]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
16 years agoThis commit was generated by cvs2svn to compensate for changes in r178848,
Coleman Kane [Thu, 8 May 2008 13:51:16 +0000 (13:51 +0000)]
This commit was generated by cvs2svn to compensate for changes in r178848,
which included commits to RCS files with non-trunk default branches.

16 years agoVirgin import (trimmed) of eXpat v2.0.1. Discussed and tested with
Coleman Kane [Thu, 8 May 2008 13:51:16 +0000 (13:51 +0000)]
Virgin import (trimmed) of eXpat v2.0.1. Discussed and tested with
sam and phk who are the two consumers of this library. If there is
any other fallout, email me and I will take care of it.

Approved by: sam, phk

16 years agoUpdate heimdal_version.
Doug Rabson [Thu, 8 May 2008 13:11:34 +0000 (13:11 +0000)]
Update heimdal_version.

Pointed out by: antoine@

16 years agoMerge from the vendor branch and resolve conflicts.
Doug Rabson [Thu, 8 May 2008 11:01:46 +0000 (11:01 +0000)]
Merge from the vendor branch and resolve conflicts.

16 years agoThis commit was generated by cvs2svn to compensate for changes in r178843,
Doug Rabson [Thu, 8 May 2008 10:58:50 +0000 (10:58 +0000)]
This commit was generated by cvs2svn to compensate for changes in r178843,
which included commits to RCS files with non-trunk default branches.

16 years agoImport com_err from heimdal-1.1
Doug Rabson [Thu, 8 May 2008 10:58:50 +0000 (10:58 +0000)]
Import com_err from heimdal-1.1

16 years agoAdd a new awk script which parses informations returned by the newly
Attilio Rao [Wed, 7 May 2008 21:50:17 +0000 (21:50 +0000)]
Add a new awk script which parses informations returned by the newly
added sysctl debug.witness.graphs and returns all the graphs involving
Giant lock creating an appropriate script in DOT format which can be
plotted immediately.

Submitted by:   Michele Dallachiesa <michele dot dallachiesa at poste dot it>

16 years agoAdd a new witness sysctl which returns the relations between any lock
Attilio Rao [Wed, 7 May 2008 21:41:36 +0000 (21:41 +0000)]
Add a new witness sysctl which returns the relations between any lock
and its children in the form:
"parent","child"
so that head and bottom of an oriented graph can be easilly detected and
various form of diagrams can be build.
The sysctl is called debug.witness.graphs and it is read-only; in order
to get the list of relations, a simple:
#sysctl debug.witness.graphs
will do the trick.

This approach has been choosen in order to support easilly things like
the DOT format and such.  Soon, an auto-explicative awk script, which
filters simple informations returned by the sysctl and converts them into
a real DOT script, will be committed to the repository between examples.

Discussed with: rwatson

16 years ago- Use the name returned by device_get_nameunit(9) for the name of the
Marius Strobl [Wed, 7 May 2008 21:22:15 +0000 (21:22 +0000)]
- Use the name returned by device_get_nameunit(9) for the name of the
  counter-timer timecounter so the associated SYSCTL nodes don't clash on
  machines having multiple U2P and U2S bridges as well as establishing a
  clear mapping between these bridges and their timecounter device.
- Don't bother setting up a "nice" name for the IOMMU, just use the name
  returned by device_get_nameunit(9), too.
- Fix some minor style(9) bugs.
- Use __FBSDID in counter.c

MFC after: 1 week

16 years agoUse a sledgehammer cast (that was in the original patch to boot) to
John Baldwin [Wed, 7 May 2008 21:00:50 +0000 (21:00 +0000)]
Use a sledgehammer cast (that was in the original patch to boot) to
quiet a warning on 64-bit platforms now that 'size' is an int and not a
size_t.

16 years agoDon't explicitly drop Giant around d_open/d_fdopen/d_close for MPSAFE
John Baldwin [Wed, 7 May 2008 19:03:57 +0000 (19:03 +0000)]
Don't explicitly drop Giant around d_open/d_fdopen/d_close for MPSAFE
drivers.  Since devfs is already marked MPSAFE it shouldn't be held
anyway.

