]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
19 years ago- Gather statistics about failed mbuf+cluster+ng_item allocations.
glebius [Thu, 12 May 2005 13:52:49 +0000 (13:52 +0000)]
- Gather statistics about failed mbuf+cluster+ng_item allocations.
- Adjust comments and variables names in nfinfo.

19 years agoMove the pc98 keymap define into pckbdtables.h because it should be used
nyan [Thu, 12 May 2005 13:39:31 +0000 (13:39 +0000)]
Move the pc98 keymap define into pckbdtables.h because it should be used
only on the pckbd driver.

19 years ago- Move the NPX_DEBUG option to options.{i386,pc98} and use opt_npx.h.
nyan [Thu, 12 May 2005 12:47:41 +0000 (12:47 +0000)]
- Move the NPX_DEBUG option to options.{i386,pc98} and use opt_npx.h.
- Move npx related defines to {i386,pc98}/include/npx.h to remove #include
  {isa,cbus}.h.

19 years agoRemove unused IO_NPX* defines.
nyan [Thu, 12 May 2005 12:36:31 +0000 (12:36 +0000)]
Remove unused IO_NPX* defines.

19 years agoMove some global variables to the correct files.
harti [Thu, 12 May 2005 11:58:39 +0000 (11:58 +0000)]
Move some global variables to the correct files.

Submitted by: Max Okumoto <okumoto@ucsd.edu> (7.232)

19 years agoUntwist the convoluted if() expressions in JobFinish(). They
harti [Thu, 12 May 2005 11:47:17 +0000 (11:47 +0000)]
Untwist the convoluted if() expressions in JobFinish(). They
could benefit from even more untwisting.

Submitted by: Max Okumoto <okumoto@ucsd.edu> (7.23[01])

19 years agoMove some comments around to where they belong and remove some unneccessary
harti [Thu, 12 May 2005 11:16:04 +0000 (11:16 +0000)]
Move some comments around to where they belong and remove some unneccessary
gcc magic.

Submitted by: Max Okumoto <okumoto@ucsd.edu> (7.229)

19 years agoConvert Compat_RunCommand() to use ProcWait(). This also gives Cmd_Exec()
harti [Thu, 12 May 2005 11:09:45 +0000 (11:09 +0000)]
Convert Compat_RunCommand() to use ProcWait(). This also gives Cmd_Exec()
the potential to handle interrupts which it doesn't currently.

Submitted by: Max Okumoto <okumoto@ucsd.edu> (7.228)

19 years agoMake a function ProcWait() that waits for the given process.
harti [Thu, 12 May 2005 10:55:37 +0000 (10:55 +0000)]
Make a function ProcWait() that waits for the given process.

Submitted by: Max Okumoto <okumoto@ucsd.edu> (7.227)

19 years agoOverhaul the fork(), exec() code. Put all the common handling that
harti [Thu, 12 May 2005 10:40:57 +0000 (10:40 +0000)]
Overhaul the fork(), exec() code. Put all the common handling that
takes place in the child process in a function ProcExec(). Make sure,
that the child does not call malloc() or other potential dangerous
functions (there are still calls to Punt() in the error case that
should go away). Allocate the argv string via malloc to overcome
the non-constness bug of the execvp prototype. Change the handling of
shell meta-characters and move the builtin list near the list of shell
builtins. Both of these lists should actuall be configurable by the .SHELL
target since they depend on the shell used.

Patch: 7.21[2-9], 7.22[0-46]
Submitted by: Max Okumoto <okumoto@ucsd.edu>

19 years agoThe final test in unlock_and_deallocate() to determine if GIANT needs to be
grehan [Thu, 12 May 2005 04:09:41 +0000 (04:09 +0000)]
The final test in unlock_and_deallocate() to determine if GIANT needs to be
unlocked wasn't updated to check for OBJ_NEEDGIANT. This caused a WITNESS
panic when debug_mpsafevm was set to 0.

Approved by: jeffr

19 years agoWe don't use 'mp' variable, but we do want to mount devfs, ehh.
pjd [Thu, 12 May 2005 01:49:51 +0000 (01:49 +0000)]
We don't use 'mp' variable, but we do want to mount devfs, ehh.

