpenFile() the ability to 'read' files out of kernel memory that have been kldloaded or pre-loaded, and disallow the use of the normal vn_open() file opening method during bootstrap (when no filesystems have been mounted yet). Some people have reported that kldloading if_ndis.ko works fine when the system is running multiuser but causes a panic when the modile is pre-loaded by /boot/loader. This happens with drivers that need to use NdisOpenFile() to access external files (i.e. firmware images). NdisOpenFile() won't work during kernel bootstrapping because no filesystems have been mounted. To get around this, you can now do the following: o Say you have a firmware file called firmware.img o Do: ndiscvt -f firmware.img -- this creates firmware.img.ko o Put the firmware.img.ko in /boot/kernel o add firmware.img_load="YES" in /boot/loader.conf o add if_ndis_load="YES" and ndis_load="YES" as well Now the loader will suck the additional file into memory as a .ko. The phony .ko has two symbols in it: filename_start and filename_end, which are generated by objcopy(1). ndis_open_file() will traverse each module in the module list looking for these symbols and, if it finds them, it'll use them to generate the file mapping address and length values that the caller of NdisOpenFile() wants. As a bonus, this will even work if the file has been statically linked into the kernel itself, since the "kernel" module is searched too. (ndiscvt(8) will generate both filename.o and filename.ko for you). - Modify the mechanism used to provide make-pretend FASTCALL support. Rather than using inline assembly to yank the first two arguments out of %ecx and %edx, we now use the __regparm__(3) attribute (and the __stdcall__ attribute) and use some macro magic to re-order the arguments and provide dummy arguments as needed so that the arguments passed in registers end up in the right place. Change taken from DragonflyBSD version of the NDISulator.
2004-08-01 20:04:31 +00:00
*status = NDIS_STATUS_SUCCESS;
*mappedbuffer = fh->nf_map;
return;
}
MAXPATHLEN is 1024, which means NdisOpenFile() and ndis_find_sym() were both consuming 1K of stack space. This is unfriendly. Allocate the buffers off the heap instead. It's a little slower, but these aren't performance critical routines. Also, add a spinlock to NdisAllocatePacketPool(), NdisAllocatePacket(), NdisFreePacketPool() and NdisFreePacket(). The pool is maintained as a linked list. I don't know for a fact that it can be corrupted, but why take chances.
2005-03-03 03:51:02 +00:00
fh->nf_map = ExAllocatePoolWithTag(NonPagedPool, fh->nf_maplen, 0);
Implement NdisOpenFile()/NdisCloseFile()/NdisMapFile()/NdisUnmapFile(). By default, we search for files in /compat/ndis. This can be changed with a systcl. These routines are used by some drivers which need to download firmware or microcode into their respective devices during initialization. Also, remove extraneous newlines from the 'built-in' sysctl/registry variables.
2004-01-09 03:57:00 +00:00
if (fh->nf_map == NULL) {
*status = NDIS_STATUS_RESOURCES;
return;
}
mtx_lock(&Giant);
error = vn_rdwr(UIO_READ, fh->nf_vp, fh->nf_map, fh->nf_maplen, 0,
UIO_SYSSPACE, 0, td->td_ucred, NOCRED, &resid, td);
mtx_unlock(&Giant);
if (error)
*status = NDIS_STATUS_FAILURE;
else {
*status = NDIS_STATUS_SUCCESS;
*mappedbuffer = fh->nf_map;
}
- Add stubs for Ndis*File() functions - Fix ndis_time(). - Implement NdisGetSystemUpTime(). - Implement RtlCopyUnicodeString() and RtlUnicodeStringToAnsiString(). - In ndis_getstate_80211(), use sc->ndis_link to determine connect status. Submitted by: Brian Feldman <green@freebsd.org>
2003-12-25 00:40:02 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisUnmapFile(filehandle)
- Add stubs for Ndis*File() functions - Fix ndis_time(). - Implement NdisGetSystemUpTime(). - Implement RtlCopyUnicodeString() and RtlUnicodeStringToAnsiString(). - In ndis_getstate_80211(), use sc->ndis_link to determine connect status. Submitted by: Brian Feldman <green@freebsd.org>
2003-12-25 00:40:02 +00:00
ndis_handle filehandle;
{
Implement NdisOpenFile()/NdisCloseFile()/NdisMapFile()/NdisUnmapFile(). By default, we search for files in /compat/ndis. This can be changed with a systcl. These routines are used by some drivers which need to download firmware or microcode into their respective devices during initialization. Also, remove extraneous newlines from the 'built-in' sysctl/registry variables.
2004-01-09 03:57:00 +00:00
ndis_fh *fh;
fh = (ndis_fh *)filehandle;
if (fh->nf_map == NULL)
return;
Big mess 'o changes: - Give ndiscvt(8) the ability to process a .SYS file directly into a .o file so that we don't have to emit big messy char arrays into the ndis_driver_data.h file. This behavior is currently optional, but may become the default some day. - Give ndiscvt(8) the ability to turn arbitrary files into .ko files so that they can be pre-loaded or kldloaded. (Both this and the previous change involve using objcopy(1)). - Give NdisOpenFile() the ability to 'read' files out of kernel memory that have been kldloaded or pre-loaded, and disallow the use of the normal vn_open() file opening method during bootstrap (when no filesystems have been mounted yet). Some people have reported that kldloading if_ndis.ko works fine when the system is running multiuser but causes a panic when the modile is pre-loaded by /boot/loader. This happens with drivers that need to use NdisOpenFile() to access external files (i.e. firmware images). NdisOpenFile() won't work during kernel bootstrapping because no filesystems have been mounted. To get around this, you can now do the following: o Say you have a firmware file called firmware.img o Do: ndiscvt -f firmware.img -- this creates firmware.img.ko o Put the firmware.img.ko in /boot/kernel o add firmware.img_load="YES" in /boot/loader.conf o add if_ndis_load="YES" and ndis_load="YES" as well Now the loader will suck the additional file into memory as a .ko. The phony .ko has two symbols in it: filename_start and filename_end, which are generated by objcopy(1). ndis_open_file() will traverse each module in the module list looking for these symbols and, if it finds them, it'll use them to generate the file mapping address and length values that the caller of NdisOpenFile() wants. As a bonus, this will even work if the file has been statically linked into the kernel itself, since the "kernel" module is searched too. (ndiscvt(8) will generate both filename.o and filename.ko for you). - Modify the mechanism used to provide make-pretend FASTCALL support. Rather than using inline assembly to yank the first two arguments out of %ecx and %edx, we now use the __regparm__(3) attribute (and the __stdcall__ attribute) and use some macro magic to re-order the arguments and provide dummy arguments as needed so that the arguments passed in registers end up in the right place. Change taken from DragonflyBSD version of the NDISulator.
2004-08-01 20:04:31 +00:00
if (fh->nf_type == NDIS_FH_TYPE_VFS)
MAXPATHLEN is 1024, which means NdisOpenFile() and ndis_find_sym() were both consuming 1K of stack space. This is unfriendly. Allocate the buffers off the heap instead. It's a little slower, but these aren't performance critical routines. Also, add a spinlock to NdisAllocatePacketPool(), NdisAllocatePacket(), NdisFreePacketPool() and NdisFreePacket(). The pool is maintained as a linked list. I don't know for a fact that it can be corrupted, but why take chances.
2005-03-03 03:51:02 +00:00
ExFreePool(fh->nf_map);
Implement NdisOpenFile()/NdisCloseFile()/NdisMapFile()/NdisUnmapFile(). By default, we search for files in /compat/ndis. This can be changed with a systcl. These routines are used by some drivers which need to download firmware or microcode into their respective devices during initialization. Also, remove extraneous newlines from the 'built-in' sysctl/registry variables.
2004-01-09 03:57:00 +00:00
fh->nf_map = NULL;
- Add stubs for Ndis*File() functions - Fix ndis_time(). - Implement NdisGetSystemUpTime(). - Implement RtlCopyUnicodeString() and RtlUnicodeStringToAnsiString(). - In ndis_getstate_80211(), use sc->ndis_link to determine connect status. Submitted by: Brian Feldman <green@freebsd.org>
2003-12-25 00:40:02 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisCloseFile(filehandle)
- Add stubs for Ndis*File() functions - Fix ndis_time(). - Implement NdisGetSystemUpTime(). - Implement RtlCopyUnicodeString() and RtlUnicodeStringToAnsiString(). - In ndis_getstate_80211(), use sc->ndis_link to determine connect status. Submitted by: Brian Feldman <green@freebsd.org>
2003-12-25 00:40:02 +00:00
ndis_handle filehandle;
{
Implement NdisOpenFile()/NdisCloseFile()/NdisMapFile()/NdisUnmapFile(). By default, we search for files in /compat/ndis. This can be changed with a systcl. These routines are used by some drivers which need to download firmware or microcode into their respective devices during initialization. Also, remove extraneous newlines from the 'built-in' sysctl/registry variables.
