openbsd
14 years agoMove the new mdoc(7) manual written by kristaps@, replacing the old one
schwarze [Sun, 1 Aug 2010 20:37:51 +0000 (20:37 +0000)]
Move the new mdoc(7) manual written by kristaps@, replacing the old one
which had much less information, such that the new one gets installed.
No change to the build system required, no text change in this commit.
Doing the big move early even though a few more improvements will follow.
The duplicate information in mdoc.samples(7) will be cleaned up post-release.
ok jmc@, and kristaps@ agreed with the plan, too

14 years agoMake parsing of /etc/remote work again (some of the options were the
nicm [Sun, 1 Aug 2010 20:27:51 +0000 (20:27 +0000)]
Make parsing of /etc/remote work again (some of the options were the
wrong type), try to make it clearer what comes from where, and add a few
XXX.

Currently there is a mess of default options set in vinit(), default
options set in getremote(), stuff read from /etc/remote and options read
from the command-line, and there remains the problems with leftover
options from rotation. This all needs to be sorted out properly later.

Noticed while trying to find a possible problem reported by miod.

14 years agosync
deraadt [Sun, 1 Aug 2010 18:57:55 +0000 (18:57 +0000)]
sync

14 years agoRead status register after all writes to the pcf8584 controller. Alegedly
kettenis [Sun, 1 Aug 2010 18:48:41 +0000 (18:48 +0000)]
Read status register after all writes to the pcf8584 controller.  Alegedly
this is what (Open)Solaris does and it makes the i2c controller on the
blade2.5k work.  From NetBSD.

Tested by sthen@, ok deraadt@

14 years agoMake __builtin_return_address(0) work with -fstack-protector. Old diff
kettenis [Sun, 1 Aug 2010 17:55:26 +0000 (17:55 +0000)]
Make __builtin_return_address(0) work with -fstack-protector.  Old diff
from Jakub Jelinek that never made it into upstream GCC.  Fixes ld.so.
Found by drahn@ and me; ok miod@

14 years agoProperly reset states when a login phase is interrupted by another one.
tobias [Sun, 1 Aug 2010 16:53:57 +0000 (16:53 +0000)]
Properly reset states when a login phase is interrupted by another one.
Prevents ftpd from letting regular users logging in during anonymous-only
mode (-A).

ok millert, ray

14 years agodocument the rw_assert_ functions
blambert [Sun, 1 Aug 2010 14:21:10 +0000 (14:21 +0000)]
document the rw_assert_ functions

ok jmc@

14 years agoSmall cleanup of the fpu switching code; don't rely on the fact the HFP_REGS
kettenis [Sun, 1 Aug 2010 14:20:41 +0000 (14:20 +0000)]
Small cleanup of the fpu switching code; don't rely on the fact the HFP_REGS
is actually zero and remove a redundant instruction.

ok jsing@

14 years agoadjust position of test so that it does something !
espie [Sun, 1 Aug 2010 10:26:27 +0000 (10:26 +0000)]
adjust position of test so that it does something !
avoid a lot of extra stupid printing.

14 years agotypo
espie [Sun, 1 Aug 2010 10:04:24 +0000 (10:04 +0000)]
typo

14 years agofix the qt 4.5 -> 4.6 update issue.
espie [Sun, 1 Aug 2010 10:03:24 +0000 (10:03 +0000)]
fix the qt 4.5 -> 4.6 update issue.
since we usually deal with directories in a global way, replacing a
directory with something else didn't work.
the fix is to record removed directories in Vstat in a "lazy" way:
we only test these objects for existence if we add something else at
the same location. Since this happens rarely, we can afford to perform
the whole SharedItems dance twice...

