]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
16 years agoDocument BOOTP_BLOCKSIZE.
benno [Fri, 16 May 2008 06:50:40 +0000 (06:50 +0000)]
Document BOOTP_BLOCKSIZE.

16 years agoAllow the block size used when booting over NFS to be overridden. It defaults
benno [Fri, 16 May 2008 06:27:03 +0000 (06:27 +0000)]
Allow the block size used when booting over NFS to be overridden.  It defaults
to 8192 bytes which is the size currently used.

16 years agoClarify that "ante meridiem" and "post meridiem" mean the same thing
grog [Fri, 16 May 2008 04:33:04 +0000 (04:33 +0000)]
Clarify that "ante meridiem" and "post meridiem" mean the same thing
as the more commonly used "a.m." and "p.m.".

Tripped over by:  Callum Gibson.

MFC after:  2 weeks

16 years agoThe beacon miss notification must run without locks held has it calls back into
thompsa [Fri, 16 May 2008 04:15:54 +0000 (04:15 +0000)]
The beacon miss notification must run without locks held has it calls back into
wpi_raw_xmit();

16 years agoSpelling and capitalization fixes.
scf [Fri, 16 May 2008 03:13:36 +0000 (03:13 +0000)]
Spelling and capitalization fixes.

MFC after: 3 days

16 years agoLO_ENROLLPEND is no more existing so just axe it (it was left out by the
attilio [Fri, 16 May 2008 02:09:13 +0000 (02:09 +0000)]
LO_ENROLLPEND is no more existing so just axe it (it was left out by the
original commit axing it).

16 years agoMake this c++ compatible.
dfr [Fri, 16 May 2008 02:06:10 +0000 (02:06 +0000)]
Make this c++ compatible.

PR: 87726

16 years agolockinit() can't accept LK_EXCLUSIVE as an initializaiton flag, so just
attilio [Thu, 15 May 2008 21:39:25 +0000 (21:39 +0000)]
lockinit() can't accept LK_EXCLUSIVE as an initializaiton flag, so just
drop it.

Reported by: Josh Carroll <josh dot carroll at gmail dot com>
Submitted by: jhb

16 years ago- Const'ify firmware and lookup-tables.
marius [Thu, 15 May 2008 20:27:18 +0000 (20:27 +0000)]
- Const'ify firmware and lookup-tables.
- Obsolete redundant inst_name and unit members of struct sym_hcb.
- Fix three more NULL vs. 0 confusions.
- Use device_set_softc(9) to tell the bus layer that this driver
  allocates a instance of struct sym_hcb itself.

16 years agoUpdate the kernel to count the number of mbufs and clusters
gnn [Thu, 15 May 2008 20:18:44 +0000 (20:18 +0000)]
Update the kernel to count the number of mbufs and clusters
(all types) used per socket buffer.

Add support to netstat to print out all of the socket buffer
statistics.

Update the netstat manual page to describe the new -x flag
which gives the extended output.

Reviewed by: rwatson, julian

16 years ago- Embed the recursion counter for any locking primitive directly in the
attilio [Thu, 15 May 2008 20:10:06 +0000 (20:10 +0000)]
- Embed the recursion counter for any locking primitive directly in the
  lock_object, using an unified field called lo_data.
- Replace lo_type usage with the w_name usage and at init time pass the
  lock "type" directly to witness_init() from the parent lock init
  function.  Handle delayed initialization before than
  witness_initialize() is called through the witness_pendhelp structure.
- Axe out LO_ENROLLPEND as it is not really needed.  The case where the
  mutex init delayed wants to be destroyed can't happen because
  witness_destroy() checks for witness_cold and panic in case.
- In enroll(), if we cannot allocate a new object from the freelist,
  notify that to userspace through a printf().
- Modify the depart function in order to return nothing as in the current
  CVS version it always returns true and adjust callers accordingly.
- Fix the witness_addgraph() argument name prototype.
- Remove unuseful code from itismychild().