2004-01-09 03:57:00 +00:00
struct thread *td = curthread;
ndis_fh *fh;
if (filehandle == NULL)
return;
fh = (ndis_fh *)filehandle;
if (fh->nf_map != NULL) {
Big mess 'o changes: - Give ndiscvt(8) the ability to process a .SYS file directly into a .o file so that we don't have to emit big messy char arrays into the ndis_driver_data.h file. This behavior is currently optional, but may become the default some day. - Give ndiscvt(8) the ability to turn arbitrary files into .ko files so that they can be pre-loaded or kldloaded. (Both this and the previous change involve using objcopy(1)). - Give NdisOpenFile() the ability to 'read' files out of kernel memory that have been kldloaded or pre-loaded, and disallow the use of the normal vn_open() file opening method during bootstrap (when no filesystems have been mounted yet). Some people have reported that kldloading if_ndis.ko works fine when the system is running multiuser but causes a panic when the modile is pre-loaded by /boot/loader. This happens with drivers that need to use NdisOpenFile() to access external files (i.e. firmware images). NdisOpenFile() won't work during kernel bootstrapping because no filesystems have been mounted. To get around this, you can now do the following: o Say you have a firmware file called firmware.img o Do: ndiscvt -f firmware.img -- this creates firmware.img.ko o Put the firmware.img.ko in /boot/kernel o add firmware.img_load="YES" in /boot/loader.conf o add if_ndis_load="YES" and ndis_load="YES" as well Now the loader will suck the additional file into memory as a .ko. The phony .ko has two symbols in it: filename_start and filename_end, which are generated by objcopy(1). ndis_open_file() will traverse each module in the module list looking for these symbols and, if it finds them, it'll use them to generate the file mapping address and length values that the caller of NdisOpenFile() wants. As a bonus, this will even work if the file has been statically linked into the kernel itself, since the "kernel" module is searched too. (ndiscvt(8) will generate both filename.o and filename.ko for you). - Modify the mechanism used to provide make-pretend FASTCALL support. Rather than using inline assembly to yank the first two arguments out of %ecx and %edx, we now use the __regparm__(3) attribute (and the __stdcall__ attribute) and use some macro magic to re-order the arguments and provide dummy arguments as needed so that the arguments passed in registers end up in the right place. Change taken from DragonflyBSD version of the NDISulator.
2004-08-01 20:04:31 +00:00
if (fh->nf_type == NDIS_FH_TYPE_VFS)
MAXPATHLEN is 1024, which means NdisOpenFile() and ndis_find_sym() were both consuming 1K of stack space. This is unfriendly. Allocate the buffers off the heap instead. It's a little slower, but these aren't performance critical routines. Also, add a spinlock to NdisAllocatePacketPool(), NdisAllocatePacket(), NdisFreePacketPool() and NdisFreePacket(). The pool is maintained as a linked list. I don't know for a fact that it can be corrupted, but why take chances.
2005-03-03 03:51:02 +00:00
ExFreePool(fh->nf_map);
Implement NdisOpenFile()/NdisCloseFile()/NdisMapFile()/NdisUnmapFile(). By default, we search for files in /compat/ndis. This can be changed with a systcl. These routines are used by some drivers which need to download firmware or microcode into their respective devices during initialization. Also, remove extraneous newlines from the 'built-in' sysctl/registry variables.
2004-01-09 03:57:00 +00:00
fh->nf_map = NULL;
}
if (fh->nf_vp == NULL)
return;
Big mess 'o changes: - Give ndiscvt(8) the ability to process a .SYS file directly into a .o file so that we don't have to emit big messy char arrays into the ndis_driver_data.h file. This behavior is currently optional, but may become the default some day. - Give ndiscvt(8) the ability to turn arbitrary files into .ko files so that they can be pre-loaded or kldloaded. (Both this and the previous change involve using objcopy(1)). - Give NdisOpenFile() the ability to 'read' files out of kernel memory that have been kldloaded or pre-loaded, and disallow the use of the normal vn_open() file opening method during bootstrap (when no filesystems have been mounted yet). Some people have reported that kldloading if_ndis.ko works fine when the system is running multiuser but causes a panic when the modile is pre-loaded by /boot/loader. This happens with drivers that need to use NdisOpenFile() to access external files (i.e. firmware images). NdisOpenFile() won't work during kernel bootstrapping because no filesystems have been mounted. To get around this, you can now do the following: o Say you have a firmware file called firmware.img o Do: ndiscvt -f firmware.img -- this creates firmware.img.ko o Put the firmware.img.ko in /boot/kernel o add firmware.img_load="YES" in /boot/loader.conf o add if_ndis_load="YES" and ndis_load="YES" as well Now the loader will suck the additional file into memory as a .ko. The phony .ko has two symbols in it: filename_start and filename_end, which are generated by objcopy(1). ndis_open_file() will traverse each module in the module list looking for these symbols and, if it finds them, it'll use them to generate the file mapping address and length values that the caller of NdisOpenFile() wants. As a bonus, this will even work if the file has been statically linked into the kernel itself, since the "kernel" module is searched too. (ndiscvt(8) will generate both filename.o and filename.ko for you). - Modify the mechanism used to provide make-pretend FASTCALL support. Rather than using inline assembly to yank the first two arguments out of %ecx and %edx, we now use the __regparm__(3) attribute (and the __stdcall__ attribute) and use some macro magic to re-order the arguments and provide dummy arguments as needed so that the arguments passed in registers end up in the right place. Change taken from DragonflyBSD version of the NDISulator.
2004-08-01 20:04:31 +00:00
if (fh->nf_type == NDIS_FH_TYPE_VFS) {
mtx_lock(&Giant);
vn_close(fh->nf_vp, FREAD, td->td_ucred, td);
mtx_unlock(&Giant);
}
Implement NdisOpenFile()/NdisCloseFile()/NdisMapFile()/NdisUnmapFile(). By default, we search for files in /compat/ndis. This can be changed with a systcl. These routines are used by some drivers which need to download firmware or microcode into their respective devices during initialization. Also, remove extraneous newlines from the 'built-in' sysctl/registry variables.
2004-01-09 03:57:00 +00:00
fh->nf_vp = NULL;
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
free(fh->nf_name, M_DEVBUF);
MAXPATHLEN is 1024, which means NdisOpenFile() and ndis_find_sym() were both consuming 1K of stack space. This is unfriendly. Allocate the buffers off the heap instead. It's a little slower, but these aren't performance critical routines. Also, add a spinlock to NdisAllocatePacketPool(), NdisAllocatePacket(), NdisFreePacketPool() and NdisFreePacket(). The pool is maintained as a linked list. I don't know for a fact that it can be corrupted, but why take chances.
2005-03-03 03:51:02 +00:00
ExFreePool(fh);
Implement NdisOpenFile()/NdisCloseFile()/NdisMapFile()/NdisUnmapFile(). By default, we search for files in /compat/ndis. This can be changed with a systcl. These routines are used by some drivers which need to download firmware or microcode into their respective devices during initialization. Also, remove extraneous newlines from the 'built-in' sysctl/registry variables.
2004-01-09 03:57:00 +00:00
- Add stubs for Ndis*File() functions - Fix ndis_time(). - Implement NdisGetSystemUpTime(). - Implement RtlCopyUnicodeString() and RtlUnicodeStringToAnsiString(). - In ndis_getstate_80211(), use sc->ndis_link to determine connect status. Submitted by: Brian Feldman <green@freebsd.org>
2003-12-25 00:40:02 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static uint8_t
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisSystemProcessorCount()
Back out the last batch of changes until I have a chance to properly evaluate them. Whatever they're meant to do, they're doing it wrong. Also: - Clean up last bits of NULL fallout in subr_pe - Don't let ndis_ifmedia_sts() do anything if the IFF_UP flag isn't set - Implement NdisSystemProcessorCount() and NdisQueryMapRegisterCount().
2003-12-26 03:31:34 +00:00
{
return(mp_ncpus);
mp_ncpus is always defined now, so no need to do an #ifdef SMP in ndis_cpu_cnt().
2004-01-14 01:16:14 +00:00
}
Back out the last batch of changes until I have a chance to properly evaluate them. Whatever they're meant to do, they're doing it wrong. Also: - Clean up last bits of NULL fallout in subr_pe - Don't let ndis_ifmedia_sts() do anything if the IFF_UP flag isn't set - Implement NdisSystemProcessorCount() and NdisQueryMapRegisterCount().
2003-12-26 03:31:34 +00:00
In ndis_attach(), report the NDIS API level that the Windows miniport driver was compiled with. Remove debug printf from ndis_assicn_pcirsc(). It doesn't serve much purpose. Implement NdisMIndicateStatus() and NdisMIndicateStatusComplete() as functions in subr_ndis.c. In NDIS 4.0, they were functions. In NDIS 5.0 and later, they're just macros. Allocate a few extra packets/buffers beyond what the driver asks for since sometimes it seems they can lie about how many they really need, and some extra stupid ones don't check to see if NdisAllocatePacket() and/or NdisAllocateBuffer() actually succeed.
2004-01-04 03:00:21 +00:00
typedef void (*ndis_statusdone_handler)(ndis_handle);
typedef void (*ndis_status_handler)(ndis_handle, ndis_status,
void *, uint32_t);
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMIndicateStatusComplete(adapter)
In ndis_attach(), report the NDIS API level that the Windows miniport driver was compiled with. Remove debug printf from ndis_assicn_pcirsc(). It doesn't serve much purpose. Implement NdisMIndicateStatus() and NdisMIndicateStatusComplete() as functions in subr_ndis.c. In NDIS 4.0, they were functions. In NDIS 5.0 and later, they're just macros. Allocate a few extra packets/buffers beyond what the driver asks for since sometimes it seems they can lie about how many they really need, and some extra stupid ones don't check to see if NdisAllocatePacket() and/or NdisAllocateBuffer() actually succeed.
2004-01-04 03:00:21 +00:00
ndis_handle adapter;
{
ndis_miniport_block *block;
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
ndis_statusdone_handler statusdonefunc;
In ndis_attach(), report the NDIS API level that the Windows miniport driver was compiled with. Remove debug printf from ndis_assicn_pcirsc(). It doesn't serve much purpose. Implement NdisMIndicateStatus() and NdisMIndicateStatusComplete() as functions in subr_ndis.c. In NDIS 4.0, they were functions. In NDIS 5.0 and later, they're just macros. Allocate a few extra packets/buffers beyond what the driver asks for since sometimes it seems they can lie about how many they really need, and some extra stupid ones don't check to see if NdisAllocatePacket() and/or NdisAllocateBuffer() actually succeed.
