This is now two seperate drivers that support (I think) all of Stallions's
range, including the high performance intelligent cards, and their older
cards.
Submitted by: Greg Ungerer (gerg@stallion.oz.au)
(User-mode parts and patches to follow)
boards by Greg Ungerer (gerg@stallion.oz.au). (v0.0.1 alpha)
This is a multiple import of all revisions available to build up
a history.
This driver supports only some of the Stallion range, in particular, not
the highly intelligent cards. That comes in shortly.
Submitted by: Greg Ungerer (gerg@stallion.oz.au)
Macroize locore.s' page table setup even more, now it's almost readable.
Rename PG_U to PG_A (so that I can...)
Rename PG_u to PG_U. "PG_u" was just too ugly...
Remove some unused vars in pmap.c
Remove PG_KR and PG_KW
Remove SSIZE
Remove SINCR
Remove BTOPKERNBASE
This concludes my spring cleaning, modulus any bug fixes for messes I
have made on the way.
(Funny to be back here in pmap.c, that's where my first significant
contribution to 386BSD was... :-)
very busy servers (eg: news, web). This is an interaction between
embryonic processes that have not yet finished forking, and happen to
cause the kernel VM space to grow, hitting the uninitialised variable.
It was possible for this to strike at any time, depending on the size of
your kernel and load patterns. One machine had paniced occasionally
when cron launches a job since before the 2.1 release.
If you had "options DIAGNOSTIC", you may have seen references to bogus
addresses like 0xdeadc142 and the like.
This is a minimal change to fix the problem, it will probably be done
better by reordering p_vmspace to be in the startzero section, but it
becomes harder to validate then.
It's been vulnerable since pmap.c rev 1.40 (Jan 9, 1995), so it's been a
cause of problems since well before 2.0.5. This was when the merged
VM/buffer cache and the dynamic growing kernel VM space were first
committed. This probably fixes a few of PR's.
time. The results are currently ignored unless certain temporary options
are used.
Added sysctls to support reading and writing the clock frequency variables
(not the frequencies themselves). Writing is supposed to atomically
adjust all related variables.
machdep.c:
Fixed spelling of a function name in a comment so that I can log this
message which should have been with the previous commit.
Initialize `cpu_class' earlier so that it can be used in startrtclock()
instead of in calibrate_cyclecounter() (which no longer exists).
Removed range checking of `cpu'. It is always initialized to CPU_XXX
so it is less likely to be out of bounds than most variables.
clock.h:
Removed I586_CYCLECTR(). Use rdtsc() instead.
clock.c:
TIMER_FREQ is now a variable timer_freq that defaults to the old value of
TIMER_FREQ. #define'ing TIMER_FREQ should still work and may be the best
way of setting the frequency.
Calibration involves counting cycles while watching the RTC for one second.
This gives values correct to within (a few ppm) + (the innaccuracy of the
RTC) on my systems.
Fixed longstanding namespace convolution involving rune_t vs wchar_t...
This change breaks similar (but more convoluted) convolutions in the
stddef.h in gcc distributions. Ports of gcc should avoid using the
gcc headers.
page dir+table index.
pmap.h: remove NUPDE, it was wrong and not used. Sanitize KSTKPTEOFF.
vmparam.h: Calculate virtual addr from PDI+PTI from pmap.h rather than
using magic math. Remove UPDT, not used.
Ensure that queued commands are not touched by the abort code by setting
the SCB status to indicate what queue it is in.
Fix deadlocks when using SCB paging by using SCBs from the assigned_scbs
queue or an SCB that completed during the same interrupt if needed.
Don't ever use insl to pull SCBs from any of the controllers. You can
only do 8bit PIO reads. This only affected SCB paging.