This commit leads to a shrinked struct lock_object and so smaller locks,
in particular on amd64 where 2 uintptr_t (16 bytes per-primitive) are
gained.

Reviewed by: jhb

16 years agoChange a use of u_int32_t to uint32_t.
brooks [Thu, 15 May 2008 20:04:36 +0000 (20:04 +0000)]
Change a use of u_int32_t to uint32_t.

PR: bin/93172
Submitted by: Robert Millan <rmh at aybabtu dot com>
MFC after: 1 week

16 years agoTest the expansion of $LINENO.
stefanf [Thu, 15 May 2008 19:58:44 +0000 (19:58 +0000)]
Test the expansion of $LINENO.

16 years agoExpand $LINENO to the current line number. This is required by SUSv3's "User
stefanf [Thu, 15 May 2008 19:55:27 +0000 (19:55 +0000)]
Expand $LINENO to the current line number.  This is required by SUSv3's "User
Portability Utilities" option.

Often configure scripts generated by the autotools test if $LINENO works and
refuse to use /bin/sh if not.

Package test run by: pav

16 years agoAdd definition for OM_uint64.
dfr [Thu, 15 May 2008 19:55:19 +0000 (19:55 +0000)]
Add definition for OM_uint64.

Submitted by: tmclaugh

16 years agogetopt.c is public domain. Add a comment to that effect.
brooks [Thu, 15 May 2008 19:27:52 +0000 (19:27 +0000)]
getopt.c is public domain.  Add a comment to that effect.

Remove confusing README.

PR: bin/98911
Submitted by: Jason McIntyre <jmc at kerhand dot co dot uk>
Obtained from: OpenBSD
MFC after: 3 days

16 years agoDon't call vm_reserv_alloc_page() on device-backed objects. Otherwise, the
alc [Thu, 15 May 2008 18:52:31 +0000 (18:52 +0000)]
Don't call vm_reserv_alloc_page() on device-backed objects.  Otherwise, the
system may panic because there is no reservation structure corresponding to
the physical address of the device memory.

Reported by: Giorgos Keramidas

16 years agoAdd the ability to do all read, all write, or random read/write.
adrian [Thu, 15 May 2008 16:10:55 +0000 (16:10 +0000)]
Add the ability to do all read, all write, or random read/write.

16 years agoAdd new heimdal-1.1 library.
dfr [Thu, 15 May 2008 15:28:18 +0000 (15:28 +0000)]
Add new heimdal-1.1 library.

16 years agoImprove the virtual scrolling mechanism to make middle clicking less
philip [Thu, 15 May 2008 15:05:02 +0000 (15:05 +0000)]
Improve the virtual scrolling mechanism to make middle clicking less
difficult.  Add a -L option (yet another option, indeed!) which changes the
speed of scrolling and change -U to only affect the scroll threshold.

This should make middle-clicking a much more pleasant experience.

PR: bin/120186
Submitted by: Aragon Gouveia <aragon -at- phat.za.net>
MFC after: 3 days

16 years agoAdd support for /conf/T/M/remount_optional.
bms [Thu, 15 May 2008 11:00:23 +0000 (11:00 +0000)]
Add support for /conf/T/M/remount_optional.

The rc.initdiskless functionality is used by NanoBSD to allow configuration
files to live on a separate configuration slice, which acts as NVRAM, whilst
the system image is mounted read-only.

Normally, if the remount command fails during boot, this is regarded as
a fatal error. If /conf/T/M/remount_optional is present, this error is
non-fatal. If the file is not present, the default behaviour is unchanged.

This is very useful for people building live CD images using FreeBSD,
where the NVRAM lives somewhere completely differently from the system image,
and may be present on removable media which is not present during the
initial boot.

16 years agoTypo
bms [Thu, 15 May 2008 10:51:30 +0000 (10:51 +0000)]
Typo

16 years agoMake the meaning of the %A format specifier, as passed to
bms [Thu, 15 May 2008 10:47:14 +0000 (10:47 +0000)]
Make the meaning of the %A format specifier, as passed to
LD_TRACE_LOADED_OBJECTS_FMT[12], more obvious for users like me.