2004-01-04 03:00:21 +00:00
block = (ndis_miniport_block *)adapter;
statusdonefunc = block->nmb_statusdone_func;
Add support for Windows/x86-64 binaries to Project Evil. Ville-Pertti Keinonen (will at exomi dot comohmygodnospampleasekthx) deserves a big thanks for submitting initial patches to make it work. I have mangled his contributions appropriately. The main gotcha with Windows/x86-64 is that Microsoft uses a different calling convention than everyone else. The standard ABI requires using 6 registers for argument passing, with other arguments on the stack. Microsoft uses only 4 registers, and requires the caller to leave room on the stack for the register arguments incase the callee needs to spill them. Unlike x86, where Microsoft uses a mix of _cdecl, _stdcall and _fastcall, all routines on Windows/x86-64 uses the same convention. This unfortunately means that all the functions we export to the driver require an intermediate translation wrapper. Similarly, we have to wrap all calls back into the driver binary itself. The original patches provided macros to wrap every single routine at compile time, providing a secondary jump table with a customized wrapper for each exported routine. I decided to use a different approach: the call wrapper for each function is created from a template at runtime, and the routine to jump to is patched into the wrapper as it is created. The subr_pe module has been modified to patch in the wrapped function instead of the original. (On x86, the wrapping routine is a no-op.) There are some minor API differences that had to be accounted for: - KeAcquireSpinLock() is a real function on amd64, not a macro wrapper around KfAcquireSpinLock() - NdisFreeBuffer() is actually IoFreeMdl(). I had to change the whole NDIS_BUFFER API a bit to accomodate this. Bugs fixed along the way: - IoAllocateMdl() always returned NULL - kern_windrv.c:windrv_unload() wasn't releasing private driver object extensions correctly (found thanks to memguard) This has only been tested with the driver for the Broadcom 802.11g chipset, which was the only Windows/x86-64 driver I could find.
2005-02-16 05:41:18 +00:00
MSCALL1(statusdonefunc, adapter);
In ndis_attach(), report the NDIS API level that the Windows miniport driver was compiled with. Remove debug printf from ndis_assicn_pcirsc(). It doesn't serve much purpose. Implement NdisMIndicateStatus() and NdisMIndicateStatusComplete() as functions in subr_ndis.c. In NDIS 4.0, they were functions. In NDIS 5.0 and later, they're just macros. Allocate a few extra packets/buffers beyond what the driver asks for since sometimes it seems they can lie about how many they really need, and some extra stupid ones don't check to see if NdisAllocatePacket() and/or NdisAllocateBuffer() actually succeed.
2004-01-04 03:00:21 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMIndicateStatus(adapter, status, sbuf, slen)
In ndis_attach(), report the NDIS API level that the Windows miniport driver was compiled with. Remove debug printf from ndis_assicn_pcirsc(). It doesn't serve much purpose. Implement NdisMIndicateStatus() and NdisMIndicateStatusComplete() as functions in subr_ndis.c. In NDIS 4.0, they were functions. In NDIS 5.0 and later, they're just macros. Allocate a few extra packets/buffers beyond what the driver asks for since sometimes it seems they can lie about how many they really need, and some extra stupid ones don't check to see if NdisAllocatePacket() and/or NdisAllocateBuffer() actually succeed.
2004-01-04 03:00:21 +00:00
ndis_handle adapter;
ndis_status status;
void *sbuf;
uint32_t slen;
{
ndis_miniport_block *block;
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
ndis_status_handler statusfunc;
In ndis_attach(), report the NDIS API level that the Windows miniport driver was compiled with. Remove debug printf from ndis_assicn_pcirsc(). It doesn't serve much purpose. Implement NdisMIndicateStatus() and NdisMIndicateStatusComplete() as functions in subr_ndis.c. In NDIS 4.0, they were functions. In NDIS 5.0 and later, they're just macros. Allocate a few extra packets/buffers beyond what the driver asks for since sometimes it seems they can lie about how many they really need, and some extra stupid ones don't check to see if NdisAllocatePacket() and/or NdisAllocateBuffer() actually succeed.
2004-01-04 03:00:21 +00:00
block = (ndis_miniport_block *)adapter;
statusfunc = block->nmb_status_func;
Add support for Windows/x86-64 binaries to Project Evil. Ville-Pertti Keinonen (will at exomi dot comohmygodnospampleasekthx) deserves a big thanks for submitting initial patches to make it work. I have mangled his contributions appropriately. The main gotcha with Windows/x86-64 is that Microsoft uses a different calling convention than everyone else. The standard ABI requires using 6 registers for argument passing, with other arguments on the stack. Microsoft uses only 4 registers, and requires the caller to leave room on the stack for the register arguments incase the callee needs to spill them. Unlike x86, where Microsoft uses a mix of _cdecl, _stdcall and _fastcall, all routines on Windows/x86-64 uses the same convention. This unfortunately means that all the functions we export to the driver require an intermediate translation wrapper. Similarly, we have to wrap all calls back into the driver binary itself. The original patches provided macros to wrap every single routine at compile time, providing a secondary jump table with a customized wrapper for each exported routine. I decided to use a different approach: the call wrapper for each function is created from a template at runtime, and the routine to jump to is patched into the wrapper as it is created. The subr_pe module has been modified to patch in the wrapped function instead of the original. (On x86, the wrapping routine is a no-op.) There are some minor API differences that had to be accounted for: - KeAcquireSpinLock() is a real function on amd64, not a macro wrapper around KfAcquireSpinLock() - NdisFreeBuffer() is actually IoFreeMdl(). I had to change the whole NDIS_BUFFER API a bit to accomodate this. Bugs fixed along the way: - IoAllocateMdl() always returned NULL - kern_windrv.c:windrv_unload() wasn't releasing private driver object extensions correctly (found thanks to memguard) This has only been tested with the driver for the Broadcom 802.11g chipset, which was the only Windows/x86-64 driver I could find.
2005-02-16 05:41:18 +00:00
MSCALL4(statusfunc, adapter, status, sbuf, slen);
In ndis_attach(), report the NDIS API level that the Windows miniport driver was compiled with. Remove debug printf from ndis_assicn_pcirsc(). It doesn't serve much purpose. Implement NdisMIndicateStatus() and NdisMIndicateStatusComplete() as functions in subr_ndis.c. In NDIS 4.0, they were functions. In NDIS 5.0 and later, they're just macros. Allocate a few extra packets/buffers beyond what the driver asks for since sometimes it seems they can lie about how many they really need, and some extra stupid ones don't check to see if NdisAllocatePacket() and/or NdisAllocateBuffer() actually succeed.
2004-01-04 03:00:21 +00:00
return;
}
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
/*
* The DDK documentation says that you should use IoQueueWorkItem()
* instead of ExQueueWorkItem(). The problem is, IoQueueWorkItem()
* is fundamentally incompatible with NdisScheduleWorkItem(), which
* depends on the API semantics of ExQueueWorkItem(). In our world,
* ExQueueWorkItem() is implemented on top of IoAllocateQueueItem()
* anyway.
Correct some problems with workitem usage. NdisScheduleWorkItem() does not use exactly the same workitem sturcture as ExQueueWorkItem() like I originally thought it did.
2005-05-16 15:29:21 +00:00
*
* There are actually three distinct APIs here. NdisScheduleWorkItem()
* takes a pointer to an NDIS_WORK_ITEM. ExQueueWorkItem() takes a pointer
* to a WORK_QUEUE_ITEM. And finally, IoQueueWorkItem() takes a pointer
* to an opaque work item thingie which you get from IoAllocateWorkItem().
* An NDIS_WORK_ITEM is not the same as a WORK_QUEUE_ITEM. However,
* the NDIS_WORK_ITEM has some opaque storage at the end of it, and we
* (ab)use this storage as a WORK_QUEUE_ITEM, which is what we submit
* to ExQueueWorkItem().
*
* Got all that? (Sheesh.)
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
*/
Implement NdisScheduleWorkItem() and RtlCompareMemory(). Also, call the libinit and libfini routines from the modevent handler in kern_ndis.c. This simplifies the initialization a little.
2004-01-04 07:47:33 +00:00
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
ndis_status
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisScheduleWorkItem(work)
Implement NdisScheduleWorkItem() and RtlCompareMemory(). Also, call the libinit and libfini routines from the modevent handler in kern_ndis.c. This simplifies the initialization a little.
2004-01-04 07:47:33 +00:00
ndis_work_item *work;
{
Correct some problems with workitem usage. NdisScheduleWorkItem() does not use exactly the same workitem sturcture as ExQueueWorkItem() like I originally thought it did.
2005-05-16 15:29:21 +00:00
work_queue_item *wqi;
wqi = (work_queue_item *)work->nwi_wraprsvd;
ExInitializeWorkItem(wqi,
(work_item_func)work->nwi_func, work->nwi_ctx);
ExQueueWorkItem(wqi, WORKQUEUE_DELAYED);
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
Implement NdisScheduleWorkItem() and RtlCompareMemory(). Also, call the libinit and libfini routines from the modevent handler in kern_ndis.c. This simplifies the initialization a little.
2004-01-04 07:47:33 +00:00
return(NDIS_STATUS_SUCCESS);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisCopyFromPacketToPacket(dpkt, doff, reqlen, spkt, soff, cpylen)
Implement NdisCopyFromPacketToPacket() and NdisCopyFromPacketToPacketSafe(). I only have one driver that references this routine (for the 3Com 3cR990) and it never gets called, but just in case, here it is.