With this checkin, SCB paging works quite a bit better, but I still have
some problems with it that may be caused by a firmware problem in my
PD1800s. It seems that using a tag number higher than the maximum number
of tags allowed by the device, confuses it. For example, if I queue
two commands, tagged 3 and 36, it never reconnects for tag 36.
mapped to semi-random place(s) depending on the content(s) of physical
address 0xA0000. This was fatal at least on my system with a some
memory-mapped devices. Console syscons somehow wasn't affected. It
bogusly hardcodes the address. Sigh.
should be <= than subordinate, not the other way around.
They are both true if the bridge is not cascaded (i.e., twin-channel
scsi/e-net adapters won't be affected by this bug), which is probably why
it was unnoticed until today.
implementations, and synchronizes us with the Nomad's latest code.
This code is based on the Nomad code, but heavily hacked by me.
Reviewed by: phk
Submitted by: the 'Nomads'
in place device drivers can now register power-down/power-up routines so
that we can use common routines to power-up/power-down cards for
insert/removals, suspend/resume, etc..
Reviewed by: phk
Submitted by: the 'Nomads'
strange results. This bug has been in here for a loooong time.
Many thanks to Pedro Salenbauch for finding this.
Submitted by: pedrosal@nce.ufrj.br (Pedro Salenbauch)
regarding apm to LINT
- Disabled the statistics clock on machines which have an APM BIOS and
have the options "APM_BROKEN_STATCLOCK" enabled (which is default
in GENERIC now)
- move around some of the code in clock.c dealing with the rtc to make
it more obvios the effects of disabling the statistics clock
Reviewed by: bde
8 not the 18 I was using during some of my own testing. Ooops.
For those that want to change the number for experimentation, you can
set the value on line 1553 of this file.
One of the manifiestations of the problem includes the -4 RSS problem
in ps.
Reviewed by: dyson
Submitted by: Stephen McKay <syssgm@devetir.qld.gov.au>
(Rev E or greater), aic7850, aic7860, aic7870, and aic7880 controllers.
SCB paging is enabled with the option "AHC_SCBPAGING_ENABLE". Full
comments on the algorithm are at the top of i386/scsi/aic7xxx.c.
options "AHC_TAGENABLE" and "AHC_QUEUE_FULL" have been removed. The
default is 4 tags without SCB paging, 8 with.
Clear the SCSIRSTI bit after throwing a bus reset. Some cards seem to get
confused otherwise.
Handle SCSIRSTI interrupts before checking to see if there is a valid
SCB in use since this can happen. (Clears PR# i386/1123)
Clean up the way we determine the number of SCBs on the card
(courtesy of Dan Eischen).
Guard against attempts to negotiate wide to a narrow controller.
Fix some comments.
Update my copyrights.
aic7770 >= Rev E, aic7850, aic7860, aic7870, and ai7880 based controllers.
Make findSCB safer for non-tagged commands when tagged commands are
active on the controller. The symptoms of this problem were
"Overlapped commands attempted" messages during error recovery
attempts.
Compact scratch ram usage. This leaves 8 bytes free for future use.
Clean up some comments.
aic7xxx_reg.h:
Update my copyright.
and B_READ before writing. This was was fatal. They also broke the
clearing of B_INVAL before doing i/o. This didn't actually matter.
Submitted by: mostly by joerg
compatibility slice. They were forgotten on last-close and then
creating them on first-open failed.
Devfs entries for slices other than the one containing the root file
system are still invisible unless you open a non-devfs inode on the
slice.
- More code cleanups
- #ifdef DEBUG debugging code
- More consistant printfs
- Better handling of the apm_int() assembly code (mostly from Bruce Evans)
Reviewed by: bde
1) Require all callers to pass a valid route pointer to ip_output()
so that we don't have to check and allocate one off the stack
as was done before. This eliminates one test and some stack
bloat from the common (UDP and TCP) case.
2) Perform the IP header checksum in-line if it's of the usual length.
This results in about a 5% speed-up in my packet-generation test.