16 years agoAdd an example of how to use ldd -f.
bms [Thu, 15 May 2008 10:43:11 +0000 (10:43 +0000)]
Add an example of how to use ldd -f.

16 years agoAdd the hx509 error table.
dfr [Thu, 15 May 2008 08:53:31 +0000 (08:53 +0000)]
Add the hx509 error table.

16 years agoFix a panic when it occurred during initializing the ndis driver because
weongyo [Thu, 15 May 2008 04:29:28 +0000 (04:29 +0000)]
Fix a panic when it occurred during initializing the ndis driver because
it try to read network address through ifnet structure which is NULL
until the ndis driver's initialization is finished.

Reviewed by: thompsa

16 years agoGo back to using the process command name (p_comm) for the file name and
jhb [Thu, 15 May 2008 03:07:34 +0000 (03:07 +0000)]
Go back to using the process command name (p_comm) for the file name and
command line arguments stored in the note at the beginning of a core dump
instead of the current thread name.

Reviewed by: julian

16 years agoFix for a bug I introduced when I cleaned up atacontrol: Don't terminate
phk [Thu, 15 May 2008 01:25:29 +0000 (01:25 +0000)]
Fix for a bug I introduced when I cleaned up atacontrol:  Don't terminate
if we are listing devices, a controller might legitimately not be there.

Submitted by: "Andrey V. Elsukov" <bu7cher@yandex.ru>

16 years agoMove speaker a lot closer to style(9)
phk [Thu, 15 May 2008 01:22:48 +0000 (01:22 +0000)]
Move speaker a lot closer to style(9)

Submitted by: Martin Voros <martin_voros@yahoo.com>

16 years agoChange the default value of synchronous_dhclient to NO.
brooks [Thu, 15 May 2008 01:06:10 +0000 (01:06 +0000)]
Change the default value of synchronous_dhclient to NO.

To preserve the existing behavior of etc/rc.d/netif, add code to wait
up to if_up_delay seconds (30 seconds by default) for a default route to
be configured if there are any dhcp interfaces.  This should be extended
to test that the interface is actually up.

X-MFC after:

16 years agoFix last commit and call childif_destroy() correctly.
brooks [Thu, 15 May 2008 00:08:02 +0000 (00:08 +0000)]
Fix last commit and call childif_destroy() correctly.

16 years agoDon't print the interface status if we only create child or destroy
brooks [Wed, 14 May 2008 23:53:39 +0000 (23:53 +0000)]
Don't print the interface status if we only create child or destroy
interfaces.

Correctly return status from childif_create().

16 years agoPopulate usage()
phk [Wed, 14 May 2008 23:29:02 +0000 (23:29 +0000)]
Populate usage()

Submitted by:   Jaakko Heinonen <jh@saunalahti.fi>

16 years agoImprove the integration of BCM5906[M] support:
marius [Wed, 14 May 2008 21:00:27 +0000 (21:00 +0000)]
Improve the integration of BCM5906[M] support:
- Rename BGE_FLAG_EEPROM to BGE_FLAG_EADDR to underline it's absence means
  "there's no chip containing an Ethernet address fitted to the BGE chip
  so we have to get it from the firmware instead" rather than "there's no
  EEPROM, but maybe NVRAM or something else".
- Don't treat BCM5906[M] generally like chips w/o BGE_FLAG_EADDR set, just
  in the two cases really necessary. This gets us line with the original
  patch for DragonFlyBSD.
- For sparc64 restore the intended behavior of obtaining the Ethernet
  address from the firmware in case BGE_FLAG_EADDR is not set, even for
  BCM5906[M].
- Fix some style(9) bugs introduced with rev. 1.208 of if_bge.c

Approved by: jhb
Additional testing by: Thomas Nystroem (BCM5906)

16 years agoClean up cpp logic and comments.
jasone [Wed, 14 May 2008 18:33:13 +0000 (18:33 +0000)]
Clean up cpp logic and comments.