2004-01-15 07:42:26 +00:00
ndis_packet *dpkt;
uint32_t doff;
uint32_t reqlen;
ndis_packet *spkt;
uint32_t soff;
uint32_t *cpylen;
{
ndis_buffer *src, *dst;
char *sptr, *dptr;
int resid, copied, len, scnt, dcnt;
*cpylen = 0;
src = spkt->np_private.npp_head;
dst = dpkt->np_private.npp_head;
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
sptr = MmGetMdlVirtualAddress(src);
dptr = MmGetMdlVirtualAddress(dst);
scnt = MmGetMdlByteCount(src);
dcnt = MmGetMdlByteCount(dst);
Implement NdisCopyFromPacketToPacket() and NdisCopyFromPacketToPacketSafe(). I only have one driver that references this routine (for the 3Com 3cR990) and it never gets called, but just in case, here it is.
2004-01-15 07:42:26 +00:00
while (soff) {
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
if (MmGetMdlByteCount(src) > soff) {
Implement NdisCopyFromPacketToPacket() and NdisCopyFromPacketToPacketSafe(). I only have one driver that references this routine (for the 3Com 3cR990) and it never gets called, but just in case, here it is.
2004-01-15 07:42:26 +00:00
sptr += soff;
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
scnt = MmGetMdlByteCount(src)- soff;
Implement NdisCopyFromPacketToPacket() and NdisCopyFromPacketToPacketSafe(). I only have one driver that references this routine (for the 3Com 3cR990) and it never gets called, but just in case, here it is.
2004-01-15 07:42:26 +00:00
break;
}
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
soff -= MmGetMdlByteCount(src);
src = src->mdl_next;
Implement NdisCopyFromPacketToPacket() and NdisCopyFromPacketToPacketSafe(). I only have one driver that references this routine (for the 3Com 3cR990) and it never gets called, but just in case, here it is.
2004-01-15 07:42:26 +00:00
if (src == NULL)
return;
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
sptr = MmGetMdlVirtualAddress(src);
Implement NdisCopyFromPacketToPacket() and NdisCopyFromPacketToPacketSafe(). I only have one driver that references this routine (for the 3Com 3cR990) and it never gets called, but just in case, here it is.
2004-01-15 07:42:26 +00:00
}
while (doff) {
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
if (MmGetMdlByteCount(dst) > doff) {
Implement NdisCopyFromPacketToPacket() and NdisCopyFromPacketToPacketSafe(). I only have one driver that references this routine (for the 3Com 3cR990) and it never gets called, but just in case, here it is.
2004-01-15 07:42:26 +00:00
dptr += doff;
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
dcnt = MmGetMdlByteCount(dst) - doff;
Implement NdisCopyFromPacketToPacket() and NdisCopyFromPacketToPacketSafe(). I only have one driver that references this routine (for the 3Com 3cR990) and it never gets called, but just in case, here it is.
2004-01-15 07:42:26 +00:00
break;
}
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
doff -= MmGetMdlByteCount(dst);
dst = dst->mdl_next;
Implement NdisCopyFromPacketToPacket() and NdisCopyFromPacketToPacketSafe(). I only have one driver that references this routine (for the 3Com 3cR990) and it never gets called, but just in case, here it is.
2004-01-15 07:42:26 +00:00
if (dst == NULL)
return;
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
dptr = MmGetMdlVirtualAddress(dst);
Implement NdisCopyFromPacketToPacket() and NdisCopyFromPacketToPacketSafe(). I only have one driver that references this routine (for the 3Com 3cR990) and it never gets called, but just in case, here it is.
2004-01-15 07:42:26 +00:00
}
resid = reqlen;
copied = 0;
while(1) {
if (resid < scnt)
len = resid;
else
len = scnt;
if (dcnt < len)
len = dcnt;
bcopy(sptr, dptr, len);
copied += len;
resid -= len;
if (resid == 0)
break;
dcnt -= len;
if (dcnt == 0) {
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
dst = dst->mdl_next;
Implement NdisCopyFromPacketToPacket() and NdisCopyFromPacketToPacketSafe(). I only have one driver that references this routine (for the 3Com 3cR990) and it never gets called, but just in case, here it is.
2004-01-15 07:42:26 +00:00
if (dst == NULL)
break;
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
dptr = MmGetMdlVirtualAddress(dst);
dcnt = MmGetMdlByteCount(dst);
Implement NdisCopyFromPacketToPacket() and NdisCopyFromPacketToPacketSafe(). I only have one driver that references this routine (for the 3Com 3cR990) and it never gets called, but just in case, here it is.
2004-01-15 07:42:26 +00:00
}
scnt -= len;
if (scnt == 0) {
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
src = src->mdl_next;
Implement NdisCopyFromPacketToPacket() and NdisCopyFromPacketToPacketSafe(). I only have one driver that references this routine (for the 3Com 3cR990) and it never gets called, but just in case, here it is.
2004-01-15 07:42:26 +00:00
if (src == NULL)
break;
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
sptr = MmGetMdlVirtualAddress(src);
scnt = MmGetMdlByteCount(src);
Implement NdisCopyFromPacketToPacket() and NdisCopyFromPacketToPacketSafe(). I only have one driver that references this routine (for the 3Com 3cR990) and it never gets called, but just in case, here it is.
2004-01-15 07:42:26 +00:00
}
}
*cpylen = copied;
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisCopyFromPacketToPacketSafe(dpkt, doff, reqlen, spkt, soff, cpylen, prio)
Implement NdisCopyFromPacketToPacket() and NdisCopyFromPacketToPacketSafe(). I only have one driver that references this routine (for the 3Com 3cR990) and it never gets called, but just in case, here it is.
2004-01-15 07:42:26 +00:00
ndis_packet *dpkt;
uint32_t doff;
uint32_t reqlen;
ndis_packet *spkt;
uint32_t soff;
uint32_t *cpylen;
uint32_t prio;
{
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisCopyFromPacketToPacket(dpkt, doff, reqlen, spkt, soff, cpylen);
Implement NdisCopyFromPacketToPacket() and NdisCopyFromPacketToPacketSafe(). I only have one driver that references this routine (for the 3Com 3cR990) and it never gets called, but just in case, here it is.
2004-01-15 07:42:26 +00:00
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMRegisterDevice(handle, devname, symname, majorfuncs, devobj, devhandle)
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
ndis_handle handle;
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
unicode_string *devname;
unicode_string *symname;
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
driver_dispatch *majorfuncs[];
void **devobj;
ndis_handle *devhandle;
Reorganize the timer code a little and implement NdisInitializeTimer() and NdisCancelTimer(). NdisInitializeTimer() doesn't accept an NDIS miniport context argument, so we have to derive it from the timer function context (which is supposed to be the adapter private context). NdisCancelTimer is now an alias for NdisMCancelTimer(). Also add stubs for NdisMRegisterDevice() and NdisMDeregisterDevice(). These are no-ops for now, but will likely get fleshed in once I start working on the Am1771/Am1772 wireless driver.
2004-01-26 21:21:53 +00:00
{
Fix some of the things I broke so that the SMC2602W (AMD Am1772) driver works again. This driver uses NdisScheduleWorkItem(), and we have to take special steps to insure that its workitems don't collide with any of the other workitems used by the NDISulator. In particular, if one of the driver's work jobs blocks, it can prevent NdisMAllocateSharedMemoryAsync() from completing when expected. The original hack to fix this was to have NdisMAllocateSharedMemoryAsync() defer its work to the DPC queue instead of the general task queue. To fix it now, I decided to add some additional workitem threads. (There's supposed to be a pool of worker threads in Windows anyway.) Currently, there are 4. There should be at least 2. One is reserved for the legacy ExQueueWorkItem() API, while the others are used in round-robin by the IoQueueWorkItem() API. NdisMAllocateSharedMemoryAsync() uses the latter API while NdisScheduleWorkItem() uses the former, so the deadlock is avoided. Fixed NdisMRegisterDevice()/NdisMDeregisterDevice() to work a little more sensibly with the new driver_object/device_object framework. It doesn't really register a working user-mode interface, but the existing code was completely wrong for the new framework. Fixed a couple of bugs dealing with the cancellation of events and DPCs. When cancelling an event that's still on the timer queue (i.e. hasn't expired yet), reset dh_inserted in its dispatch header to FALSE. Previously, it was left set to TRUE, which would make a cancelled timer appear to have not been cancelled. Also, when removing a DPC from a queue, reset its list pointers, otherwise a cancelled DPC might mistakenly be treated as still pending. Lastly, fix the behavior of ntoskrnl_wakeup() when dealing with objects that have nobody waiting on them: sync event objects get their signalled state reset to FALSE, but notification objects should still be set to TRUE.
2005-05-19 04:44:26 +00:00
uint32_t status;
device_object *dobj;
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
Fix some of the things I broke so that the SMC2602W (AMD Am1772) driver works again. This driver uses NdisScheduleWorkItem(), and we have to take special steps to insure that its workitems don't collide with any of the other workitems used by the NDISulator. In particular, if one of the driver's work jobs blocks, it can prevent NdisMAllocateSharedMemoryAsync() from completing when expected. The original hack to fix this was to have NdisMAllocateSharedMemoryAsync() defer its work to the DPC queue instead of the general task queue. To fix it now, I decided to add some additional workitem threads. (There's supposed to be a pool of worker threads in Windows anyway.) Currently, there are 4. There should be at least 2. One is reserved for the legacy ExQueueWorkItem() API, while the others are used in round-robin by the IoQueueWorkItem() API. NdisMAllocateSharedMemoryAsync() uses the latter API while NdisScheduleWorkItem() uses the former, so the deadlock is avoided. Fixed NdisMRegisterDevice()/NdisMDeregisterDevice() to work a little more sensibly with the new driver_object/device_object framework. It doesn't really register a working user-mode interface, but the existing code was completely wrong for the new framework. Fixed a couple of bugs dealing with the cancellation of events and DPCs. When cancelling an event that's still on the timer queue (i.e. hasn't expired yet), reset dh_inserted in its dispatch header to FALSE. Previously, it was left set to TRUE, which would make a cancelled timer appear to have not been cancelled. Also, when removing a DPC from a queue, reset its list pointers, otherwise a cancelled DPC might mistakenly be treated as still pending. Lastly, fix the behavior of ntoskrnl_wakeup() when dealing with objects that have nobody waiting on them: sync event objects get their signalled state reset to FALSE, but notification objects should still be set to TRUE.
