d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
1a37aa566b
freebsd-dev/sys/dev/pci/pci.c

1536 lines
39 KiB
C
Raw Normal View History

Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
/*
* Copyright (c) 1997, Stefan Esser <se@freebsd.org>
* 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 unmodified, 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.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``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 THE AUTHOR 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.
*
$Id$ -> $FreeBSD$
1999-08-28 01:08:13 +00:00
* $FreeBSD$
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
*
*/
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
#include "opt_bus.h"
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
#include <sys/param.h>
#include <sys/systm.h>
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> Bug fixed, that caused system hang on first interrupt on some motherboards. New version of PCI bus configuration code, now supports dynamic interrupt configuration (using BIOS supplied values). NCR SCSI and DEC Ethernet driver patched to use this feature. *** Remove PCI IRQ specifications from your kernel config file ! ***
1994-10-12 02:33:23 +00:00
#include <sys/malloc.h>
Replace the pcidevice_set linker set based configuration mechanism for old style pci drivers with a simple one-line change to use a module that registers itself under new-bus and should in theory enable just about all of the pci drivers to be loadable (kldload and loader(8)) but without having the impact of converting the APIs yet. This also fixes the problem of having undefined variables when only new-style pci drivers are present.
1999-04-24 19:59:20 +00:00
#include <sys/module.h>
Next phase in the PCI subsystem cleanup. - Move PCI core code to dev/pci. - Split bridge code out into separate modules. - Remove the descriptive strings from the bridge drivers. If you want to know what a device is, use pciconf. Add support for broadly identifying devices based on class/subclass, and for parsing a preloaded device identification database so that if you want to waste the memory, you can identify *anything* we know about. - Remove machine-dependant code from the core PCI code. APIC interrupt mapping is performed by shadowing the intline register in machine- dependant code. - Bring interrupt routing support to the Alpha (although many platforms don't yet support routing or mapping interrupts entirely correctly). This resulted in spamming <sys/bus.h> into more places than it really should have gone. - Put sys/dev on the kernel/modules include path. This avoids having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
#include <sys/linker.h>
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
#include <sys/fcntl.h>
Add support for ioctl() accesses to PCI config space registers. Garrett Wollman sent me this code a few weeks ago for review, and I made some significant changes, which he in turn accepted ... In order to make use of these changes, a device entry has to added to /dev. Submitted by: wollman
1996-10-22 20:20:14 +00:00
#include <sys/conf.h>
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
#include <sys/kernel.h>
Revive PCIConf. Submitted by: "Kenneth D. Merry" <ken@plutotech.com>
1998-09-15 08:21:13 +00:00
#include <sys/queue.h>
#include <sys/types.h>
Completely new PCI code: 1) Supports PCI to PCI bridge devices (and tries to initialise them, even if the BIOS is brain dead). 2) Supports shared PCI interrupts. Interrupt handlers now MUST return '0' if they found nothing to do, '1' otherwise. New features tested with i486 systems based on the Intel Saturn and a DEC 4channel Ethernet card only, but expected to work on most systems. The option PCI_REMAP has been removed ! Submitted by: Wolfgang Stanglmeier <wolf@kintaro.cologne.de>
1995-03-21 23:01:06 +00:00
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
#include <vm/vm.h>
Untangled the vm.h include file spaghetti.
1995-12-07 12:48:31 +00:00
#include <vm/pmap.h>
Revive PCIConf. Submitted by: "Kenneth D. Merry" <ken@plutotech.com>
1998-09-15 08:21:13 +00:00
#include <vm/vm_extern.h>
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
#include <sys/bus.h>
#include <machine/bus.h>
#include <sys/rman.h>
#include <machine/resource.h>
Encapsulate the old PCI compatability support and APIs completely under "options COMPAT_OLDPCI". This option already existed, but now also tidies up the declarations in #include <pci/pci*.h>. It is amazing how much stuff was using the old pre-FreeBSD 3.x names and going silently undetected.
2000-05-28 16:35:57 +00:00
#include <sys/pciio.h>
Completely new PCI code: 1) Supports PCI to PCI bridge devices (and tries to initialise them, even if the BIOS is brain dead). 2) Supports shared PCI interrupts. Interrupt handlers now MUST return '0' if they found nothing to do, '1' otherwise. New features tested with i486 systems based on the Intel Saturn and a DEC 4channel Ethernet card only, but expected to work on most systems. The option PCI_REMAP has been removed ! Submitted by: Wolfgang Stanglmeier <wolf@kintaro.cologne.de>
1995-03-21 23:01:06 +00:00
#include <pci/pcireg.h>
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
#include <pci/pcivar.h>
Completely new PCI code: 1) Supports PCI to PCI bridge devices (and tries to initialise them, even if the BIOS is brain dead). 2) Supports shared PCI interrupts. Interrupt handlers now MUST return '0' if they found nothing to do, '1' otherwise. New features tested with i486 systems based on the Intel Saturn and a DEC 4channel Ethernet card only, but expected to work on most systems. The option PCI_REMAP has been removed ! Submitted by: Wolfgang Stanglmeier <wolf@kintaro.cologne.de>
1995-03-21 23:01:06 +00:00
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
#include "pcib_if.h"
Next phase in the PCI subsystem cleanup. - Move PCI core code to dev/pci. - Split bridge code out into separate modules. - Remove the descriptive strings from the bridge drivers. If you want to know what a device is, use pciconf. Add support for broadly identifying devices based on class/subclass, and for parsing a preloaded device identification database so that if you want to waste the memory, you can identify *anything* we know about. - Remove machine-dependant code from the core PCI code. APIC interrupt mapping is performed by shadowing the intline register in machine- dependant code. - Bring interrupt routing support to the Alpha (although many platforms don't yet support routing or mapping interrupts entirely correctly). This resulted in spamming <sys/bus.h> into more places than it really should have gone. - Put sys/dev on the kernel/modules include path. This avoids having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
static char *pci_vendordata;
static size_t pci_vendordata_size;
static char *pci_describe_device(device_t dev);
Modified to use renamed get_pci_apic_irq() -> pci_apic_pin() function.
1997-06-25 20:56:29 +00:00
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
static devclass_t pci_devclass;
Fix some resource allocation peculiarities of the intpm device.
1999-10-28 08:06:59 +00:00
struct pci_quirk {
u_int32_t devid; /* Vendor/device of the card */
int type;
#define PCI_QUIRK_MAP_REG 1 /* PCI map register in wierd place */
int arg1;
int arg2;
};
struct pci_quirk pci_quirks[] = {
/*
* The Intel 82371AB has a map register at offset 0x90.
*/
{ 0x71138086, PCI_QUIRK_MAP_REG, 0x90, 0 },
{ 0 }
};
* Implement bus_set/get/delete_resource for pci. * Change the hack used on the alpha for mapping devices into DENSE or BWX memory spaces to a simpler one. Its still a hack and should be a seperate api to explicitly map the resource. * Add $FreeBSD$ as necessary.
1999-10-14 21:38:33 +00:00
/* map register information */
#define PCI_MAPMEM 0x01 /* memory map */
#define PCI_MAPMEMP 0x02 /* prefetchable memory map */
#define PCI_MAPPORT 0x04 /* port map */
Back out the previous change to the queue(3) interface. It was not discussed and should probably not happen. Requested by: msmith and others
2000-05-26 02:09:24 +00:00
static STAILQ_HEAD(devlist, pci_devinfo) pci_devq;
Revive PCIConf. Submitted by: "Kenneth D. Merry" <ken@plutotech.com>
1998-09-15 08:21:13 +00:00
u_int32_t pci_numdevs = 0;
Staticize.
1999-04-11 02:50:42 +00:00
static u_int32_t pci_generation = 0;
Revive PCIConf. Submitted by: "Kenneth D. Merry" <ken@plutotech.com>
1998-09-15 08:21:13 +00:00
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
/* return base address of memory or port map */
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> Bug fixed, that caused system hang on first interrupt on some motherboards. New version of PCI bus configuration code, now supports dynamic interrupt configuration (using BIOS supplied values). NCR SCSI and DEC Ethernet driver patched to use this feature. *** Remove PCI IRQ specifications from your kernel config file ! ***
1994-10-12 02:33:23 +00:00
* Implement bus_set/get/delete_resource for pci. * Change the hack used on the alpha for mapping devices into DENSE or BWX memory spaces to a simpler one. Its still a hack and should be a seperate api to explicitly map the resource. * Add $FreeBSD$ as necessary.
1999-10-14 21:38:33 +00:00
static u_int32_t
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
pci_mapbase(unsigned mapreg)
Completely new PCI code: 1) Supports PCI to PCI bridge devices (and tries to initialise them, even if the BIOS is brain dead). 2) Supports shared PCI interrupts. Interrupt handlers now MUST return '0' if they found nothing to do, '1' otherwise. New features tested with i486 systems based on the Intel Saturn and a DEC 4channel Ethernet card only, but expected to work on most systems. The option PCI_REMAP has been removed ! Submitted by: Wolfgang Stanglmeier <wolf@kintaro.cologne.de>
1995-03-21 23:01:06 +00:00
{
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
int mask = 0x03;
if ((mapreg & 0x01) == 0)
mask = 0x0f;
return (mapreg & ~mask);
Completely new PCI code: 1) Supports PCI to PCI bridge devices (and tries to initialise them, even if the BIOS is brain dead). 2) Supports shared PCI interrupts. Interrupt handlers now MUST return '0' if they found nothing to do, '1' otherwise. New features tested with i486 systems based on the Intel Saturn and a DEC 4channel Ethernet card only, but expected to work on most systems. The option PCI_REMAP has been removed ! Submitted by: Wolfgang Stanglmeier <wolf@kintaro.cologne.de>
1995-03-21 23:01:06 +00:00
}
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
/* return map type of memory or port map */
Update PCI bus code from my current sources: - always use pci_conf_read() and pci_conf_write(). (This is required to simulate non-existant devices in my system for PCI bridge code tests.) - reorder some functions (put the main functions at the end). - correct off by one bug in the code dealing with unitialized PCI to PCI bridge chips. (Bug found by ASAMI Satoshi.) - print function number for multi-function devices.
1996-04-14 20:14:36 +00:00
static int
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
pci_maptype(unsigned mapreg)
{
static u_int8_t maptype[0x10] = {
PCI_MAPMEM, PCI_MAPPORT,
PCI_MAPMEM, 0,
PCI_MAPMEM, PCI_MAPPORT,
0, 0,
PCI_MAPMEM|PCI_MAPMEMP, PCI_MAPPORT,
PCI_MAPMEM|PCI_MAPMEMP, 0,
PCI_MAPMEM|PCI_MAPMEMP, PCI_MAPPORT,
0, 0,
Add PCI LKM support.
1997-01-21 23:23:40 +00:00
};
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
return maptype[mapreg & 0x0f];
Add PCI LKM support.
1997-01-21 23:23:40 +00:00
}
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
/* return log2 of map size decoded for memory or port map */
Completely new PCI code: 1) Supports PCI to PCI bridge devices (and tries to initialise them, even if the BIOS is brain dead). 2) Supports shared PCI interrupts. Interrupt handlers now MUST return '0' if they found nothing to do, '1' otherwise. New features tested with i486 systems based on the Intel Saturn and a DEC 4channel Ethernet card only, but expected to work on most systems. The option PCI_REMAP has been removed ! Submitted by: Wolfgang Stanglmeier <wolf@kintaro.cologne.de>
1995-03-21 23:01:06 +00:00
Add heuristic to detect multi-function devices that don't announce this feature in the header type register, though it is required by the PCI spec. This should correctly probe both functions of the Intel 82371FB chip, without the need for a special case based on the device ID.
1996-01-30 01:14:29 +00:00
static int
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
pci_mapsize(unsigned testval)
Add heuristic to detect multi-function devices that don't announce this feature in the header type register, though it is required by the PCI spec. This should correctly probe both functions of the Intel 82371FB chip, without the need for a special case based on the device ID.
1996-01-30 01:14:29 +00:00
{
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
int ln2size;
Add heuristic to detect multi-function devices that don't announce this feature in the header type register, though it is required by the PCI spec. This should correctly probe both functions of the Intel 82371FB chip, without the need for a special case based on the device ID.
1996-01-30 01:14:29 +00:00
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
testval = pci_mapbase(testval);
Two minor changes to the code that builds the pci map array: 1) Stop at the first map register that contains a zero value. 2) When testing for the map size work up from low values, since this works around a bug in some BusLogic SCSI card, which has the 16 upper port base address bits hardwired to zero. The config register dump printed in the bootverbose case has been slightly rearranged.
1997-05-28 10:01:03 +00:00
ln2size = 0;
if (testval != 0) {
while ((testval & 1) == 0)
{
ln2size++;
testval >>= 1;
}
Make mapping messages depend on bootverbose flag. Add PCI subclass to unknown device message.
1995-09-07 15:20:53 +00:00
}
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
return (ln2size);
Completely new PCI code: 1) Supports PCI to PCI bridge devices (and tries to initialise them, even if the BIOS is brain dead). 2) Supports shared PCI interrupts. Interrupt handlers now MUST return '0' if they found nothing to do, '1' otherwise. New features tested with i486 systems based on the Intel Saturn and a DEC 4channel Ethernet card only, but expected to work on most systems. The option PCI_REMAP has been removed ! Submitted by: Wolfgang Stanglmeier <wolf@kintaro.cologne.de>
1995-03-21 23:01:06 +00:00
}
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
/* return log2 of address range supported by map register */
Completely new PCI code: 1) Supports PCI to PCI bridge devices (and tries to initialise them, even if the BIOS is brain dead). 2) Supports shared PCI interrupts. Interrupt handlers now MUST return '0' if they found nothing to do, '1' otherwise. New features tested with i486 systems based on the Intel Saturn and a DEC 4channel Ethernet card only, but expected to work on most systems. The option PCI_REMAP has been removed ! Submitted by: Wolfgang Stanglmeier <wolf@kintaro.cologne.de>
1995-03-21 23:01:06 +00:00
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
static int
pci_maprange(unsigned mapreg)
{
int ln2range = 0;
switch (mapreg & 0x07) {
case 0x00:
case 0x01:
case 0x05:
ln2range = 32;
break;
case 0x02:
ln2range = 20;
break;
case 0x04:
ln2range = 64;
break;
Update PCI bus code from my current sources: - always use pci_conf_read() and pci_conf_write(). (This is required to simulate non-existant devices in my system for PCI bridge code tests.) - reorder some functions (put the main functions at the end). - correct off by one bug in the code dealing with unitialized PCI to PCI bridge chips. (Bug found by ASAMI Satoshi.) - print function number for multi-function devices.
1996-04-14 20:14:36 +00:00
}
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
return (ln2range);
Reapplied all of Stefan's changes. What a mess - the files were modified and moved at the same time. This made it *very* difficult to fix the revision log lossage that happend when the files were moved. SIGH.
1995-02-02 13:12:18 +00:00
}
Completely new PCI code: 1) Supports PCI to PCI bridge devices (and tries to initialise them, even if the BIOS is brain dead). 2) Supports shared PCI interrupts. Interrupt handlers now MUST return '0' if they found nothing to do, '1' otherwise. New features tested with i486 systems based on the Intel Saturn and a DEC 4channel Ethernet card only, but expected to work on most systems. The option PCI_REMAP has been removed ! Submitted by: Wolfgang Stanglmeier <wolf@kintaro.cologne.de>
1995-03-21 23:01:06 +00:00
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
/* adjust some values from PCI 1.0 devices to match 2.0 standards ... */
Add PCI LKM support.
1997-01-21 23:23:40 +00:00
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
static void
pci_fixancient(pcicfgregs *cfg)
Add PCI LKM support.
1997-01-21 23:23:40 +00:00
{
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
if (cfg->hdrtype != 0)
return;
Add PCI LKM support.
1997-01-21 23:23:40 +00:00
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
/* PCI to PCI bridges use header type 1 */
Use "baseclass" instead of "class" for storing the contents of PCI register 0xB so that C++ programs can use the PCI conf interface.
1998-08-13 19:12:20 +00:00
if (cfg->baseclass == PCIC_BRIDGE && cfg->subclass == PCIS_BRIDGE_PCI)
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
cfg->hdrtype = 1;
Add PCI LKM support.