16 years agoAdd support for bzip2-compressed manpages.
ru [Wed, 14 May 2008 17:18:12 +0000 (17:18 +0000)]
Add support for bzip2-compressed manpages.

PR: bin/120730
Submitted by: ighighi AT gmail DOT com
MFC after: 1 week

16 years agoAdd an ability to run man(1) on local files (the argument should
ru [Wed, 14 May 2008 17:08:31 +0000 (17:08 +0000)]
Add an ability to run man(1) on local files (the argument should
contain a `/' character); based on the submission in the PR.

PR: bin/120730
MFC after: 1 week

16 years agoMake -t <tty> optional in server mode. If not specified use stdin/stdout.
emax [Wed, 14 May 2008 16:47:30 +0000 (16:47 +0000)]
Make -t <tty> optional in server mode. If not specified use stdin/stdout.
Document this. Do not require channel number in server mode. If not
specified - bind to ''wildcard'' channel zero. Real channel number will
be obtained automatically and registered with local sdpd(8). While I'm
here fix serial port service registration.

Submitted by: luigi
Tested by: Helge Oldach <freebsd-bluetooth at oldach dot net>
MFC after: 3 days

16 years agoAdd the devctl notifications for the cdev create/destroy events.
kib [Wed, 14 May 2008 14:29:54 +0000 (14:29 +0000)]
Add the devctl notifications for the cdev create/destroy events.

Based on the submission by: Andriy Gapon <avg icyb net ua>
MFC after: 2 weeks

16 years agostyle(9): remove FreeBSD CVS ID from the initial license comment.
rpaulo [Wed, 14 May 2008 10:02:25 +0000 (10:02 +0000)]
style(9): remove FreeBSD CVS ID from the initial license comment.

MFC after: 1 day

16 years agoActually, don't rely on the unsafe MAX() macro. Use imax() as provided
rpaulo [Wed, 14 May 2008 09:57:21 +0000 (09:57 +0000)]
Actually, don't rely on the unsafe MAX() macro. Use imax() as provided
in the PR patch.

Pointed out by: bde
PR: 123542

16 years agoFix pstat behaviour when using coredumps. The reference to tp was
remko [Wed, 14 May 2008 00:22:57 +0000 (00:22 +0000)]
Fix pstat behaviour when using coredumps. The reference to tp was
incorrect and should have been poining to &tty, tp is a virtual
address from the coredump, while we should obtain the address through
the tty struct.

Approved by: imp (mentor, implicit trivial changes)
MFC after: 1 week
Submitted by: Ed Schouten (ed at 80836 dot nl)

16 years agoAdd a -8 switch to syslogd to prevent it from mangling 8-bit data.
brian [Wed, 14 May 2008 00:22:21 +0000 (00:22 +0000)]
Add a -8 switch to syslogd to prevent it from mangling 8-bit data.

16 years agoFix #2.
adrian [Tue, 13 May 2008 23:24:06 +0000 (23:24 +0000)]
Fix #2.

16 years agoFix whitespace bug introduced a couple commits ago.
adrian [Tue, 13 May 2008 23:07:42 +0000 (23:07 +0000)]
Fix whitespace bug introduced a couple commits ago.

16 years agoBSDCan update #2.
adrian [Tue, 13 May 2008 22:46:13 +0000 (22:46 +0000)]
BSDCan update #2.

16 years agoBSDCan calendar file update #1.
adrian [Tue, 13 May 2008 22:27:32 +0000 (22:27 +0000)]
BSDCan calendar file update #1.

16 years agoBump modification date.
remko [Tue, 13 May 2008 21:51:35 +0000 (21:51 +0000)]
Bump modification date.

