]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
14 years agoDon't leak a file descriptor when ejecting a CDROM.
cperciva [Mon, 26 Oct 2009 23:24:59 +0000 (23:24 +0000)]
Don't leak a file descriptor when ejecting a CDROM.

Submitted by: Ronald Klop, trhodes
Note to self: don't do commits while half-asleep

14 years agoSync with the other archs and wrapper the prototype of in_cksum_skip(9)
marius [Mon, 26 Oct 2009 22:00:26 +0000 (22:00 +0000)]
Sync with the other archs and wrapper the prototype of in_cksum_skip(9)
in #ifdef _KERNEL.

Submitted by: Ulrich Spoerlein
MFC after: 1 month

14 years agoRevert r198500 for now, this will break situations when
thompsa [Mon, 26 Oct 2009 21:47:16 +0000 (21:47 +0000)]
Revert r198500 for now, this will break situations when
hw.pci.usb_early_takeover is set to zero and the SMM release
is never done.

Pointed out by: marcel

14 years agoRemove usb controller takeover code now that it is handled by the pci code.
thompsa [Mon, 26 Oct 2009 20:23:15 +0000 (20:23 +0000)]
Remove usb controller takeover code now that it is handled by the pci code.

Reminded by: jhb
Reviewed by: HPS

14 years agoImprove the round robin stream scheduler.
tuexen [Mon, 26 Oct 2009 19:23:34 +0000 (19:23 +0000)]
Improve the round robin stream scheduler.

Approved by: rrs (mentor)
MFC after: 3 days

14 years agoTrim empty mergeinfo.
jhb [Mon, 26 Oct 2009 19:10:54 +0000 (19:10 +0000)]
Trim empty mergeinfo.

14 years agoAnother nit that both I and ispell missed.
jhb [Mon, 26 Oct 2009 18:32:06 +0000 (18:32 +0000)]
Another nit that both I and ispell missed.

Submitted by: Ben Kaduk  minimarmot of gmail

14 years agoThere is no need to "busy" a page when the object is locked for the duration
alc [Mon, 26 Oct 2009 18:02:05 +0000 (18:02 +0000)]
There is no need to "busy" a page when the object is locked for the duration
of the operation.

14 years agoFix some spelling nits.
jhb [Mon, 26 Oct 2009 17:42:03 +0000 (17:42 +0000)]
Fix some spelling nits.

14 years agoFix typo in bluetooth.3
emax [Mon, 26 Oct 2009 17:27:30 +0000 (17:27 +0000)]
Fix typo in bluetooth.3
Do not use reserved C++ keyword "new"

MFC after: 1 month

14 years agoFix parsing of mount options specified with -o in case an option with
jh [Mon, 26 Oct 2009 14:57:33 +0000 (14:57 +0000)]
Fix parsing of mount options specified with -o in case an option with
value is preceded by an option without value (for example -o
option1,option2=value). Options must be separated before searching for
'='. Also compare pnextopt explicitly against NULL.

PR: bin/134069
Approved by: trasz (mentor)

14 years agoMerge r198489 from vendor/ncurses/dist:
rafan [Mon, 26 Oct 2009 13:03:52 +0000 (13:03 +0000)]
Merge r198489 from vendor/ncurses/dist:

 Pull upstream patch to fix ee(1) crash when received SIGWINCH:

   modify _nc_wgetch() to check for a -1 in the fifo, e.g., after a
   SIGWINCH, and discard that value, to avoid confusing application
   (patch by Eygene Ryabinkin, FreeBSD bin/136223).

PR: 136223
Submitted by: Eygene Ryabinkin
Obtained from: ncurses-5.7-20091024 snapshot
MFC after: 3 days

14 years agoReport SATA speeds to CAM, to not confuse users with low numbers logged.
mav [Mon, 26 Oct 2009 11:26:49 +0000 (11:26 +0000)]
Report SATA speeds to CAM, to not confuse users with low numbers logged.

14 years agoRound timeout up when converting CAM milliseconds to ATA seconds.
mav [Mon, 26 Oct 2009 11:23:41 +0000 (11:23 +0000)]
Round timeout up when converting CAM milliseconds to ATA seconds.