MFC after: 2 weeks
Discussed with: phk

16 years agoFix reading the address of a znode_phys from a znode on 64-bit platforms
John Baldwin [Wed, 7 May 2008 18:27:38 +0000 (18:27 +0000)]
Fix reading the address of a znode_phys from a znode on 64-bit platforms
where sizeof(pointer) != sizeof(int).

MFC after: 1 week
PR: amd64/123456
Submitted by: KOIE Hidetaka | hide koie.org

16 years agoThe debug.sizeof.znode sysctl returns an int, not a size_t. This can cause
John Baldwin [Wed, 7 May 2008 17:55:28 +0000 (17:55 +0000)]
The debug.sizeof.znode sysctl returns an int, not a size_t.  This can cause
a hang on 64-bit platforms.

MFC after: 1 week
PR: amd64/123456
Submitted by: KOIE Hidetaka | hide koie.org

16 years agoOnly output details about the current working directory of a process if
John Baldwin [Wed, 7 May 2008 17:49:31 +0000 (17:49 +0000)]
Only output details about the current working directory of a process if
the vnode pointer is not NULL.  This avoids spurious warnings in fstat -v
output for kernel processes.

MFC after: 1 week
PR: amd64/123456
Submitted by: KOIE Hidetaka | hide koie.org

16 years agoKeep versions on a dependency chain to exclude even remote possiblity
Alexander Kabaev [Wed, 7 May 2008 15:39:34 +0000 (15:39 +0000)]
Keep versions on a dependency chain to exclude even remote possiblity
of private version ever getting index 2.

16 years agoTag FILE's _bf as being part of the public ABI as well due to the in-tree
John Baldwin [Wed, 7 May 2008 15:12:45 +0000 (15:12 +0000)]
Tag FILE's _bf as being part of the public ABI as well due to the in-tree
sort(1) referencing it.

16 years agoFix conflicts after heimdal-1.1 import and add build infrastructure. Import
Doug Rabson [Wed, 7 May 2008 13:53:12 +0000 (13:53 +0000)]
Fix conflicts after heimdal-1.1 import and add build infrastructure. Import
all non-style changes made by heimdal to our own libgssapi.

16 years agoThis commit was generated by cvs2svn to compensate for changes in r178825,
Doug Rabson [Wed, 7 May 2008 13:39:42 +0000 (13:39 +0000)]
This commit was generated by cvs2svn to compensate for changes in r178825,
which included commits to RCS files with non-trunk default branches.

16 years agoVendor import of Heimdal 1.1
Doug Rabson [Wed, 7 May 2008 13:39:42 +0000 (13:39 +0000)]
Vendor import of Heimdal 1.1

16 years agoInclude a very basic (and beta) tool for stressing disks using the POSIX
Adrian Chadd [Wed, 7 May 2008 07:23:47 +0000 (07:23 +0000)]
Include a very basic (and beta) tool for stressing disks using the POSIX
AIO calls.

This small program queues up a controllable number of concurrent AIO
read operations w/ controllable io size against a disk or regular file.
There are a few other things to add (notably optional write support!)
but it works well enough at the present time to stress the AIO code out
relatively harshly in the disk IO case.

16 years ago- change function name from *_vdir to *_vnode because
Daichi GOTO [Wed, 7 May 2008 05:32:55 +0000 (05:32 +0000)]
- change function name from *_vdir to *_vnode because
  VSOCK has been added as cache target. Now they process
  not only VDIR but also VSOCK.
- fixed panic issue caused by cache incorrect free process
  by "umount -f"

Submitted by: Masanori OZAWA <ozawa@ongs.co.jp>
MFC after: 1 week

16 years agoInstall the mpilib headers from mpt(4) into /usr/include/dev/mpt/mpilib.
John Baldwin [Wed, 7 May 2008 04:11:21 +0000 (04:11 +0000)]
Install the mpilib headers from mpt(4) into /usr/include/dev/mpt/mpilib.
This allows <sys/mpt_ioctl.h> to be used from userland.

Prodded by: scottl

16 years agoFix spelling in comment.
Julian Elischer [Tue, 6 May 2008 22:41:23 +0000 (22:41 +0000)]
Fix spelling in comment.