3) Use ip_vhl field rather than ip_v and ip_hl bitfields.
forked child to be dissociated from the parent).
Cleanup fork1(), implement vfork() and fork() in terms of rfork() flags.
Remove RFENVG, RFNOTEG, RFCNAMEG, RFCENVG which are Plan9 specific and cannot
possibly be implemented in FreeBSD.
Renumbered the flags to make up for the removal of the above flags.
Reviewed by: peter, smpatel
Submitted by: Mike Grupenhoff <kashmir@umiacs.umd.edu>
1) Set the persist timer to help time-out connections in the CLOSING state.
2) Honor the keep-alive timer in the CLOSING state.
This fixes problems with connections getting "stuck" due to incompletion
of the final connection shutdown which can be a BIG problem on busy WWW
servers.
- 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.
assignment to avoid one bug and several pessimizations.
In the old version, gcc-2.6.3 (i386 version) generates 16 bytes
of static data and copies it using 4 4-byte load-stores. gcc-2.7.2
generates 2 1-byte stores and calls memset() to zero 14 bytes.
Linking fails because memset() doesn't exist in the kernel.
In both versions, the 2 bytes stored directly are all that is
actually used unless the null padding at the end is used, since
the 3 4-byte words in the middle are initialized again by struct
assignment. These words are misaligned. gcc generates misaligned
load-stores for (small) misaligned struct copies.
Submitted by: Bruce Evans
an application #defines `noreturn'.
Changed one instance of `const' similarly. This is less like to be a
problem since applications shouldn't #define `const', and the common
hack of #defining `const' as nothing gives harmless (?) null attributes
instead of syntax errors.
Fixed comments on #endifs to match code.
common labels for LINT. There are still some common declarations for the
!KERNEL case in tcp_debug.h and spx_debug.h. trpt depends on the ones in
tcp_debug.h.
common labels for LINT. There are still some common declarations for the
!KERNEL case in tcp_debug.h and spx_debug.h. trpt depends on the ones in
tcp_debug.h.
bug only affected FPU emulators. It might have caused bogus FPU states
in core dumps and in the child pcb after a fork. Emulated FPU states
in core dumps don't work for other reasons, and the child FPU state
is reinitialized by exec, so the problem might not have caused any
noticeable affects.
Cleaned up #includes.
was duplicated until the canq filled up, and write() normally returned 0.
This case is apparently rare. It was reported for Jove's shell buffer in
PR 1130.
Clarified GUS DMA Settings.
Other misc. changes.
This should hold us over until I can finish cleaning up TASD, and finish
reintegrating all of the FreeBSD changes to the sound driver. At that time
this document will be removed, and it's information moved to the handbook.
header for use in decompressing subsequant packets. If cslip gets garbage
(such as what happens when there is a port speed mismatch or modem line
noise), it will occasionally mistake the packet as a valid uncompressed
packet. When it tries to save off the header, it doesn't bother to check
for the validity of the header length and will happily clobber not only
the cslip data structure, but parts of other kernel memory that happens
to follow it...causing, ahem, undesired behavior.
is needed because of the vm_fault used to bring the page table page
for the kernel stack (UPAGES) back in. The consequence of the
previous incorrect change was a system hang.
channel at the same time. The functions isa_dma_acquire() and
isa_dma_release() should be used in all ISA drivers which call isa_dmastart().
This can be used more generally to register the usage of DMA channels in
any driver, but it is required for drivers using isa_dmastart() and friends.
Clean up sanity checks, error messages, etc.
Remove isa_dmadone_nobounce(), it is no longer needed
Reviewed by: bde
Moved declaration of bootverbose to a better place. It isn't
machine-dependent.
proc.h:
Moved declaration of cpu_fork() to a better place. Only its
implementation is machine-dependent.
hd essentially wired the FreeBSD drive number to 0 without changing
the BIOS drive number. Now the numbers can be specified independently.