2005-05-19 04:44:26 +00:00
status = IoCreateDevice(handle, 0, devname,
FILE_DEVICE_UNKNOWN, 0, FALSE, &dobj);
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
Fix some of the things I broke so that the SMC2602W (AMD Am1772) driver works again. This driver uses NdisScheduleWorkItem(), and we have to take special steps to insure that its workitems don't collide with any of the other workitems used by the NDISulator. In particular, if one of the driver's work jobs blocks, it can prevent NdisMAllocateSharedMemoryAsync() from completing when expected. The original hack to fix this was to have NdisMAllocateSharedMemoryAsync() defer its work to the DPC queue instead of the general task queue. To fix it now, I decided to add some additional workitem threads. (There's supposed to be a pool of worker threads in Windows anyway.) Currently, there are 4. There should be at least 2. One is reserved for the legacy ExQueueWorkItem() API, while the others are used in round-robin by the IoQueueWorkItem() API. NdisMAllocateSharedMemoryAsync() uses the latter API while NdisScheduleWorkItem() uses the former, so the deadlock is avoided. Fixed NdisMRegisterDevice()/NdisMDeregisterDevice() to work a little more sensibly with the new driver_object/device_object framework. It doesn't really register a working user-mode interface, but the existing code was completely wrong for the new framework. Fixed a couple of bugs dealing with the cancellation of events and DPCs. When cancelling an event that's still on the timer queue (i.e. hasn't expired yet), reset dh_inserted in its dispatch header to FALSE. Previously, it was left set to TRUE, which would make a cancelled timer appear to have not been cancelled. Also, when removing a DPC from a queue, reset its list pointers, otherwise a cancelled DPC might mistakenly be treated as still pending. Lastly, fix the behavior of ntoskrnl_wakeup() when dealing with objects that have nobody waiting on them: sync event objects get their signalled state reset to FALSE, but notification objects should still be set to TRUE.
2005-05-19 04:44:26 +00:00
if (status == STATUS_SUCCESS) {
*devobj = dobj;
*devhandle = dobj;
}
return(status);
Reorganize the timer code a little and implement NdisInitializeTimer() and NdisCancelTimer(). NdisInitializeTimer() doesn't accept an NDIS miniport context argument, so we have to derive it from the timer function context (which is supposed to be the adapter private context). NdisCancelTimer is now an alias for NdisMCancelTimer(). Also add stubs for NdisMRegisterDevice() and NdisMDeregisterDevice(). These are no-ops for now, but will likely get fleshed in once I start working on the Am1771/Am1772 wireless driver.
2004-01-26 21:21:53 +00:00
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMDeregisterDevice(handle)
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
ndis_handle handle;
Reorganize the timer code a little and implement NdisInitializeTimer() and NdisCancelTimer(). NdisInitializeTimer() doesn't accept an NDIS miniport context argument, so we have to derive it from the timer function context (which is supposed to be the adapter private context). NdisCancelTimer is now an alias for NdisMCancelTimer(). Also add stubs for NdisMRegisterDevice() and NdisMDeregisterDevice(). These are no-ops for now, but will likely get fleshed in once I start working on the Am1771/Am1772 wireless driver.
2004-01-26 21:21:53 +00:00
{
Fix some of the things I broke so that the SMC2602W (AMD Am1772) driver works again. This driver uses NdisScheduleWorkItem(), and we have to take special steps to insure that its workitems don't collide with any of the other workitems used by the NDISulator. In particular, if one of the driver's work jobs blocks, it can prevent NdisMAllocateSharedMemoryAsync() from completing when expected. The original hack to fix this was to have NdisMAllocateSharedMemoryAsync() defer its work to the DPC queue instead of the general task queue. To fix it now, I decided to add some additional workitem threads. (There's supposed to be a pool of worker threads in Windows anyway.) Currently, there are 4. There should be at least 2. One is reserved for the legacy ExQueueWorkItem() API, while the others are used in round-robin by the IoQueueWorkItem() API. NdisMAllocateSharedMemoryAsync() uses the latter API while NdisScheduleWorkItem() uses the former, so the deadlock is avoided. Fixed NdisMRegisterDevice()/NdisMDeregisterDevice() to work a little more sensibly with the new driver_object/device_object framework. It doesn't really register a working user-mode interface, but the existing code was completely wrong for the new framework. Fixed a couple of bugs dealing with the cancellation of events and DPCs. When cancelling an event that's still on the timer queue (i.e. hasn't expired yet), reset dh_inserted in its dispatch header to FALSE. Previously, it was left set to TRUE, which would make a cancelled timer appear to have not been cancelled. Also, when removing a DPC from a queue, reset its list pointers, otherwise a cancelled DPC might mistakenly be treated as still pending. Lastly, fix the behavior of ntoskrnl_wakeup() when dealing with objects that have nobody waiting on them: sync event objects get their signalled state reset to FALSE, but notification objects should still be set to TRUE.
2005-05-19 04:44:26 +00:00
IoDeleteDevice(handle);
Reorganize the timer code a little and implement NdisInitializeTimer() and NdisCancelTimer(). NdisInitializeTimer() doesn't accept an NDIS miniport context argument, so we have to derive it from the timer function context (which is supposed to be the adapter private context). NdisCancelTimer is now an alias for NdisMCancelTimer(). Also add stubs for NdisMRegisterDevice() and NdisMDeregisterDevice(). These are no-ops for now, but will likely get fleshed in once I start working on the Am1771/Am1772 wireless driver.
2004-01-26 21:21:53 +00:00
return(NDIS_STATUS_SUCCESS);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static ndis_status
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMQueryAdapterInstanceName(name, handle)
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
unicode_string *name;
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
ndis_handle handle;
{
ndis_miniport_block *block;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
device_t dev;
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
ansi_string as;
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
block = (ndis_miniport_block *)handle;
Next step on the road to IRPs: create and use an imitation of the Windows DRIVER_OBJECT and DEVICE_OBJECT mechanism so that we can simulate driver stacking. In Windows, each loaded driver image is attached to a DRIVER_OBJECT structure. Windows uses the registry to match up a given vendor/device ID combination with a corresponding DRIVER_OBJECT. When a driver image is first loaded, its DriverEntry() routine is invoked, which sets up the AddDevice() function pointer in the DRIVER_OBJECT and creates a dispatch table (based on IRP major codes). When a Windows bus driver detects a new device, it creates a Physical Device Object (PDO) for it. This is a DEVICE_OBJECT structure, with semantics analagous to that of a device_t in FreeBSD. The Windows PNP manager will invoke the driver's AddDevice() function and pass it pointers to the DRIVER_OBJECT and the PDO. The AddDevice() function then creates a new DRIVER_OBJECT structure of its own. This is known as the Functional Device Object (FDO) and corresponds roughly to a private softc instance. The driver uses IoAttachDeviceToDeviceStack() to add this device object to the driver stack for this PDO. Subsequent drivers (called filter drivers in Windows-speak) can be loaded which add themselves to the stack. When someone issues an IRP to a device, it travel along the stack passing through several possible filter drivers until it reaches the functional driver (which actually knows how to talk to the hardware) at which point it will be completed. This is how Windows achieves driver layering. Project Evil now simulates most of this. if_ndis now has a modevent handler which will use MOD_LOAD and MOD_UNLOAD events to drive the creation and destruction of DRIVER_OBJECTs. (The load event also does the relocation/dynalinking of the image.) We don't have a registry, so the DRIVER_OBJECTS are stored in a linked list for now. Eventually, the list entry will contain the vendor/device ID list extracted from the .INF file. When ndis_probe() is called and detectes a supported device, it will create a PDO for the device instance and attach it to the DRIVER_OBJECT just as in Windows. ndis_attach() will then call our NdisAddDevice() handler to create the FDO. The NDIS miniport block is now a device extension hung off the FDO, just as it is in Windows. The miniport characteristics table is now an extension hung off the DRIVER_OBJECT as well (the characteristics are the same for all devices handled by a given driver, so they don't need to be per-instance.) We also do an IoAttachDeviceToDeviceStack() to put the FDO on the stack for the PDO. There are a couple of fake bus drivers created for the PCI and pccard buses. Eventually, there will be one for USB, which will actually accept USB IRP.s Things should still work just as before, only now we do things in the proper order and maintain the correct framework to support passing IRPs between drivers. Various changes: - corrected the comments about IRQL handling in subr_hal.c to more accurately reflect reality - update ndiscvt to make the drv_data symbol in ndis_driver_data.h a global so that if_ndis_pci.o and/or if_ndis_pccard.o can see it. - Obtain the softc pointer from the miniport block by referencing the PDO rather than a private pointer of our own (nmb_ifp is no longer used) - implement IoAttachDeviceToDeviceStack(), IoDetachDevice(), IoGetAttachedDevice(), IoAllocateDriverObjectExtension(), IoGetDriverObjectExtension(), IoCreateDevice(), IoDeleteDevice(), IoAllocateIrp(), IoReuseIrp(), IoMakeAssociatedIrp(), IoFreeIrp(), IoInitializeIrp() - fix a few mistakes in the driver_object and device_object definitions - add a new module, kern_windrv.c, to handle the driver registration and relocation/dynalinkign duties (which don't really belong in kern_ndis.c). - made ndis_block and ndis_chars in the ndis_softc stucture pointers and modified all references to it - fixed NdisMRegisterMiniport() and NdisInitializeWrapper() so they work correctly with the new driver_object mechanism - changed ndis_attach() to call NdisAddDevice() instead of ndis_load_driver() (which is now deprecated) - used ExAllocatePoolWithTag()/ExFreePool() in lookaside list routines instead of kludged up alloc/free routines - added kern_windrv.c to sys/modules/ndis/Makefile and files.i386.