16 years agoAdd a new personality to mpt(4) devices to allow userland applications to
John Baldwin [Tue, 6 May 2008 20:49:53 +0000 (20:49 +0000)]
Add a new personality to mpt(4) devices to allow userland applications to
perform various operations on a controller.  Specifically, for each mpt(4)
device, create a character device in devfs which accepts ioctl requests for
reading and writing configuration pages and performing RAID actions.

MFC after: 1 week
Reviewed by: scottl

16 years agofix build
Kip Macy [Tue, 6 May 2008 17:45:54 +0000 (17:45 +0000)]
fix build

16 years agoUpdate the date on the man-page to reflect the date that the '-u name'
Garance A Drosehn [Tue, 6 May 2008 16:06:02 +0000 (16:06 +0000)]
Update the date on the man-page to reflect the date that the '-u name'
change was committed, instead of when I had first started writing it...

16 years agoSpecify the full path to the md5(1) binary so the script will
Mike Makonnen [Tue, 6 May 2008 10:40:20 +0000 (10:40 +0000)]
Specify the full path to the md5(1) binary so the script will
still work even if it's not in the shell's path.

PR: conf/122215
MFC after: 1 week

16 years agoFix the problem with the C++ exception handling for the multithreaded
Konstantin Belousov [Tue, 6 May 2008 09:27:41 +0000 (09:27 +0000)]
Fix the problem with the C++ exception handling for the multithreaded
programs.

From the PR description:
The gcc runtime's _Unwind_Find_FDE function, invoked during exception
handling's stack unwinding, is not safe to execute from within multiple
threads. FreeBSD' s dl_iterate_phdr() however permits multiple threads
to pass through it though. The result is surprisingly reliable infinite
looping of one or more threads if they just happen to be unwinding at
the same time.

Introduce the new lock that is write locked around the dl_iterate_pdr,
thus providing required exclusion for the stack unwinders.

PR: threads/123062
Submitted by: Andy Newman <an at atrn org>
Reviewed by: kan
MFC after: 2 weeks

16 years agoRemove libc_r's remnant code.
David Xu [Tue, 6 May 2008 07:27:11 +0000 (07:27 +0000)]
Remove libc_r's remnant code.

16 years agoonly build iw_cxgb on i386 and amd64
Kip Macy [Tue, 6 May 2008 02:31:27 +0000 (02:31 +0000)]
only build iw_cxgb on i386 and amd64

16 years agoAdd a comment stating not to bump the FBSDprivate version.
Daniel Eischen [Tue, 6 May 2008 01:41:55 +0000 (01:41 +0000)]
Add a comment stating not to bump the FBSDprivate version.

Don't inherit the public namespace from the private namespace.

16 years agoWhen testing whether to enter the audit argument gathering code, rather
Robert Watson [Tue, 6 May 2008 00:32:23 +0000 (00:32 +0000)]
When testing whether to enter the audit argument gathering code, rather
than checking whether audit is enabled globally, instead check whether
the current thread has an audit record.  This avoids entering the audit
code to collect argument data if auditing is enabled but the current
system call is not of interest to audit.

MFC after: 1 week
Sponsored by: Apple, Inc.

16 years agoreplace spaces added in last change with tabs
Kip Macy [Mon, 5 May 2008 23:13:27 +0000 (23:13 +0000)]
replace spaces added in last change with tabs

16 years agoconditionally define PANIC_IF, remove 'unlikely'
Kip Macy [Mon, 5 May 2008 22:37:21 +0000 (22:37 +0000)]
conditionally define PANIC_IF, remove 'unlikely'

16 years agoadd rdma to build for i386 and amd64
Kip Macy [Mon, 5 May 2008 20:41:54 +0000 (20:41 +0000)]
add rdma to build for i386 and amd64

16 years agoLINT fixes
Kip Macy [Mon, 5 May 2008 20:41:10 +0000 (20:41 +0000)]
LINT fixes

16 years agoadd iw_cxgb to the build
Kip Macy [Mon, 5 May 2008 20:21:20 +0000 (20:21 +0000)]
add iw_cxgb to the build