14 years agoIncrease ATA command timeouts. Some drives need more then 5s to spin-up.
mav [Mon, 26 Oct 2009 11:20:14 +0000 (11:20 +0000)]
Increase ATA command timeouts. Some drives need more then 5s to spin-up.

PR: kern/111023

14 years agoDocument new modularised ATA kernel options.
mav [Mon, 26 Oct 2009 10:35:16 +0000 (10:35 +0000)]
Document new modularised ATA kernel options.

PR: kern/133162
MFC after: 3 days

14 years agoAdd IDs for PATA part also.
mav [Mon, 26 Oct 2009 10:07:52 +0000 (10:07 +0000)]
Add IDs for PATA part also.

14 years agoAdd two more VIA SATA chip IDs.
mav [Mon, 26 Oct 2009 10:00:59 +0000 (10:00 +0000)]
Add two more VIA SATA chip IDs.

PR: kern/135057

14 years agoDocument atapci kernel module split.
mav [Mon, 26 Oct 2009 09:16:08 +0000 (09:16 +0000)]
Document atapci kernel module split.

PR: amd64/139859
MFC after: 3 days

14 years agoFix SATA on nVidia MCP55 chipset. It needs some short time to allow BAR(5)
mav [Mon, 26 Oct 2009 08:41:10 +0000 (08:41 +0000)]
Fix SATA on nVidia MCP55 chipset. It needs some short time to allow BAR(5)
memory access.

PR: amd64/128686, amd64/132372, amd64/139156
MFC after: 3 days

14 years ago- Initialize variable in order to avoid GCC warning and enable WARNS=6.
lulf [Mon, 26 Oct 2009 07:43:41 +0000 (07:43 +0000)]
- Initialize variable in order to avoid GCC warning and enable WARNS=6.

PR: bin/139970
Submitted by: Ulrich Spörlein <uqs -at- spoerlein.net>

14 years agoEject CDROM after installation if used as source media.
cperciva [Mon, 26 Oct 2009 06:51:20 +0000 (06:51 +0000)]
Eject CDROM after installation if used as source media.

Submitted by: randi
MFC after: 1 month

14 years agoSimplify the inner loop of vm_fault_copy_entry().
alc [Mon, 26 Oct 2009 00:01:52 +0000 (00:01 +0000)]
Simplify the inner loop of vm_fault_copy_entry().

Reviewed by: kib

14 years ago- Add support for Marvell Yukon 88E8042 device.
lulf [Sun, 25 Oct 2009 21:46:38 +0000 (21:46 +0000)]
- Add support for Marvell Yukon 88E8042 device.

Submitted by: Mario Lobo <mlobo -at- digiart.art.br>
Approved by: yongari

14 years agoSpell out the name of the month.
brueffer [Sun, 25 Oct 2009 17:47:52 +0000 (17:47 +0000)]
Spell out the name of the month.

14 years agoEliminate an unnecessary check from vm_fault_prefault().
alc [Sun, 25 Oct 2009 17:30:50 +0000 (17:30 +0000)]
Eliminate an unnecessary check from vm_fault_prefault().

14 years agoCompile libgcov without stack protection. It can be linked into
kan [Sun, 25 Oct 2009 15:52:31 +0000 (15:52 +0000)]
Compile libgcov without stack protection. It can be linked into
both static and dynamic binaries compiled with or without stack
protection and should not depend on libssp_nonshared.a symbols.

Discussed with: kib
PR: bin/139052

14 years agoUpdate according to the latest iwn(4) driver import.
rpaulo [Sun, 25 Oct 2009 11:01:53 +0000 (11:01 +0000)]
Update according to the latest iwn(4) driver import.

14 years agoExplain that iwn was updated and the firmware images are now split.
rpaulo [Sun, 25 Oct 2009 10:29:37 +0000 (10:29 +0000)]
Explain that iwn was updated and the firmware images are now split.