2005-02-08 17:23:25 +00:00
dev = block->nmb_physdeviceobj->do_devext;
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
RtlInitAnsiString(&as, __DECONST(char *, device_get_nameunit(dev)));
if (RtlAnsiStringToUnicodeString(name, &as, TRUE))
return(NDIS_STATUS_RESOURCES);
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
return(NDIS_STATUS_SUCCESS);
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Begin the first phase of trying to add IRP support (and ultimately USB device support): - Convert all of my locally chosen function names to their actual Windows equivalents, where applicable. This is a big no-op change since it doesn't affect functionality, but it helps avoid a bit of confusion (it's now a lot easier to see which functions are emulated Windows API routines and which are just locally defined). - Turn ndis_buffer into an mdl, like it should have been. The structure is the same, but now it belongs to the subr_ntoskrnl module. - Implement a bunch of MDL handling macros from Windows and use them where applicable. - Correct the implementation of IoFreeMdl(). - Properly implement IoAllocateMdl() and MmBuildMdlForNonPagedPool(). - Add the definitions for struct irp and struct driver_object. - Add IMPORT_FUNC() and IMPORT_FUNC_MAP() macros to make formatting the module function tables a little cleaner. (Should also help with AMD64 support later on.) - Fix if_ndis.c to use KeRaiseIrql() and KeLowerIrql() instead of the previous calls to hal_raise_irql() and hal_lower_irql() which have been renamed. The function renaming generated a lot of churn here, but there should be very little operational effect.
2005-01-24 18:18:12 +00:00
NdisMRegisterUnloadHandler(handle, func)
Add a whole bunch of new stuff to make the driver for the AMD Am1771/Am1772 802.11b chipset work. This chip is present on the SMC2602W version 3 NIC, which is what was used for testing. This driver creates kernel threads (12 of them!) for various purposes, and required the following routines: PsCreateSystemThread() PsTerminateSystemThread() KeInitializeEvent() KeSetEvent() KeResetEvent() KeInitializeMutex() KeReleaseMutex() KeWaitForSingleObject() KeWaitForMultipleObjects() IoGetDeviceProperty() and several more. Also, this driver abuses the fact that NDIS events and timers are actually Windows events and timers, and uses NDIS events with KeWaitForSingleObject(). The NDIS event routines have been rewritten to interface with the ntoskrnl module. Many routines with incorrect prototypes have been cleaned up. Also, this driver puts jobs on the NDIS taskqueue (via NdisScheduleWorkItem()) which block on events, and this interferes with the operation of NdisMAllocateSharedMemoryAsync(), which was also being put on the NDIS taskqueue. To avoid the deadlock, NdisMAllocateSharedMemoryAsync() is now performed in the NDIS SWI thread instead. There's still room for some cleanups here, and I really should implement KeInitializeTimer() and friends.
2004-02-07 06:44:13 +00:00
ndis_handle handle;
void *func;
{
return;
}
Reorganize the timer code a little and implement NdisInitializeTimer() and NdisCancelTimer(). NdisInitializeTimer() doesn't accept an NDIS miniport context argument, so we have to derive it from the timer function context (which is supposed to be the adapter private context). NdisCancelTimer is now an alias for NdisMCancelTimer(). Also add stubs for NdisMRegisterDevice() and NdisMDeregisterDevice(). These are no-ops for now, but will likely get fleshed in once I start working on the Am1771/Am1772 wireless driver.
2004-01-26 21:21:53 +00:00
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
static void
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
dummy()
{
printf ("NDIS dummy called...\n");
return;
}
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
/*
* Note: a couple of entries in this table specify the
* number of arguments as "foo + 1". These are routines
* that accept a 64-bit argument, passed by value. On
* x86, these arguments consume two longwords on the stack,
* so we lie and say there's one additional argument so
* that the wrapping routines will do the right thing.
*/
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
image_patch_table ndis_functbl[] = {
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
IMPORT_SFUNC(NdisCopyFromPacketToPacket, 6),
IMPORT_SFUNC(NdisCopyFromPacketToPacketSafe, 7),
IMPORT_SFUNC(NdisScheduleWorkItem, 1),
IMPORT_SFUNC(NdisMIndicateStatusComplete, 1),
IMPORT_SFUNC(NdisMIndicateStatus, 4),
IMPORT_SFUNC(NdisSystemProcessorCount, 0),
IMPORT_SFUNC(NdisUnchainBufferAtBack, 2),
IMPORT_SFUNC(NdisGetFirstBufferFromPacket, 5),
IMPORT_SFUNC(NdisGetFirstBufferFromPacketSafe, 6),
IMPORT_SFUNC(NdisGetBufferPhysicalArraySize, 2),
IMPORT_SFUNC(NdisMGetDeviceProperty, 6),
IMPORT_SFUNC(NdisInitAnsiString, 2),
IMPORT_SFUNC(NdisInitUnicodeString, 2),
IMPORT_SFUNC(NdisWriteConfiguration, 4),
IMPORT_SFUNC(NdisAnsiStringToUnicodeString, 2),
IMPORT_SFUNC(NdisTerminateWrapper, 2),
IMPORT_SFUNC(NdisOpenConfigurationKeyByName, 4),
IMPORT_SFUNC(NdisOpenConfigurationKeyByIndex, 5),
IMPORT_SFUNC(NdisMRemoveMiniport, 1),
IMPORT_SFUNC(NdisInitializeString, 2),
IMPORT_SFUNC(NdisFreeString, 1),
IMPORT_SFUNC(NdisGetCurrentSystemTime, 1),
IMPORT_SFUNC(NdisGetSystemUpTime, 1),
IMPORT_SFUNC(NdisMSynchronizeWithInterrupt, 3),
IMPORT_SFUNC(NdisMAllocateSharedMemoryAsync, 4),
IMPORT_SFUNC(NdisInterlockedInsertHeadList, 3),
IMPORT_SFUNC(NdisInterlockedInsertTailList, 3),
IMPORT_SFUNC(NdisInterlockedRemoveHeadList, 2),
IMPORT_SFUNC(NdisInitializeWrapper, 4),
IMPORT_SFUNC(NdisMRegisterMiniport, 3),
IMPORT_SFUNC(NdisAllocateMemoryWithTag, 3),
IMPORT_SFUNC(NdisAllocateMemory, 4 + 1),
IMPORT_SFUNC(NdisMSetAttributesEx, 5),
IMPORT_SFUNC(NdisCloseConfiguration, 1),
IMPORT_SFUNC(NdisReadConfiguration, 5),
IMPORT_SFUNC(NdisOpenConfiguration, 3),
IMPORT_SFUNC(NdisAcquireSpinLock, 1),
IMPORT_SFUNC(NdisReleaseSpinLock, 1),
IMPORT_SFUNC(NdisDprAcquireSpinLock, 1),
IMPORT_SFUNC(NdisDprReleaseSpinLock, 1),
IMPORT_SFUNC(NdisAllocateSpinLock, 1),
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
IMPORT_SFUNC(NdisInitializeReadWriteLock, 1),
IMPORT_SFUNC(NdisAcquireReadWriteLock, 3),
IMPORT_SFUNC(NdisReleaseReadWriteLock, 2),
Create new i386 windows/bsd thunking layer, similar to the amd64 thunking layer, but with a twist. The twist has to do with the fact that Microsoft supports structured exception handling in kernel mode. On the i386 arch, exception handling is implemented by hanging an exception registration list off the Thread Environment Block (TEB), and the TEB is accessed via the %fs register. The problem is, we use %fs as a pointer to the pcpu stucture, which means any driver that tries to write through %fs:0 will overwrite the curthread pointer and make a serious mess of things. To get around this, Project Evil now creates a special entry in the GDT on each processor. When we call into Windows code, a context switch routine will fix up %fs so it points to our new descriptor, which in turn points to a fake TEB. When the Windows code returns, or calls out to an external routine, we swap %fs back again. Currently, Project Evil makes use of GDT slot 7, which is all 0s by default. I fully expect someone to jump up and say I can't do that, but I couldn't find any code that makes use of this entry anywhere. Sadly, this was the only method I could come up with that worked on both UP and SMP. (Modifying the LDT works on UP, but becomes incredibly complicated on SMP.) If necessary, the context switching stuff can be yanked out while preserving the convention calling wrappers. (Fortunately, it looks like Microsoft uses some special epilog/prolog code on amd64 to implement exception handling, so the same nastiness won't be necessary on that arch.) The advantages are: - Any driver that uses %fs as though it were a TEB pointer won't clobber pcpu. - All the __stdcall/__fastcall/__regparm stuff that's specific to gcc goes away. Also, while I'm here, switch NdisGetSystemUpTime() back to using nanouptime() again. It turns out nanouptime() is way more accurate than just using ticks(). On slower machines, the Atheros drivers I tested seem to take a long time to associate due to the loss in accuracy.