19 years agoWhen looking for the next hole to retransmit from the scoreboard,
ps [Wed, 11 May 2005 21:37:42 +0000 (21:37 +0000)]
When looking for the next hole to retransmit from the scoreboard,
or to compute the total retransmitted bytes in this sack recovery
episode, the scoreboard is traversed. While in sack recovery, this
traversal occurs on every call to tcp_output(), every dupack and
every partial ack. The scoreboard could potentially get quite large,
making this traversal expensive.

This change optimizes this by storing hints (for the next hole to
retransmit and the total retransmitted bytes in this sack recovery
episode) reducing the complexity to find these values from O(n) to
constant time.

The debug code that sanity checks the hints against the computed
value will be removed eventually.

Submitted by:   Mohan Srinivasan, Noritoshi Demizu, Raja Mukerji.

19 years agoDon't forget to copy the TP when forking, or bad things will happen to the
cognet [Wed, 11 May 2005 21:20:42 +0000 (21:20 +0000)]
Don't forget to copy the TP when forking, or bad things will happen to the
child process if it tries to use threads.

19 years agoFix three typos in comments. Two of them obtained from OpenBSD.
brueffer [Wed, 11 May 2005 21:10:35 +0000 (21:10 +0000)]
Fix three typos in comments.  Two of them obtained from OpenBSD.

MFC after: 3 days

19 years agoRemove unised variable introduced by accident in rev 1.168.
pjd [Wed, 11 May 2005 19:50:34 +0000 (19:50 +0000)]
Remove unised variable introduced by accident in rev 1.168.

Found by: Coverity Prevent analysis tool

19 years agocp can't be NULL.
pjd [Wed, 11 May 2005 19:36:56 +0000 (19:36 +0000)]
cp can't be NULL.

Noticed by: Coverity Prevent analysis tool

19 years agogp can't be NULL.
pjd [Wed, 11 May 2005 19:35:43 +0000 (19:35 +0000)]
gp can't be NULL.

Noticed by: Coverity Prevent analysis tool

19 years agoPlug memory leaks.
pjd [Wed, 11 May 2005 19:27:38 +0000 (19:27 +0000)]
Plug memory leaks.

Found by: Coverity Prevent analysis tool

19 years agoDo not dereference dvp pointer before doing a NULL check.
kan [Wed, 11 May 2005 19:08:38 +0000 (19:08 +0000)]
Do not dereference dvp pointer before doing a NULL check.

Noticed by: Coverity Prevent analysis tool.

19 years agoPlug a potential memory leak.
kan [Wed, 11 May 2005 18:53:14 +0000 (18:53 +0000)]
Plug a potential memory leak.

Noticed by: Coverity Prevent analysis tool.

19 years agoAdd new Intel PCI ID's for the ICH7 chips (untested).
sos [Wed, 11 May 2005 18:29:15 +0000 (18:29 +0000)]
Add new Intel PCI ID's for the ICH7 chips (untested).

19 years agoHandle theoretical case of vfs_export being called with both MNT_DELEXPORT and
kan [Wed, 11 May 2005 18:25:42 +0000 (18:25 +0000)]
Handle theoretical case of vfs_export being called with both MNT_DELEXPORT and
MNT_EXPORT flags set. Do not reuse the memory that has just been freed.

19 years agoAdd KASSERT() to be sure there is an active component.
pjd [Wed, 11 May 2005 18:13:51 +0000 (18:13 +0000)]
Add KASSERT() to be sure there is an active component.

Suggested by: Coverity Prevent analysis tool

19 years agoCheck return value.
pjd [Wed, 11 May 2005 18:07:39 +0000 (18:07 +0000)]
Check return value.

Found by: Coverity Prevent analysis tool

19 years agoDo not initialize path variable with useless value just before
kan [Wed, 11 May 2005 17:39:33 +0000 (17:39 +0000)]
Do not initialize path variable with useless value just before
xpt_create_path overwrites it anyway.