16 years agoadd makefiles for rdma
Kip Macy [Mon, 5 May 2008 20:19:33 +0000 (20:19 +0000)]
add makefiles for rdma

16 years agoadd rcv_nxt, snd_nxt, and toe offload id to FreeBSD-specific
Kip Macy [Mon, 5 May 2008 20:13:31 +0000 (20:13 +0000)]
add rcv_nxt, snd_nxt, and toe offload id to FreeBSD-specific
extension fields for tcp_info

16 years agoadd malloc flag to blist so that it can be used in ithread context
Kip Macy [Mon, 5 May 2008 19:48:54 +0000 (19:48 +0000)]
add malloc flag to blist so that it can be used in ithread context

Reviewed by: alc, bsdimp

16 years agoconditionally define PANIC_IF
Kip Macy [Mon, 5 May 2008 19:39:20 +0000 (19:39 +0000)]
conditionally define PANIC_IF

16 years agoRevert back to accessing FILE internals directly.
John Baldwin [Mon, 5 May 2008 19:38:25 +0000 (19:38 +0000)]
Revert back to accessing FILE internals directly.

(Sorry, forgot to commit this earlier.)

16 years agoFix a few edge cases with error handling in cpufreq(4)'s CPUFREQ_GET()
John Baldwin [Mon, 5 May 2008 19:13:52 +0000 (19:13 +0000)]
Fix a few edge cases with error handling in cpufreq(4)'s CPUFREQ_GET()
method:
- If the last of the child cpufreq drivers returns an error while trying to
  fetch its list of supported frequencies but an earlier driver found the
  requested frequency, don't return an error to the caller.
- If all of the child cpufreq drivers fail and the attempt to match the
  frequency based on 'cpu_est_clockrate()' fails, return ENXIO rather than
  returning success and returning a frequency of CPUFREQ_VAL_UNKNOWN.

MFC after: 3 days
PR: kern/121433
Reported by: Eugene Grosbein  eugen ! kuzbass dot ru

16 years agoimport support for iwarp on Chelsio T3 card
Kip Macy [Mon, 5 May 2008 18:46:18 +0000 (18:46 +0000)]
import support for iwarp on Chelsio T3 card

Supported by Chelsio Inc.

16 years agoUse a better approach to force the interrupt which should work for
Bjoern A. Zeeb [Mon, 5 May 2008 18:42:17 +0000 (18:42 +0000)]
Use a better approach to force the interrupt which should work for
all cards/modes.
In addition to the intr forcing added with rev. 1.205 adopt the other
places to use the same logic.

We need to exclude a few chips/revisions (5700, 5788) from using the
enhanced version and fall back to the old way as that is the only
method they support.

Tested by: phk
Suggested by: davidch, Broadcom (thanks a lot for the help!)
MFC after: 16 days

16 years agoImport basic common and iwarp kernel RDMA infrastructure.
Kip Macy [Mon, 5 May 2008 18:35:55 +0000 (18:35 +0000)]
Import basic common and iwarp kernel RDMA infrastructure.

Supported by: Chelsio Inc.

16 years agoRetire the __fgetcookie(), __fgetpendout(), and __fsetfileno() accessors
John Baldwin [Mon, 5 May 2008 16:14:02 +0000 (16:14 +0000)]
Retire the __fgetcookie(), __fgetpendout(), and __fsetfileno() accessors
as we aren't hiding FILE's internals anymore.

16 years agoGo back to fondling FILE internals.
John Baldwin [Mon, 5 May 2008 16:12:28 +0000 (16:12 +0000)]
Go back to fondling FILE internals.

16 years agoNote that FILE's __cookie is also part of the public ABI.
John Baldwin [Mon, 5 May 2008 16:04:58 +0000 (16:04 +0000)]
Note that FILE's __cookie is also part of the public ABI.

16 years agoExpose FILE's internals to the world again in all their glory. Restore
John Baldwin [Mon, 5 May 2008 16:03:52 +0000 (16:03 +0000)]
Expose FILE's internals to the world again in all their glory.  Restore
all the previous inline optimizations as well.  FILE is back to using
__mbstate_t, struct pthread *, and struct pthread_mutex *.