2005-04-11 02:02:35 +00:00
IMPORT_SFUNC(NdisFreeSpinLock, 1),
IMPORT_SFUNC(NdisFreeMemory, 3),
IMPORT_SFUNC(NdisReadPciSlotInformation, 5),
IMPORT_SFUNC(NdisWritePciSlotInformation, 5),
IMPORT_SFUNC_MAP(NdisImmediateReadPciSlotInformation,
NdisReadPciSlotInformation, 5),
IMPORT_SFUNC_MAP(NdisImmediateWritePciSlotInformation,
NdisWritePciSlotInformation, 5),
IMPORT_CFUNC(NdisWriteErrorLogEntry, 0),
IMPORT_SFUNC(NdisMStartBufferPhysicalMapping, 6),
IMPORT_SFUNC(NdisMCompleteBufferPhysicalMapping, 3),
IMPORT_SFUNC(NdisMInitializeTimer, 4),
IMPORT_SFUNC(NdisInitializeTimer, 3),
IMPORT_SFUNC(NdisSetTimer, 2),
IMPORT_SFUNC(NdisMCancelTimer, 2),
IMPORT_SFUNC_MAP(NdisCancelTimer, NdisMCancelTimer, 2),
IMPORT_SFUNC(NdisMSetPeriodicTimer, 2),
IMPORT_SFUNC(NdisMQueryAdapterResources, 4),
IMPORT_SFUNC(NdisMRegisterIoPortRange, 4),
IMPORT_SFUNC(NdisMDeregisterIoPortRange, 4),
IMPORT_SFUNC(NdisReadNetworkAddress, 4),
IMPORT_SFUNC(NdisQueryMapRegisterCount, 2),
IMPORT_SFUNC(NdisMAllocateMapRegisters, 5),
IMPORT_SFUNC(NdisMFreeMapRegisters, 1),
IMPORT_SFUNC(NdisMAllocateSharedMemory, 5),
IMPORT_SFUNC(NdisMMapIoSpace, 4 + 1),
IMPORT_SFUNC(NdisMUnmapIoSpace, 3),
IMPORT_SFUNC(NdisGetCacheFillSize, 0),
IMPORT_SFUNC(NdisMGetDmaAlignment, 1),
IMPORT_SFUNC(NdisMInitializeScatterGatherDma, 3),
IMPORT_SFUNC(NdisAllocatePacketPool, 4),
IMPORT_SFUNC(NdisAllocatePacketPoolEx, 5),
IMPORT_SFUNC(NdisAllocatePacket, 3),
IMPORT_SFUNC(NdisFreePacket, 1),
IMPORT_SFUNC(NdisFreePacketPool, 1),
IMPORT_SFUNC_MAP(NdisDprAllocatePacket, NdisAllocatePacket, 3),
IMPORT_SFUNC_MAP(NdisDprFreePacket, NdisFreePacket, 1),
IMPORT_SFUNC(NdisAllocateBufferPool, 3),
IMPORT_SFUNC(NdisAllocateBuffer, 5),
IMPORT_SFUNC(NdisQueryBuffer, 3),
IMPORT_SFUNC(NdisQueryBufferSafe, 4),
IMPORT_SFUNC(NdisBufferVirtualAddress, 1),
IMPORT_SFUNC(NdisBufferVirtualAddressSafe, 2),
IMPORT_SFUNC(NdisBufferLength, 1),
IMPORT_SFUNC(NdisFreeBuffer, 1),
IMPORT_SFUNC(NdisFreeBufferPool, 1),
IMPORT_SFUNC(NdisInterlockedIncrement, 1),
IMPORT_SFUNC(NdisInterlockedDecrement, 1),
IMPORT_SFUNC(NdisInitializeEvent, 1),
IMPORT_SFUNC(NdisSetEvent, 1),
IMPORT_SFUNC(NdisResetEvent, 1),
IMPORT_SFUNC(NdisWaitEvent, 2),
IMPORT_SFUNC(NdisUnicodeStringToAnsiString, 2),
IMPORT_SFUNC(NdisMPciAssignResources, 3),
IMPORT_SFUNC(NdisMFreeSharedMemory, 5 + 1),
IMPORT_SFUNC(NdisMRegisterInterrupt, 7),
IMPORT_SFUNC(NdisMDeregisterInterrupt, 1),
IMPORT_SFUNC(NdisMRegisterAdapterShutdownHandler, 3),
IMPORT_SFUNC(NdisMDeregisterAdapterShutdownHandler, 1),
IMPORT_SFUNC(NDIS_BUFFER_TO_SPAN_PAGES, 1),
IMPORT_SFUNC(NdisQueryBufferOffset, 3),
IMPORT_SFUNC(NdisAdjustBufferLength, 2),
IMPORT_SFUNC(NdisPacketPoolUsage, 1),
IMPORT_SFUNC(NdisMSleep, 1),
IMPORT_SFUNC(NdisUnchainBufferAtFront, 2),
IMPORT_SFUNC(NdisReadPcmciaAttributeMemory, 4),
IMPORT_SFUNC(NdisWritePcmciaAttributeMemory, 4),
IMPORT_SFUNC(NdisOpenFile, 5 + 1),
IMPORT_SFUNC(NdisMapFile, 3),
IMPORT_SFUNC(NdisUnmapFile, 1),
IMPORT_SFUNC(NdisCloseFile, 1),
IMPORT_SFUNC(NdisMRegisterDevice, 6),
IMPORT_SFUNC(NdisMDeregisterDevice, 1),
IMPORT_SFUNC(NdisMQueryAdapterInstanceName, 2),
IMPORT_SFUNC(NdisMRegisterUnloadHandler, 2),
IMPORT_SFUNC(ndis_timercall, 4),
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
IMPORT_SFUNC(ndis_asyncmem_complete, 2),
This commit makes a big round of updates and fixes many, many things. First and most importantly, I threw out the thread priority-twiddling implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in favor of a new scheme that uses sleep mutexes. The old scheme was really very naughty and sought to provide the same behavior as Windows spinlocks (i.e. blocking pre-emption) but in a way that wouldn't raise the ire of WITNESS. The new scheme represents 'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If a thread on cpu0 acquires the 'dispatcher mutex,' it will block any other thread on the same processor that tries to acquire it, in effect only allowing one thread on the processor to be at 'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit and spin' routine on the spinlock variable itself. If a thread on cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher mutex' for cpu1 and then it too does an atomic sit and spin to try acquiring the spinlock. Unlike real spinlocks, this does not disable pre-emption of all threads on the CPU, but it does put any threads involved with the NDISulator to sleep, which is just as good for our purposes. This means I can now play nice with WITNESS, and I can safely do things like call malloc() when I'm at 'DISPATCH_LEVEL,' which you're allowed to do in Windows. Next, I completely re-wrote most of the event/timer/mutex handling and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects() have been re-written to use condition variables instead of msleep(). This allows us to use the Windows convention whereby thread A can tell thread B "wake up with a boosted priority." (With msleep(), you instead have thread B saying "when I get woken up, I'll use this priority here," and thread A can't tell it to do otherwise.) The new KeWaitForMultipleObjects() has been better tested and better duplicates the semantics of its Windows counterpart. I also overhauled the IoQueueWorkItem() API and underlying code. Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the same work item isn't put on the queue twice. ExQueueWorkItem(), which in my implementation is built on top of IoQueueWorkItem(), was also modified to perform a similar test. I renamed the doubly-linked list macros to give them the same names as their Windows counterparts and fixed RemoveListTail() and RemoveListHead() so they properly return the removed item. I also corrected the list handling code in ntoskrnl_dpc_thread() and ntoskrnl_workitem_thread(). I realized that the original logic did not correctly handle the case where a DPC callout tries to queue up another DPC. It works correctly now. I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to use them. I also tried to duplicate the interrupt handling scheme used in Windows. The interrupt handling is now internal to ndis.ko, and the ndis_intr() function has been removed from if_ndis.c. (In the USB case, interrupt handling isn't needed in if_ndis.c anyway.) NdisMSleep() has been rewritten to use a KeWaitForSingleObject() and a KeTimer, which is how it works in Windows. (This is mainly to insure that the NDISulator uses the KeTimer API so I can spot any problems with it that may arise.) KeCancelTimer() has been changed so that it only cancels timers, and does not attempt to cancel a DPC if the timer managed to fire and queue one up before KeCancelTimer() was called. The Windows DDK documentation seems to imply that KeCantelTimer() will also call KeRemoveQueueDpc() if necessary, but it really doesn't. The KeTimer implementation has been rewritten to use the callout API directly instead of timeout()/untimeout(). I still cheat a little in that I have to manage my own small callout timer wheel, but the timer code works more smoothly now. I discovered a race condition using timeout()/untimeout() with periodic timers where untimeout() fails to actually cancel a timer. I don't quite understand where the race is, using callout_init()/callout_reset()/callout_stop() directly seems to fix it. I also discovered and fixed a bug in winx32_wrap.S related to translating _stdcall calls. There are a couple of routines (i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that return 64-bit quantities. On the x86 arch, 64-bit values are returned in the %eax and %edx registers. However, it happens that the ctxsw_utow() routine uses %edx as a scratch register, and x86_stdcall_wrap() and x86_stdcall_call() were only preserving %eax before branching to ctxsw_utow(). This means %edx was getting clobbered in some cases. Curiously, the most noticeable effect of this bug is that the driver for the TI AXC110 chipset would constantly drop and reacquire its link for no apparent reason. Both %eax and %edx are preserved on the stack now. The _fastcall and _regparm wrappers already handled everything correctly. I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem() instead of the NdisScheduleWorkItem() API. This is to avoid possible deadlocks with any drivers that use NdisScheduleWorkItem() themselves. The unicode/ansi conversion handling code has been cleaned up. The internal routines have been moved to subr_ntoskrnl and the RtlXXX routines have been exported so that subr_ndis can call them. This removes the incestuous relationship between the two modules regarding this code and fixes the implementation so that it honors the 'maxlen' fields correctly. (Previously it was possible for NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't own, which was causing many mysterious crashes in the Marvell 8335 driver.) The registry handling code (NdisOpen/Close/ReadConfiguration()) has been fixed to allocate memory for all the parameters it hands out to callers and delete whem when NdisCloseConfiguration() is called. (Previously, it would secretly use a single static buffer.) I also substantially updated if_ndis so that the source can now be built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled. The original WPA code has been updated to fit in more cleanly with the net80211 API, and to eleminate the use of magic numbers. The ndis_80211_setstate() routine now sets a default authmode of OPEN and initializes the RTS threshold and fragmentation threshold. The WPA routines were changed so that the authentication mode is always set first, followed by the cipher. Some drivers depend on the operations being performed in this order. I also added passthrough ioctls that allow application code to directly call the MiniportSetInformation()/MiniportQueryInformation() methods via ndis_set_info() and ndis_get_info(). The ndis_linksts() routine also caches the last 4 events signalled by the driver via NdisMIndicateStatus(), and they can be queried by an application via a separate ioctl. This is done to allow wpa_supplicant to directly program the various crypto and key management options in the driver, allowing things like WPA2 support to work. Whew.