Noticed by:     Coverity Prevent analysis tool

19 years ago- check for geom_uzip module presence using kldstat -m.
fjoe [Wed, 11 May 2005 17:02:38 +0000 (17:02 +0000)]
- check for geom_uzip module presence using kldstat -m.
kldstat -m finds geom_uzip module even if it is compiled in statically.
- create output file with x bit set.
- build mkuzip on all architectures (verified with "make universe").
- fix typo in info message.

19 years agoAdd support for AHCI compliant ATA devices.
sos [Wed, 11 May 2005 16:10:08 +0000 (16:10 +0000)]
Add support for AHCI compliant ATA devices.
For now just support the Intel ICH6 as that the HW at hand.

Sponsored by: pair.com

19 years agoTake out Giant in uhub_child_* in giant. There's one place where we
imp [Wed, 11 May 2005 15:21:22 +0000 (15:21 +0000)]
Take out Giant in uhub_child_* in giant.  There's one place where we
could sleep which I think can lead to races.  However, there are fewer
with this code than without it.

Submitted by: Hans Petter Selasky

19 years agoAdd back one line that got deleted in the last commit.
sos [Wed, 11 May 2005 14:36:26 +0000 (14:36 +0000)]
Add back one line that got deleted in the last commit.
Change the comment there so it better describes whats going on.

This should fix the boot problems of late.

19 years agoConnect MLINKS for ttyname_r(3), and add prototype into unistd.h.
delphij [Wed, 11 May 2005 14:07:25 +0000 (14:07 +0000)]
Connect MLINKS for ttyname_r(3), and add prototype into unistd.h.

19 years agoAvoid (unnecessarily) casting away const within uuid_is_nil.
delphij [Wed, 11 May 2005 13:18:10 +0000 (13:18 +0000)]
Avoid (unnecessarily) casting away const within uuid_is_nil.

19 years ago'ngtee' also depends on net.inet.ip.fw.one_pass.
glebius [Wed, 11 May 2005 12:58:15 +0000 (12:58 +0000)]
'ngtee' also depends on net.inet.ip.fw.one_pass.

19 years agoFix ATAPI DMA. We need to set the proper flags for DMA modes.
sos [Wed, 11 May 2005 12:41:35 +0000 (12:41 +0000)]
Fix ATAPI DMA. We need to set the proper flags for DMA modes.

19 years agoDescribe newly implemented bug.
glebius [Wed, 11 May 2005 11:53:37 +0000 (11:53 +0000)]
Describe newly implemented bug.

19 years agoA new version of NetFlow node.
glebius [Wed, 11 May 2005 11:26:24 +0000 (11:26 +0000)]
A new version of NetFlow node.

The most significant changes are:
- Use UMA zone instead of own chunk of memory.
- Lock each hash entry separately.
- Expire items "actively" - interrupt method can expire flows
  from hash slot, when it searches through it.
- Remove global tailqueue. Make callout thread search through
  every hash slot.
- Export datagram is detached from private data and filled. If
  it is incomplete, it is attached back. Another thread will
  continue working with it.

Lesser, but also important speedups:
- Flows in hash slot are stored in tailqueue. Whenever a flow is
  hit, it is moved to the begging, so it can be located quicker.
- When callout thread works with hash slot it bails out if
  slot mutex is contested.

19 years agoAdd MacOS 10.4.
maxim [Wed, 11 May 2005 11:11:29 +0000 (11:11 +0000)]
Add MacOS 10.4.

Submitted by: Jung-uk Kim

19 years agoRun /etc/rc.d/syscons restart when a usb keyboard is attached so that the
brian [Wed, 11 May 2005 10:24:20 +0000 (10:24 +0000)]
Run /etc/rc.d/syscons restart when a usb keyboard is attached so that the
keymap and other settings are correct.

19 years agoMove assertion below initializer.
glebius [Wed, 11 May 2005 08:28:09 +0000 (08:28 +0000)]
Move assertion below initializer.