16 years agoo Convert whitespaces to tabs.
Maxim Konovalov [Mon, 5 May 2008 15:52:54 +0000 (15:52 +0000)]
o Convert whitespaces to tabs.

16 years agoo Terminate "case" with "esac" not "fi".
Maxim Konovalov [Mon, 5 May 2008 15:50:20 +0000 (15:50 +0000)]
o Terminate "case" with "esac" not "fi".

Reported by: Randy Bush

16 years agoDo not read away the target directory entry when encountering deleted
Konstantin Belousov [Mon, 5 May 2008 14:05:23 +0000 (14:05 +0000)]
Do not read away the target directory entry when encountering deleted
files after a seekdir().

The seekdir shall set the position for the next readdir operation.
When the _readdir_unlocked() encounters deleted entry, dd_loc is
already advanced. Continuing the loop leads to premature read of
the target entry.

Submitted by: Marc Balmer <mbalmer at openbsd org>
Obtained from: OpenBSD
MFC after: 2 weeks

16 years agoSpell KDB_REQ_REBOOT correctly.
Peter Wemm [Mon, 5 May 2008 08:25:46 +0000 (08:25 +0000)]
Spell KDB_REQ_REBOOT correctly.

16 years agoFix improper use of checkyesno routine.
Mike Makonnen [Mon, 5 May 2008 07:43:48 +0000 (07:43 +0000)]
Fix improper use of checkyesno routine.

Noticed by: oliver
MFC after: 1 week

16 years agoo Change the warning dialog for the 'W' command in both the label
Mike Makonnen [Mon, 5 May 2008 06:31:41 +0000 (06:31 +0000)]
o Change the warning dialog for the 'W' command in both the label
  and partition editors to reflect the fact that this is a stand-alone
  application, not sysinstall(8).
o Change an instance of sade(8) refering to itself as sysinstall(8) in
  a confirmation dialog.

MFC after: 1 week

16 years agoMips ttys file. Copied from i386 version with removal of the vga
Warner Losh [Mon, 5 May 2008 05:35:47 +0000 (05:35 +0000)]
Mips ttys file.  Copied from i386 version with removal of the vga
entries.

16 years agoMFSVN:
Kip Macy [Mon, 5 May 2008 01:41:53 +0000 (01:41 +0000)]
MFSVN:
 - add / remove clients from cxgb_main.c now
 - change ifdef TOE_ENABLED to TCP_OFFLOAD_DISABLE
 - update copyrights
 - fix transmit data mismatch bug caused by not setting SB_NOCOALESCE
   on tx sockbuf on passive connections
 - fix receive sequence mismatch bug caused by not setting SB_NOCOALESCE
   on rx sockbuf on passive connections
 - don't sleep without checking SBS_CANTRCVMORE first
 - various ddp ordering fixes

Supported by: Chelsio Inc.

16 years agoExpand kdb_alt_break a little, most commonly used with the option
Peter Wemm [Sun, 4 May 2008 23:29:38 +0000 (23:29 +0000)]
Expand kdb_alt_break a little, most commonly used with the option
ALT_BREAK_TO_DEBUGGER.  In addition to "Enter ~ ctrl-B" (to enter the
debugger), there is now "Enter ~ ctrl-P" (force panic) and
"Enter ~ ctrl-R" (request clean reboot, ala ctrl-alt-del on syscons).

We've used variations of this at work.  The force panic sequence is
best used with KDB_UNATTENDED for when you just want it to dump and
get on with it.

The reboot request is a safer way of getting into single user than
a power cycle.  eg: you've hosed the ability to log in (pam, rtld, etc).
It gives init the reboot signal, which causes an orderly reboot.

I've taken my best guess at what the !x86 and non-sio code changes
should be.

This also makes sio release its spinlock before calling KDB/DDB.

16 years ago Add MIPS support to libdisk
Oleksandr Tymoshenko [Sun, 4 May 2008 22:24:40 +0000 (22:24 +0000)]
  Add MIPS support to libdisk

  Approved by: cognet (mentor)

16 years agoRestore SUBDIR+= accidentally removed in the previous revision.
Marius Strobl [Sun, 4 May 2008 16:02:45 +0000 (16:02 +0000)]
Restore SUBDIR+= accidentally removed in the previous revision.