1997-01-21 23:23:40 +00:00
}
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
/* read config data specific to header type 2 device (PCI to CardBus bridge) */
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
static void *
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
pci_readpcb(device_t pcib, int b, int s, int f)
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
{
pcih2cfgregs *p;
New PCI attach code: PCI BIOS mappings are retained, except if option PCI_REMAP is specified in the kernel config file. There is now a list of attach addresses, and the first address that seems to make some device registers appear is chosen. Reviewed by: se Submitted by: wolf
1995-02-22 14:17:15 +00:00
Next phase in the PCI subsystem cleanup. - Move PCI core code to dev/pci. - Split bridge code out into separate modules. - Remove the descriptive strings from the bridge drivers. If you want to know what a device is, use pciconf. Add support for broadly identifying devices based on class/subclass, and for parsing a preloaded device identification database so that if you want to waste the memory, you can identify *anything* we know about. - Remove machine-dependant code from the core PCI code. APIC interrupt mapping is performed by shadowing the intline register in machine- dependant code. - Bring interrupt routing support to the Alpha (although many platforms don't yet support routing or mapping interrupts entirely correctly). This resulted in spamming <sys/bus.h> into more places than it really should have gone. - Put sys/dev on the kernel/modules include path. This avoids having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
p = malloc(sizeof (pcih2cfgregs), M_DEVBUF, M_WAITOK | M_ZERO);
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
if (p == NULL)
return (NULL);
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + <se> Improved bus probing, symbolic names for registers.. Chip set parameters get dumped for intel PCI chip sets (82424+82434 only, for now).
1994-09-16 00:33:29 +00:00
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
p->secstat = PCIB_READ_CONFIG(pcib, b, s, f, PCIR_SECSTAT_2, 2);
p->bridgectl = PCIB_READ_CONFIG(pcib, b, s, f, PCIR_BRIDGECTL_2, 2);
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
p->seclat = PCIB_READ_CONFIG(pcib, b, s, f, PCIR_SECLAT_2, 1);
Completely new PCI code: 1) Supports PCI to PCI bridge devices (and tries to initialise them, even if the BIOS is brain dead). 2) Supports shared PCI interrupts. Interrupt handlers now MUST return '0' if they found nothing to do, '1' otherwise. New features tested with i486 systems based on the Intel Saturn and a DEC 4channel Ethernet card only, but expected to work on most systems. The option PCI_REMAP has been removed ! Submitted by: Wolfgang Stanglmeier <wolf@kintaro.cologne.de>
1995-03-21 23:01:06 +00:00
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
p->membase0 = PCIB_READ_CONFIG(pcib, b, s, f, PCIR_MEMBASE0_2, 4);
p->memlimit0 = PCIB_READ_CONFIG(pcib, b, s, f, PCIR_MEMLIMIT0_2, 4);
p->membase1 = PCIB_READ_CONFIG(pcib, b, s, f, PCIR_MEMBASE1_2, 4);
p->memlimit1 = PCIB_READ_CONFIG(pcib, b, s, f, PCIR_MEMLIMIT1_2, 4);
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
p->iobase0 = PCIB_READ_CONFIG(pcib, b, s, f, PCIR_IOBASE0_2, 4);
p->iolimit0 = PCIB_READ_CONFIG(pcib, b, s, f, PCIR_IOLIMIT0_2, 4);
p->iobase1 = PCIB_READ_CONFIG(pcib, b, s, f, PCIR_IOBASE1_2, 4);
p->iolimit1 = PCIB_READ_CONFIG(pcib, b, s, f, PCIR_IOLIMIT1_2, 4);
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + <se> Improved bus probing, symbolic names for registers.. Chip set parameters get dumped for intel PCI chip sets (82424+82434 only, for now).
1994-09-16 00:33:29 +00:00
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
p->pccardif = PCIB_READ_CONFIG(pcib, b, s, f, PCIR_PCCARDIF_2, 4);
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
return p;
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
}
First try to add support for PCI-PCI bridge chips (written for the DEC 21050 chip in particular, don't have specs of other such chips). This should add support for Multiple-Ethernet PCI cards (e.g. Znyx 314). Reviewed by: se Submitted by: <wolf@kintaro.cologne.de> Wolfgang Stanglmeier
1995-02-27 17:17:14 +00:00
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
/* extract header type specific config data */
Completely new PCI code: 1) Supports PCI to PCI bridge devices (and tries to initialise them, even if the BIOS is brain dead). 2) Supports shared PCI interrupts. Interrupt handlers now MUST return '0' if they found nothing to do, '1' otherwise. New features tested with i486 systems based on the Intel Saturn and a DEC 4channel Ethernet card only, but expected to work on most systems. The option PCI_REMAP has been removed ! Submitted by: Wolfgang Stanglmeier <wolf@kintaro.cologne.de>
1995-03-21 23:01:06 +00:00
static void
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
pci_hdrtypedata(device_t pcib, int b, int s, int f, pcicfgregs *cfg)
Completely new PCI code: 1) Supports PCI to PCI bridge devices (and tries to initialise them, even if the BIOS is brain dead). 2) Supports shared PCI interrupts. Interrupt handlers now MUST return '0' if they found nothing to do, '1' otherwise. New features tested with i486 systems based on the Intel Saturn and a DEC 4channel Ethernet card only, but expected to work on most systems. The option PCI_REMAP has been removed ! Submitted by: Wolfgang Stanglmeier <wolf@kintaro.cologne.de>
1995-03-21 23:01:06 +00:00
{
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
#define REG(n, w) PCIB_READ_CONFIG(pcib, b, s, f, n, w)
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
switch (cfg->hdrtype) {
case 0:
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
cfg->subvendor = REG(PCIR_SUBVEND_0, 2);
cfg->subdevice = REG(PCIR_SUBDEV_0, 2);
* Implement bus_set/get/delete_resource for pci. * Change the hack used on the alpha for mapping devices into DENSE or BWX memory spaces to a simpler one. Its still a hack and should be a seperate api to explicitly map the resource. * Add $FreeBSD$ as necessary.
1999-10-14 21:38:33 +00:00
cfg->nummaps = PCI_MAXMAPS_0;
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
break;
case 1:
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
cfg->subvendor = REG(PCIR_SUBVEND_1, 2);
cfg->subdevice = REG(PCIR_SUBDEV_1, 2);
* Implement bus_set/get/delete_resource for pci. * Change the hack used on the alpha for mapping devices into DENSE or BWX memory spaces to a simpler one. Its still a hack and should be a seperate api to explicitly map the resource. * Add $FreeBSD$ as necessary.
1999-10-14 21:38:33 +00:00
cfg->nummaps = PCI_MAXMAPS_1;
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
break;
case 2:
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
cfg->subvendor = REG(PCIR_SUBVEND_2, 2);
cfg->subdevice = REG(PCIR_SUBDEV_2, 2);
cfg->secondarybus = REG(PCIR_SECBUS_2, 1);
cfg->subordinatebus = REG(PCIR_SUBBUS_2, 1);
* Implement bus_set/get/delete_resource for pci. * Change the hack used on the alpha for mapping devices into DENSE or BWX memory spaces to a simpler one. Its still a hack and should be a seperate api to explicitly map the resource. * Add $FreeBSD$ as necessary.
1999-10-14 21:38:33 +00:00
cfg->nummaps = PCI_MAXMAPS_2;
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
cfg->hdrspec = pci_readpcb(pcib, b, s, f);
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
break;
Make mapping messages depend on bootverbose flag. Add PCI subclass to unknown device message.
1995-09-07 15:20:53 +00:00
}
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
#undef REG
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
}
Completely new PCI code: 1) Supports PCI to PCI bridge devices (and tries to initialise them, even if the BIOS is brain dead). 2) Supports shared PCI interrupts. Interrupt handlers now MUST return '0' if they found nothing to do, '1' otherwise. New features tested with i486 systems based on the Intel Saturn and a DEC 4channel Ethernet card only, but expected to work on most systems. The option PCI_REMAP has been removed ! Submitted by: Wolfgang Stanglmeier <wolf@kintaro.cologne.de>
1995-03-21 23:01:06 +00:00
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
/* read configuration header into pcicfgrect structure */
Revive PCIConf. Submitted by: "Kenneth D. Merry" <ken@plutotech.com>
1998-09-15 08:21:13 +00:00
static struct pci_devinfo *
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
pci_read_device(device_t pcib, int b, int s, int f)
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
{
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
#define REG(n, w) PCIB_READ_CONFIG(pcib, b, s, f, n, w)
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
pcicfgregs *cfg = NULL;
Revive PCIConf. Submitted by: "Kenneth D. Merry" <ken@plutotech.com>
1998-09-15 08:21:13 +00:00
struct pci_devinfo *devlist_entry;
struct devlist *devlist_head;
devlist_head = &pci_devq;
devlist_entry = NULL;
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
if (PCIB_READ_CONFIG(pcib, b, s, f, PCIR_DEVVENDOR, 4) != -1) {
Revive PCIConf. Submitted by: "Kenneth D. Merry" <ken@plutotech.com>
1998-09-15 08:21:13 +00:00
devlist_entry = malloc(sizeof(struct pci_devinfo),
Next phase in the PCI subsystem cleanup. - Move PCI core code to dev/pci. - Split bridge code out into separate modules. - Remove the descriptive strings from the bridge drivers. If you want to know what a device is, use pciconf. Add support for broadly identifying devices based on class/subclass, and for parsing a preloaded device identification database so that if you want to waste the memory, you can identify *anything* we know about. - Remove machine-dependant code from the core PCI code. APIC interrupt mapping is performed by shadowing the intline register in machine- dependant code. - Bring interrupt routing support to the Alpha (although many platforms don't yet support routing or mapping interrupts entirely correctly). This resulted in spamming <sys/bus.h> into more places than it really should have gone. - Put sys/dev on the kernel/modules include path. This avoids having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
M_DEVBUF, M_WAITOK | M_ZERO);
Revive PCIConf. Submitted by: "Kenneth D. Merry" <ken@plutotech.com>
1998-09-15 08:21:13 +00:00
if (devlist_entry == NULL)
return (NULL);
cfg = &devlist_entry->cfg;
Add support for multiple PCI "hoses" used on various alpha platforms. The specific intent of this commit is to pave the way for importing Compaq XP1000 support. These changes should not affect the i386 port. Reviewed by: Doug Rabson <dfr@nlsystems.com> (actually, he walked me through most of it & deserves more than reviewd-by credit )
1999-05-20 15:33:33 +00:00
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
cfg->bus = b;
cfg->slot = s;
cfg->func = f;
cfg->vendor = REG(PCIR_VENDOR, 2);
cfg->device = REG(PCIR_DEVICE, 2);
cfg->cmdreg = REG(PCIR_COMMAND, 2);
cfg->statreg = REG(PCIR_STATUS, 2);
cfg->baseclass = REG(PCIR_CLASS, 1);
cfg->subclass = REG(PCIR_SUBCLASS, 1);
cfg->progif = REG(PCIR_PROGIF, 1);
cfg->revid = REG(PCIR_REVID, 1);
cfg->hdrtype = REG(PCIR_HEADERTYPE, 1);
cfg->cachelnsz = REG(PCIR_CACHELNSZ, 1);
cfg->lattimer = REG(PCIR_LATTIMER, 1);
cfg->intpin = REG(PCIR_INTPIN, 1);
cfg->intline = REG(PCIR_INTLINE, 1);
Add support for APIC_IO to pci IRQ configuration. The support for APIC_IO was lost in the new set of pci modules. This patch restores the ability to build SMP/APIC_IO kernels.
1997-05-27 04:09:01 +00:00
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
cfg->mingnt = REG(PCIR_MINGNT, 1);
cfg->maxlat = REG(PCIR_MAXLAT, 1);
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
cfg->mfdev = (cfg->hdrtype & PCIM_MFDEV) != 0;
cfg->hdrtype &= ~PCIM_MFDEV;
pci_fixancient(cfg);
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
pci_hdrtypedata(pcib, b, s, f, cfg);
Revive PCIConf. Submitted by: "Kenneth D. Merry" <ken@plutotech.com>
1998-09-15 08:21:13 +00:00
STAILQ_INSERT_TAIL(devlist_head, devlist_entry, pci_links);
devlist_entry->conf.pc_sel.pc_bus = cfg->bus;
devlist_entry->conf.pc_sel.pc_dev = cfg->slot;
devlist_entry->conf.pc_sel.pc_func = cfg->func;
devlist_entry->conf.pc_hdr = cfg->hdrtype;
devlist_entry->conf.pc_subvendor = cfg->subvendor;
devlist_entry->conf.pc_subdevice = cfg->subdevice;
devlist_entry->conf.pc_vendor = cfg->vendor;
devlist_entry->conf.pc_device = cfg->device;
devlist_entry->conf.pc_class = cfg->baseclass;
devlist_entry->conf.pc_subclass = cfg->subclass;
devlist_entry->conf.pc_progif = cfg->progif;
devlist_entry->conf.pc_revid = cfg->revid;
pci_numdevs++;
pci_generation++;
Make mapping messages depend on bootverbose flag. Add PCI subclass to unknown device message.
1995-09-07 15:20:53 +00:00
}
Revive PCIConf. Submitted by: "Kenneth D. Merry" <ken@plutotech.com>
1998-09-15 08:21:13 +00:00
return (devlist_entry);
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
#undef REG
Completely new PCI code: 1) Supports PCI to PCI bridge devices (and tries to initialise them, even if the BIOS is brain dead). 2) Supports shared PCI interrupts. Interrupt handlers now MUST return '0' if they found nothing to do, '1' otherwise. New features tested with i486 systems based on the Intel Saturn and a DEC 4channel Ethernet card only, but expected to work on most systems. The option PCI_REMAP has been removed ! Submitted by: Wolfgang Stanglmeier <wolf@kintaro.cologne.de>
1995-03-21 23:01:06 +00:00
}
Remove a bunch of variables which were unused both in GENERIC and LINT. Found by: -Wunused
1997-11-07 08:53:44 +00:00
#if 0
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
/* free pcicfgregs structure and all depending data structures */
Completely new PCI code: 1) Supports PCI to PCI bridge devices (and tries to initialise them, even if the BIOS is brain dead). 2) Supports shared PCI interrupts. Interrupt handlers now MUST return '0' if they found nothing to do, '1' otherwise. New features tested with i486 systems based on the Intel Saturn and a DEC 4channel Ethernet card only, but expected to work on most systems. The option PCI_REMAP has been removed ! Submitted by: Wolfgang Stanglmeier <wolf@kintaro.cologne.de>
1995-03-21 23:01:06 +00:00
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
static int
Revive PCIConf. Submitted by: "Kenneth D. Merry" <ken@plutotech.com>
1998-09-15 08:21:13 +00:00
pci_freecfg(struct pci_devinfo *dinfo)
Completely new PCI code: 1) Supports PCI to PCI bridge devices (and tries to initialise them, even if the BIOS is brain dead). 2) Supports shared PCI interrupts. Interrupt handlers now MUST return '0' if they found nothing to do, '1' otherwise. New features tested with i486 systems based on the Intel Saturn and a DEC 4channel Ethernet card only, but expected to work on most systems. The option PCI_REMAP has been removed ! Submitted by: Wolfgang Stanglmeier <wolf@kintaro.cologne.de>
1995-03-21 23:01:06 +00:00
{
Revive PCIConf. Submitted by: "Kenneth D. Merry" <ken@plutotech.com>
1998-09-15 08:21:13 +00:00
struct devlist *devlist_head;
devlist_head = &pci_devq;
if (dinfo->cfg.hdrspec != NULL)
free(dinfo->cfg.hdrspec, M_DEVBUF);
if (dinfo->cfg.map != NULL)
free(dinfo->cfg.map, M_DEVBUF);
/* XXX this hasn't been tested */
Back out the previous change to the queue(3) interface. It was not discussed and should probably not happen. Requested by: msmith and others
2000-05-26 02:09:24 +00:00
STAILQ_REMOVE(devlist_head, dinfo, pci_devinfo, pci_links);
Revive PCIConf. Submitted by: "Kenneth D. Merry" <ken@plutotech.com>
1998-09-15 08:21:13 +00:00
free(dinfo, M_DEVBUF);
/* increment the generation count */
pci_generation++;
/* we're losing one device */
pci_numdevs--;
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
return (0);
Completely new PCI code: 1) Supports PCI to PCI bridge devices (and tries to initialise them, even if the BIOS is brain dead). 2) Supports shared PCI interrupts. Interrupt handlers now MUST return '0' if they found nothing to do, '1' otherwise. New features tested with i486 systems based on the Intel Saturn and a DEC 4channel Ethernet card only, but expected to work on most systems. The option PCI_REMAP has been removed ! Submitted by: Wolfgang Stanglmeier <wolf@kintaro.cologne.de>
1995-03-21 23:01:06 +00:00
}
Remove a bunch of variables which were unused both in GENERIC and LINT. Found by: -Wunused
1997-11-07 08:53:44 +00:00
#endif
Completely new PCI code: 1) Supports PCI to PCI bridge devices (and tries to initialise them, even if the BIOS is brain dead). 2) Supports shared PCI interrupts. Interrupt handlers now MUST return '0' if they found nothing to do, '1' otherwise. New features tested with i486 systems based on the Intel Saturn and a DEC 4channel Ethernet card only, but expected to work on most systems. The option PCI_REMAP has been removed ! Submitted by: Wolfgang Stanglmeier <wolf@kintaro.cologne.de>
1995-03-21 23:01:06 +00:00
Add support for multiple PCI "hoses" used on various alpha platforms. The specific intent of this commit is to pave the way for importing Compaq XP1000 support. These changes should not affect the i386 port. Reviewed by: Doug Rabson <dfr@nlsystems.com> (actually, he walked me through most of it & deserves more than reviewd-by credit )
1999-05-20 15:33:33 +00:00
Add support for ioctl() accesses to PCI config space registers. Garrett Wollman sent me this code a few weeks ago for review, and I made some significant changes, which he in turn accepted ... In order to make use of these changes, a device entry has to added to /dev. Submitted by: wollman
1996-10-22 20:20:14 +00:00
/*
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
* This is the user interface to PCI configuration space.