Submitted by: Noritoshi Demizu
Pointy hat to:  glebius

19 years agoDocument zlib update procedure and provide future maintainers a
kientzle [Wed, 11 May 2005 03:50:50 +0000 (03:50 +0000)]
Document zlib update procedure and provide future maintainers a
snapshot of the current state of the import.

MFC after: 2 weeks

19 years agoSince the FreeBSD local modifications are mostly trivial (consisting
kientzle [Wed, 11 May 2005 03:47:48 +0000 (03:47 +0000)]
Since the FreeBSD local modifications are mostly trivial (consisting
primarily of pointless $FreeBSD$ tags), sync most files in HEAD with
those in the ZLIB branch.  This minimizes the differences between
HEAD and ZLIB and should simplify future imports.

After this, there are only three files with local modifications
(gzio.c, minigzip.c, and zconf.h) and two non-vendor files
(Makefile, zopen.c).  The rest exactly match the vendor distribution.

PR: i386/76294
MFC after: 2 weeks

19 years agoFix some incorrectly swapped fields in an ICB.
mjacob [Wed, 11 May 2005 03:00:50 +0000 (03:00 +0000)]
Fix some incorrectly swapped fields in an ICB.
Access a PCI register with correct width.

Obtained from: Dmitry Valeryevich Trikoz

19 years agoshutdown(2)'s how argument as 0 is nowhere near as cool as SHUT_RD
jmallett [Wed, 11 May 2005 02:50:41 +0000 (02:50 +0000)]
shutdown(2)'s how argument as 0 is nowhere near as cool as SHUT_RD

19 years agoSpell 1 as SHUT_WR in argument to shutdown(2).
jmallett [Wed, 11 May 2005 02:49:03 +0000 (02:49 +0000)]
Spell 1 as SHUT_WR in argument to shutdown(2).

19 years agoSpell out shutdown(2)'s how argument in #ifdef notdef'd code.
jmallett [Wed, 11 May 2005 02:47:54 +0000 (02:47 +0000)]
Spell out shutdown(2)'s how argument in #ifdef notdef'd code.

19 years agoSpell SHUT_RD as SHUT_RD not 0 as the how argument to shutdown(2).
jmallett [Wed, 11 May 2005 02:45:49 +0000 (02:45 +0000)]
Spell SHUT_RD as SHUT_RD not 0 as the how argument to shutdown(2).

19 years agoSpell SHUT_RDWR as SHUT_RDWR not 2 as the how argument to shutdown(2).
jmallett [Wed, 11 May 2005 02:43:04 +0000 (02:43 +0000)]
Spell SHUT_RDWR as SHUT_RDWR not 2 as the how argument to shutdown(2).

19 years agoSpell SHUT_RDWR as SHUT_RDWR not 1+1 as the how argument to shutdown(2).
jmallett [Wed, 11 May 2005 02:41:39 +0000 (02:41 +0000)]
Spell SHUT_RDWR as SHUT_RDWR not 1+1 as the how argument to shutdown(2).

19 years agoRefactor isp_prt declaration so that platform
mjacob [Wed, 11 May 2005 00:22:17 +0000 (00:22 +0000)]
Refactor isp_prt declaration so that platform
requirements can stay in platform files.

19 years agoAdd FreeBSD 5.4.
maxim [Tue, 10 May 2005 17:51:18 +0000 (17:51 +0000)]
Add FreeBSD 5.4.

19 years agoManual page for the "mn" Siemens Easy321-R1 E1/T1 device driver.
brueffer [Tue, 10 May 2005 16:48:03 +0000 (16:48 +0000)]
Manual page for the "mn" Siemens Easy321-R1 E1/T1 device driver.
This is based on the comments in if_mn.c.

MFC after: 3 days

19 years agoMark AVM USB Bluetooth-Adapter BlueFritz! as "broken". This device is not
emax [Tue, 10 May 2005 16:25:58 +0000 (16:25 +0000)]
Mark AVM USB Bluetooth-Adapter BlueFritz! as "broken". This device is not
supported by ng_ubt(4) driver and needs its own driver.