14 years agoUpdate firmware images according to the latest iwn updated.
rpaulo [Sun, 25 Oct 2009 10:23:11 +0000 (10:23 +0000)]
Update firmware images according to the latest iwn updated.
"device iwnfw" includes all firmware images, but you can pick just one
by using the model number, e.g.: "device iwn4965fw".

14 years agoUnconditionally call the setsockopt for IPV6_V6ONLY for v6 linux sockets
bz [Sun, 25 Oct 2009 09:58:56 +0000 (09:58 +0000)]
Unconditionally call the setsockopt for IPV6_V6ONLY for v6 linux sockets
no matter whether we are compiled as module or if our default of the
net.inet6.ip6.v6only sysctl already matches what we would set.

This avoids unnecessary complications with modules, VIMAGES, INET6 and
the sysctl value, especially considering that most users will use
linux compat as a module.

Discussed with: kib, rwatson (weeks ago)
Reviewed by: rwatson
MFC after: 6 weeks

14 years agoOutput a comment on top of each generated file explaining where it came
rpaulo [Sun, 25 Oct 2009 09:48:21 +0000 (09:48 +0000)]
Output a comment on top of each generated file explaining where it came
from.

14 years agoInform hwpmc(4) of a thread's impending demise prior to invoking sched_throw().
jkoshy [Sun, 25 Oct 2009 04:34:47 +0000 (04:34 +0000)]
Inform hwpmc(4) of a thread's impending demise prior to invoking sched_throw().

Debugging help: fabient
Review and testing by: fabient

14 years agoUpdate a comment to reflect the previous change.
alc [Sun, 25 Oct 2009 02:48:29 +0000 (02:48 +0000)]
Update a comment to reflect the previous change.

14 years agosh: Exempt $@ and $* from set -u
jilles [Sat, 24 Oct 2009 21:20:04 +0000 (21:20 +0000)]
sh: Exempt $@ and $* from set -u

This seems more useful and will likely be in the next POSIX standard.

Also document more precisely in the man page what set -u does (note that
$@, $* and $! are the only special parameters that can ever be unset, all
the others are always set, although they may be empty).

14 years agoAdd some tests for ${var?} and set -u.
jilles [Sat, 24 Oct 2009 20:57:11 +0000 (20:57 +0000)]
Add some tests for ${var?} and set -u.

14 years agoAdd PRINTF_BUFR_SIZE=128, since we have SMP by default.
marcel [Sat, 24 Oct 2009 20:35:34 +0000 (20:35 +0000)]
Add PRINTF_BUFR_SIZE=128, since we have SMP by default.
While here, fix tabulation.

14 years agoA 32KB kernel stack is not quite enough. The new USB stack is a bit
marcel [Sat, 24 Oct 2009 20:28:42 +0000 (20:28 +0000)]
A 32KB kernel stack is not quite enough. The new USB stack is a bit
more stack hungry as compared to the old one that my RX2660 gets
a machine check and spontaneously reboots at the time the USB DVD
drive is found and attached to CAM as a mass storage device. This
doesn't happen always, but definitely varies per kernel build.
Likewise when using a 128-byte printf buffer. The additional 128
bytes that printf needs seems to be enough to have the memory stack
and register stack collide and causing a machine check.

Thus: Bump KSTACK_PAGES from 4 to 5.

14 years agoImplement _umtx_op_err() for ia64.
marcel [Sat, 24 Oct 2009 20:07:17 +0000 (20:07 +0000)]
Implement _umtx_op_err() for ia64.

14 years ago- Rename tunable kern.ipc.shmmaxpgs to kern.ipc.shmall.
ru [Sat, 24 Oct 2009 19:00:58 +0000 (19:00 +0000)]
- Rename tunable kern.ipc.shmmaxpgs to kern.ipc.shmall.
- Explain the fuss when initializing shmmax.

PR: 75542 (mistakenly closed instead of PR 75541)

14 years agoSpell DIAGNOSTIC correctly.
ru [Sat, 24 Oct 2009 18:49:17 +0000 (18:49 +0000)]
Spell DIAGNOSTIC correctly.