Replaced the BOOT_HD compile time flag with with BOOT_HD_BIAS. Defining
the new flag as 1 should give the same behaviour as defining the old
flag as anything. I haven't tested defining these flags.
This fixes a panic that occurs when ifconfig ioctl(s) were interrupted
by IP traffic at the wrong time - resulting in a NULL pointer dereference.
This was originally noticed on a FreeBSD 1.0 system, but the problem still
exists in current sources.
#include <i386/include/clock.h> to get sysbeep() prototype
pcic.c:
add appropriate #ifdef around a prototype to quiet GCC because
fn decl. is also #ifdef'd.
They don't have BUSY de-asserted by the time they ACK (and thus cause
an interrupt). The workaround is to try seeing if the BUSY will be
de-asserted soon, and if not, to use an incremental backoff and
semi-polled mode instead of the fixed timeout with 1/2 s we've been
using previously (that caused the printer to run really slooow).
Printers that have been working previously should not be affected by
this.
keepalive on all tcp sessions. Setsockopt(2) cannot override this setting.
Maybe another one is needed that just changes the default for SO_KEEPALIVE ?
Requested by: Joe Greco <jgreco@brasil.moneng.mei.com>
in a suboptimal manner. I had also noticed some panics that appeared
to be at least superficially caused by this problem. Also, included
are some minor mods to support more general handling of page table page
faulting. More details in a future commit.
Changed DEVFS structure devfs_token so that adding the devices is
a simple matter of a 4 line for loop versus 16 lines of code
Reviewed by: julian@freebsd.org
Always delay using one inb(0x84) after each i/o in rtcin() - don't
do this conditional on the bogus option DUMMY_NOPS not being defined.
If you want an optionally slightly faster rtcin() again, then inline
it and use a better named option or sysctl variable. It only needs
to be fast in rtcintr().
aesthetics of using the 4.4 queue macros without paying undo space or time
in scenartios where a singly-linked list works fine.
From queue.h:
/*
* A singly-linked list is headed by a single forward pointer. The elements
* are singly linked for minimum space and pointer manipulation overhead at
* the expense of O(n) removal for arbitrary elements. New elements can be
* added to the list after an existing element or at the head of the list.
* Elements being removed from the head of the list should use the explicit
* macro for this purpose for optimum efficiency. A singly-linked list may
* only be traversed in the forward direction. Singly-linked lists are ideal
* for applications with large datasets and few or no removals or for
* implementing a LIFO queue.
*
* A singly-linked tail queue is headed by a pair of pointers, one to the
* head of the list and the other to the tail of the list. The elements are
* singly linked for minimum space and pointer manipulation overhead at the
* expense of O(n) removal for arbitrary elements. New elements can be added
* to the list after an existing element, at the head of the list, or at the
* end of the list. Elements being removed from the head of the tail queue
* should use the explicit macro for this purpose for optimum efficiency.
* A singly-linked tail queue may only be traversed in the forward direction.
* Singly-linked tail queues are ideal for applications with large datasets
* and few or no removals or for implementing a FIFO queue.
*/
Use new XS_SELTIMEOUT error code for selection timeouts.
aic7870.c:
Move SCB walking code to aic7xxx.c and make it work for all card types.
The flag AHC_EXTSCB is no longer needed since the SCBs are walked in
all cases now.
QINCNT. The 7850 puts random garbage in the high bits and all my attempts
to determine the cause of this failed. This approach does seem to work
around the problem.
Go back to relying on the SCSIPERR interrupt instead of having the sequencer
interrupt at the beginning of ITloop after a parity error occured.
Determine the number of SCBs on a card automatically and base the qcntmask
on the number of SCBs.
Add entries for 11.4MHz, 8.8MHz, 8.0MHz, and 7.2MHz to ULTRA portion of
the syncrate table. They seem to work fine on the 2940UW I have here and
will allow more non-ultra devices (like my tape drive) to run sync while
the adapter is in ULTRA mode.