16 years agoThe first of thirteen patches by Ed that resolves a documentation
remko [Tue, 13 May 2008 21:51:02 +0000 (21:51 +0000)]
The first of thirteen patches by Ed that resolves a documentation
issue in termios.4, ttydefaults.h does not exist in /usr/include
but in /usr/include/sys.

MFC after: 3 days
Submitted by: Ed Schouten (ed at 80386 dot nl)

16 years agoDon't let hacksync() call bus_dmamap_sync(9) on DMA maps which
marius [Tue, 13 May 2008 20:58:08 +0000 (20:58 +0000)]
Don't let hacksync() call bus_dmamap_sync(9) on DMA maps which
are not initialized. This fixes a panic on sparc64 where calling
bus_dmamap_sync(9) on NULL DMA maps is fatal.

Approved by: sam

16 years agoRegen.
rdivacky [Tue, 13 May 2008 20:02:26 +0000 (20:02 +0000)]
Regen.

Approved by: kib (mentor)

16 years agoImplement robust futexes. Most of the code is modelled after
rdivacky [Tue, 13 May 2008 20:01:27 +0000 (20:01 +0000)]
Implement robust futexes. Most of the code is modelled after
what Linux does. This is because robust futexes are mostly
userspace thing which we cannot alter. Two syscalls maintain
pointer to userspace list and when process exits a routine
walks this list waking up processes sleeping on futexes
from that list.

Reviewed by: kib (mentor)
MFC after: 1 month

16 years agoFrom the OpenBSD commit log :
cognet [Tue, 13 May 2008 14:00:09 +0000 (14:00 +0000)]
From the OpenBSD commit log :
Add support for the Apple USB Ethernet adapter.
Work around the "latch in at the first working PHY address hack",
that fails for this adapter because it returns 0xffff when reading
from lower PHY addresses. Also add more debugging printfs

Obtained from: OpenBSD
MFC After: 3 days

16 years agoReplace the ee with the ed to resolve disk overflow.
nyan [Tue, 13 May 2008 11:51:19 +0000 (11:51 +0000)]
Replace the ee with the ed to resolve disk overflow.

16 years agoImprove temporary file handling
kevlo [Tue, 13 May 2008 09:42:03 +0000 (09:42 +0000)]
Improve temporary file handling

Obtained from: OpenBSD

16 years ago- Set sc->dev to the new-bus device_t so all the device_printf()s work.
jhb [Mon, 12 May 2008 21:34:52 +0000 (21:34 +0000)]
- Set sc->dev to the new-bus device_t so all the device_printf()s work.
- Add a missing newline to a printf.

MFC after: 1 week
Submitted by: Andriy Gapon  avg <> icyb.net.ua

16 years agoFor the sake of the log, revision 1.353 was bumped for setfib(2) and
ceri [Mon, 12 May 2008 20:12:34 +0000 (20:12 +0000)]
For the sake of the log, revision 1.353 was bumped for setfib(2) and
multiple routing table support.

16 years agoAdd support for management apps. Work around an apparent firmware bug that
scottl [Mon, 12 May 2008 14:09:19 +0000 (14:09 +0000)]
Add support for management apps.  Work around an apparent firmware bug that
results in hung i/o if more than 128 commands are scheduled for an array.

16 years agoDon't require a configuration file. Ntpd will be perfectly happy if there's
flz [Mon, 12 May 2008 11:49:16 +0000 (11:49 +0000)]
Don't require a configuration file. Ntpd will be perfectly happy if there's
none or if the file doesn't exist (there's no ntp.conf in the base install).

PR: conf/119592
Submitted by: Renaud Waldura <renaud+freebsd@waldura.org>
MFC after: 1 week

16 years agoTypo fixes.
brueffer [Mon, 12 May 2008 08:53:02 +0000 (08:53 +0000)]
Typo fixes.

16 years agofix typo in runz_fuzz
julian [Mon, 12 May 2008 06:42:06 +0000 (06:42 +0000)]
fix typo in runz_fuzz
noticed by:Elijah Buck

16 years agoFix the loopback interface. Cleaning up some code with new macros
gnn [Mon, 12 May 2008 02:44:53 +0000 (02:44 +0000)]
Fix the loopback interface.  Cleaning up some code with new macros
was a tad too aggressive.