14 years agoTurn on NAP mode on G5 systems, and refactor the HID0 setup code a little.
nwhitehorn [Sat, 24 Oct 2009 18:33:01 +0000 (18:33 +0000)]
Turn on NAP mode on G5 systems, and refactor the HID0 setup code a little.
This makes my G5 Xserve sound slightly less like it is filled with
howling banshees.

14 years agoAllow Heathrow-based machines to boot a kernel containing option SMP
nwhitehorn [Sat, 24 Oct 2009 18:31:22 +0000 (18:31 +0000)]
Allow Heathrow-based machines to boot a kernel containing option SMP
without panicing.

14 years agoAdd more obsolete man pages.
antoine [Sat, 24 Oct 2009 12:59:29 +0000 (12:59 +0000)]
Add more obsolete man pages.

14 years agoReplace ocurrences of FreeBSD CURRENT by FreeBSD 8.0.
rpaulo [Sat, 24 Oct 2009 11:32:02 +0000 (11:32 +0000)]
Replace ocurrences of FreeBSD CURRENT by FreeBSD 8.0.

MFC after: 2 days

14 years agoUpdates to the iwn driver:
rpaulo [Sat, 24 Oct 2009 09:55:11 +0000 (09:55 +0000)]
Updates to the iwn driver:
* iwnfw has now been split into individual modules so autoloading of
  firmware module(s) does work again.
* Changes have been made to RUN -> AUTH transition, this should fix the
  issue reported by Glen and others.
* Brandon reported issues in iwn_cmd() with large commands, those have
  been fixed to.
* DEAUTH is now handled correctly.

Submitted by: Bernhard Schmidt <bschmidt at techwires.net>

14 years agoCorrect spelling typo in ip_input comment.
rwatson [Sat, 24 Oct 2009 09:18:26 +0000 (09:18 +0000)]
Correct spelling typo in ip_input comment.

Pointed out by: N.J. Mann <njm at njm.me.uk>,
John Nielsen <john at jnielsen.net>, julian (!), lstewart
MFC after: 2 days

14 years agoNot all Intel Core (TM) CPUs implement PMC_CLASS_IAF fixed-function
jkoshy [Sat, 24 Oct 2009 04:11:40 +0000 (04:11 +0000)]
Not all Intel Core (TM) CPUs implement PMC_CLASS_IAF fixed-function
counters.  For such CPUs, use an alternate mapping of convenience
names to events supported by PMC_CLASS_IAP programmable counters.

Testing and review by: fabient

14 years agoOnly claim that the PMC_CLASS_IAF PMCs are supported by a CPU if
jkoshy [Sat, 24 Oct 2009 01:58:10 +0000 (01:58 +0000)]
Only claim that the PMC_CLASS_IAF PMCs are supported by a CPU if
there are PMCs on the CPU that belong to the class.

Review and testing by: fabient

14 years agoBIOSes, buggy or otherwise, are i386 or amd64 specific.
marcel [Fri, 23 Oct 2009 22:53:01 +0000 (22:53 +0000)]
BIOSes, buggy or otherwise, are i386 or amd64 specific.
Have the early USB takeover enabled for i386 and amd64
by default.
This also avoids a panic on PowerPC where the resource
isn't released properly and we find a busy resource
when the USB host controller wants to allocate it...

14 years agoUpdated iwn(4) driver supporting the newer series, 5000, 5150 and 5300.
rpaulo [Fri, 23 Oct 2009 22:04:18 +0000 (22:04 +0000)]
Updated iwn(4) driver supporting the newer series, 5000, 5150 and 5300.

Submitted by: Bernhard Schmidt <bschmidt at techwires.net>

14 years agoRemove debugging printf that snuck in here.
nwhitehorn [Fri, 23 Oct 2009 21:44:46 +0000 (21:44 +0000)]
Remove debugging printf that snuck in here.

Pointy hat to: me

14 years agoAdd some more paranoia to setting HID registers, and update the AIM
nwhitehorn [Fri, 23 Oct 2009 21:36:33 +0000 (21:36 +0000)]
Add some more paranoia to setting HID registers, and update the AIM
clock routines to work better with SMP. This makes SMP work fully and
stably on an Xserve G5.