Return XS_SELTIMEOUT instead of XS_TIMEOUT for selection timeouts. I was
getting sick of waiting for the SCSI code to retry each non-existant unit
multiple times during boot and XS_SELTIMEOUT bypasses all retries.
Use new SLIST queue macros. This was inspired by NetBSD using TAILQs in
their SCSI drivers. For optimum cache hits, the free scb list should
be LIFO which is what the old and new code does. NetBSD implemented a
FIFO queue for some reason.
Spaces -> tabs.
Fix support for the aic7850 by looking only at the relavent bits of the
QINCNT. The 7850 puts random garbage in the high bits and all my attempts
to determine the cause of this failed. This approach does seem to work
around the problem.
Don't trust SCSIPERR to tell us when there is a parity error. On
some revs of the 7870 and the 7880, this bit follows the parity of
the current byte. Instead of using a SEQINT to tell the kernel,
re-enable the standard parity error interrupt since it seems to pause
the sequencer right at the time of the error which is the effect we were
looking for anyway.
aic7xxx_reg.h:
Remove PARITY_ERROR seqeuncer interrupt type, its no longer used.
Define QCOUNTMASK as the SRAM location for the mask to use on the
QINCNT register. QCOUNTMASK is determined by the number of SCBs
supported by the device we're working on.
aic7xxx_asm.c
Properly check the return value of fopen, and define the arg list
in getopt correctly.
Submitted by: Pete Bentley <pete@demon.net>
Make the SA_NODEFER handling more correct, previously if you called
sigaction to set a handler and had SA_NODEFER set, and manually masked
the signal itself in sa_mask, and when you read the settings back later,
you'd find SA_NODEFER incorrectly cleared.
Pointed out by: bde
the process's memory, it was possible for the procfs_domem() call to
return a residual leftover, but with no errno. Since this is no good for
ptrace which ignored the the residual, remap a leftover amount into an
errno rather than fooling the caller into thinking it was successful when
in fact it was not.
Submitted by: bde (a very long time ago :-)
be called with the directory referenced, and this reference will
be dropped iff relookup() fails, so the value returned must not be
ignored.
Reviewed by: davidg
when rt == rt->rt_gwroute . rt == rt->gwroute shouldn't happen
in the first place, but that's another problem.
(try "route add -host <hostonmynet> <hostonmynet>; ping <hostonmynet>;
route delete <hostonmynet>")
can and will overflow on large machines - especially on machines with
filesystems with lots of files (like netnews servers), and the result
is a "free vnode isn't" panic or worse.
This fixes one of the causes of these panics that I've been experiancing on
wcarchive.
counter before loading the performance-monitor control register. I'm
not sure I believe this, but we'll follow their lead for the moment.
As a result of this commit, the performance-monitoring test program that
I wrote now works (the program will find its way to share/examples).
Made the devfs `fd' devices bug for bug compatible with the ones created
by MAKEDEV:
- ownership is bin.bin, not root.wheel, except for std*. The devfsext
interface doesn't seem to allow specifying the ownership of /devfs/fd,
so it's still incompatible.
- std* aren't links to fd/[0-2].
Added scsi control devices.
Converted almost everything that I changed to use devfs_add_devswf()
and verbose id macros.
st.c:
Renamed enrst* to erst* since that's what the current name is (enrst
seems to be an old name).
is incorrectly set to 0, for the purpose of "ignoring" the signal.
This does not ignore the signal, but rather, executes the function
at location 0 in kernel mode, which shortly thereafter causes a panic.
The sv_sensig entry for ibcs2 emulation should be set to the system's
normal sendsig routine.
- legitimate null frames from idle() (traceback was aborted after a null
pointer trap)
- second instruction of normal function prologue, and last instruction of
a function (caller wasn't reported).
Reviewed by: davidg
(This code is as yet untested; to come after man page is written.)