PR: kern/123568
Submitted by: Vladimir Ermakov <samflanker at gmail dot com>
Obtained from: antoine

16 years agodeclare ieee80211_phymode_name with an array size
sam [Mon, 12 May 2008 00:34:28 +0000 (00:34 +0000)]
declare ieee80211_phymode_name with an array size

16 years agoUpdate PLCP<->rate mapping support:
sam [Mon, 12 May 2008 00:32:52 +0000 (00:32 +0000)]
Update PLCP<->rate mapping support:
o correct mapping of CCK rates to PLCP; was using nonstandard Ralink
  values which just happened to also be used by Zydas (so went unnoticed)
o change ieee80211_plcp2rate api to take a phy type instead of a flag
  that indicates ofdm/!ofdm
o update drivers to match (restore per-driver code to map rate->PLCP)

Reviewed by: sephe, weongyo, thompsa

16 years agoMinor cleanup of vap create work:
sam [Mon, 12 May 2008 00:15:30 +0000 (00:15 +0000)]
Minor cleanup of vap create work:
o add IEEE80211_C_STA capability to indicate sta mode is supported
  (was previously assumed) and mark drivers as capable
o add ieee80211_opcap array to map an opmode to the equivalent capability bit
o move IEEE80211_C_OPMODE definition to where capabilities are defined so it's
  clear it should be kept in sync (on future additions)
o check device capabilities in clone create before trying to create a vap;
  this makes driver checks unneeded
o make error codes return on failed clone request unique
o temporarily add console printfs on clone request failures to aid in
  debugging; these will move under DIAGNOSTIC or similar before release

16 years agowe have a DFS capability now so enable check
sam [Sun, 11 May 2008 23:36:58 +0000 (23:36 +0000)]
we have a DFS capability now so enable check

16 years agouse c99-style initialization for ieee80211_phymode_name
sam [Sun, 11 May 2008 23:33:56 +0000 (23:33 +0000)]
use c99-style initialization for ieee80211_phymode_name

16 years agoadd DFS capability bit and use it to auto-enable DFS support
sam [Sun, 11 May 2008 23:32:07 +0000 (23:32 +0000)]
add DFS capability bit and use it to auto-enable DFS support

16 years agouse the current left edge of the BA window when forming ADDBA request
sam [Sun, 11 May 2008 23:27:57 +0000 (23:27 +0000)]
use the current left edge of the BA window when forming ADDBA request
so an existing session is re-established with the correct seq#

16 years agoadd PLCP service bit definitions
sam [Sun, 11 May 2008 23:20:26 +0000 (23:20 +0000)]
add PLCP service bit definitions

16 years agomove inline keyword to silence compiler complaints
sam [Sun, 11 May 2008 23:18:11 +0000 (23:18 +0000)]
move inline keyword to silence compiler complaints

16 years agoChange the check for cpu_high to actually match CPUID 0x06.
rpaulo [Sun, 11 May 2008 23:17:57 +0000 (23:17 +0000)]
Change the check for cpu_high to actually match CPUID 0x06.

Submitted by: Arthur Hartwig <arthur.hartwig at nokia.com>
PR: 122878
MFC after: 3 days

16 years agoDon't use libkern's max() function as that's for unsigned numbers only.
rpaulo [Sun, 11 May 2008 23:14:07 +0000 (23:14 +0000)]
Don't use libkern's max() function as that's for unsigned numbers only.
Instead use the worldwide known MAX() function.
This should fix problems with negative values showing up on
dev.cpu.%d.temperature.
This is slightly different from the fix in the PR.