Pointed out by: ariff

16 years agoDon't build unused SBus front-ends for sun4v, don't build EBus front-ends
Marius Strobl [Sun, 4 May 2008 14:59:25 +0000 (14:59 +0000)]
Don't build unused SBus front-ends for sun4v, don't build EBus front-ends
which are also likely to be irrelevant for sun4v (there's no SBus on sun4v
and only some EBus devices). While at it fix some style bugs according to
style.Makefile(5) where appropriate.

MFC after: 3 days

16 years agosync_vnode() has some messy code about locking in order to deal with
Attilio Rao [Sun, 4 May 2008 13:54:55 +0000 (13:54 +0000)]
sync_vnode() has some messy code about locking in order to deal with
mount fs needing Giant to be held when processing bufobjs.
Use a different subqueue for pending workitems on filesystems requiring
Giant. This simplifies the code notably and also reduces the number of
Giant acquisitions (and the whole processing cost).

Suggested by: jeff
Reviewed by: kib
Tested by: pho

16 years agoLower WARNS to avoid warning about type-punned casting.
Marcel Moolenaar [Sun, 4 May 2008 04:13:35 +0000 (04:13 +0000)]
Lower WARNS to avoid warning about type-punned casting.

16 years agoAdd __fgetcookie(), __fgetpendout() and __fsetfileno() to the private
Marcel Moolenaar [Sun, 4 May 2008 04:11:01 +0000 (04:11 +0000)]
Add __fgetcookie(), __fgetpendout() and __fsetfileno() to the private
name space.

16 years agoUnbreak build: gnu sort has been configured to grope inside struct
Marcel Moolenaar [Sat, 3 May 2008 23:36:00 +0000 (23:36 +0000)]
Unbreak build: gnu sort has been configured to grope inside struct
__sFILE. It's opaque now, so add a function that returns the pending
output bytes.

Pointy hat: jhb

16 years ago- Backout 1.15, it was committed by accident
Pav Lucistnik [Sat, 3 May 2008 23:17:37 +0000 (23:17 +0000)]
- Backout 1.15, it was committed by accident

Pointy hat to: pav

16 years ago- Restore functionality broken in previous commit; we need to be able to report
Pav Lucistnik [Sat, 3 May 2008 22:56:50 +0000 (22:56 +0000)]
- Restore functionality broken in previous commit; we need to be able to report
  multiple installed packages with the same PKGORIGIN.

Reported by: marcus
MFC after: 1 month

16 years agoo unbreak handling of TKIP tx-only keys for splitmic chips
Sam Leffler [Sat, 3 May 2008 21:55:02 +0000 (21:55 +0000)]
o unbreak handling of TKIP tx-only keys for splitmic chips
o yank compat support for hal's older than 0.9.20.3; leave a
  CTASSERT in place just in case

16 years agoadd back sysctl's to display the regdomain and country code from eeprom;
Sam Leffler [Sat, 3 May 2008 21:52:05 +0000 (21:52 +0000)]
add back sysctl's to display the regdomain and country code from eeprom;
useful for debugging

16 years ago Add MIPS to the list of "no TLS" architectures.
Oleksandr Tymoshenko [Sat, 3 May 2008 21:24:28 +0000 (21:24 +0000)]
  Add MIPS to the list of "no TLS" architectures.

  Approved by: cognet (mentor)

16 years ago Symbol.map is handled by cpp, so use C-style comments
Oleksandr Tymoshenko [Sat, 3 May 2008 21:16:08 +0000 (21:16 +0000)]
  Symbol.map is handled by cpp, so use C-style comments

  Approved by: cognet (mentor)

16 years ago Add FLT_EVAL_METHOD and DECIMAL_DIG, required by C99 standard.
Oleksandr Tymoshenko [Sat, 3 May 2008 21:04:21 +0000 (21:04 +0000)]
  Add FLT_EVAL_METHOD and DECIMAL_DIG, required by C99 standard.