Obtained from: Book-E (clock bits)

14 years agoReimplement device reset sequence in more controller-specific way.
mav [Fri, 23 Oct 2009 21:33:26 +0000 (21:33 +0000)]
Reimplement device reset sequence in more controller-specific way.

14 years agoRemove a redundant byte swapping in the previous commit.
jkim [Fri, 23 Oct 2009 19:02:53 +0000 (19:02 +0000)]
Remove a redundant byte swapping in the previous commit.

14 years agoTry hiding annoying text cursor after the video controller is reset.
jkim [Fri, 23 Oct 2009 18:57:52 +0000 (18:57 +0000)]
Try hiding annoying text cursor after the video controller is reset.

14 years agoSearch for default 800x600 graphics mode from supported VESA mode list.
jkim [Fri, 23 Oct 2009 18:53:21 +0000 (18:53 +0000)]
Search for default 800x600 graphics mode from supported VESA mode list.
Many video controllers do not support 800x600x24 mode any more.

14 years agoCorrect some issues with zfs boot.
rnoland [Fri, 23 Oct 2009 18:44:53 +0000 (18:44 +0000)]
Correct some issues with zfs boot.

 - Teach it to read gang blocks. (essentially untested)
   If you see "ZFS: gang block detected!", please let
   me know, so we can either remove the printf if it
   works, or fix it if it doesn't.

 - If multiple partitions exist on a disk, probe them all.
   We also need to reset dsk->start to 0 to read the right
   sector here.

 - With GPT, we can have 128 partitions.

 - If the bootfs property has ever been set on a pool
   it seems that it never goes away.  zpool won't allow
   you to add to the pool with the bootfs property set.
   However, if you clear the property back to default
   we end up getting 0 for the object number and read
   a bogus block pointer and fail to boot.

 - Fix some error printfs. The printf in the loader is
   only capable of c,s and u formats.

 - Teach printf how to display %llu

Reviewed by: dfr, jhb
MFC after: 2 weeks

14 years ago- When we restore VESA state, try BIOS POST earlier. VESA restore state
jkim [Fri, 23 Oct 2009 18:41:00 +0000 (18:41 +0000)]
- When we restore VESA state, try BIOS POST earlier.  VESA restore state
function may not work properly if we don't.  Turn off hardware cursor as
vesa_set_mode() does.
- Add VBE 3.0 specific fields in VESA mode structure and pack it.  Note
the padding is 190 bytes although VBE 3.0 says 189 bytes.  It must be wrong
because the size of structure becomes 255 bytes and the specification says
it must be 256 bytes in total.  In fact, an example code in the spec. does
it right, though.  While we are at it, fix some i386-isms.
- Remove state buffer size limitation.  It is no longer necessary since
sys/compat/x86bios/x86bios.c r198251.
- Move int 0x10 vector test into vesa_bios_post() as we always do it anyway.

14 years agoUse the correct option name in the preprocessor command to enable
qingli [Fri, 23 Oct 2009 18:27:34 +0000 (18:27 +0000)]
Use the correct option name in the preprocessor command to enable
or disable diagnostic messages.

Reviewed by: ru
MFC after: 3 days

14 years agoRemove unneeded blank line from bpf_drvinit().
rwatson [Fri, 23 Oct 2009 17:26:29 +0000 (17:26 +0000)]
Remove unneeded blank line from bpf_drvinit().

MFC after: 3 days

14 years ago- Fix several off-by-one errors when using MAXCOMLEN. The p_comm[] and
jhb [Fri, 23 Oct 2009 15:14:54 +0000 (15:14 +0000)]
- Fix several off-by-one errors when using MAXCOMLEN.  The p_comm[] and
  td_name[] arrays are actually MAXCOMLEN + 1 in size and a few places that
  created shadow copies of these arrays were just using MAXCOMLEN.
- Prefer using sizeof() of an array type to explicit constants for the
  array length in a few places.
- Ensure that all of p_comm[] and td_name[] is always zero'd during
  execve() to guard against any possible information leaks.  Previously
  trailing garbage in p_comm[] could be leaked to userland in ktrace
  record headers via td_name[].