PR: kern/76205
Submitted by: Tim Hemel < bsd AT timit DOT nl >
MFC after: 1 week

19 years agoMove the definitions of the OP_* constants from make.h into GNode.h
harti [Tue, 10 May 2005 14:27:04 +0000 (14:27 +0000)]
Move the definitions of the OP_* constants from make.h into GNode.h
where they actually belong to. Move the definitions of the strings
for special macros like "$*" from make.h to parse.h - they're used
only in the parser.

Submitted by: Max Okumoto <okumoto@ucsd.edu> (7.211)

19 years agoset R_MUSTRESEND flag in mark_for_reconnect so re-connected requests get
rees [Tue, 10 May 2005 14:25:14 +0000 (14:25 +0000)]
set R_MUSTRESEND flag in mark_for_reconnect so re-connected requests get
re-sent instead of timing out.

don't log an error message on reconnection, which is not an error.

remove unused nfs_mrep_before_tsleep.

Reviewed by: Mohan Srinivasan
Approved by: alfred

19 years agoPath_Expand() expects its first argument to be writeable so put the
harti [Tue, 10 May 2005 14:23:31 +0000 (14:23 +0000)]
Path_Expand() expects its first argument to be writeable so put the
default system directory into a writeable character array before passing
it to Path_Expand().

Submitted by: Max Okumoto <okumoto@ucsd.edu> (7.210)

19 years ago- Assert that mbuf length equals packet length.
glebius [Tue, 10 May 2005 14:19:10 +0000 (14:19 +0000)]
- Assert that mbuf length equals packet length.
- Tell libalias, that we have MCLBYTES to play with.
- Obtain length of data in mbuf from updated IP header.

19 years agoMove declarations of variables belonging to the variable module
harti [Tue, 10 May 2005 14:16:40 +0000 (14:16 +0000)]
Move declarations of variables belonging to the variable module
from globals.h into var.h.

Submitted by: Max Okumoto <okumoto@ucsd.edu> (7.209)

19 years agoStyle: remove spaces before tabs.
harti [Tue, 10 May 2005 14:13:20 +0000 (14:13 +0000)]
Style: remove spaces before tabs.

Patch: 7.207
Submitted by: Max Okumoto <okumoto@ucsd.edu>

19 years agoMove the remaining two prototypes from nonints.h to make.h and
harti [Tue, 10 May 2005 14:10:44 +0000 (14:10 +0000)]
Move the remaining two prototypes from nonints.h to make.h and
remove nonints.h.

Patch: 7.204
Submitted by: Max Okumoto <okumoto@ucsd.edu>

19 years agoMake shellPath and shellName static - they're used in job.c only.
harti [Tue, 10 May 2005 14:06:19 +0000 (14:06 +0000)]
Make shellPath and shellName static - they're used in job.c only.
Move definition of parsing error levels from make.h into parse.h.

Patches: 7.202,7.203
Submitted by: Max Okumoto <okumoto@ucsd.edu>

19 years agoMake the three versions of fork() more uniform.
harti [Tue, 10 May 2005 14:03:28 +0000 (14:03 +0000)]
Make the three versions of fork() more uniform.

Patch: 7.201 (idea taken, but done differently)
Submitted by: Max Okumoto <okumoto@ucsd.edu>

19 years agoMove the defines for TMPPAT and KQUEUE from the public job.h into job.c
harti [Tue, 10 May 2005 13:48:15 +0000 (13:48 +0000)]
Move the defines for TMPPAT and KQUEUE from the public job.h into job.c
since they are used only there.

Patch: 7.200
Submitted by: Max Okumoto <okumoto@ucsd.edu>

19 years agoMerge compat.c into job.c. There is a lot in common between these
harti [Tue, 10 May 2005 13:45:29 +0000 (13:45 +0000)]
Merge compat.c into job.c. There is a lot in common between these
files and we are going to factor this out.

Patch: 7.199 (slightly changed)
Submitted by: Max Okumoto <okumoto@ucsd.edu>

19 years agoAdd option -l, to make output line buffered.
glebius [Tue, 10 May 2005 13:40:50 +0000 (13:40 +0000)]
Add option -l, to make output line buffered.