  Approved by: cognet (mentor)

16 years agoUnbreak build: libftpio gropes inside struct __sFILE. Implement
Marcel Moolenaar [Sat, 3 May 2008 20:09:44 +0000 (20:09 +0000)]
Unbreak build: libftpio gropes inside struct __sFILE. Implement
accessor functions for its benefit now thaat FILE is opaque.
I'm sure there's a better way. I leave that for people to work
on in a src tree that isn't broken.

Pointy hat: jhb

16 years agoA couple of cameras that I recently bought
Warner Losh [Sat, 3 May 2008 20:06:15 +0000 (20:06 +0000)]
A couple of cameras that I recently bought

16 years agoThese files are unused, so remove them for now. If they turn out to
Warner Losh [Sat, 3 May 2008 19:57:45 +0000 (19:57 +0000)]
These files are unused, so remove them for now.  If they turn out to
be needed later, they can be restored.

16 years agoFix a comment.
Jason Evans [Sat, 3 May 2008 17:49:16 +0000 (17:49 +0000)]
Fix a comment.

16 years agoenable IEEE80211_AMDPU_AGE by default
Sam Leffler [Sat, 3 May 2008 17:06:59 +0000 (17:06 +0000)]
enable IEEE80211_AMDPU_AGE by default

16 years agoenable IEEE80211_DEBUG and IEEE80211_AMPDU_AGE by default
Sam Leffler [Sat, 3 May 2008 17:05:38 +0000 (17:05 +0000)]
enable IEEE80211_DEBUG and IEEE80211_AMPDU_AGE by default

16 years ago Bring C runtime bits for FreeBSD/mips from p4 mips2-jnpr branch.
Oleksandr Tymoshenko [Sat, 3 May 2008 11:16:32 +0000 (11:16 +0000)]
  Bring C runtime bits for FreeBSD/mips from p4 mips2-jnpr branch.

  Approved by: cognet (mentor)

16 years agoReplace a couple mentions of the soon to be removed vaps_<ifn>
Brooks Davis [Sat, 3 May 2008 07:06:48 +0000 (07:06 +0000)]
Replace a couple mentions of the soon to be removed vaps_<ifn>
variable form with wlans_<ifn>.

16 years agosade(8) does not need FTP I/O.
Xin LI [Sat, 3 May 2008 02:42:57 +0000 (02:42 +0000)]
sade(8) does not need FTP I/O.

16 years agoRelinquish exclusive TTY access when tip(1) or cu(1) exit.
Bruce M Simpson [Sat, 3 May 2008 02:29:02 +0000 (02:29 +0000)]
Relinquish exclusive TTY access when tip(1) or cu(1) exit.
Previously they would have left TIOCEXCL enabled, requiring
either a reboot or use of tip/cu as the root user.

Observed when running QEMU with character devices redirected to pty instances.

MFC after: 2 weeks

16 years agoFix build, together with a bit of style breakage.
Dmitry Morozovsky [Fri, 2 May 2008 18:54:36 +0000 (18:54 +0000)]
Fix build, together with a bit of style breakage.

16 years agoThis file is unused, so remove it for now.
Warner Losh [Fri, 2 May 2008 18:03:00 +0000 (18:03 +0000)]
This file is unused, so remove it for now.

16 years agoRemove an header which is unused for sun4v.
Marius Strobl [Fri, 2 May 2008 17:44:18 +0000 (17:44 +0000)]
Remove an header which is unused for sun4v.

MFC after: 3 days

16 years agoDon't built the unused counter-timer abstraction.
Marius Strobl [Fri, 2 May 2008 17:41:52 +0000 (17:41 +0000)]
Don't built the unused counter-timer abstraction.

MFC after: 3 days

16 years agoFix type name: struct auditpipe_preselect is actually struct
Robert Watson [Fri, 2 May 2008 17:36:22 +0000 (17:36 +0000)]
Fix type name: struct auditpipe_preselect is actually struct
auditpipe_ioctl_preselect.

MFC after: 3 days
Submitted by: Stacey D. Son <stacey at son.org>

16 years agoRestore multi-release tradition of the driver.
Jung-uk Kim [Fri, 2 May 2008 17:02:35 +0000 (17:02 +0000)]
Restore multi-release tradition of the driver.

Reviewed by: mjacob

16 years agoAdd AUTHORS section[*] and fix HISTORY section.
Hiroki Sato [Fri, 2 May 2008 16:23:47 +0000 (16:23 +0000)]
Add AUTHORS section[*] and fix HISTORY section.