Reviewed by: bde

14 years agoStyle fix.
jhb [Fri, 23 Oct 2009 15:10:41 +0000 (15:10 +0000)]
Style fix.

14 years agoDon't bother copying the name of a kproc or kthread out into a temporary
jhb [Fri, 23 Oct 2009 15:09:51 +0000 (15:09 +0000)]
Don't bother copying the name of a kproc or kthread out into a temporary
array just to pass that array to printf().  kproc and kthread names are
NUL-terminated and can be printed using printf() directly.

Reviewed by: bde

14 years agoMFp4:
mav [Fri, 23 Oct 2009 14:56:29 +0000 (14:56 +0000)]
MFp4:
Do not differentiate 12/16 bytes ATAPI CCB formats when it is not needed.

14 years agowordexp(3): fix some bugs with signals and long outputs
jilles [Fri, 23 Oct 2009 14:50:11 +0000 (14:50 +0000)]
wordexp(3): fix some bugs with signals and long outputs
* retry various system calls on EINTR
* retry the rest after a short read (common if there is more than about 1K
  of output)
* block SIGCHLD like system(3) does (note that this does not and cannot
  work fully in threaded programs, they will need to be careful with wait
  functions)

PR: 90580
MFC after: 1 month

14 years agoDo not map the trap vectors into the kernel's address space. They are
nwhitehorn [Fri, 23 Oct 2009 14:27:40 +0000 (14:27 +0000)]
Do not map the trap vectors into the kernel's address space. They are
only used in real mode and keeping them mapped only serves to make NULL
a valid address, which results in silent NULL pointer deferences.

Suggested by:   Patrick Kerharo
Obtained from: projects/ppc64

14 years agoMake "Retrying Command" to be printed before actual retrying.
mav [Fri, 23 Oct 2009 13:39:30 +0000 (13:39 +0000)]
Make "Retrying Command" to be printed before actual retrying.
It should make debug/error log a bit more readable.

14 years agoImprove grammar in ip_input comment while attempting to maintain what
rwatson [Fri, 23 Oct 2009 13:35:00 +0000 (13:35 +0000)]
Improve grammar in ip_input comment while attempting to maintain what
might be its meaning.

MFC after: 3 days

14 years agoProperly sort the intr_event_describe_handler() prototype.
jhb [Fri, 23 Oct 2009 13:28:33 +0000 (13:28 +0000)]
Properly sort the intr_event_describe_handler() prototype.

Submitted by: bde

14 years agoRevert interrupt reason check order back.
mav [Fri, 23 Oct 2009 13:07:22 +0000 (13:07 +0000)]
Revert interrupt reason check order back.
ATAPI errors may set IF bit together with TFE.

14 years agoMFp4:
mav [Fri, 23 Oct 2009 12:36:42 +0000 (12:36 +0000)]
MFp4:
Move Port Multiplier support code out of ATA XPT into pmp periph driver.
This is convinient, as PMP itself is a bus target and has own state.

14 years agoFix the build.
mav [Fri, 23 Oct 2009 11:26:58 +0000 (11:26 +0000)]
Fix the build.

14 years agoThe draft spec doesn't say beacon frames need to have a wildcard BSSID,
rpaulo [Fri, 23 Oct 2009 11:13:08 +0000 (11:13 +0000)]
The draft spec doesn't say beacon frames need to have a wildcard BSSID,
so remove the mesh code necessary for that.

MFC after: 2 days

14 years agoUse double-quotation marks to fix the unexpanded variable issue.
hrs [Fri, 23 Oct 2009 09:30:19 +0000 (09:30 +0000)]
Use double-quotation marks to fix the unexpanded variable issue.

Spotted by: swell.k

14 years agoReplace most of priority numbers with defines. No logical changes.
mav [Fri, 23 Oct 2009 08:27:55 +0000 (08:27 +0000)]
Replace most of priority numbers with defines. No logical changes.

14 years agoRemove some obsoleted comments.
mav [Fri, 23 Oct 2009 07:54:15 +0000 (07:54 +0000)]
Remove some obsoleted comments.