PR: bin/78692
Submitted by: Eugene Grosbein
MFC after: 2 weeks

19 years agoMove Cmd_Exec() from main.c to job.c and fix its prototype. This
harti [Tue, 10 May 2005 13:18:58 +0000 (13:18 +0000)]
Move Cmd_Exec() from main.c to job.c and fix its prototype. This
results in a warning that will go away soon.

Patch: 7.198
Submitted by: Max Okumoto <okumoto@ucsd.edu>

19 years agoThis commit was generated by cvs2svn to compensate for changes in r146052,
tjr [Tue, 10 May 2005 13:01:37 +0000 (13:01 +0000)]
This commit was generated by cvs2svn to compensate for changes in r146052,
which included commits to RCS files with non-trunk default branches.

19 years ago- Move lptreg.h into pc98/cbus and rename to olptreg.h.
nyan [Tue, 10 May 2005 12:30:30 +0000 (12:30 +0000)]
- Move lptreg.h into pc98/cbus and rename to olptreg.h.
- Remove ifdef pc98.

19 years agoUpdate instructions to match recent import.
tjr [Tue, 10 May 2005 12:04:01 +0000 (12:04 +0000)]
Update instructions to match recent import.

19 years agoChange a directory layout for pc98.
nyan [Tue, 10 May 2005 12:02:18 +0000 (12:02 +0000)]
Change a directory layout for pc98.
  - Move MD files into <arch>/<arch>.
  - Move bus dependent files into <arch>/<bus>.
Rename some files to more suitable names.

Repo-copied by: peter
Discussed with: imp

19 years agoVar_SubstOnly() is only used to subsitute a variable from the global
harti [Tue, 10 May 2005 12:02:15 +0000 (12:02 +0000)]
Var_SubstOnly() is only used to subsitute a variable from the global
context (and only in one place to substitute the .for variable). Therefor
there is no need to pass the context as a parameter.

Patch: 7.197
Submitted by: Max Okumoto <okumoto@ucsd.edu>

19 years agoMove the declaration of VAR_CMD and VAR_GLOBAL from globals.h to var.h
harti [Tue, 10 May 2005 11:58:52 +0000 (11:58 +0000)]
Move the declaration of VAR_CMD and VAR_GLOBAL from globals.h to var.h

Patch: 7.196
Submitted by: Max Okumoto <okumoto@ucsd.edu>

19 years agoClean up comments. No code changes.
harti [Tue, 10 May 2005 11:55:46 +0000 (11:55 +0000)]
Clean up comments. No code changes.

Patch: 7.195
Submitted by: Max Okumoto <okumoto@ucsd.edu>

19 years agoMerge var_modify.c into var.c and move types and function declarations
harti [Tue, 10 May 2005 11:53:20 +0000 (11:53 +0000)]
Merge var_modify.c into var.c and move types and function declarations
that are now used only in var.c from var.h to var.c

Patches: 7.193,7.194
Submitted by: Max Okumoto <okumoto@ucsd.edu>

19 years agoUpdate build infrastructure for new GNU regex bits.
tjr [Tue, 10 May 2005 11:37:37 +0000 (11:37 +0000)]
Update build infrastructure for new GNU regex bits.

19 years agoMerge. Discard local bracket expression collating order patch, since
tjr [Tue, 10 May 2005 10:52:30 +0000 (10:52 +0000)]
Merge. Discard local bracket expression collating order patch, since
modern versions of GNU regex already do things the way we want when
multibyte character support is enabled.

19 years agoThis commit was generated by cvs2svn to compensate for changes in r146040,
tjr [Tue, 10 May 2005 10:39:53 +0000 (10:39 +0000)]
This commit was generated by cvs2svn to compensate for changes in r146040,
which included commits to RCS files with non-trunk default branches.

19 years agoImport of regex bits from fedora-glibc-2_3_4-21 tag in glibc CVS.
tjr [Tue, 10 May 2005 10:39:53 +0000 (10:39 +0000)]
Import of regex bits from fedora-glibc-2_3_4-21 tag in glibc CVS.