Submitted by: KOIE Hidetaka <hide at koie.org>
PR: 123542

16 years agofix typo's that broke duration calculation on protection frames
sam [Sun, 11 May 2008 22:11:01 +0000 (22:11 +0000)]
fix typo's that broke duration calculation on protection frames

16 years agoCorrect an error in pmap_align_superpage(). Specifically, correctly
alc [Sun, 11 May 2008 20:33:47 +0000 (20:33 +0000)]
Correct an error in pmap_align_superpage().  Specifically, correctly
handle the case where the mapping is greater than a superpage in size
but the alignment of the physical pages spans a superpage boundary.

16 years agoBetter to just statically set the name vs. determine at run time.
obrien [Sun, 11 May 2008 17:23:57 +0000 (17:23 +0000)]
Better to just statically set the name vs. determine at run time.

16 years agoAdd the ed and remove the vi.
nyan [Sun, 11 May 2008 14:25:24 +0000 (14:25 +0000)]
Add the ed and remove the vi.
This fixes a disk-full error.

16 years agoAdd manpage links to krb5_principal.3.
dfr [Sun, 11 May 2008 10:32:37 +0000 (10:32 +0000)]
Add manpage links to krb5_principal.3.

16 years agoDon't try to make links to manpages that no longer exist. Fixes installworld
dfr [Sun, 11 May 2008 08:27:17 +0000 (08:27 +0000)]
Don't try to make links to manpages that no longer exist. Fixes installworld

Submitted by: phk

16 years agoSync program name agnostic changes with SADE.
obrien [Sun, 11 May 2008 07:18:22 +0000 (07:18 +0000)]
Sync program name agnostic changes with SADE.

16 years agoClean up several instances of SADE calling itself sysinstall.
obrien [Sun, 11 May 2008 07:13:08 +0000 (07:13 +0000)]
Clean up several instances of SADE calling itself sysinstall.
(do so generically so the same set of changes can be applied to sysinstall)

16 years agoUpdate HARDWARE section.
joel [Sun, 11 May 2008 06:40:04 +0000 (06:40 +0000)]
Update HARDWARE section.

16 years agoCommit missing mips libthr support that I thought I'd committed earlier
imp [Sun, 11 May 2008 05:54:52 +0000 (05:54 +0000)]
Commit missing mips libthr support that I thought I'd committed earlier

16 years agoProvide the new argument to kmem_suballoc().
alc [Sat, 10 May 2008 23:39:27 +0000 (23:39 +0000)]
Provide the new argument to kmem_suballoc().

16 years agoIntroduce a new parameter "superpage_align" to kmem_suballoc() that is
alc [Sat, 10 May 2008 21:46:20 +0000 (21:46 +0000)]
Introduce a new parameter "superpage_align" to kmem_suballoc() that is
used to request superpage alignment for the submap.

Request superpage alignment for the kmem_map.

Pass VMFS_ANY_SPACE instead of TRUE to vm_map_find().  (They are currently
equivalent but VMFS_ANY_SPACE is the new preferred spelling.)

Remove a stale comment from kmem_malloc().

16 years agoAdd -mno-sse3 for amd64 case too
ache [Sat, 10 May 2008 20:46:07 +0000 (20:46 +0000)]
Add -mno-sse3 for amd64 case too

PR:             123518
Submitted by:   Marc Olzheim <marcolz@stack.nl>

16 years agoOnly start the vaps if the init routine completed.
thompsa [Sat, 10 May 2008 20:25:59 +0000 (20:25 +0000)]
Only start the vaps if the init routine completed.

16 years ago- Associate from a taskq as we can deadlock on the ndis hal and the com lock.
thompsa [Sat, 10 May 2008 20:12:43 +0000 (20:12 +0000)]
- Associate from a taskq as we can deadlock on the ndis hal and the com lock.
- Remove double vap init (ieee80211_start_all)
- Keep ic_curchan in sync with the scan results.

16 years agoPartially revert the last rev. Do call ndis_setstate_80211() when we up the
thompsa [Sat, 10 May 2008 20:07:00 +0000 (20:07 +0000)]
Partially revert the last rev. Do call ndis_setstate_80211() when we up the
interface but break out the associate code into a separate function. This fixes
association with an 11b Apple Airport.