14 years agoAdd SMP support on U3-based G5 systems. This does not yet work perfectly:
nwhitehorn [Fri, 23 Oct 2009 03:17:02 +0000 (03:17 +0000)]
Add SMP support on U3-based G5 systems. This does not yet work perfectly:
at least on my Xserve, getting the decrementer and timebase on APs to tick
requires setting up a clock chip over I2C, which is not yet done.

While here, correct the 64-bit tlbie function to set the CPU to 64-bit
mode correctly.

Hardware donated by: grehan

14 years agoAfter thinking again, implement cam_ccbq_fini().
mav [Thu, 22 Oct 2009 21:07:32 +0000 (21:07 +0000)]
After thinking again, implement cam_ccbq_fini().
This is effectively NULL change, but makes this API a bit more consistent.

14 years agoPrevent wraparound of the timeout variable.
thompsa [Thu, 22 Oct 2009 21:01:41 +0000 (21:01 +0000)]
Prevent wraparound of the timeout variable.

Submitted by: HPS

14 years agoMove comments to the beginning of the line to make it look better.
edwin [Thu, 22 Oct 2009 20:59:51 +0000 (20:59 +0000)]
Move comments to the beginning of the line to make it look better.
Thank to ru@ for his noticing of it.

MFC after: 1 week

14 years agoRename default to default_function, for compatibility with GNU awk.
emaste [Thu, 22 Oct 2009 20:57:17 +0000 (20:57 +0000)]
Rename default to default_function, for compatibility with GNU awk.
(For cross-compiling out-of-tree kernel modules, for example.)

14 years agoAllow dumping the USB mouse reports via 'sysctl -b dev.ums.N.parseinfo',
thompsa [Thu, 22 Oct 2009 20:54:01 +0000 (20:54 +0000)]
Allow dumping the USB mouse reports via 'sysctl -b dev.ums.N.parseinfo',
previously only available via bootverbose.

PR: usb/137191
Submitted by: Eygene Ryabinkin

14 years agocam_ccbq_fini() declared for 11 years, but never implemented. Remove it.
mav [Thu, 22 Oct 2009 20:44:55 +0000 (20:44 +0000)]
cam_ccbq_fini() declared for 11 years, but never implemented. Remove it.

14 years agoInitialize rann_flags properly.
rpaulo [Thu, 22 Oct 2009 17:30:22 +0000 (17:30 +0000)]
Initialize rann_flags properly.

MFC after: 2 days

14 years agoSet the devclass_t pointer specified in the DRIVER_MODULE() macro
jhb [Thu, 22 Oct 2009 14:53:44 +0000 (14:53 +0000)]
Set the devclass_t pointer specified in the DRIVER_MODULE() macro
sooner so it is always valid when a driver's identify routine is
called.  Previously, new-bus would attempt to create the devclass for
a newly loaded driver in two separate places, once in
devclass_add_driver(), and again after devclass_add_driver() returned
in driver_module_handler().  Only the second lookup attempted to set a
device class' parent and set the devclass_t pointer specified in the
DRIVER_MODULE() macro.  However, by the time it was executed, the
driver was already added to existing instances of the parent driver at
which point in time the new driver's identify routine would have been
invoked.  The fix is to merge the two attempts and only create the
devclass once in devclass_add_driver() including setting the
devclass_t pointer passed to DRIVER_MODULE() before the driver is
added to any existing bus devices.

Reported by: avg
Reviewed by: imp
MFC after: 2 weeks

14 years agoDon't use BUS_DMA_ALLOCNOW as that causes the attachment to fail on
rpaulo [Thu, 22 Oct 2009 12:48:17 +0000 (12:48 +0000)]
Don't use BUS_DMA_ALLOCNOW as that causes the attachment to fail on
Cambria boards.

MFC after: 2 months

14 years agoUnbreak NO_WARNS, keeping CSTD effect on CFLAGS out of its control.
ru [Thu, 22 Oct 2009 11:45:35 +0000 (11:45 +0000)]
Unbreak NO_WARNS, keeping CSTD effect on CFLAGS out of its control.
Unbreak compiles with icc.