19 years agoMove some debugging code from targ.c to var.c where it actually belongs.
harti [Tue, 10 May 2005 08:14:26 +0000 (08:14 +0000)]
Move some debugging code from targ.c to var.c where it actually belongs.

Patch: 7.192
Submitted by: Max Okumoto <okumoto@ucsd.edu>

19 years agoMake make a little bit more POSIXish with regard to option parsing:
harti [Tue, 10 May 2005 08:06:13 +0000 (08:06 +0000)]
Make make a little bit more POSIXish with regard to option parsing:
take everything after -- as either a macro assignment or a target.
Note that make still reorders arguments before --: anything starting
with a dash is considered an option, anything which contains an equal
sign is considered a macro assignment and everything else a target.
This still is not POSIX with regard to the options, but it will probably
not change because it has been make's behaviour for ages.
Add a new function Var_Match() that correctly skips a macro call by just
doing the same as Var_Subst() but without producing output. This will help
making the parser more robust.

Patches: 7.190,7.191
Submitted by: Max Okumoto <okumoto@ucsd.edu>

19 years agoDon't define _MACHINE_BUS_MEMIO_H_ nor _MACHINE_BUS_PIO_H_.
marcel [Tue, 10 May 2005 02:59:24 +0000 (02:59 +0000)]
Don't define _MACHINE_BUS_MEMIO_H_ nor _MACHINE_BUS_PIO_H_.

19 years agoRemove reference to seedrandom, since it is now gone, and was in there
imp [Mon, 9 May 2005 16:44:22 +0000 (16:44 +0000)]
Remove reference to seedrandom, since it is now gone, and was in there
just in case.

19 years agoSplit Var_Subst() into two functions: Var_SubstOnly() which substitutes
harti [Mon, 9 May 2005 14:06:04 +0000 (14:06 +0000)]
Split Var_Subst() into two functions: Var_SubstOnly() which substitutes
only one variable and Var_Subst() which substitutes all. Split out the
test whether a variable should not be expanded into match_var().
Make access to the input string consistently using str[]. Remove two
unused functions: Var_GetTail() and Var_GetHead().

Patches: 7.184-7.189
Submitted by: Max Okumoto <okumoto@ucsd.edu>

19 years agoSync with Linux acpi4asus, adding support for loads of new models.
philip [Mon, 9 May 2005 12:51:48 +0000 (12:51 +0000)]
Sync with Linux acpi4asus, adding support for loads of new models.

MFC after: 3 days

19 years agoPut the time traveling machine back where I found it before anyone sees. ;-)
philip [Mon, 9 May 2005 09:01:13 +0000 (09:01 +0000)]
Put the time traveling machine back where I found it before anyone sees. ;-)

Pointy hat to:     philip

19 years agoAdd support for Asus V6V[1] and W1N notebooks.
philip [Mon, 9 May 2005 08:43:49 +0000 (08:43 +0000)]
Add support for Asus V6V[1] and W1N notebooks.

Submitted by: Jean-Sébastien Pédron <jspedron@club-internet.fr> [1]
MFC after: 3 days

19 years agoSimplify the ACPI taskqueue implementation. Use a thread queue type instead
marks [Mon, 9 May 2005 07:34:04 +0000 (07:34 +0000)]
Simplify the ACPI taskqueue implementation. Use a thread queue type instead
of swi. This allows us to use the taskqueue_thread_* functions instead of
rolling our own. It also avoids a double trip through the queue.

Submitted by: njl
Reviewed by: sam

19 years agoFix separate module build.
marks [Mon, 9 May 2005 07:26:16 +0000 (07:26 +0000)]
Fix separate module build.

Submitted by: njl

19 years agoMove if_wl_i82586.h into sys/dev/wl.
nyan [Mon, 9 May 2005 04:47:58 +0000 (04:47 +0000)]
Move if_wl_i82586.h into sys/dev/wl.