This also adds inlines to cpufunc.h for the RDTSC, RDMSR, WRMSR, and RDPMC
instructions. The user-mode interface is via a subdevice of mem.c;
there is also a kernel-size interface which might be used to aid
profiling.
Move ipip_input() and rsvp_input() prototypes to ip_var.h
Remove unused prototype for rip_ip_input() from ip_var.h
Remove unused variable *opts from rip_output()
mailing list.
When initiating a write, ccdbuffer() returns two "struct ccdbuf *"s
linked together by the cb_mirror field. "cb_pflags &
CCDPF_MIRROR_DONE" is set to 0 on both of them.
When a component returns to ccdiodone(), it checks if "cb_pflags &
CCDPF_MIRROR_DONE" is set or not. If not, it sets the partner's
flag and returns. If it is, it means its partner has already
returned, so it will go to the regular cleanup (which is in the
fallthrough code).
There should be no performance or functionality changes unless the
higher-level scsi driver does something with the resid value. The change
is purely aesthetical and prepares us for the parity implementation.
determine if the interface had been assigned an IP address.
This code prevented the interface from receiving ethernet
broadcasts if it had no IP address assigned, and appeared
to be an optimization that is not completely needed.
Get rid of ac->ac_ipaddr and arpwhohas() since they assume that
an interface has only one address.
Obtained from: BSD/OS 2.1, via Rich Stevens <rstevens@noao.edu>
from Larry Peterson &co. at Arizona:
- Header prediction for ACKs did not exclude Fast Retransmit/Recovery.
- srtt calculation tended to get ``stuck'' and could never decrease
when below 8. It still can't, but the scaling factors are adjusted
so that this artifact does not cause as bad an effect on the RTO
value as it used to.
The paper also points out the incr/8 error that has been long since fixed,
and the problems with ACKing frequency resulting from the use of options
which I suspect to be fixed already as well (as part of the T/TCP work).
Obtained from: Brakmo & Peterson, ``Performance Problems in BSD4.4 TCP''
Add support for LKM operation.
Change M_NOWAIT on buffer memory allocation to M_WAIT in hopes we'll be
able to get ourselves a nice fat buffer from the kernel if we suspend.
Note: The LKM support looks kinda screwy in two areas, where I found
problems with the kernel proper. First, calling dev_attach()
at module load time will cause a panic. I haven't investigated.
Secondly, I had to manually call qcam_drvinit() to register the
device softc structure by hand at module load time. This seems
bogus, it should be called as a core part of the module load
process for character/block device drivers.
vm_offset_t is currently unsigned long but should probably be plain
unsigned for i386's to match the choice of minimal types to represent
for fixed-width types in Lite2. Anyway, it shouldn't be assumed
to be unsigned long.
I only fixed the type mismatches that were detected when I changed
vm_offset_t to unsigned. Only pointer type mismatches were detected.
device have reference count problems. We mark the underlying object
ono-persistent, and account for the reference count that the VM system
maintainsfor the special device close. This should fix the removable
device problem.
isn't supplying all the proper header info here! Last commit of fe0
entry should have had the following Submitted by line also).
Submitted-by: Masahiro SEKIGUCHI <seki@sysrap.cs.fujitsu.co.jp>
1. Create 2 x 8k transmit buffer blocks in place of the 16k block previously.
With this change the speed as tested with ttcp on a 2Mbit link went up
from 206kbyte/s to 236kbyte/s.
2. Change the rest of the functions to also have the definition of the
return value on a sepperate line.
3. Remove some unused variables.
4. Add code to recover from DMA underruns.
5. Reorder ar_get_packets() to handle errors better.
6. Only allocate a mbuf cluster if the data is more than the mbuf.
(and in a second diff in addition to the above)
7. Stops the occasional DMA underruns that occurred when 2 channels
are running at 2Mbit/s.
Submitted by: John Hay <jhay@mikom.csir.co.za>