Add support for ioctl() accesses to PCI config space registers. Garrett Wollman sent me this code a few weeks ago for review, and I made some significant changes, which he in turn accepted ... In order to make use of these changes, a device entry has to added to /dev. Submitted by: wollman
1996-10-22 20:20:14 +00:00
*/
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
Add support for ioctl() accesses to PCI config space registers. Garrett Wollman sent me this code a few weeks ago for review, and I made some significant changes, which he in turn accepted ... In order to make use of these changes, a device entry has to added to /dev. Submitted by: wollman
1996-10-22 20:20:14 +00:00
static int
pci_open(dev_t dev, int oflags, int devtype, struct proc *p)
{
if ((oflags & FWRITE) && securelevel > 0) {
return EPERM;
}
return 0;
}
static int
pci_close(dev_t dev, int flag, int devtype, struct proc *p)
{
return 0;
}
Revive PCIConf. Submitted by: "Kenneth D. Merry" <ken@plutotech.com>
1998-09-15 08:21:13 +00:00
/*
* Match a single pci_conf structure against an array of pci_match_conf
* structures. The first argument, 'matches', is an array of num_matches
* pci_match_conf structures. match_buf is a pointer to the pci_conf
* structure that will be compared to every entry in the matches array.
* This function returns 1 on failure, 0 on success.
*/
static int
pci_conf_match(struct pci_match_conf *matches, int num_matches,
struct pci_conf *match_buf)
{
int i;
if ((matches == NULL) || (match_buf == NULL) || (num_matches <= 0))
return(1);
for (i = 0; i < num_matches; i++) {
/*
* I'm not sure why someone would do this...but...
*/
if (matches[i].flags == PCI_GETCONF_NO_MATCH)
continue;
/*
* Look at each of the match flags. If it's set, do the
* comparison. If the comparison fails, we don't have a
* match, go on to the next item if there is one.
*/
if (((matches[i].flags & PCI_GETCONF_MATCH_BUS) != 0)
&& (match_buf->pc_sel.pc_bus != matches[i].pc_sel.pc_bus))
continue;
if (((matches[i].flags & PCI_GETCONF_MATCH_DEV) != 0)
&& (match_buf->pc_sel.pc_dev != matches[i].pc_sel.pc_dev))
continue;
if (((matches[i].flags & PCI_GETCONF_MATCH_FUNC) != 0)
&& (match_buf->pc_sel.pc_func != matches[i].pc_sel.pc_func))
continue;
if (((matches[i].flags & PCI_GETCONF_MATCH_VENDOR) != 0)
&& (match_buf->pc_vendor != matches[i].pc_vendor))
continue;
if (((matches[i].flags & PCI_GETCONF_MATCH_DEVICE) != 0)
&& (match_buf->pc_device != matches[i].pc_device))
continue;
if (((matches[i].flags & PCI_GETCONF_MATCH_CLASS) != 0)
&& (match_buf->pc_class != matches[i].pc_class))
continue;
if (((matches[i].flags & PCI_GETCONF_MATCH_UNIT) != 0)
&& (match_buf->pd_unit != matches[i].pd_unit))
continue;
if (((matches[i].flags & PCI_GETCONF_MATCH_NAME) != 0)
&& (strncmp(matches[i].pd_name, match_buf->pd_name,
sizeof(match_buf->pd_name)) != 0))
continue;
return(0);
}
return(1);
}
On the new Meteor cards, the Philips SAA 7116 is connected to the PCI bus via an IBM PCI-PCI bridge (82351 or 82352 or 82353) The driver must identify if it is on a secondary PCI bus, which is created via the IBM PCI-PCI bridge. If it is, then it must initialise the IBM PCI-PCI bridge correctly. To do this, the following new functions are added. Because they use the pcici_t tag, they are considered 2.2 compatibility APIs pcici_t * pci_get_parent_from_tag(pcici_t tag); int pci_get_bus_from_tag(pcici_t tag); (The _from_tag suffix is used to prevent clashes with similarly named newbus PCI API functions) Submitted by: Anton Berezin <tobez@plab.ku.dk> Reviewed by: Doug Rabson <dfr@nlsystems.com> Reworked by: Me (roger)
1999-05-31 22:13:37 +00:00
/*
* Locate the parent of a PCI device by scanning the PCI devlist
* and return the entry for the parent.
* For devices on PCI Bus 0 (the host bus), this is the PCI Host.
* For devices on secondary PCI busses, this is that bus' PCI-PCI Bridge.
*/
pcicfgregs *
pci_devlist_get_parent(pcicfgregs *cfg)
{
struct devlist *devlist_head;
struct pci_devinfo *dinfo;
pcicfgregs *bridge_cfg;
int i;
dinfo = STAILQ_FIRST(devlist_head = &pci_devq);
/* If the device is on PCI bus 0, look for the host */
if (cfg->bus == 0) {
for (i = 0; (dinfo != NULL) && (i < pci_numdevs);
dinfo = STAILQ_NEXT(dinfo, pci_links), i++) {
bridge_cfg = &dinfo->cfg;
if (bridge_cfg->baseclass == PCIC_BRIDGE
&& bridge_cfg->subclass == PCIS_BRIDGE_HOST
&& bridge_cfg->bus == cfg->bus) {
return bridge_cfg;
}
}
}
/* If the device is not on PCI bus 0, look for the PCI-PCI bridge */
if (cfg->bus > 0) {
for (i = 0; (dinfo != NULL) && (i < pci_numdevs);
dinfo = STAILQ_NEXT(dinfo, pci_links), i++) {
bridge_cfg = &dinfo->cfg;
if (bridge_cfg->baseclass == PCIC_BRIDGE
&& bridge_cfg->subclass == PCIS_BRIDGE_PCI
&& bridge_cfg->secondarybus == cfg->bus) {
return bridge_cfg;
}
}
}
return NULL;
}
Add support for ioctl() accesses to PCI config space registers. Garrett Wollman sent me this code a few weeks ago for review, and I made some significant changes, which he in turn accepted ... In order to make use of these changes, a device entry has to added to /dev. Submitted by: wollman
1996-10-22 20:20:14 +00:00
static int
This commit fixes various 64bit portability problems required for FreeBSD/alpha. The most significant item is to change the command argument to ioctl functions from int to u_long. This change brings us inline with various other BSD versions. Driver writers may like to use (__FreeBSD_version == 300003) to detect this change. The prototype FreeBSD/alpha machdep will follow in a couple of days time.
1998-06-07 17:13:14 +00:00
pci_ioctl(dev_t dev, u_long cmd, caddr_t data, int flag, struct proc *p)
Add support for ioctl() accesses to PCI config space registers. Garrett Wollman sent me this code a few weeks ago for review, and I made some significant changes, which he in turn accepted ... In order to make use of these changes, a device entry has to added to /dev. Submitted by: wollman
1996-10-22 20:20:14 +00:00
{
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
device_t pci, pcib;
Add support for ioctl() accesses to PCI config space registers. Garrett Wollman sent me this code a few weeks ago for review, and I made some significant changes, which he in turn accepted ... In order to make use of these changes, a device entry has to added to /dev. Submitted by: wollman
1996-10-22 20:20:14 +00:00
struct pci_io *io;
For the ioctl that reads the pci configuration, look up the name and unit on the fly so that we can see the driver assignment of new pci devices as well in the 'pciconf -l' display.
1999-05-09 20:27:26 +00:00
const char *name;
Add support for ioctl() accesses to PCI config space registers. Garrett Wollman sent me this code a few weeks ago for review, and I made some significant changes, which he in turn accepted ... In order to make use of these changes, a device entry has to added to /dev. Submitted by: wollman
1996-10-22 20:20:14 +00:00
int error;
Revive PCIConf. Submitted by: "Kenneth D. Merry" <ken@plutotech.com>
1998-09-15 08:21:13 +00:00
if (!(flag & FWRITE))
Add support for ioctl() accesses to PCI config space registers. Garrett Wollman sent me this code a few weeks ago for review, and I made some significant changes, which he in turn accepted ... In order to make use of these changes, a device entry has to added to /dev. Submitted by: wollman
1996-10-22 20:20:14 +00:00
return EPERM;
Revive PCIConf. Submitted by: "Kenneth D. Merry" <ken@plutotech.com>
1998-09-15 08:21:13 +00:00
Add support for ioctl() accesses to PCI config space registers. Garrett Wollman sent me this code a few weeks ago for review, and I made some significant changes, which he in turn accepted ... In order to make use of these changes, a device entry has to added to /dev. Submitted by: wollman
1996-10-22 20:20:14 +00:00
switch(cmd) {
case PCIOCGETCONF:
Revive PCIConf. Submitted by: "Kenneth D. Merry" <ken@plutotech.com>
1998-09-15 08:21:13 +00:00
{
struct pci_devinfo *dinfo;
struct pci_conf_io *cio;
struct devlist *devlist_head;
struct pci_match_conf *pattern_buf;
int num_patterns;
size_t iolen;
int ionum, i;
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
Add support for ioctl() accesses to PCI config space registers. Garrett Wollman sent me this code a few weeks ago for review, and I made some significant changes, which he in turn accepted ... In order to make use of these changes, a device entry has to added to /dev. Submitted by: wollman
1996-10-22 20:20:14 +00:00
cio = (struct pci_conf_io *)data;
Revive PCIConf. Submitted by: "Kenneth D. Merry" <ken@plutotech.com>
1998-09-15 08:21:13 +00:00
num_patterns = 0;
dinfo = NULL;
/*
* Hopefully the user won't pass in a null pointer, but it
* can't hurt to check.
*/
if (cio == NULL) {
error = EINVAL;
break;
}
/*
* If the user specified an offset into the device list,
* but the list has changed since they last called this
* ioctl, tell them that the list has changed. They will
* have to get the list from the beginning.
*/
if ((cio->offset != 0)
&& (cio->generation != pci_generation)){
cio->num_matches = 0;
cio->status = PCI_GETCONF_LIST_CHANGED;
error = 0;
break;
}
/*
* Check to see whether the user has asked for an offset
* past the end of our list.
*/
if (cio->offset >= pci_numdevs) {
cio->num_matches = 0;
cio->status = PCI_GETCONF_LAST_DEVICE;
error = 0;
break;
}
/* get the head of the device queue */
devlist_head = &pci_devq;
/*
* Determine how much room we have for pci_conf structures.
* Round the user's buffer size down to the nearest
* multiple of sizeof(struct pci_conf) in case the user
* didn't specify a multiple of that size.
*/
iolen = min(cio->match_buf_len -
(cio->match_buf_len % sizeof(struct pci_conf)),
pci_numdevs * sizeof(struct pci_conf));
/*
* Since we know that iolen is a multiple of the size of
* the pciconf union, it's okay to do this.
*/
ionum = iolen / sizeof(struct pci_conf);
/*
* If this test is true, the user wants the pci_conf
* structures returned to match the supplied entries.
*/
if ((cio->num_patterns > 0)
&& (cio->pat_buf_len > 0)) {
/*
* pat_buf_len needs to be:
* num_patterns * sizeof(struct pci_match_conf)
* While it is certainly possible the user just
* allocated a large buffer, but set the number of
* matches correctly, it is far more likely that
* their kernel doesn't match the userland utility
* they're using. It's also possible that the user
* forgot to initialize some variables. Yes, this
* may be overly picky, but I hazard to guess that
* it's far more likely to just catch folks that
* updated their kernel but not their userland.
*/
if ((cio->num_patterns *
sizeof(struct pci_match_conf)) != cio->pat_buf_len){
/* The user made a mistake, return an error*/
cio->status = PCI_GETCONF_ERROR;
printf("pci_ioctl: pat_buf_len %d != "
"num_patterns (%d) * sizeof(struct "
"pci_match_conf) (%d)\npci_ioctl: "
"pat_buf_len should be = %d\n",
cio->pat_buf_len, cio->num_patterns,
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
(int)sizeof(struct pci_match_conf),
(int)sizeof(struct pci_match_conf) *
Revive PCIConf. Submitted by: "Kenneth D. Merry" <ken@plutotech.com>
1998-09-15 08:21:13 +00:00
cio->num_patterns);
printf("pci_ioctl: do your headers match your "
"kernel?\n");
cio->num_matches = 0;
error = EINVAL;
break;
}
/*
* Check the user's buffer to make sure it's readable.
*/
Don't test boolean return against != 1. Don't needlessly assign the error variable in an if statement.
1999-10-29 19:03:18 +00:00
if (!useracc((caddr_t)cio->patterns,
Change useracc() and kernacc() to use VM_PROT_{READ|WRITE|EXECUTE} for the "rw" argument, rather than hijacking B_{READ|WRITE}. Fix two bugs (physio & cam) resulting by the confusion caused by this. Submitted by: Tor.Egge@fast.no Reviewed by: alc, ken (partly)
1999-10-30 06:32:05 +00:00
cio->pat_buf_len, VM_PROT_READ)) {
s/%#p/%p/ to fix a warning. This looks like a leftover of once being %#x, which causes a leading 0x to be prepended. %p has this by default and the '#' is ignored by the kernel prinf code for %p.
1998-11-09 08:08:06 +00:00
printf("pci_ioctl: pattern buffer %p, "
Revive PCIConf. Submitted by: "Kenneth D. Merry" <ken@plutotech.com>
1998-09-15 08:21:13 +00:00
"length %u isn't user accessible for"
Some workarounds for a common DEC pci-pci bridge found on alphas. Fix printf format errors which show up on the alpha.
1998-09-16 08:28:11 +00:00
" READ\n", cio->patterns,
Revive PCIConf. Submitted by: "Kenneth D. Merry" <ken@plutotech.com>
1998-09-15 08:21:13 +00:00
cio->pat_buf_len);
error = EACCES;
break;
}
/*
* Allocate a buffer to hold the patterns.
*/
pattern_buf = malloc(cio->pat_buf_len, M_TEMP,
M_WAITOK);
error = copyin(cio->patterns, pattern_buf,
cio->pat_buf_len);
if (error != 0)
break;
num_patterns = cio->num_patterns;
} else if ((cio->num_patterns > 0)
|| (cio->pat_buf_len > 0)) {
/*
* The user made a mistake, spit out an error.