19 years agoEnable debug_mpsafevm on ia64 due to the severe functional regression
marcel [Sun, 8 May 2005 23:56:16 +0000 (23:56 +0000)]
Enable debug_mpsafevm on ia64 due to the severe functional regression
caused by recent locking changes when it's off. Revert the logic to
trim down the conditional.

Clued-in by: alc@

19 years agoMore fixes for multibus drivers. When calling out to the match
wpaul [Sun, 8 May 2005 23:19:20 +0000 (23:19 +0000)]
More fixes for multibus drivers. When calling out to the match
function in if_ndis_pci.c and if_ndis_pccard.c, provide the bustype
too so the stubs can ignore devlists that don't concern them.

19 years agoFix support for Windows drivers that support both PCI and PCMCIA devices at
wpaul [Sun, 8 May 2005 23:07:51 +0000 (23:07 +0000)]
Fix support for Windows drivers that support both PCI and PCMCIA devices at
the same time.

Fix if_ndis_pccard.c so that it sets sc->ndis_dobj and sc->ndis_regvals.

Correct IMPORT_SFUNC() macros for the READ_PORT_BUFFER_xxx() routines,
which take 3 arguments, not 2.

This fixes it so that the Windows driver for my Cisco Aironet 340 PCMCIA
card works again. (Yes, I know the an(4) driver supports this card natively,
but it's the only PCMCIA device I have with a Windows XP driver.)

19 years agoOne more variant branding of 3C574 family: the plain old 3c574
imp [Sun, 8 May 2005 22:36:10 +0000 (22:36 +0000)]
One more variant branding of 3C574 family: the plain old 3c574

19 years agoIn uart_cnprobe(), fill in the cn_name field of the consdev structure.
marcel [Sun, 8 May 2005 20:25:09 +0000 (20:25 +0000)]
In uart_cnprobe(), fill in the cn_name field of the consdev structure.
The core console code checks this field when a console is added and
emits a warning if it's empty. In practice the warning is harmless for
uart(4), because the cn_name is filled in as soon as the device name is
known; which is when the device is enumerated.
To avoid the warning, to avoid possible complications caused by emitting
the warning without there (possibly) being a console selected yet and to
avoid complications when the UART isn't found during bus enumeration, we
just preset the cn_name field here to the name of the driver.

19 years agoarchive_entry_set_link is supposed to update whichever link field
kientzle [Sun, 8 May 2005 19:10:41 +0000 (19:10 +0000)]
archive_entry_set_link is supposed to update whichever link field
(symlink or hardlink) is already set.  Instead, it was always setting
the hardlink field.  In particular, this caused GNU tar format long
symlinks to be interpreted as hardlinks.

Thanks to: Brooks Davis
MFC after: 7 days

19 years agoRemove ifdef PC98.
nyan [Sun, 8 May 2005 14:17:28 +0000 (14:17 +0000)]
Remove ifdef PC98.

19 years agoMerged from libi386/biosdisk.c revisions 1.37 and 1.41.
nyan [Sun, 8 May 2005 13:30:14 +0000 (13:30 +0000)]
Merged from libi386/biosdisk.c revisions 1.37 and 1.41.

19 years agoChange DOSPTYP_386BSD to 0xc494.
nyan [Sun, 8 May 2005 11:30:26 +0000 (11:30 +0000)]
Change DOSPTYP_386BSD to 0xc494.

19 years agoUse DOSMID_386BSD rather than DOSPTYP_386BSD.
nyan [Sun, 8 May 2005 11:15:38 +0000 (11:15 +0000)]
Use DOSMID_386BSD rather than DOSPTYP_386BSD.

19 years agoThe Nocona may be AMD64 compatible, but it is still an Intel part.
des [Sun, 8 May 2005 10:11:33 +0000 (10:11 +0000)]
The Nocona may be AMD64 compatible, but it is still an Intel part.

Pointy hat to: trhodes
MFC after: 1 week

19 years agoAvoid trying to queue up an interrupt handler DPC if the driver
wpaul [Sun, 8 May 2005 09:36:16 +0000 (09:36 +0000)]
Avoid trying to queue up an interrupt handler DPC if the driver
hasn't called NdisMRegisterInterrupt() yet.