2005-10-10 16:46:39 +00:00
IMPORT_SFUNC(ndis_intr, 1),
IMPORT_SFUNC(ndis_intrhand, 4),
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
/*
* This last entry is a catch-all for any function we haven't
* implemented yet. The PE import list patching routine will
* use it for any function that doesn't have an explicit match
* in this table.
*/
This commit makes a bunch of changes, some big, some not so big. - Remove the old task threads from kern_ndis.c and reimplement them in subr_ntoskrnl.c, in order to more properly emulate the Windows DPC API. Each CPU gets its own DPC queue/thread, and each queue can have low, medium and high importance DPCs. New APIs implemented: KeSetTargetProcessorDpc(), KeSetImportanceDpc() and KeFlushQueuedDpcs(). (This is the biggest change.) - Fix a bug in NdisMInitializeTimer(): the k_dpc pointer in the nmt_timer embedded in the ndis_miniport_timer struct must be set to point to the DPC, also embedded in the struct. Failing to do this breaks dequeueing of DPCs submitted via timers, and in turn breaks cancelling timers. - Fix a bug in KeCancelTimer(): if the timer is interted in the timer queue (i.e. the timeout callback is still pending), we have to both untimeout() the timer _and_ call KeRemoveQueueDpc() to nuke the DPC that might be pending. Failing to do this breaks cancellation of periodic timers, which always appear to be inserted in the timer queue. - Make use of the nmt_nexttimer field in ndis_miniport_timer: keep a queue of pending timers and cancel them all in ndis_halt_nic(), prior to calling MiniportHalt(). Also call KeFlushQueuedDpcs() to make sure any DPCs queued by the timers have expired. - Modify NdisMAllocateSharedMemory() and NdisMFreeSharedMemory() to keep track of both the virtual and physical addresses of the shared memory buffers that get handed out. The AirGo MIMO driver appears to have a bug in it: for one of the segments is allocates, it returns the wrong virtual address. This would confuse NdisMFreeSharedMemory() and cause a crash. Why it doesn't crash Windows too I have no idea (from reading the documentation for NdisMFreeSharedMemory(), it appears to be a violation of the API). - Implement strstr(), strchr() and MmIsAddressValid(). - Implement IoAllocateWorkItem(), IoFreeWorkItem(), IoQueueWorkItem() and ExQueueWorkItem(). (This is the second biggest change.) - Make NdisScheduleWorkItem() call ExQueueWorkItem(). (Note that the ExQueueWorkItem() API is deprecated by Microsoft, but NDIS still uses it, since NdisScheduleWorkItem() is incompatible with the IoXXXWorkItem() API.) - Change if_ndis.c to use the NdisScheduleWorkItem() interface for scheduling tasks. With all these changes and fixes, the AirGo MIMO driver for the Belkin F5D8010 Pre-N card now works. Special thanks to Paul Robinson (paul dawt robinson at pwermedia dawt net) for the loan of a card for testing.
2005-05-05 03:56:09 +00:00
{ NULL, (FUNC)dummy, NULL, 0, WINDRV_WRAP_STDCALL },
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
/* End of list. */
Add support for Windows/x86-64 binaries to Project Evil. Ville-Pertti Keinonen (will at exomi dot comohmygodnospampleasekthx) deserves a big thanks for submitting initial patches to make it work. I have mangled his contributions appropriately. The main gotcha with Windows/x86-64 is that Microsoft uses a different calling convention than everyone else. The standard ABI requires using 6 registers for argument passing, with other arguments on the stack. Microsoft uses only 4 registers, and requires the caller to leave room on the stack for the register arguments incase the callee needs to spill them. Unlike x86, where Microsoft uses a mix of _cdecl, _stdcall and _fastcall, all routines on Windows/x86-64 uses the same convention. This unfortunately means that all the functions we export to the driver require an intermediate translation wrapper. Similarly, we have to wrap all calls back into the driver binary itself. The original patches provided macros to wrap every single routine at compile time, providing a secondary jump table with a customized wrapper for each exported routine. I decided to use a different approach: the call wrapper for each function is created from a template at runtime, and the routine to jump to is patched into the wrapper as it is created. The subr_pe module has been modified to patch in the wrapped function instead of the original. (On x86, the wrapping routine is a no-op.) There are some minor API differences that had to be accounted for: - KeAcquireSpinLock() is a real function on amd64, not a macro wrapper around KfAcquireSpinLock() - NdisFreeBuffer() is actually IoFreeMdl(). I had to change the whole NDIS_BUFFER API a bit to accomodate this. Bugs fixed along the way: - IoAllocateMdl() always returned NULL - kern_windrv.c:windrv_unload() wasn't releasing private driver object extensions correctly (found thanks to memguard) This has only been tested with the driver for the Broadcom 802.11g chipset, which was the only Windows/x86-64 driver I could find.
2005-02-16 05:41:18 +00:00
{ NULL, NULL, NULL }
Commit the first cut of Project Evil, also known as the NDISulator. Yes, it's what you think it is. Yes, you should run away now. This is a special compatibility module for allowing Windows NDIS miniport network drivers to be used with FreeBSD/x86. This provides _binary_ NDIS compatibility (not source): you can run NDIS driver code, but you can't build it. There are three main parts: sys/compat/ndis: the NDIS compat API, which provides binary compatibility functions for many routines in NDIS.SYS, HAL.dll and ntoskrnl.exe in Windows (these are the three modules that most NDIS miniport drivers use). The compat module also contains a small PE relocator/dynalinker which relocates the Windows .SYS image and then patches in our native routines. sys/dev/if_ndis: the if_ndis driver wrapper. This module makes use of the ndis compat API and can be compiled with a specially prepared binary image file (ndis_driver_data.h) containing the Windows .SYS image and registry key information parsed out of the accompanying .INF file. Once if_ndis.ko is built, it can be loaded and unloaded just like a native FreeBSD kenrel module. usr.sbin/ndiscvt: a special utility that converts foo.sys and foo.inf into an ndis_driver_data.h file that can be compiled into if_ndis.o. Contains an .inf file parser graciously provided by Matt Dodd (and mercilessly hacked upon by me) that strips out device ID info and registry key info from a .INF file and packages it up with a binary image array. The ndiscvt(8) utility also does some manipulation of the segments within the .sys file to make life easier for the kernel loader. (Doing the manipulation here saves the kernel code from having to move things around later, which would waste memory.) ndiscvt is only built for the i386 arch. Only files.i386 has been updated, and none of this is turned on in GENERIC. It should probably work on pc98. I have no idea about amd64 or ia64 at this point. This is still a work in progress. I estimate it's about %85 done, but I want it under CVS control so I can track subsequent changes. It has been tested with exactly three drivers: the LinkSys LNE100TX v4 driver (Lne100v4.sys), the sample Intel 82559 driver from the Windows DDK (e100bex.sys) and the Broadcom BCM43xx wireless driver (bcmwl5.sys). It still needs to have a net80211 stuff added to it. To use it, you would do something like this: # cd /sys/modules/ndis # make; make load # cd /sys/modules/if_ndis # ndiscvt -i /path/to/foo.inf -s /path/to/foo.sys -o ndis_driver_data.h # make; make load # sysctl -a | grep ndis All registry keys are mapped to sysctl nodes. Sometimes drivers refer to registry keys that aren't mentioned in foo.inf. If this happens, the NDIS API module creates sysctl nodes for these keys on the fly so you can tweak them. An example usage of the Broadcom wireless driver would be: # sysctl hw.ndis0.EnableAutoConnect=1 # sysctl hw.ndis0.SSID="MY_SSID" # sysctl hw.ndis0.NetworkType=0 (0 for bss, 1 for adhoc) # ifconfig ndis0 <my ipaddr> netmask 0xffffff00 up Things to be done: - get rid of debug messages - add in ndis80211 support - defer transmissions until after a status update with NDIS_STATUS_CONNECTED occurs - Create smarter lookaside list support - Split off if_ndis_pci.c and if_ndis_pccard.c attachments - Make sure PCMCIA support works - Fix ndiscvt to properly parse PCMCIA device IDs from INF files - write ndisapi.9 man page
2003-12-11 22:34:37 +00:00
};
Reference in New Issue Copy Permalink