*/
cio->status = PCI_GETCONF_ERROR;
cio->num_matches = 0;
printf("pci_ioctl: invalid GETCONF arguments\n");
error = EINVAL;
break;
} else
pattern_buf = NULL;
/*
* Make sure we can write to the match buffer.
*/
Don't test boolean return against != 1. Don't needlessly assign the error variable in an if statement.
1999-10-29 19:03:18 +00:00
if (!useracc((caddr_t)cio->matches,
Change useracc() and kernacc() to use VM_PROT_{READ|WRITE|EXECUTE} for the "rw" argument, rather than hijacking B_{READ|WRITE}. Fix two bugs (physio & cam) resulting by the confusion caused by this. Submitted by: Tor.Egge@fast.no Reviewed by: alc, ken (partly)
1999-10-30 06:32:05 +00:00
cio->match_buf_len, VM_PROT_WRITE)) {
s/%#p/%p/ to fix a warning. This looks like a leftover of once being %#x, which causes a leading 0x to be prepended. %p has this by default and the '#' is ignored by the kernel prinf code for %p.
1998-11-09 08:08:06 +00:00
printf("pci_ioctl: match buffer %p, length %u "
Revive PCIConf. Submitted by: "Kenneth D. Merry" <ken@plutotech.com>
1998-09-15 08:21:13 +00:00
"isn't user accessible for WRITE\n",
Some workarounds for a common DEC pci-pci bridge found on alphas. Fix printf format errors which show up on the alpha.
1998-09-16 08:28:11 +00:00
cio->matches, cio->match_buf_len);
Revive PCIConf. Submitted by: "Kenneth D. Merry" <ken@plutotech.com>
1998-09-15 08:21:13 +00:00
error = EACCES;
break;
}
/*
* Go through the list of devices and copy out the devices
* that match the user's criteria.
*/
for (cio->num_matches = 0, error = 0, i = 0,
dinfo = STAILQ_FIRST(devlist_head);
(dinfo != NULL) && (cio->num_matches < ionum)
&& (error == 0) && (i < pci_numdevs);
dinfo = STAILQ_NEXT(dinfo, pci_links), i++) {
if (i < cio->offset)
continue;
For the ioctl that reads the pci configuration, look up the name and unit on the fly so that we can see the driver assignment of new pci devices as well in the 'pciconf -l' display.
1999-05-09 20:27:26 +00:00
/* Populate pd_name and pd_unit */
name = NULL;
if (dinfo->cfg.dev && dinfo->conf.pd_name[0] == '\0')
name = device_get_name(dinfo->cfg.dev);
if (name) {
strncpy(dinfo->conf.pd_name, name,
sizeof(dinfo->conf.pd_name));
dinfo->conf.pd_name[PCI_MAXNAMELEN] = 0;
dinfo->conf.pd_unit =
device_get_unit(dinfo->cfg.dev);
}
Revive PCIConf. Submitted by: "Kenneth D. Merry" <ken@plutotech.com>
1998-09-15 08:21:13 +00:00
if ((pattern_buf == NULL) ||
(pci_conf_match(pattern_buf, num_patterns,
&dinfo->conf) == 0)) {
/*
* If we've filled up the user's buffer,
* break out at this point. Since we've
* got a match here, we'll pick right back
* up at the matching entry. We can also
* tell the user that there are more matches
* left.
*/
if (cio->num_matches >= ionum)
break;
error = copyout(&dinfo->conf,
&cio->matches[cio->num_matches],
sizeof(struct pci_conf));
cio->num_matches++;
}
}
/*
* Set the pointer into the list, so if the user is getting
* n records at a time, where n < pci_numdevs,
*/
cio->offset = i;
/*
* Set the generation, the user will need this if they make
* another ioctl call with offset != 0.
*/
cio->generation = pci_generation;
Add support for ioctl() accesses to PCI config space registers. Garrett Wollman sent me this code a few weeks ago for review, and I made some significant changes, which he in turn accepted ... In order to make use of these changes, a device entry has to added to /dev. Submitted by: wollman
1996-10-22 20:20:14 +00:00
Revive PCIConf. Submitted by: "Kenneth D. Merry" <ken@plutotech.com>
1998-09-15 08:21:13 +00:00
/*
* If this is the last device, inform the user so he won't
* bother asking for more devices. If dinfo isn't NULL, we
* know that there are more matches in the list because of
* the way the traversal is done.
*/
if (dinfo == NULL)
cio->status = PCI_GETCONF_LAST_DEVICE;
else
cio->status = PCI_GETCONF_MORE_DEVS;
if (pattern_buf != NULL)
free(pattern_buf, M_TEMP);
break;
}
Add support for ioctl() accesses to PCI config space registers. Garrett Wollman sent me this code a few weeks ago for review, and I made some significant changes, which he in turn accepted ... In order to make use of these changes, a device entry has to added to /dev. Submitted by: wollman
1996-10-22 20:20:14 +00:00
case PCIOCREAD:
io = (struct pci_io *)data;
switch(io->pi_width) {
case 4:
case 2:
case 1:
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
/*
* Assume that the user-level bus number is
* actually the pciN instance number. We map
* from that to the real pcib+bus combination.
*/
pci = devclass_get_device(pci_devclass,
io->pi_sel.pc_bus);
if (pci) {
int b = pcib_get_bus(pci);
pcib = device_get_parent(pci);
io->pi_data =
PCIB_READ_CONFIG(pcib,
b,
io->pi_sel.pc_dev,
io->pi_sel.pc_func,
io->pi_reg,
io->pi_width);
error = 0;
} else {
error = ENODEV;
}
Add support for ioctl() accesses to PCI config space registers. Garrett Wollman sent me this code a few weeks ago for review, and I made some significant changes, which he in turn accepted ... In order to make use of these changes, a device entry has to added to /dev. Submitted by: wollman
1996-10-22 20:20:14 +00:00
break;
default:
error = ENODEV;
break;
}
break;
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
case PCIOCWRITE:
Add PCI LKM support.
1997-01-21 23:23:40 +00:00
io = (struct pci_io *)data;
switch(io->pi_width) {
case 4:
case 2:
case 1:
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
/*
* Assume that the user-level bus number is
* actually the pciN instance number. We map
* from that to the real pcib+bus combination.
*/
pci = devclass_get_device(pci_devclass,
io->pi_sel.pc_bus);
if (pci) {
int b = pcib_get_bus(pci);
pcib = device_get_parent(pci);
PCIB_WRITE_CONFIG(pcib,
b,
io->pi_sel.pc_dev,
io->pi_sel.pc_func,
io->pi_reg,
io->pi_data,
io->pi_width);
error = 0;
} else {
error = ENODEV;
}
Completely replace the PCI bus driver code to make it better reflect reality. There will be a new call interface, but for now the file pci_compat.c (which is to be deleted, after all drivers are converted) provides an emulation of the old PCI bus driver functions. The only change that might be visible to drivers is, that the type pcici_t (which had been meant to be just a handle, whose exact definition should not be relied on), has been converted into a pcicfgregs* . The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t and has been converted to just call the PCI drivers functions to access configuration space register, instead of inventing its own ... This code is by no means complete, but assumed to be fully operational, and brings the official code base more in line with my development code. A new generic device descriptor data type has to be agreed on. The PCI code will then use that data type to provide new functionality: 1) userconfig support 2) "wired" PCI devices 3) conflicts checking against ISA/EISA 4) maps will depend on the command register enable bits 5) PCI to Anything bridges can be defined as devices, and are probed like any "standard" PCI device. The following features are currently missing, but will be added back, soon: 1) unknown device probe message 2) suppression of "mirrored" devices caused by ancient, broken chip-sets This code relies on generic shared interrupt support just commited to kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
break;
Add PCI LKM support.
1997-01-21 23:23:40 +00:00
default:
error = ENODEV;
break;
}
break;
Add support for ioctl() accesses to PCI config space registers. Garrett Wollman sent me this code a few weeks ago for review, and I made some significant changes, which he in turn accepted ... In order to make use of these changes, a device entry has to added to /dev. Submitted by: wollman
1996-10-22 20:20:14 +00:00
default:
error = ENOTTY;
break;
}
return (error);
}
#define PCI_CDEV 78
static struct cdevsw pcicdev = {
This commit should be a extensive NO-OP: Reformat and initialize correctly all "struct cdevsw". Initialize the d_maj and d_bmaj fields. The d_reset field was not removed, although it is never used. I used a program to do most of this, so all the files now use the same consistent format. Please keep it that way. Vinum and i4b not modified, patches emailed to respective authors.
1999-05-30 16:53:49 +00:00
/* open */ pci_open,
/* close */ pci_close,
/* read */ noread,
/* write */ nowrite,
/* ioctl */ pci_ioctl,
/* poll */ nopoll,
/* mmap */ nommap,
/* strategy */ nostrategy,
/* name */ "pci",
/* maj */ PCI_CDEV,
/* dump */ nodump,
/* psize */ nopsize,
/* flags */ 0,
/* bmaj */ -1
Add support for ioctl() accesses to PCI config space registers. Garrett Wollman sent me this code a few weeks ago for review, and I made some significant changes, which he in turn accepted ... In order to make use of these changes, a device entry has to added to /dev. Submitted by: wollman
1996-10-22 20:20:14 +00:00
};
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
#include "pci_if.h"
/*
* New style pci driver. Parent device is either a pci-host-bridge or a
* pci-pci-bridge. Both kinds are represented by instances of pcib.
*/
static void
pci_print_verbose(struct pci_devinfo *dinfo)
{
if (bootverbose) {
pcicfgregs *cfg = &dinfo->cfg;
printf("found->\tvendor=0x%04x, dev=0x%04x, revid=0x%02x\n",
cfg->vendor, cfg->device, cfg->revid);
When dumping the 'found devices' list in verbose mode, actually show the bus/slot/function numbers. The old PCI code used other markers or something, but without it here under the new pci code it is very hard to tell which device is which (this only affects bootverbose mode).
2000-09-05 00:32:19 +00:00
printf("\tbus=%d, slot=%d, func=%d\n",
cfg->bus, cfg->slot, cfg->func);
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
printf("\tclass=%02x-%02x-%02x, hdrtype=0x%02x, mfdev=%d\n",
cfg->baseclass, cfg->subclass, cfg->progif,
cfg->hdrtype, cfg->mfdev);
printf("\tsubordinatebus=%x \tsecondarybus=%x\n",
cfg->subordinatebus, cfg->secondarybus);
#ifdef PCI_DEBUG
printf("\tcmdreg=0x%04x, statreg=0x%04x, cachelnsz=%d (dwords)\n",
cfg->cmdreg, cfg->statreg, cfg->cachelnsz);
printf("\tlattimer=0x%02x (%d ns), mingnt=0x%02x (%d ns), maxlat=0x%02x (%d ns)\n",
cfg->lattimer, cfg->lattimer * 30,
cfg->mingnt, cfg->mingnt * 250, cfg->maxlat, cfg->maxlat * 250);
#endif /* PCI_DEBUG */
if (cfg->intpin > 0)
printf("\tintpin=%c, irq=%d\n", cfg->intpin +'a' -1, cfg->intline);
* Implement bus_set/get/delete_resource for pci. * Change the hack used on the alpha for mapping devices into DENSE or BWX memory spaces to a simpler one. Its still a hack and should be a seperate api to explicitly map the resource. * Add $FreeBSD$ as necessary.
1999-10-14 21:38:33 +00:00
}
}
static int
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
pci_porten(device_t pcib, int b, int s, int f)
* Implement bus_set/get/delete_resource for pci. * Change the hack used on the alpha for mapping devices into DENSE or BWX memory spaces to a simpler one. Its still a hack and should be a seperate api to explicitly map the resource. * Add $FreeBSD$ as necessary.
1999-10-14 21:38:33 +00:00
{
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
return (PCIB_READ_CONFIG(pcib, b, s, f, PCIR_COMMAND, 2)
& PCIM_CMD_PORTEN) != 0;
* Implement bus_set/get/delete_resource for pci. * Change the hack used on the alpha for mapping devices into DENSE or BWX memory spaces to a simpler one. Its still a hack and should be a seperate api to explicitly map the resource. * Add $FreeBSD$ as necessary.
1999-10-14 21:38:33 +00:00
}
static int
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
pci_memen(device_t pcib, int b, int s, int f)
* Implement bus_set/get/delete_resource for pci. * Change the hack used on the alpha for mapping devices into DENSE or BWX memory spaces to a simpler one. Its still a hack and should be a seperate api to explicitly map the resource. * Add $FreeBSD$ as necessary.
1999-10-14 21:38:33 +00:00
{
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
return (PCIB_READ_CONFIG(pcib, b, s, f, PCIR_COMMAND, 2)
& PCIM_CMD_MEMEN) != 0;
* Implement bus_set/get/delete_resource for pci. * Change the hack used on the alpha for mapping devices into DENSE or BWX memory spaces to a simpler one. Its still a hack and should be a seperate api to explicitly map the resource. * Add $FreeBSD$ as necessary.
1999-10-14 21:38:33 +00:00
}
Fix some resource allocation peculiarities of the intpm device.
1999-10-28 08:06:59 +00:00
/*
* Add a resource based on a pci map register. Return 1 if the map
* register is a 32bit map register or 2 if it is a 64bit register.
*/
static int
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
pci_add_map(device_t pcib, int b, int s, int f, int reg,
struct resource_list *rl)
* Implement bus_set/get/delete_resource for pci. * Change the hack used on the alpha for mapping devices into DENSE or BWX memory spaces to a simpler one. Its still a hack and should be a seperate api to explicitly map the resource. * Add $FreeBSD$ as necessary.
1999-10-14 21:38:33 +00:00
{
Fix some resource allocation peculiarities of the intpm device.
1999-10-28 08:06:59 +00:00
u_int32_t map;
u_int64_t base;
u_int8_t ln2size;
u_int8_t ln2range;
u_int32_t testval;
fix warning compile error about unused variable
2000-10-29 10:07:43 +00:00
#ifdef PCI_ENABLE_IO_MODES
If a base address register has been set up by the BIOS, but the relevant enable bit hasn't been set in the command register, set the bit and honour the register. It seems that quite a few lazy BIOS writers aren't bothering to do this, which upsets the existing code and causes us to miss out on properly-configured devices.
2000-09-01 23:09:02 +00:00
u_int16_t cmd;
fix warning compile error about unused variable
2000-10-29 10:07:43 +00:00
#endif
Fix some resource allocation peculiarities of the intpm device.
1999-10-28 08:06:59 +00:00
int type;
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
map = PCIB_READ_CONFIG(pcib, b, s, f, reg, 4);
Fix some resource allocation peculiarities of the intpm device.
1999-10-28 08:06:59 +00:00
if (map == 0 || map == 0xffffffff)
return 1; /* skip invalid entry */
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
PCIB_WRITE_CONFIG(pcib, b, s, f, reg, 0xffffffff, 4);
testval = PCIB_READ_CONFIG(pcib, b, s, f, reg, 4);
PCIB_WRITE_CONFIG(pcib, b, s, f, reg, map, 4);
Fix some resource allocation peculiarities of the intpm device.
1999-10-28 08:06:59 +00:00
base = pci_mapbase(map);
if (pci_maptype(map) & PCI_MAPMEM)
type = SYS_RES_MEMORY;
else
type = SYS_RES_IOPORT;
ln2size = pci_mapsize(testval);
ln2range = pci_maprange(testval);
if (ln2range == 64) {
/* Read the other half of a 64bit map register */
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
base |= (u_int64_t) PCIB_READ_CONFIG(pcib, b, s, f, reg + 4, 4) << 32;
Fix some resource allocation peculiarities of the intpm device.
1999-10-28 08:06:59 +00:00
}
* Implement bus_set/get/delete_resource for pci. * Change the hack used on the alpha for mapping devices into DENSE or BWX memory spaces to a simpler one. Its still a hack and should be a seperate api to explicitly map the resource. * Add $FreeBSD$ as necessary.
1999-10-14 21:38:33 +00:00
Print the PCI resources even if they are disabled. This shows more clearly when the BIOS is forgetful about initialising the USB controllers.