14 years agochange buf_release() prototype accordingly to buf.c
zinovik [Sun, 1 Aug 2010 09:55:40 +0000 (09:55 +0000)]
change buf_release() prototype accordingly to buf.c

unbreaks opencvs build

14 years agochange returning type to void like it is done in OpenRCS:buf_release()
zinovik [Sun, 1 Aug 2010 09:19:29 +0000 (09:19 +0000)]
change returning type to void like it is done in OpenRCS:buf_release()

ok nicm@

14 years agoANSIfy
kettenis [Sun, 1 Aug 2010 09:01:45 +0000 (09:01 +0000)]
ANSIfy

14 years agodon't rely on mbuf.h to pull in pool.h here either
blambert [Sun, 1 Aug 2010 05:30:13 +0000 (05:30 +0000)]
don't rely on mbuf.h to pull in pool.h here either

ok miod@

14 years agofix use of uninitialized variable.
chl [Sun, 1 Aug 2010 03:09:48 +0000 (03:09 +0000)]
fix use of uninitialized variable.

reported by gcc.

ok canacar@

14 years agoremove unused variable
chl [Sun, 1 Aug 2010 02:51:03 +0000 (02:51 +0000)]
remove unused variable

ok thib@

14 years agoadd missing header needed by strcmp()
chl [Sun, 1 Aug 2010 02:49:07 +0000 (02:49 +0000)]
add missing header needed by strcmp()
add missing prototype

ok stsp@

14 years agoSync to bsd.lv:
schwarze [Sun, 1 Aug 2010 00:09:48 +0000 (00:09 +0000)]
Sync to bsd.lv:
State that we parse and ignore \s and \m font size and colour control
sequences, and remove the detailed, irrelevant documentation.

14 years agoCorrect Copyright lines in the rarely touched files:
schwarze [Sat, 31 Jul 2010 23:42:04 +0000 (23:42 +0000)]
Correct Copyright lines in the rarely touched files:
* add missing years (only where substantial changes were committed)
* update Kristaps' email address in the remaining places
No code changes.

14 years agoMajor cleanup (but there is still more to come):
schwarze [Sat, 31 Jul 2010 22:07:11 +0000 (22:07 +0000)]
Major cleanup (but there is still more to come):
* rewrite .An, .Bd, .Bk, .Bl, .Ex descriptions
* correct "parsable" to "parsed"
* and various formatting and wording tweaks
This commit includes a patch from kristaps@ explaining empty .Dd.
Feedback and OK jmc@ and kristaps@.

14 years agodocument .in; from kristaps@
schwarze [Sat, 31 Jul 2010 21:51:33 +0000 (21:51 +0000)]
document .in; from kristaps@

14 years agoMerge bsd.lv version 1.10.5: last larger batch of bug fixes before release.
schwarze [Sat, 31 Jul 2010 21:43:07 +0000 (21:43 +0000)]
Merge bsd.lv version 1.10.5: last larger batch of bug fixes before release.
NOT including Kristaps' .Bd -literal changes which cause regressions.
Features:
* -Tpdf now fully working
Bugfixes:
* proper handling of quoted strings by .ds in roff(7)
* allow empty .Dd
* make .Sm start no-spacing after the first output word
* underline .Ad
* minor fixes in -Thtml
and some optimisations in terminal output.

14 years agoremove references to ``opl'', since it doesn't exist any longer
ratchov [Sat, 31 Jul 2010 21:24:55 +0000 (21:24 +0000)]
remove references to ``opl'', since it doesn't exist any longer
noticed by brad

14 years agoIgnore the PSL_VEC bit in the sigcontext and always copy the bit out from
kettenis [Sat, 31 Jul 2010 21:15:05 +0000 (21:15 +0000)]
Ignore the PSL_VEC bit in the sigcontext and always copy the bit out from
the trap frame.  The Altivec unit might have been stolen away from us while
we were running the signal handler, so it is ok that the two don't agree.
Fixes the X server mysteriously exiting (and seemingly freeze).

ok miod@, drahn@

14 years agoanother pair of files which relied on mbuf.h for pool.h; should be the last,
blambert [Sat, 31 Jul 2010 20:48:32 +0000 (20:48 +0000)]
another pair of files which relied on mbuf.h for pool.h; should be the last,
as the kernel now compiles w/o ``#include <sys/pool.h>'' in mbuf.h; removing
that line, though, is for another day, as a userland program (systat, IIRC)
fails to build without it there.

ok krw@

14 years agozap a comma; from Joachim Schipper
jmc [Sat, 31 Jul 2010 20:26:21 +0000 (20:26 +0000)]
zap a comma; from Joachim Schipper

14 years agoSync
krw [Sat, 31 Jul 2010 19:51:26 +0000 (19:51 +0000)]
Sync

14 years agoAdd NVIDIA 0x07e1 == GeForce 7100, now that new driver will support
krw [Sat, 31 Jul 2010 19:49:56 +0000 (19:49 +0000)]
Add NVIDIA 0x07e1 == GeForce 7100, now that new driver will support
it. Seen on my Acer Aspire X1800.