14 years agoList more dependencies for these drivers. While here, convert
brueffer [Thu, 22 Oct 2009 11:35:12 +0000 (11:35 +0000)]
List more dependencies for these drivers. While here, convert
atapicam(4) to use our standard section 4 SYNOPSIS layout.

PR: 132525
Submitted by: gcooper
MFC after: 3 days

14 years agoRemove self-reference.
pjd [Thu, 22 Oct 2009 08:38:27 +0000 (08:38 +0000)]
Remove self-reference.

14 years agoCheck pointer for NULL before dereferencing it, not after.
brueffer [Thu, 22 Oct 2009 06:51:29 +0000 (06:51 +0000)]
Check pointer for NULL before dereferencing it, not after.

PR: 138387, 138388
Submitted by: Patroklos Argyroudis <argp@census-labs.com>
MFC after: 1 week

14 years agoCheck pointer for NULL before dereferencing it, not after.
brueffer [Thu, 22 Oct 2009 06:17:04 +0000 (06:17 +0000)]
Check pointer for NULL before dereferencing it, not after.

PR: 138390
Submitted by: Patroklos Argyroudis <argp@census-labs.com>
MFC after: 1 week

14 years agoFix a memory leak in an error case.
brueffer [Thu, 22 Oct 2009 06:13:07 +0000 (06:13 +0000)]
Fix a memory leak in an error case.

PR: 138376
Submitted by: Patroklos Argyroudis <argp@census-labs.com>
Reviewed by: scottl
MFC after: 1 week

14 years agoVerify "smp_started" is true before calling
qingli [Thu, 22 Oct 2009 00:32:01 +0000 (00:32 +0000)]
Verify "smp_started" is true before calling
sched_bind() and sched_unbind().

Reviewed by: kmacy
MFC after: 3 days

14 years agoMake dhclient use bootpc (68) as the source port for unicast DHCPREQUEST
philip [Wed, 21 Oct 2009 23:50:35 +0000 (23:50 +0000)]
Make dhclient use bootpc (68) as the source port for unicast DHCPREQUEST
packets instead of allowing the protocol stack to pick a random source port.

This fixes the behaviour where dhclient would never transition from RENEWING
to BOUND without going through REBINDING in networks which are paranoid about
DHCP spoofing, such as most mainstream cable-broadband ISP networks.

Reviewed by: brooks
Obtained from: OpenBSD (partly - I'm not convinced their solution can work)
MFC after: 1 week (pending re approval)

14 years agoAfter the installation of the /usr/share/zoneinfo, run tzsetup if
edwin [Wed, 21 Oct 2009 20:59:12 +0000 (20:59 +0000)]
After the installation of the /usr/share/zoneinfo, run tzsetup if
/var/db/zoneinfo exists.

MFC after: 1 week

14 years ago- Add support for chrooted installs.
edwin [Wed, 21 Oct 2009 20:55:04 +0000 (20:55 +0000)]
- Add support for chrooted installs.
- Add examples to the man-page.

MFC after: 1 week

14 years agoMove sed(1) from cross-tools to bootstrap-tools.
ru [Wed, 21 Oct 2009 19:39:34 +0000 (19:39 +0000)]
Move sed(1) from cross-tools to bootstrap-tools.

14 years agoDon't call the newstate callback as that's dangerous. Rely no ENETRESET
rpaulo [Wed, 21 Oct 2009 19:31:23 +0000 (19:31 +0000)]
Don't call the newstate callback as that's dangerous. Rely no ENETRESET
to DTRT.

MFC after: 3 days

14 years agoChange gcc to assume a default machine architecture of 486 instead of 386
jhb [Wed, 21 Oct 2009 19:26:12 +0000 (19:26 +0000)]
Change gcc to assume a default machine architecture of 486 instead of 386
on "i386".  Doing it in the compiler is deemed to be less fragile then
attempting to provide a default -march setting via bsd.cpu.mk.  FreeBSD
itself has not supported plain 386 CPUs since 5.x.

Suggested by: kan
Requested by: rdivacky
MFC after: 1 month