2000-03-18 19:18:36 +00:00
if (bootverbose) {
printf("\tmap[%02x]: type %x, range %2d, base %08x, size %2d",
Print the correct value for the map type on a verbose boot. PR: kern/18662 Submitted by: tamaru@ap.t.u-tokyo.ac.jp
2000-05-18 20:28:15 +00:00
reg, pci_maptype(map), ln2range,
Print the PCI resources even if they are disabled. This shows more clearly when the BIOS is forgetful about initialising the USB controllers.
2000-03-18 19:18:36 +00:00
(unsigned int) base, ln2size);
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
if (type == SYS_RES_IOPORT && !pci_porten(pcib, b, s, f))
Print the PCI resources even if they are disabled. This shows more clearly when the BIOS is forgetful about initialising the USB controllers.
2000-03-18 19:18:36 +00:00
printf(", port disabled\n");
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
else if (type == SYS_RES_MEMORY && !pci_memen(pcib, b, s, f))
Print the PCI resources even if they are disabled. This shows more clearly when the BIOS is forgetful about initialising the USB controllers.
2000-03-18 19:18:36 +00:00
printf(", memory disabled\n");
else
printf(", enabled\n");
}
Unconditionally turning on the I/O and memory enable bits in the PCI command register is too aggressive. Revert to the previous behaviour, but leave the new behaviour available as an undocumented option. It's not clear what the Right, Right Thing is to do here, but the more conservative approach is safer.
2000-10-28 23:07:13 +00:00
/*
* This code theoretically does the right thing, but has
* undesirable side effects in some cases where
* peripherals respond oddly to having these bits
* enabled. Leave them alone by default.
*/
#ifdef PCI_ENABLE_IO_MODES
If a base address register has been set up by the BIOS, but the relevant enable bit hasn't been set in the command register, set the bit and honour the register. It seems that quite a few lazy BIOS writers aren't bothering to do this, which upsets the existing code and causes us to miss out on properly-configured devices.
2000-09-01 23:09:02 +00:00
/* Turn on resources that have been left off by a lazy BIOS */
if (type == SYS_RES_IOPORT && !pci_porten(pcib, b, s, f)) {
cmd = PCIB_READ_CONFIG(pcib, b, s, f, PCIR_COMMAND, 2);
cmd |= PCIM_CMD_PORTEN;
PCIB_WRITE_CONFIG(pcib, b, s, f, PCIR_COMMAND, cmd, 2);
}
if (type == SYS_RES_MEMORY && !pci_memen(pcib, b, s, f)) {
cmd = PCIB_READ_CONFIG(pcib, b, s, f, PCIR_COMMAND, 2);
cmd |= PCIM_CMD_MEMEN;
PCIB_WRITE_CONFIG(pcib, b, s, f, PCIR_COMMAND, cmd, 2);
}
Unconditionally turning on the I/O and memory enable bits in the PCI command register is too aggressive. Revert to the previous behaviour, but leave the new behaviour available as an undocumented option. It's not clear what the Right, Right Thing is to do here, but the more conservative approach is safer.
2000-10-28 23:07:13 +00:00
#else
Fix params passed to pci_porten() and pci_memen().
2000-10-29 09:59:28 +00:00
if (type == SYS_RES_IOPORT && !pci_porten(pcib, b, s, f))
Unconditionally turning on the I/O and memory enable bits in the PCI command register is too aggressive. Revert to the previous behaviour, but leave the new behaviour available as an undocumented option. It's not clear what the Right, Right Thing is to do here, but the more conservative approach is safer.
2000-10-28 23:07:13 +00:00
return 1;
Fix params passed to pci_porten() and pci_memen().
2000-10-29 09:59:28 +00:00
if (type == SYS_RES_MEMORY && !pci_memen(pcib, b, s, f))
Unconditionally turning on the I/O and memory enable bits in the PCI command register is too aggressive. Revert to the previous behaviour, but leave the new behaviour available as an undocumented option. It's not clear what the Right, Right Thing is to do here, but the more conservative approach is safer.
2000-10-28 23:07:13 +00:00
return 1;
#endif
* Implement bus_set/get/delete_resource for pci. * Change the hack used on the alpha for mapping devices into DENSE or BWX memory spaces to a simpler one. Its still a hack and should be a seperate api to explicitly map the resource. * Add $FreeBSD$ as necessary.
1999-10-14 21:38:33 +00:00
Fix some resource allocation peculiarities of the intpm device.
1999-10-28 08:06:59 +00:00
resource_list_add(rl, type, reg,
base, base + (1 << ln2size) - 1,
(1 << ln2size));
* Implement bus_set/get/delete_resource for pci. * Change the hack used on the alpha for mapping devices into DENSE or BWX memory spaces to a simpler one. Its still a hack and should be a seperate api to explicitly map the resource. * Add $FreeBSD$ as necessary.
1999-10-14 21:38:33 +00:00
Fix some resource allocation peculiarities of the intpm device.
1999-10-28 08:06:59 +00:00
return (ln2range == 64) ? 2 : 1;
}
static void
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
pci_add_resources(device_t pcib, int b, int s, int f, device_t dev)
Fix some resource allocation peculiarities of the intpm device.
1999-10-28 08:06:59 +00:00
{
struct pci_devinfo *dinfo = device_get_ivars(dev);
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
pcicfgregs *cfg = &dinfo->cfg;
Fix some resource allocation peculiarities of the intpm device.
1999-10-28 08:06:59 +00:00
struct resource_list *rl = &dinfo->resources;
struct pci_quirk *q;
int i;
for (i = 0; i < cfg->nummaps;) {
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
i += pci_add_map(pcib, b, s, f, PCIR_MAPS + i*4, rl);
Fix some resource allocation peculiarities of the intpm device.
1999-10-28 08:06:59 +00:00
}
for (q = &pci_quirks[0]; q->devid; q++) {
if (q->devid == ((cfg->device << 16) | cfg->vendor)
&& q->type == PCI_QUIRK_MAP_REG)
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
pci_add_map(pcib, b, s, f, q->arg1, rl);
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
}
Fix some resource allocation peculiarities of the intpm device.
1999-10-28 08:06:59 +00:00
Show the port/mem/irq of pci devices too.
2000-01-08 10:12:21 +00:00
if (cfg->intpin > 0 && cfg->intline != 255)
* Implement bus_set/get/delete_resource for pci. * Change the hack used on the alpha for mapping devices into DENSE or BWX memory spaces to a simpler one. Its still a hack and should be a seperate api to explicitly map the resource. * Add $FreeBSD$ as necessary.
1999-10-14 21:38:33 +00:00
resource_list_add(rl, SYS_RES_IRQ, 0,
cfg->intline, cfg->intline, 1);
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
}
* Implement bus_set/get/delete_resource for pci. * Change the hack used on the alpha for mapping devices into DENSE or BWX memory spaces to a simpler one. Its still a hack and should be a seperate api to explicitly map the resource. * Add $FreeBSD$ as necessary.
1999-10-14 21:38:33 +00:00
static void
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
pci_add_children(device_t dev, int busno)
{
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
device_t pcib = device_get_parent(dev);
int maxslots;
int s, f;
Add support for multiple PCI "hoses" used on various alpha platforms. The specific intent of this commit is to pave the way for importing Compaq XP1000 support. These changes should not affect the i386 port. Reviewed by: Doug Rabson <dfr@nlsystems.com> (actually, he walked me through most of it & deserves more than reviewd-by credit )
1999-05-20 15:33:33 +00:00
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
maxslots = PCIB_MAXSLOTS(pcib);
for (s = 0; s <= maxslots; s++) {
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
int pcifunchigh = 0;
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
for (f = 0; f <= pcifunchigh; f++) {
struct pci_devinfo *dinfo =
pci_read_device(pcib, busno, s, f);
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
if (dinfo != NULL) {
if (dinfo->cfg.mfdev)
pcifunchigh = 7;
pci_print_verbose(dinfo);
Remove the 'ivars' arguement to device_add_child() and device_add_child_ordered(). 'ivars' may now be set using the device_set_ivars() function. This makes it easier for us to change how arbitrary data structures are associated with a device_t. Eventually we won't be modifying device_t to add additional pointers for ivars, softc data etc. Despite my best efforts I've probably forgotten something so let me know if this breaks anything. I've been running with this change for months and its been quite involved actually isolating all the changes from the rest of the local changes in my tree. Reviewed by: peter, dfr
1999-12-03 08:41:24 +00:00
dinfo->cfg.dev = device_add_child(dev, NULL, -1);
device_set_ivars(dinfo->cfg.dev, dinfo);
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
pci_add_resources(pcib, busno, s, f,
dinfo->cfg.dev);
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
}
}
}
}
static int
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
pci_probe(device_t dev)
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
{
Validate the PCI bus number that we fetch from our parent, since there's no guarantee that everything attached to *it* is a PCI bus.
2000-10-09 00:43:45 +00:00
static int once, busno;
Next phase in the PCI subsystem cleanup. - Move PCI core code to dev/pci. - Split bridge code out into separate modules. - Remove the descriptive strings from the bridge drivers. If you want to know what a device is, use pciconf. Add support for broadly identifying devices based on class/subclass, and for parsing a preloaded device identification database so that if you want to waste the memory, you can identify *anything* we know about. - Remove machine-dependant code from the core PCI code. APIC interrupt mapping is performed by shadowing the intline register in machine- dependant code. - Bring interrupt routing support to the Alpha (although many platforms don't yet support routing or mapping interrupts entirely correctly). This resulted in spamming <sys/bus.h> into more places than it really should have gone. - Put sys/dev on the kernel/modules include path. This avoids having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
caddr_t vendordata, info;
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
Convert DEVFS hooks in (most) drivers to make_dev(). Diskslice/label code not yet handled. Vinum, i4b, alpha, pc98 not dealt with (left to respective Maintainers) Add the correct hook for devfs to kern_conf.c The net result of this excercise is that a lot less files depends on DEVFS, and devtoname() gets more sensible output in many cases. A few drivers had minor additional cleanups performed relating to cdevsw registration. A few drivers don't register a cdevsw{} anymore, but only use make_dev().
1999-08-23 20:59:21 +00:00
device_set_desc(dev, "PCI bus");
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
if (bootverbose)
device_printf(dev, "physical bus=%d\n", pcib_get_bus(dev));
/*
* Since there can be multiple independantly numbered PCI
* busses on some large alpha systems, we can't use the unit
* number to decide what bus we are probing. We ask the parent
* pcib what our bus number is.
*/
Validate the PCI bus number that we fetch from our parent, since there's no guarantee that everything attached to *it* is a PCI bus.
2000-10-09 00:43:45 +00:00
busno = pcib_get_bus(dev);
if (busno < 0)
return ENXIO;
pci_add_children(dev, busno);
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
Convert DEVFS hooks in (most) drivers to make_dev(). Diskslice/label code not yet handled. Vinum, i4b, alpha, pc98 not dealt with (left to respective Maintainers) Add the correct hook for devfs to kern_conf.c The net result of this excercise is that a lot less files depends on DEVFS, and devtoname() gets more sensible output in many cases. A few drivers had minor additional cleanups performed relating to cdevsw registration. A few drivers don't register a cdevsw{} anymore, but only use make_dev().
1999-08-23 20:59:21 +00:00
if (!once) {
make_dev(&pcicdev, 0, UID_ROOT, GID_WHEEL, 0644, "pci");
Next phase in the PCI subsystem cleanup. - Move PCI core code to dev/pci. - Split bridge code out into separate modules. - Remove the descriptive strings from the bridge drivers. If you want to know what a device is, use pciconf. Add support for broadly identifying devices based on class/subclass, and for parsing a preloaded device identification database so that if you want to waste the memory, you can identify *anything* we know about. - Remove machine-dependant code from the core PCI code. APIC interrupt mapping is performed by shadowing the intline register in machine- dependant code. - Bring interrupt routing support to the Alpha (although many platforms don't yet support routing or mapping interrupts entirely correctly). This resulted in spamming <sys/bus.h> into more places than it really should have gone. - Put sys/dev on the kernel/modules include path. This avoids having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
if ((vendordata = preload_search_by_type("pci_vendor_data")) != NULL) {
info = preload_search_info(vendordata, MODINFO_ADDR);
pci_vendordata = *(char **)info;
info = preload_search_info(vendordata, MODINFO_SIZE);
pci_vendordata_size = *(size_t *)info;
/* terminate the database */
pci_vendordata[pci_vendordata_size] = '\n';
}
Convert DEVFS hooks in (most) drivers to make_dev(). Diskslice/label code not yet handled. Vinum, i4b, alpha, pc98 not dealt with (left to respective Maintainers) Add the correct hook for devfs to kern_conf.c The net result of this excercise is that a lot less files depends on DEVFS, and devtoname() gets more sensible output in many cases. A few drivers had minor additional cleanups performed relating to cdevsw registration. A few drivers don't register a cdevsw{} anymore, but only use make_dev().
1999-08-23 20:59:21 +00:00
once++;
}
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
return 0;
}
Show the port/mem/irq of pci devices too.
2000-01-08 10:12:21 +00:00
static int
pci_print_resources(struct resource_list *rl, const char *name, int type,
const char *format)
{
struct resource_list_entry *rle;
int printed, retval;
printed = 0;
retval = 0;
/* Yes, this is kinda cheating */
SLIST_FOREACH(rle, rl, link) {
if (rle->type == type) {
if (printed == 0)
retval += printf(" %s ", name);
else if (printed > 0)
retval += printf(",");
printed++;
retval += printf(format, rle->start);
if (rle->count > 1) {
retval += printf("-");
retval += printf(format, rle->start +
rle->count - 1);
}
}
}
return retval;
}
Alter the behavior of sys/kern/subr_bus.c:device_print_child() - device_print_child() either lets the BUS_PRINT_CHILD method produce the entire device announcement message or it prints "foo0: not found\n" Alter sys/kern/subr_bus.c:bus_generic_print_child() to take on the previous behavior of device_print_child() (printing the "foo0: <FooDevice 1.1>" bit of the announce message.) Provide bus_print_child_header() and bus_print_child_footer() to actually print the output for bus_generic_print_child(). These functions should be used whenever possible (unless you can just use bus_generic_print_child()) The BUS_PRINT_CHILD method now returns int instead of void. Modify everything else that defines or uses a BUS_PRINT_CHILD method to comply with the above changes. - Devices are 'on' a bus, not 'at' it. - If a custom BUS_PRINT_CHILD method does the same thing as bus_generic_print_child(), use bus_generic_print_child() - Use device_get_nameunit() instead of both device_get_name() and device_get_unit() - All BUS_PRINT_CHILD methods return the number of characters output. Reviewed by: dfr, peter
1999-07-29 01:03:04 +00:00
static int
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
pci_print_child(device_t dev, device_t child)
{
Print 'irq nn' on the device attach line like the old pci code did. However, we are not printing 'int a/b/c/d' yet, is it worth it on non-SMP systems? (It's useful when tracing PCI->IO-APIC routing on SMP systems)
1999-05-08 20:28:01 +00:00
struct pci_devinfo *dinfo;
Show the port/mem/irq of pci devices too.
2000-01-08 10:12:21 +00:00
struct resource_list *rl;
Print 'irq nn' on the device attach line like the old pci code did. However, we are not printing 'int a/b/c/d' yet, is it worth it on non-SMP systems? (It's useful when tracing PCI->IO-APIC routing on SMP systems)
1999-05-08 20:28:01 +00:00
pcicfgregs *cfg;
Alter the behavior of sys/kern/subr_bus.c:device_print_child() - device_print_child() either lets the BUS_PRINT_CHILD method produce the entire device announcement message or it prints "foo0: not found\n" Alter sys/kern/subr_bus.c:bus_generic_print_child() to take on the previous behavior of device_print_child() (printing the "foo0: <FooDevice 1.1>" bit of the announce message.) Provide bus_print_child_header() and bus_print_child_footer() to actually print the output for bus_generic_print_child(). These functions should be used whenever possible (unless you can just use bus_generic_print_child()) The BUS_PRINT_CHILD method now returns int instead of void. Modify everything else that defines or uses a BUS_PRINT_CHILD method to comply with the above changes. - Devices are 'on' a bus, not 'at' it. - If a custom BUS_PRINT_CHILD method does the same thing as bus_generic_print_child(), use bus_generic_print_child() - Use device_get_nameunit() instead of both device_get_name() and device_get_unit() - All BUS_PRINT_CHILD methods return the number of characters output. Reviewed by: dfr, peter
1999-07-29 01:03:04 +00:00
int retval = 0;
Print 'irq nn' on the device attach line like the old pci code did. However, we are not printing 'int a/b/c/d' yet, is it worth it on non-SMP systems? (It's useful when tracing PCI->IO-APIC routing on SMP systems)
1999-05-08 20:28:01 +00:00
dinfo = device_get_ivars(child);
cfg = &dinfo->cfg;
Show the port/mem/irq of pci devices too.