14 years agomention M-Audio Audiophile 192k
ratchov [Sat, 31 Jul 2010 16:56:42 +0000 (16:56 +0000)]
mention M-Audio Audiophile 192k

14 years agoadd support for M-audio Audiophile 192k
ratchov [Sat, 31 Jul 2010 16:52:37 +0000 (16:52 +0000)]
add support for M-audio Audiophile 192k

14 years agointroduce macro SIZE_LEFT and function buf_grow() from OpenRCS
zinovik [Sat, 31 Jul 2010 16:28:31 +0000 (16:28 +0000)]
introduce macro SIZE_LEFT and function buf_grow() from OpenRCS

buf_grow() replaces macro BUF_GROW()

ok nicm@, ray@

14 years agoregen
miod [Sat, 31 Jul 2010 16:13:28 +0000 (16:13 +0000)]
regen

14 years agoSupport extra F-keys (up to F24) on HID keyboards.
miod [Sat, 31 Jul 2010 16:12:37 +0000 (16:12 +0000)]
Support extra F-keys (up to F24) on HID keyboards.

14 years agoFactorization of the bluetooth and usb hid input driver logic, to reduce code
miod [Sat, 31 Jul 2010 16:04:46 +0000 (16:04 +0000)]
Factorization of the bluetooth and usb hid input driver logic, to reduce code
duplication and divergence.
Thanks to mlarkin@ for bluetooth devices tests.

14 years agoadds a remark about how printing routing labels to -r, it is the
sobrado [Sat, 31 Jul 2010 14:25:37 +0000 (14:25 +0000)]
adds a remark about how printing routing labels to -r, it is the
more logical place to find this information; improve description
for the -v option (stolen from NetBSD).

written with excellent advice from bluhm@ and jmc@.

ok jmc@

14 years agoThe code referencing urndis_watchdog() is #if 0'd out, but the actual
mk [Sat, 31 Jul 2010 11:51:45 +0000 (11:51 +0000)]
The code referencing urndis_watchdog() is #if 0'd out, but the actual
function is not.

This change #if 0's urndis_watchdog() and some functions that are only
called from here.

At some point we probably want to use the watchdog functionality but the
current code is completely untested so disable it entirely rather than
enabling it this close to release.

14 years agoWhen looking for revision, break out of the loop at the end rather than
nicm [Sat, 31 Jul 2010 11:37:37 +0000 (11:37 +0000)]
When looking for revision, break out of the loop at the end rather than
looping forever.

This prevents the server going into an infinite loop when the repository
is behind the checkout, but the behaviour is not correct (it is an error
rather than ignoring the file), so mark with an XXX for now. Some
underlying OpenCVS functions are too fatal()-happy and will probably
need to be changed in order to fix this properly.

Problem reported by sthen.

ok xsa ray

14 years agovery old thinko, not really a problem, obviously, but groups and users are
espie [Sat, 31 Jul 2010 11:30:50 +0000 (11:30 +0000)]
very old thinko, not really a problem, obviously, but groups and users are
categories, the corresponding keywords are newgroup and newuser.

14 years agoadd SA3_XXX macros removed by mistake with opl, but used by ym mixer
ratchov [Sat, 31 Jul 2010 11:25:38 +0000 (11:25 +0000)]
add SA3_XXX macros removed by mistake with opl, but used by ym mixer
from mk@

14 years agokeep location open while we grab updateinfo.
espie [Sat, 31 Jul 2010 11:17:22 +0000 (11:17 +0000)]
keep location open while we grab updateinfo.
deal with error fallout and need to forget unneeded stuff
(halves number of connections, roughly)

14 years agounbreak ym.c, opl3sa3reg.h doesn't exist any longer
ratchov [Sat, 31 Jul 2010 11:00:08 +0000 (11:00 +0000)]
unbreak ym.c, opl3sa3reg.h doesn't exist any longer
noticed by mk@

14 years agoBreak some really long lines.
mk [Sat, 31 Jul 2010 10:59:37 +0000 (10:59 +0000)]
Break some really long lines.