Requested by: Dave Yost (original author)[*]
History checked by: The CSRG Archives
MFC after: 3 days

16 years agoAxe now-empty __BSD_VISIBLE block that held renameat().
John Baldwin [Fri, 2 May 2008 15:59:22 +0000 (15:59 +0000)]
Axe now-empty __BSD_VISIBLE block that held renameat().

Reported by: kib
Pointy hat: jhb

16 years ago- Move declaration of renameat() to the existing section of BSD_VISIBLE
John Baldwin [Fri, 2 May 2008 15:28:23 +0000 (15:28 +0000)]
- Move declaration of renameat() to the existing section of BSD_VISIBLE
  function prototypes.
- Fix a few whitespace inconsistencies in prototypes.

MFC after: 1 month

16 years agoNext round of stdio changes: Remove all inlining of stdio operations and
John Baldwin [Fri, 2 May 2008 15:25:07 +0000 (15:25 +0000)]
Next round of stdio changes: Remove all inlining of stdio operations and
move the definition of the type backing FILE (struct __sFILE) into an
internal header.
- Remove macros to inline certain operations from stdio.h.  Applications
  will now always call the functions instead.
- Move the various foo_unlocked() functions from unlocked.c into foo.c.
  This lets some of the inlining macros (e.g. __sfeof()) move into
  foo.c.
- Update a few comments.
- struct __sFILE can now go back to using mbstate_t, pthread_t, and
  pthread_mutex_t instead of knowing about their private, backing types.

MFC after: 1 month
Reviewed by: kan

16 years agoInclude libc_private.h for the declaration of __isthreaded instead of
John Baldwin [Fri, 2 May 2008 14:51:22 +0000 (14:51 +0000)]
Include libc_private.h for the declaration of __isthreaded instead of
relying on namespace pollution in stdio.h.

MFC after: 3 days

16 years agoRemove unused variable saved_id16.
Rui Paulo [Fri, 2 May 2008 10:16:41 +0000 (10:16 +0000)]
Remove unused variable saved_id16.

Pointy hat to: me
Pointed out by: jhb
MFC after: 1 week

16 years agoDocumentation updates:
Tim Kientzle [Fri, 2 May 2008 05:40:05 +0000 (05:40 +0000)]
Documentation updates:
  * --format can be used with -r or -u
  * -o is a synonym for --format=ustar when used with -c, -r, or -u
Also, fix the erroneous sanity check that suppressed --format with -r or -u.

16 years agobsdtar --version should succeed.
Tim Kientzle [Fri, 2 May 2008 05:18:47 +0000 (05:18 +0000)]
bsdtar --version should succeed.

16 years agoNew bsdtar test harness. Still rather skimpy, but a lot easier
Tim Kientzle [Fri, 2 May 2008 05:17:16 +0000 (05:17 +0000)]
New bsdtar test harness.  Still rather skimpy, but a lot easier
to run and maintain than the old scripts that used to be here.

16 years agoAllow -r with -T even if there are no files on the command line.
Tim Kientzle [Fri, 2 May 2008 05:14:58 +0000 (05:14 +0000)]
Allow -r with -T even if there are no files on the command line.

PR: bin/123246
MFC after: 3 days

16 years ago- Change how the vmcore target maps FreeBSD thread IDs to GDB ptids. We
John Baldwin [Thu, 1 May 2008 20:36:48 +0000 (20:36 +0000)]
- Change how the vmcore target maps FreeBSD thread IDs to GDB ptids.  We
  now only use the TID and ignore the PID and use pid_to_ptid() to build a
  ptid treating the TID as a PID.  The benefit of this is that the vmcore
  target now uses the same scheme as GDB's remote targets.  As a result,
  the 'tid' command now works for remote targets (however, it only accepts
  TIDs and not addresses of 'struct thread' objects).
- Use gdb_thread_select() to do the actual thread switch for the 'tid' and
  'proc' commands.  This now gives the same UI feedback when switching
  threads as the GDB 'thread' command rather than providing no visual
  output at all.

MFC after: 1 week