2000-01-08 10:12:21 +00:00
rl = &dinfo->resources;
Alter the behavior of sys/kern/subr_bus.c:device_print_child() - device_print_child() either lets the BUS_PRINT_CHILD method produce the entire device announcement message or it prints "foo0: not found\n" Alter sys/kern/subr_bus.c:bus_generic_print_child() to take on the previous behavior of device_print_child() (printing the "foo0: <FooDevice 1.1>" bit of the announce message.) Provide bus_print_child_header() and bus_print_child_footer() to actually print the output for bus_generic_print_child(). These functions should be used whenever possible (unless you can just use bus_generic_print_child()) The BUS_PRINT_CHILD method now returns int instead of void. Modify everything else that defines or uses a BUS_PRINT_CHILD method to comply with the above changes. - Devices are 'on' a bus, not 'at' it. - If a custom BUS_PRINT_CHILD method does the same thing as bus_generic_print_child(), use bus_generic_print_child() - Use device_get_nameunit() instead of both device_get_name() and device_get_unit() - All BUS_PRINT_CHILD methods return the number of characters output. Reviewed by: dfr, peter
1999-07-29 01:03:04 +00:00
retval += bus_print_child_header(dev, child);
Show the port/mem/irq of pci devices too.
2000-01-08 10:12:21 +00:00
retval += pci_print_resources(rl, "port", SYS_RES_IOPORT, "%#lx");
retval += pci_print_resources(rl, "mem", SYS_RES_MEMORY, "%#lx");
retval += pci_print_resources(rl, "irq", SYS_RES_IRQ, "%ld");
if (device_get_flags(dev))
retval += printf(" flags %#x", device_get_flags(dev));
Alter the behavior of sys/kern/subr_bus.c:device_print_child() - device_print_child() either lets the BUS_PRINT_CHILD method produce the entire device announcement message or it prints "foo0: not found\n" Alter sys/kern/subr_bus.c:bus_generic_print_child() to take on the previous behavior of device_print_child() (printing the "foo0: <FooDevice 1.1>" bit of the announce message.) Provide bus_print_child_header() and bus_print_child_footer() to actually print the output for bus_generic_print_child(). These functions should be used whenever possible (unless you can just use bus_generic_print_child()) The BUS_PRINT_CHILD method now returns int instead of void. Modify everything else that defines or uses a BUS_PRINT_CHILD method to comply with the above changes. - Devices are 'on' a bus, not 'at' it. - If a custom BUS_PRINT_CHILD method does the same thing as bus_generic_print_child(), use bus_generic_print_child() - Use device_get_nameunit() instead of both device_get_name() and device_get_unit() - All BUS_PRINT_CHILD methods return the number of characters output. Reviewed by: dfr, peter
1999-07-29 01:03:04 +00:00
retval += printf(" at device %d.%d", pci_get_slot(child),
pci_get_function(child));
retval += bus_print_child_footer(dev, child);
return (retval);
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
}
Next phase in the PCI subsystem cleanup. - Move PCI core code to dev/pci. - Split bridge code out into separate modules. - Remove the descriptive strings from the bridge drivers. If you want to know what a device is, use pciconf. Add support for broadly identifying devices based on class/subclass, and for parsing a preloaded device identification database so that if you want to waste the memory, you can identify *anything* we know about. - Remove machine-dependant code from the core PCI code. APIC interrupt mapping is performed by shadowing the intline register in machine- dependant code. - Bring interrupt routing support to the Alpha (although many platforms don't yet support routing or mapping interrupts entirely correctly). This resulted in spamming <sys/bus.h> into more places than it really should have gone. - Put sys/dev on the kernel/modules include path. This avoids having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
static struct
{
int class;
int subclass;
char *desc;
} pci_nomatch_tab[] = {
{PCIC_OLD, -1, "old"},
{PCIC_OLD, PCIS_OLD_NONVGA, "non-VGA display device"},
{PCIC_OLD, PCIS_OLD_VGA, "VGA-compatible display device"},
{PCIC_STORAGE, -1, "mass storage"},
{PCIC_STORAGE, PCIS_STORAGE_SCSI, "SCSI"},
{PCIC_STORAGE, PCIS_STORAGE_IDE, "ATA"},
{PCIC_STORAGE, PCIS_STORAGE_FLOPPY, "floppy disk"},
{PCIC_STORAGE, PCIS_STORAGE_IPI, "IPI"},
{PCIC_STORAGE, PCIS_STORAGE_RAID, "RAID"},
{PCIC_NETWORK, -1, "network"},
{PCIC_NETWORK, PCIS_NETWORK_ETHERNET, "ethernet"},
{PCIC_NETWORK, PCIS_NETWORK_TOKENRING, "token ring"},
{PCIC_NETWORK, PCIS_NETWORK_FDDI, "fddi"},
{PCIC_NETWORK, PCIS_NETWORK_ATM, "ATM"},
{PCIC_DISPLAY, -1, "display"},
{PCIC_DISPLAY, PCIS_DISPLAY_VGA, "VGA"},
{PCIC_DISPLAY, PCIS_DISPLAY_XGA, "XGA"},
{PCIC_MULTIMEDIA, -1, "multimedia"},
{PCIC_MULTIMEDIA, PCIS_MULTIMEDIA_VIDEO, "video"},
{PCIC_MULTIMEDIA, PCIS_MULTIMEDIA_AUDIO, "audio"},
{PCIC_MEMORY, -1, "memory"},
{PCIC_MEMORY, PCIS_MEMORY_RAM, "RAM"},
{PCIC_MEMORY, PCIS_MEMORY_FLASH, "flash"},
{PCIC_BRIDGE, -1, "bridge"},
{PCIC_BRIDGE, PCIS_BRIDGE_HOST, "HOST-PCI"},
{PCIC_BRIDGE, PCIS_BRIDGE_ISA, "PCI-ISA"},
{PCIC_BRIDGE, PCIS_BRIDGE_EISA, "PCI-EISA"},
{PCIC_BRIDGE, PCIS_BRIDGE_MCA, "PCI-MCA"},
{PCIC_BRIDGE, PCIS_BRIDGE_PCI, "PCI-PCI"},
{PCIC_BRIDGE, PCIS_BRIDGE_PCMCIA, "PCI-PCMCIA"},
{PCIC_BRIDGE, PCIS_BRIDGE_NUBUS, "PCI-NuBus"},
{PCIC_BRIDGE, PCIS_BRIDGE_CARDBUS, "PCI-CardBus"},
{PCIC_BRIDGE, PCIS_BRIDGE_OTHER, "PCI-unknown"},
{PCIC_SIMPLECOMM, -1, "simple comms"},
{PCIC_SIMPLECOMM, PCIS_SIMPLECOMM_UART, "UART"}, /* could detect 16550 */
{PCIC_SIMPLECOMM, PCIS_SIMPLECOMM_PAR, "parallel port"},
{PCIC_BASEPERIPH, -1, "base peripheral"},
{PCIC_BASEPERIPH, PCIS_BASEPERIPH_PIC, "interrupt controller"},
{PCIC_BASEPERIPH, PCIS_BASEPERIPH_DMA, "DMA controller"},
{PCIC_BASEPERIPH, PCIS_BASEPERIPH_TIMER, "timer"},
{PCIC_BASEPERIPH, PCIS_BASEPERIPH_RTC, "realtime clock"},
{PCIC_INPUTDEV, -1, "input device"},
{PCIC_INPUTDEV, PCIS_INPUTDEV_KEYBOARD, "keyboard"},
{PCIC_INPUTDEV, PCIS_INPUTDEV_DIGITIZER,"digitizer"},
{PCIC_INPUTDEV, PCIS_INPUTDEV_MOUSE, "mouse"},
{PCIC_DOCKING, -1, "docking station"},
{PCIC_PROCESSOR, -1, "processor"},
{PCIC_SERIALBUS, -1, "serial bus"},
{PCIC_SERIALBUS, PCIS_SERIALBUS_FW, "FireWire"},
{PCIC_SERIALBUS, PCIS_SERIALBUS_ACCESS, "AccessBus"},
{PCIC_SERIALBUS, PCIS_SERIALBUS_SSA, "SSA"},
{PCIC_SERIALBUS, PCIS_SERIALBUS_USB, "USB"},
{PCIC_SERIALBUS, PCIS_SERIALBUS_FC, "Fibre Channel"},
{PCIC_SERIALBUS, PCIS_SERIALBUS_SMBUS, "SMBus"},
{0, 0, NULL}
};
Implement the BUS_PROBE_NOMATCH method for the PCI bus. This function is called for each device for which no driver was found. Output is similar to the eisa_probe_nomatch() function but with the added benefit of displaying the assigned IRQ (since PCI gives us this information up front.) Output is like so: pci0: unknown card CPQ0508 (vendor=0x0e11, dev=0x0508) at 11.0 irq 9 pci0: unknown card DFZ0508 (vendor=0x10da, dev=0x0508) at 11.0 irq 9 pci0: unknown card DBL0508 (vendor=0x104c, dev=0x0508) at 11.0 irq 9 pci0: unknown card DDM0011 (vendor=0x108d, dev=0x0011) at 11.0 irq 9 I'm not happy with the 3 lines of macro cruft that got added but I consider it a temporary annoyance as those bits will be moved to some place where PCI, EISA and ISAPNP code will be able to use them. (Not surprisingly, this message is longer than the code in question.) Reviewed by: peter, dfr
1999-07-27 04:28:14 +00:00
static void
pci_probe_nomatch(device_t dev, device_t child)
{
Next phase in the PCI subsystem cleanup. - Move PCI core code to dev/pci. - Split bridge code out into separate modules. - Remove the descriptive strings from the bridge drivers. If you want to know what a device is, use pciconf. Add support for broadly identifying devices based on class/subclass, and for parsing a preloaded device identification database so that if you want to waste the memory, you can identify *anything* we know about. - Remove machine-dependant code from the core PCI code. APIC interrupt mapping is performed by shadowing the intline register in machine- dependant code. - Bring interrupt routing support to the Alpha (although many platforms don't yet support routing or mapping interrupts entirely correctly). This resulted in spamming <sys/bus.h> into more places than it really should have gone. - Put sys/dev on the kernel/modules include path. This avoids having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
int i;
char *cp, *scp, *device;
/*
* Look for a listing for this device in a loaded device database.
*/
if ((device = pci_describe_device(child)) != NULL) {
- Fix the device database parsing code so that it actually works. - Improve the formatting for devices identified by the database. - Fix the pcib_route_interrupt method definition, as an old version snuck in here somehow 8( - Remove a couple of the vendor/device IDs for PCI:ISA bridges which correctly identify themselves. Submitted by: peter
2000-12-09 09:15:38 +00:00
device_printf(dev, "<%s>", device);
Next phase in the PCI subsystem cleanup. - Move PCI core code to dev/pci. - Split bridge code out into separate modules. - Remove the descriptive strings from the bridge drivers. If you want to know what a device is, use pciconf. Add support for broadly identifying devices based on class/subclass, and for parsing a preloaded device identification database so that if you want to waste the memory, you can identify *anything* we know about. - Remove machine-dependant code from the core PCI code. APIC interrupt mapping is performed by shadowing the intline register in machine- dependant code. - Bring interrupt routing support to the Alpha (although many platforms don't yet support routing or mapping interrupts entirely correctly). This resulted in spamming <sys/bus.h> into more places than it really should have gone. - Put sys/dev on the kernel/modules include path. This avoids having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
free(device, M_DEVBUF);
} else {
/*
* Scan the class/subclass descriptions for a general description.
*/
cp = "unknown";
scp = NULL;
for (i = 0; pci_nomatch_tab[i].desc != NULL; i++) {
if (pci_nomatch_tab[i].class == pci_get_class(child)) {
if (pci_nomatch_tab[i].subclass == -1) {
cp = pci_nomatch_tab[i].desc;
} else if (pci_nomatch_tab[i].subclass == pci_get_subclass(child)) {
scp = pci_nomatch_tab[i].desc;
}
}
}
device_printf(dev, "<%s%s%s>",
cp ? : "",
((cp != NULL) && (scp != NULL)) ? ", " : "",
scp ? : "");
}
printf(" at %d:%d (no driver attached)\n",
pci_get_slot(child),
pci_get_function(child));
return;
}
Move the pretty printing of the description for USB controllers to pci_probe_nomatch, so it won't be in the way when loading USB as a module. The reason for them being there in the first place is that every motherboard comes with USB kit and this way it looks more pretty (peter). The real solution will be to define some method of detaching a driver after it has attached.
1999-11-22 03:34:43 +00:00
Next phase in the PCI subsystem cleanup. - Move PCI core code to dev/pci. - Split bridge code out into separate modules. - Remove the descriptive strings from the bridge drivers. If you want to know what a device is, use pciconf. Add support for broadly identifying devices based on class/subclass, and for parsing a preloaded device identification database so that if you want to waste the memory, you can identify *anything* we know about. - Remove machine-dependant code from the core PCI code. APIC interrupt mapping is performed by shadowing the intline register in machine- dependant code. - Bring interrupt routing support to the Alpha (although many platforms don't yet support routing or mapping interrupts entirely correctly). This resulted in spamming <sys/bus.h> into more places than it really should have gone. - Put sys/dev on the kernel/modules include path. This avoids having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
/*
* Parse the PCI device database, if loaded, and return a pointer to a
* description of the device.
*
* The database is flat text formatted as follows:
*
* Any line not in a valid format is ignored.
* Lines are terminated with newline '\n' characters.
*
* A VENDOR line consists of the 4 digit (hex) vendor code, a TAB, then
* the vendor name.
*
* A DEVICE line is entered immediately below the corresponding VENDOR ID.
* - devices cannot be listed without a corresponding VENDOR line.
* A DEVICE line consists of a TAB, the 4 digit (hex) device code,
* another TAB, then the device name.
*/
/*
* Assuming (ptr) points to the beginning of a line in the database,
* return the vendor or device and description of the next entry.
* The value of (vendor) or (device) inappropriate for the entry type
* is set to -1. Returns nonzero at the end of the database.
*
* Note that this is slightly unrobust in the face of corrupt data;
* we attempt to safeguard against this by spamming the end of the
* database with a newline when we initialise.