Reported by: Ted Lindgreen

16 years agoGeneralize vm_map_find(9)'s parameter "find_space". Specifically, add
alc [Sat, 10 May 2008 18:55:35 +0000 (18:55 +0000)]
Generalize vm_map_find(9)'s parameter "find_space".  Specifically, add
support for VMFS_ALIGNED_SPACE, which requests the allocation of an
address range best suited to superpages.  The old options TRUE and FALSE
are mapped to VMFS_ANY_SPACE and VMFS_NO_SPACE, so that there is no
immediate need to update all of vm_map_find(9)'s callers.

While I'm here, correct a misstatement about vm_map_find(9)'s return
values in the man page.

16 years agoDocument freebsd extensions to netcat a bit better:
antoine [Sat, 10 May 2008 18:50:45 +0000 (18:50 +0000)]
Document freebsd extensions to netcat a bit better:
- sort the options
- document -o and -O everywhere

Reviewed by: delphij
Approved by: rwatson (mentor)

16 years agoAdd more obsolete files.
antoine [Sat, 10 May 2008 18:43:22 +0000 (18:43 +0000)]
Add more obsolete files.

Approved by: rwatson (mentor)

16 years agoRemove useless call to getdtablesize(2) in fdopen(3) and its useless
antoine [Sat, 10 May 2008 18:39:20 +0000 (18:39 +0000)]
Remove useless call to getdtablesize(2) in fdopen(3) and its useless
variable nofile.

PR: 123109
Submitted by: Christoph Mallon
Approved by: rwatson (mentor)
MFC after: 1 month

16 years agoAdd missing braces in #if 0ed code.
antoine [Sat, 10 May 2008 18:33:38 +0000 (18:33 +0000)]
Add missing braces in #if 0ed code.

Approved by: rwatson (mentor)
MFC after: 1 month

16 years agoChange two variables to size_t to improve portability.
julian [Sat, 10 May 2008 15:02:56 +0000 (15:02 +0000)]
Change two variables to size_t to improve portability.
Submitted by: Xin Li

16 years ago- Fix panic on detach.
simokawa [Sat, 10 May 2008 13:40:42 +0000 (13:40 +0000)]
- Fix panic on detach.
- Fix a comment.

MFC after: 2 weeks

16 years agoKqueue_scan() may sleep when encountered the influx knotes. On the other
kib [Sat, 10 May 2008 11:37:05 +0000 (11:37 +0000)]
Kqueue_scan() may sleep when encountered the influx knotes. On the other
hand, it may cause other threads to sleep since kqueue_scan() may mark
some knotes as infux. This could lead to the deadlock.

Before kqueue_scan() sleeps, wakeup the threads that are waiting for the
influx knotes produced by this thread.

Tested by: pho (previous version)
Reviewed by: jmg
MFC after: 2 weeks

16 years agoThe kqueue_close() encountering the KN_INFLUX knotes on the kq being
kib [Sat, 10 May 2008 11:35:32 +0000 (11:35 +0000)]
The kqueue_close() encountering the KN_INFLUX knotes on the kq being
closed is the legitimate situation. For instance, filedescriptor with
registered events may be closed in parallel with closing the kqueue.
Properly handle the case instead of asserting that this cannot happen.

Reported and tested by: pho
Reviewed by: jmg
MFC after: 2 weeks

16 years agoFix build.
delphij [Sat, 10 May 2008 09:22:17 +0000 (09:22 +0000)]
Fix build.

16 years ago- Disable interrupts on suspend to eliminate excessive
simokawa [Sat, 10 May 2008 09:22:06 +0000 (09:22 +0000)]
- Disable interrupts on suspend to eliminate excessive
  'device physically ejected?' message on resume.
- Fix memory leak on resume reported by kiyohara at netbsd.org.

MFC after: 2 weeks