*/
static int
pci_describe_parse_line(char **ptr, int *vendor, int *device, char **desc)
{
char *cp = *ptr;
- Fix the device database parsing code so that it actually works. - Improve the formatting for devices identified by the database. - Fix the pcib_route_interrupt method definition, as an old version snuck in here somehow 8( - Remove a couple of the vendor/device IDs for PCI:ISA bridges which correctly identify themselves. Submitted by: peter
2000-12-09 09:15:38 +00:00
int left;
Next phase in the PCI subsystem cleanup. - Move PCI core code to dev/pci. - Split bridge code out into separate modules. - Remove the descriptive strings from the bridge drivers. If you want to know what a device is, use pciconf. Add support for broadly identifying devices based on class/subclass, and for parsing a preloaded device identification database so that if you want to waste the memory, you can identify *anything* we know about. - Remove machine-dependant code from the core PCI code. APIC interrupt mapping is performed by shadowing the intline register in machine- dependant code. - Bring interrupt routing support to the Alpha (although many platforms don't yet support routing or mapping interrupts entirely correctly). This resulted in spamming <sys/bus.h> into more places than it really should have gone. - Put sys/dev on the kernel/modules include path. This avoids having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
*device = -1;
*vendor = -1;
- Fix the device database parsing code so that it actually works. - Improve the formatting for devices identified by the database. - Fix the pcib_route_interrupt method definition, as an old version snuck in here somehow 8( - Remove a couple of the vendor/device IDs for PCI:ISA bridges which correctly identify themselves. Submitted by: peter
2000-12-09 09:15:38 +00:00
**desc = '\0';
Next phase in the PCI subsystem cleanup. - Move PCI core code to dev/pci. - Split bridge code out into separate modules. - Remove the descriptive strings from the bridge drivers. If you want to know what a device is, use pciconf. Add support for broadly identifying devices based on class/subclass, and for parsing a preloaded device identification database so that if you want to waste the memory, you can identify *anything* we know about. - Remove machine-dependant code from the core PCI code. APIC interrupt mapping is performed by shadowing the intline register in machine- dependant code. - Bring interrupt routing support to the Alpha (although many platforms don't yet support routing or mapping interrupts entirely correctly). This resulted in spamming <sys/bus.h> into more places than it really should have gone. - Put sys/dev on the kernel/modules include path. This avoids having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
for (;;) {
- Fix the device database parsing code so that it actually works. - Improve the formatting for devices identified by the database. - Fix the pcib_route_interrupt method definition, as an old version snuck in here somehow 8( - Remove a couple of the vendor/device IDs for PCI:ISA bridges which correctly identify themselves. Submitted by: peter
2000-12-09 09:15:38 +00:00
left = pci_vendordata_size - (cp - pci_vendordata);
if (left <= 0) {
*ptr = cp;
Next phase in the PCI subsystem cleanup. - Move PCI core code to dev/pci. - Split bridge code out into separate modules. - Remove the descriptive strings from the bridge drivers. If you want to know what a device is, use pciconf. Add support for broadly identifying devices based on class/subclass, and for parsing a preloaded device identification database so that if you want to waste the memory, you can identify *anything* we know about. - Remove machine-dependant code from the core PCI code. APIC interrupt mapping is performed by shadowing the intline register in machine- dependant code. - Bring interrupt routing support to the Alpha (although many platforms don't yet support routing or mapping interrupts entirely correctly). This resulted in spamming <sys/bus.h> into more places than it really should have gone. - Put sys/dev on the kernel/modules include path. This avoids having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
return(1);
- Fix the device database parsing code so that it actually works. - Improve the formatting for devices identified by the database. - Fix the pcib_route_interrupt method definition, as an old version snuck in here somehow 8( - Remove a couple of the vendor/device IDs for PCI:ISA bridges which correctly identify themselves. Submitted by: peter
2000-12-09 09:15:38 +00:00
}
Next phase in the PCI subsystem cleanup. - Move PCI core code to dev/pci. - Split bridge code out into separate modules. - Remove the descriptive strings from the bridge drivers. If you want to know what a device is, use pciconf. Add support for broadly identifying devices based on class/subclass, and for parsing a preloaded device identification database so that if you want to waste the memory, you can identify *anything* we know about. - Remove machine-dependant code from the core PCI code. APIC interrupt mapping is performed by shadowing the intline register in machine- dependant code. - Bring interrupt routing support to the Alpha (although many platforms don't yet support routing or mapping interrupts entirely correctly). This resulted in spamming <sys/bus.h> into more places than it really should have gone. - Put sys/dev on the kernel/modules include path. This avoids having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
/* vendor entry? */
- Fix the device database parsing code so that it actually works. - Improve the formatting for devices identified by the database. - Fix the pcib_route_interrupt method definition, as an old version snuck in here somehow 8( - Remove a couple of the vendor/device IDs for PCI:ISA bridges which correctly identify themselves. Submitted by: peter
2000-12-09 09:15:38 +00:00
if (*cp != '\t' && sscanf(cp, "%x\t%80[^\n]", vendor, *desc) == 2)
Next phase in the PCI subsystem cleanup. - Move PCI core code to dev/pci. - Split bridge code out into separate modules. - Remove the descriptive strings from the bridge drivers. If you want to know what a device is, use pciconf. Add support for broadly identifying devices based on class/subclass, and for parsing a preloaded device identification database so that if you want to waste the memory, you can identify *anything* we know about. - Remove machine-dependant code from the core PCI code. APIC interrupt mapping is performed by shadowing the intline register in machine- dependant code. - Bring interrupt routing support to the Alpha (although many platforms don't yet support routing or mapping interrupts entirely correctly). This resulted in spamming <sys/bus.h> into more places than it really should have gone. - Put sys/dev on the kernel/modules include path. This avoids having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
break;
/* device entry? */
- Fix the device database parsing code so that it actually works. - Improve the formatting for devices identified by the database. - Fix the pcib_route_interrupt method definition, as an old version snuck in here somehow 8( - Remove a couple of the vendor/device IDs for PCI:ISA bridges which correctly identify themselves. Submitted by: peter
2000-12-09 09:15:38 +00:00
if (*cp == '\t' && sscanf(cp, "%x\t%80[^\n]", device, *desc) == 2)
Next phase in the PCI subsystem cleanup. - Move PCI core code to dev/pci. - Split bridge code out into separate modules. - Remove the descriptive strings from the bridge drivers. If you want to know what a device is, use pciconf. Add support for broadly identifying devices based on class/subclass, and for parsing a preloaded device identification database so that if you want to waste the memory, you can identify *anything* we know about. - Remove machine-dependant code from the core PCI code. APIC interrupt mapping is performed by shadowing the intline register in machine- dependant code. - Bring interrupt routing support to the Alpha (although many platforms don't yet support routing or mapping interrupts entirely correctly). This resulted in spamming <sys/bus.h> into more places than it really should have gone. - Put sys/dev on the kernel/modules include path. This avoids having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
break;
/* skip to next line */
- Fix the device database parsing code so that it actually works. - Improve the formatting for devices identified by the database. - Fix the pcib_route_interrupt method definition, as an old version snuck in here somehow 8( - Remove a couple of the vendor/device IDs for PCI:ISA bridges which correctly identify themselves. Submitted by: peter
2000-12-09 09:15:38 +00:00
while (*cp != '\n' && left > 0) {
cp++;
left--;
}
if (*cp == '\n') {
Next phase in the PCI subsystem cleanup. - Move PCI core code to dev/pci. - Split bridge code out into separate modules. - Remove the descriptive strings from the bridge drivers. If you want to know what a device is, use pciconf. Add support for broadly identifying devices based on class/subclass, and for parsing a preloaded device identification database so that if you want to waste the memory, you can identify *anything* we know about. - Remove machine-dependant code from the core PCI code. APIC interrupt mapping is performed by shadowing the intline register in machine- dependant code. - Bring interrupt routing support to the Alpha (although many platforms don't yet support routing or mapping interrupts entirely correctly). This resulted in spamming <sys/bus.h> into more places than it really should have gone. - Put sys/dev on the kernel/modules include path. This avoids having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
cp++;
- Fix the device database parsing code so that it actually works. - Improve the formatting for devices identified by the database. - Fix the pcib_route_interrupt method definition, as an old version snuck in here somehow 8( - Remove a couple of the vendor/device IDs for PCI:ISA bridges which correctly identify themselves. Submitted by: peter
2000-12-09 09:15:38 +00:00
left--;
}
}
/* skip to next line */
while (*cp != '\n' && left > 0) {
Next phase in the PCI subsystem cleanup. - Move PCI core code to dev/pci. - Split bridge code out into separate modules. - Remove the descriptive strings from the bridge drivers. If you want to know what a device is, use pciconf. Add support for broadly identifying devices based on class/subclass, and for parsing a preloaded device identification database so that if you want to waste the memory, you can identify *anything* we know about. - Remove machine-dependant code from the core PCI code. APIC interrupt mapping is performed by shadowing the intline register in machine- dependant code. - Bring interrupt routing support to the Alpha (although many platforms don't yet support routing or mapping interrupts entirely correctly). This resulted in spamming <sys/bus.h> into more places than it really should have gone. - Put sys/dev on the kernel/modules include path. This avoids having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
cp++;
- Fix the device database parsing code so that it actually works. - Improve the formatting for devices identified by the database. - Fix the pcib_route_interrupt method definition, as an old version snuck in here somehow 8( - Remove a couple of the vendor/device IDs for PCI:ISA bridges which correctly identify themselves. Submitted by: peter
2000-12-09 09:15:38 +00:00
left--;
o No need to print the vendor/device ID for things that matched succesfully. o ``<device name>'' versus ``device name'' for things that fall under nomatch. Reviewed by: dfr (in principle) Approved by: Baron von Hubbard
2000-02-22 21:44:39 +00:00
}
- Fix the device database parsing code so that it actually works. - Improve the formatting for devices identified by the database. - Fix the pcib_route_interrupt method definition, as an old version snuck in here somehow 8( - Remove a couple of the vendor/device IDs for PCI:ISA bridges which correctly identify themselves. Submitted by: peter
2000-12-09 09:15:38 +00:00
if (*cp == '\n' && left > 0)
cp++;
Next phase in the PCI subsystem cleanup. - Move PCI core code to dev/pci. - Split bridge code out into separate modules. - Remove the descriptive strings from the bridge drivers. If you want to know what a device is, use pciconf. Add support for broadly identifying devices based on class/subclass, and for parsing a preloaded device identification database so that if you want to waste the memory, you can identify *anything* we know about. - Remove machine-dependant code from the core PCI code. APIC interrupt mapping is performed by shadowing the intline register in machine- dependant code. - Bring interrupt routing support to the Alpha (although many platforms don't yet support routing or mapping interrupts entirely correctly). This resulted in spamming <sys/bus.h> into more places than it really should have gone. - Put sys/dev on the kernel/modules include path. This avoids having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
*ptr = cp;
return(0);
}
static char *
pci_describe_device(device_t dev)
{
int vendor, device;
char *desc, *vp, *dp, *line;
desc = vp = dp = NULL;
/*
* If we have no vendor data, we can't do anything.
*/
if (pci_vendordata == NULL)
goto out;
/*
* Scan the vendor data looking for this device
*/
line = pci_vendordata;
if ((vp = malloc(80, M_DEVBUF, M_NOWAIT)) == NULL)
goto out;
for (;;) {
if (pci_describe_parse_line(&line, &vendor, &device, &vp))
goto out;
if (vendor == pci_get_vendor(dev))
break;
o No need to print the vendor/device ID for things that matched succesfully. o ``<device name>'' versus ``device name'' for things that fall under nomatch. Reviewed by: dfr (in principle) Approved by: Baron von Hubbard
2000-02-22 21:44:39 +00:00
}
Next phase in the PCI subsystem cleanup. - Move PCI core code to dev/pci. - Split bridge code out into separate modules. - Remove the descriptive strings from the bridge drivers. If you want to know what a device is, use pciconf. Add support for broadly identifying devices based on class/subclass, and for parsing a preloaded device identification database so that if you want to waste the memory, you can identify *anything* we know about. - Remove machine-dependant code from the core PCI code. APIC interrupt mapping is performed by shadowing the intline register in machine- dependant code. - Bring interrupt routing support to the Alpha (although many platforms don't yet support routing or mapping interrupts entirely correctly). This resulted in spamming <sys/bus.h> into more places than it really should have gone. - Put sys/dev on the kernel/modules include path. This avoids having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
if ((dp = malloc(80, M_DEVBUF, M_NOWAIT)) == NULL)
goto out;
for (;;) {
- Fix the device database parsing code so that it actually works. - Improve the formatting for devices identified by the database. - Fix the pcib_route_interrupt method definition, as an old version snuck in here somehow 8( - Remove a couple of the vendor/device IDs for PCI:ISA bridges which correctly identify themselves. Submitted by: peter
2000-12-09 09:15:38 +00:00
if (pci_describe_parse_line(&line, &vendor, &device, &dp)) {
*dp = 0;
break;
}
Next phase in the PCI subsystem cleanup. - Move PCI core code to dev/pci. - Split bridge code out into separate modules. - Remove the descriptive strings from the bridge drivers. If you want to know what a device is, use pciconf. Add support for broadly identifying devices based on class/subclass, and for parsing a preloaded device identification database so that if you want to waste the memory, you can identify *anything* we know about. - Remove machine-dependant code from the core PCI code. APIC interrupt mapping is performed by shadowing the intline register in machine- dependant code. - Bring interrupt routing support to the Alpha (although many platforms don't yet support routing or mapping interrupts entirely correctly). This resulted in spamming <sys/bus.h> into more places than it really should have gone. - Put sys/dev on the kernel/modules include path. This avoids having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
if (vendor != -1) {
*dp = 0;
break;
}
if (device == pci_get_device(dev))
break;
Implement the BUS_PROBE_NOMATCH method for the PCI bus. This function is called for each device for which no driver was found. Output is similar to the eisa_probe_nomatch() function but with the added benefit of displaying the assigned IRQ (since PCI gives us this information up front.) Output is like so: pci0: unknown card CPQ0508 (vendor=0x0e11, dev=0x0508) at 11.0 irq 9 pci0: unknown card DFZ0508 (vendor=0x10da, dev=0x0508) at 11.0 irq 9 pci0: unknown card DBL0508 (vendor=0x104c, dev=0x0508) at 11.0 irq 9 pci0: unknown card DDM0011 (vendor=0x108d, dev=0x0011) at 11.0 irq 9 I'm not happy with the 3 lines of macro cruft that got added but I consider it a temporary annoyance as those bits will be moved to some place where PCI, EISA and ISAPNP code will be able to use them. (Not surprisingly, this message is longer than the code in question.) Reviewed by: peter, dfr
1999-07-27 04:28:14 +00:00
}
- Fix the device database parsing code so that it actually works. - Improve the formatting for devices identified by the database. - Fix the pcib_route_interrupt method definition, as an old version snuck in here somehow 8( - Remove a couple of the vendor/device IDs for PCI:ISA bridges which correctly identify themselves. Submitted by: peter
2000-12-09 09:15:38 +00:00
if (dp[0] == '\0')
snprintf(dp, 80, "0x%x", pci_get_device(dev));
Next phase in the PCI subsystem cleanup. - Move PCI core code to dev/pci. - Split bridge code out into separate modules. - Remove the descriptive strings from the bridge drivers. If you want to know what a device is, use pciconf. Add support for broadly identifying devices based on class/subclass, and for parsing a preloaded device identification database so that if you want to waste the memory, you can identify *anything* we know about. - Remove machine-dependant code from the core PCI code. APIC interrupt mapping is performed by shadowing the intline register in machine- dependant code. - Bring interrupt routing support to the Alpha (although many platforms don't yet support routing or mapping interrupts entirely correctly). This resulted in spamming <sys/bus.h> into more places than it really should have gone. - Put sys/dev on the kernel/modules include path. This avoids having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
if ((desc = malloc(strlen(vp) + strlen(dp) + 3, M_DEVBUF, M_NOWAIT)) != NULL)
- Fix the device database parsing code so that it actually works. - Improve the formatting for devices identified by the database. - Fix the pcib_route_interrupt method definition, as an old version snuck in here somehow 8( - Remove a couple of the vendor/device IDs for PCI:ISA bridges which correctly identify themselves. Submitted by: peter
2000-12-09 09:15:38 +00:00
sprintf(desc, "%s, %s", vp, dp);
Next phase in the PCI subsystem cleanup. - Move PCI core code to dev/pci. - Split bridge code out into separate modules. - Remove the descriptive strings from the bridge drivers. If you want to know what a device is, use pciconf. Add support for broadly identifying devices based on class/subclass, and for parsing a preloaded device identification database so that if you want to waste the memory, you can identify *anything* we know about. - Remove machine-dependant code from the core PCI code. APIC interrupt mapping is performed by shadowing the intline register in machine- dependant code. - Bring interrupt routing support to the Alpha (although many platforms don't yet support routing or mapping interrupts entirely correctly). This resulted in spamming <sys/bus.h> into more places than it really should have gone. - Put sys/dev on the kernel/modules include path. This avoids having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
out:
if (vp != NULL)
free(vp, M_DEVBUF);
if (dp != NULL)
free(dp, M_DEVBUF);
return(desc);
Implement the BUS_PROBE_NOMATCH method for the PCI bus. This function is called for each device for which no driver was found. Output is similar to the eisa_probe_nomatch() function but with the added benefit of displaying the assigned IRQ (since PCI gives us this information up front.) Output is like so: pci0: unknown card CPQ0508 (vendor=0x0e11, dev=0x0508) at 11.0 irq 9 pci0: unknown card DFZ0508 (vendor=0x10da, dev=0x0508) at 11.0 irq 9 pci0: unknown card DBL0508 (vendor=0x104c, dev=0x0508) at 11.0 irq 9 pci0: unknown card DDM0011 (vendor=0x108d, dev=0x0011) at 11.0 irq 9 I'm not happy with the 3 lines of macro cruft that got added but I consider it a temporary annoyance as those bits will be moved to some place where PCI, EISA and ISAPNP code will be able to use them. (Not surprisingly, this message is longer than the code in question.) Reviewed by: peter, dfr
1999-07-27 04:28:14 +00:00
}
Next phase in the PCI subsystem cleanup. - Move PCI core code to dev/pci. - Split bridge code out into separate modules. - Remove the descriptive strings from the bridge drivers. If you want to know what a device is, use pciconf. Add support for broadly identifying devices based on class/subclass, and for parsing a preloaded device identification database so that if you want to waste the memory, you can identify *anything* we know about. - Remove machine-dependant code from the core PCI code. APIC interrupt mapping is performed by shadowing the intline register in machine- dependant code. - Bring interrupt routing support to the Alpha (although many platforms don't yet support routing or mapping interrupts entirely correctly). This resulted in spamming <sys/bus.h> into more places than it really should have gone. - Put sys/dev on the kernel/modules include path. This avoids having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
static int
Fixed the type of some ivar access functions. Ivars have type uintptr_t, not u_long. On i386's with 64-bit longs, returning u_longs indirectly in (more than) the space reserved for uintptr_t's tended to corrupt the previous frame pointer in the stack frame, so it was not easy to debug. The type mismatches are hidden by the bogus cast in DEVMETHOD().
2000-04-30 10:01:56 +00:00
pci_read_ivar(device_t dev, device_t child, int which, uintptr_t *result)
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
{
struct pci_devinfo *dinfo;
pcicfgregs *cfg;
dinfo = device_get_ivars(child);
cfg = &dinfo->cfg;
switch (which) {
case PCI_IVAR_SUBVENDOR:
*result = cfg->subvendor;
break;
case PCI_IVAR_SUBDEVICE:
*result = cfg->subdevice;
break;
case PCI_IVAR_VENDOR:
*result = cfg->vendor;
break;
case PCI_IVAR_DEVICE:
*result = cfg->device;
break;
case PCI_IVAR_DEVID:
*result = (cfg->device << 16) | cfg->vendor;
break;
case PCI_IVAR_CLASS:
*result = cfg->baseclass;
break;
case PCI_IVAR_SUBCLASS:
*result = cfg->subclass;
break;
case PCI_IVAR_PROGIF:
*result = cfg->progif;
break;
case PCI_IVAR_REVID:
*result = cfg->revid;
break;
case PCI_IVAR_INTPIN:
*result = cfg->intpin;
break;
case PCI_IVAR_IRQ:
*result = cfg->intline;
break;
case PCI_IVAR_BUS:
*result = cfg->bus;
break;
case PCI_IVAR_SLOT:
*result = cfg->slot;
break;
case PCI_IVAR_FUNCTION:
*result = cfg->func;
break;
case PCI_IVAR_SECONDARYBUS:
*result = cfg->secondarybus;
break;
case PCI_IVAR_SUBORDINATEBUS:
*result = cfg->subordinatebus;
break;
default:
return ENOENT;
}
return 0;
}
static int
pci_write_ivar(device_t dev, device_t child, int which, uintptr_t value)
{
struct pci_devinfo *dinfo;
pcicfgregs *cfg;
dinfo = device_get_ivars(child);
cfg = &dinfo->cfg;
switch (which) {
case PCI_IVAR_SUBVENDOR:
case PCI_IVAR_SUBDEVICE:
case PCI_IVAR_VENDOR:
case PCI_IVAR_DEVICE:
case PCI_IVAR_DEVID:
case PCI_IVAR_CLASS:
case PCI_IVAR_SUBCLASS:
case PCI_IVAR_PROGIF:
case PCI_IVAR_REVID:
case PCI_IVAR_INTPIN:
case PCI_IVAR_IRQ:
case PCI_IVAR_BUS:
case PCI_IVAR_SLOT:
case PCI_IVAR_FUNCTION:
return EINVAL; /* disallow for now */
case PCI_IVAR_SECONDARYBUS:
cfg->secondarybus = value;
break;
case PCI_IVAR_SUBORDINATEBUS:
cfg->subordinatebus = value;
break;
default:
return ENOENT;
}
return 0;
}
* Implement bus_set/get/delete_resource for pci. * Change the hack used on the alpha for mapping devices into DENSE or BWX memory spaces to a simpler one. Its still a hack and should be a seperate api to explicitly map the resource. * Add $FreeBSD$ as necessary.
1999-10-14 21:38:33 +00:00
static struct resource *
pci_alloc_resource(device_t dev, device_t child, int type, int *rid,
u_long start, u_long end, u_long count, u_int flags)
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
{
* Implement bus_set/get/delete_resource for pci. * Change the hack used on the alpha for mapping devices into DENSE or BWX memory spaces to a simpler one. Its still a hack and should be a seperate api to explicitly map the resource. * Add $FreeBSD$ as necessary.
1999-10-14 21:38:33 +00:00
struct pci_devinfo *dinfo = device_get_ivars(child);
struct resource_list *rl = &dinfo->resources;
When a pci device hasn't had an interrupt routed to it (signified by the intline of 255) go ahead and route the interrupt when we allocate an interrupt. Submitted by: msmith
2000-10-16 07:24:00 +00:00
pcicfgregs *cfg = &dinfo->cfg;
/*
* Perform lazy resource allocation
*
* XXX add support here for SYS_RES_IOPORT and SYS_RES_MEMORY
*/
if (device_get_parent(child) == dev) {
Next phase in the PCI subsystem cleanup. - Move PCI core code to dev/pci. - Split bridge code out into separate modules. - Remove the descriptive strings from the bridge drivers. If you want to know what a device is, use pciconf. Add support for broadly identifying devices based on class/subclass, and for parsing a preloaded device identification database so that if you want to waste the memory, you can identify *anything* we know about. - Remove machine-dependant code from the core PCI code. APIC interrupt mapping is performed by shadowing the intline register in machine- dependant code. - Bring interrupt routing support to the Alpha (although many platforms don't yet support routing or mapping interrupts entirely correctly). This resulted in spamming <sys/bus.h> into more places than it really should have gone. - Put sys/dev on the kernel/modules include path. This avoids having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
/*
* If device doesn't have an interrupt routed, and is deserving of
* an interrupt, try to assign it one.
*/
if ((type == SYS_RES_IRQ) && (cfg->intline == 255) && (cfg->intpin != 0)) {
cfg->intline = PCIB_ROUTE_INTERRUPT(device_get_parent(dev), child,
cfg->intpin);
When a pci device hasn't had an interrupt routed to it (signified by the intline of 255) go ahead and route the interrupt when we allocate an interrupt. Submitted by: msmith
2000-10-16 07:24:00 +00:00
if (cfg->intline != 255) {
Write the routed interrupt back to PCI configuration space.
2000-10-19 08:07:23 +00:00
pci_write_config(child, PCIR_INTLINE, cfg->intline, 1);
When a pci device hasn't had an interrupt routed to it (signified by the intline of 255) go ahead and route the interrupt when we allocate an interrupt. Submitted by: msmith
2000-10-16 07:24:00 +00:00
resource_list_add(rl, SYS_RES_IRQ, 0,
Next phase in the PCI subsystem cleanup. - Move PCI core code to dev/pci. - Split bridge code out into separate modules. - Remove the descriptive strings from the bridge drivers. If you want to know what a device is, use pciconf. Add support for broadly identifying devices based on class/subclass, and for parsing a preloaded device identification database so that if you want to waste the memory, you can identify *anything* we know about. - Remove machine-dependant code from the core PCI code. APIC interrupt mapping is performed by shadowing the intline register in machine- dependant code. - Bring interrupt routing support to the Alpha (although many platforms don't yet support routing or mapping interrupts entirely correctly). This resulted in spamming <sys/bus.h> into more places than it really should have gone. - Put sys/dev on the kernel/modules include path. This avoids having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
cfg->intline, cfg->intline, 1);
When a pci device hasn't had an interrupt routed to it (signified by the intline of 255) go ahead and route the interrupt when we allocate an interrupt. Submitted by: msmith
2000-10-16 07:24:00 +00:00
}
}
}
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
* Implement bus_set/get/delete_resource for pci. * Change the hack used on the alpha for mapping devices into DENSE or BWX memory spaces to a simpler one. Its still a hack and should be a seperate api to explicitly map the resource. * Add $FreeBSD$ as necessary.
1999-10-14 21:38:33 +00:00
return resource_list_alloc(rl, dev, child, type, rid,
start, end, count, flags);
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
}
Reduce code duplication by using the GET_RESOURCE_LIST bus method and related generic resource_list management functions. I'll deal with the EISA bits later. Not objected to by: new-bus
2000-11-28 07:12:12 +00:00
static void
pci_delete_resource(device_t dev, device_t child, int type, int rid)
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
{
Reduce code duplication by using the GET_RESOURCE_LIST bus method and related generic resource_list management functions. I'll deal with the EISA bits later. Not objected to by: new-bus
2000-11-28 07:12:12 +00:00
printf("pci_delete_resource: PCI resources can not be deleted\n");
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
}
Reduce code duplication by using the GET_RESOURCE_LIST bus method and related generic resource_list management functions. I'll deal with the EISA bits later. Not objected to by: new-bus
2000-11-28 07:12:12 +00:00
static struct resource_list *
pci_get_resource_list (device_t dev, device_t child)
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
{
Reduce code duplication by using the GET_RESOURCE_LIST bus method and related generic resource_list management functions. I'll deal with the EISA bits later. Not objected to by: new-bus
2000-11-28 07:12:12 +00:00
struct pci_devinfo * dinfo = device_get_ivars(child);
struct resource_list * rl = &dinfo->resources;
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
Reduce code duplication by using the GET_RESOURCE_LIST bus method and related generic resource_list management functions. I'll deal with the EISA bits later. Not objected to by: new-bus
2000-11-28 07:12:12 +00:00
if (!rl)
return (NULL);
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
Reduce code duplication by using the GET_RESOURCE_LIST bus method and related generic resource_list management functions. I'll deal with the EISA bits later. Not objected to by: new-bus
2000-11-28 07:12:12 +00:00
return (rl);
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
}
static u_int32_t
pci_read_config_method(device_t dev, device_t child, int reg, int width)
{
struct pci_devinfo *dinfo = device_get_ivars(child);
pcicfgregs *cfg = &dinfo->cfg;
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
return PCIB_READ_CONFIG(device_get_parent(dev),
cfg->bus, cfg->slot, cfg->func,
reg, width);
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
}
static void
pci_write_config_method(device_t dev, device_t child, int reg,
u_int32_t val, int width)
{
struct pci_devinfo *dinfo = device_get_ivars(child);
pcicfgregs *cfg = &dinfo->cfg;
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
PCIB_WRITE_CONFIG(device_get_parent(dev),
cfg->bus, cfg->slot, cfg->func,
reg, val, width);
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
}
static int
pci_modevent(module_t mod, int what, void *arg)
{
switch (what) {
case MOD_LOAD:
Argh, don't clobber the pci device list if there are multiple busses! (An AGP counts as a PCI bus, it seems...) This stopped 'pciconf -l' from working on AGP or PCI->PCI bridge systems.
1999-05-09 15:54:04 +00:00
STAILQ_INIT(&pci_devq);
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
break;
case MOD_UNLOAD:
break;
}
return 0;
}
static device_method_t pci_methods[] = {
/* Device interface */
* Completely rewrite the alpha busspace to hide the implementation from the drivers. * Remove legacy inx/outx support from chipset and replace with macros which call busspace. * Rework pci config accesses to route through the pcib device instead of calling a MD function directly. With these changes it is possible to cleanly support machines which have more than one independantly numbered PCI busses. As a bonus, the new busspace implementation should be measurably faster than the old one.
2000-08-28 21:48:13 +00:00
DEVMETHOD(device_probe, pci_probe),
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
DEVMETHOD(device_attach, bus_generic_attach),
DEVMETHOD(device_shutdown, bus_generic_shutdown),
Add missing suspend/resume methods.
1999-05-10 17:56:23 +00:00
DEVMETHOD(device_suspend, bus_generic_suspend),
DEVMETHOD(device_resume, bus_generic_resume),
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
/* Bus interface */
DEVMETHOD(bus_print_child, pci_print_child),
Case matters. DEv_METHOD to DEV_METHOD.
1999-07-27 05:08:36 +00:00
DEVMETHOD(bus_probe_nomatch, pci_probe_nomatch),
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
DEVMETHOD(bus_read_ivar, pci_read_ivar),
DEVMETHOD(bus_write_ivar, pci_write_ivar),
DEVMETHOD(bus_driver_added, bus_generic_driver_added),
DEVMETHOD(bus_setup_intr, bus_generic_setup_intr),
DEVMETHOD(bus_teardown_intr, bus_generic_teardown_intr),
Reduce code duplication by using the GET_RESOURCE_LIST bus method and related generic resource_list management functions. I'll deal with the EISA bits later. Not objected to by: new-bus
2000-11-28 07:12:12 +00:00
DEVMETHOD(bus_get_resource_list,pci_get_resource_list),
DEVMETHOD(bus_set_resource, bus_generic_rl_set_resource),
DEVMETHOD(bus_get_resource, bus_generic_rl_get_resource),
* Implement bus_set/get/delete_resource for pci. * Change the hack used on the alpha for mapping devices into DENSE or BWX memory spaces to a simpler one. Its still a hack and should be a seperate api to explicitly map the resource. * Add $FreeBSD$ as necessary.
1999-10-14 21:38:33 +00:00
DEVMETHOD(bus_delete_resource, pci_delete_resource),
Reduce code duplication by using the GET_RESOURCE_LIST bus method and related generic resource_list management functions. I'll deal with the EISA bits later. Not objected to by: new-bus
2000-11-28 07:12:12 +00:00
DEVMETHOD(bus_alloc_resource, pci_alloc_resource),
DEVMETHOD(bus_release_resource, bus_generic_rl_release_resource),
DEVMETHOD(bus_activate_resource, bus_generic_activate_resource),
DEVMETHOD(bus_deactivate_resource, bus_generic_deactivate_resource),
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
/* PCI interface */
DEVMETHOD(pci_read_config, pci_read_config_method),
DEVMETHOD(pci_write_config, pci_write_config_method),
{ 0, 0 }
};
static driver_t pci_driver = {
"pci",
pci_methods,
Next phase in the PCI subsystem cleanup. - Move PCI core code to dev/pci. - Split bridge code out into separate modules. - Remove the descriptive strings from the bridge drivers. If you want to know what a device is, use pciconf. Add support for broadly identifying devices based on class/subclass, and for parsing a preloaded device identification database so that if you want to waste the memory, you can identify *anything* we know about. - Remove machine-dependant code from the core PCI code. APIC interrupt mapping is performed by shadowing the intline register in machine- dependant code. - Bring interrupt routing support to the Alpha (although many platforms don't yet support routing or mapping interrupts entirely correctly). This resulted in spamming <sys/bus.h> into more places than it really should have gone. - Put sys/dev on the kernel/modules include path. This avoids having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
0, /* no softc */
Bring the 'new-bus' to the i386. This extensively changes the way the i386 platform boots, it is no longer ISA-centric, and is fully dynamic. Most old drivers compile and run without modification via 'compatability shims' to enable a smoother transition. eisa, isapnp and pccard* are not yet using the new resource manager. Once fully converted, all drivers will be loadable, including PCI and ISA. (Some other changes appear to have snuck in, including a port of Soren's ATA driver to the Alpha. Soren, back this out if you need to.) This is a checkpoint of work-in-progress, but is quite functional. The bulk of the work was done over the last few years by Doug Rabson and Garrett Wollman. Approved by: core
1999-04-16 21:22:55 +00:00
};
DRIVER_MODULE(pci, pcib, pci_driver, pci_devclass, pci_modevent, 0);
Allow PCI busses to be connected to host bridges detected by ACPI as well.
2000-10-28 07:04:07 +00:00
DRIVER_MODULE(pci, acpi_pcib, pci_driver, pci_devclass, pci_modevent, 0);
Reference in New Issue Copy Permalink