14 years agogrammar fix;
jmc [Sat, 31 Jul 2010 09:38:03 +0000 (09:38 +0000)]
grammar fix;

14 years agodelete IIJ local functionalities.
yasuoka [Sat, 31 Jul 2010 09:33:09 +0000 (09:33 +0000)]
delete IIJ local functionalities.

14 years agoChecking for -1 as an overflow value is pretty absurd. Returning
ray [Sat, 31 Jul 2010 08:54:42 +0000 (08:54 +0000)]
Checking for -1 as an overflow value is pretty absurd. Returning
the unix epoch is equally absurd. Just return Start for now. We may
add overflow checks in the future.

OK xsa nicm

14 years agoremove 2 years old compatibility mode
ratchov [Sat, 31 Jul 2010 08:48:01 +0000 (08:48 +0000)]
remove 2 years old compatibility mode

14 years agoremove forgotten debug message
ratchov [Sat, 31 Jul 2010 08:46:56 +0000 (08:46 +0000)]
remove forgotten debug message

14 years agoremove the ``midisyn'' framework and anything using it, i.e., the opl
ratchov [Sat, 31 Jul 2010 08:33:19 +0000 (08:33 +0000)]
remove the ``midisyn'' framework and anything using it, i.e., the opl
device and the midi interface to pcppi.

14 years agofix typo preventing compilation with AUDIO_DEBUG
ratchov [Sat, 31 Jul 2010 08:08:18 +0000 (08:08 +0000)]
fix typo preventing compilation with AUDIO_DEBUG

14 years agofix a capitalisation error;
jmc [Sat, 31 Jul 2010 07:09:37 +0000 (07:09 +0000)]
fix a capitalisation error;

14 years agocheck for audio stream and control interface descriptors in the
jakemsr [Sat, 31 Jul 2010 02:52:16 +0000 (02:52 +0000)]
check for audio stream and control interface descriptors in the
device configuration descriptor in the match() function.  stops
midi devices from attaching as uaudio then bailing out, leaving
a phantom uaudio.

14 years agoMention the horde port module. Not documented yet.
sthen [Fri, 30 Jul 2010 22:51:03 +0000 (22:51 +0000)]
Mention the horde port module. Not documented yet.

14 years agoCheck date_parse return values.
ray [Fri, 30 Jul 2010 21:47:18 +0000 (21:47 +0000)]
Check date_parse return values.

OK xsa

14 years agoAdd "-S" to library installs to avoid breakage when one
ckuethe [Fri, 30 Jul 2010 19:23:38 +0000 (19:23 +0000)]
Add "-S" to library installs to avoid breakage when one
"make -j8 build" job is installing libc and another job
is expecting to use libc (for example). ok phessler

14 years agoWhen converting a floating value to an integer, properly raise the "invalid"
kettenis [Fri, 30 Jul 2010 18:05:23 +0000 (18:05 +0000)]
When converting a floating value to an integer, properly raise the "invalid"
exception when the floating value is infinite or NaN or if the integral part
of the floating value exceeds the range of the integer type, as required by
the C99/IEEE754 standard.

Fixes Python 2.6 build on hppa.

ok miod@

14 years agoDon't reset pms/pmsi devices when attempting to set the mode. Makes nicm@'s,
krw [Fri, 30 Jul 2010 17:15:14 +0000 (17:15 +0000)]
Don't reset pms/pmsi devices when attempting to set the mode. Makes nicm@'s,
todd@'s and my machines all work as expected. Various other success stories
and no reports of failures. Also fix DEBUG compiles by removing a return
that was unneeded anyway.

"gud" miod@

14 years agoThe device table entry for HD3470 was incorrectly using a logical or
jsg [Fri, 30 Jul 2010 13:31:47 +0000 (13:31 +0000)]
The device table entry for HD3470 was incorrectly using a logical or
instead of a binary or.  Found via lint.
ok oga@

14 years agoDo not print "Label" caption for "netstat -rvA".
bluhm [Fri, 30 Jul 2010 12:55:12 +0000 (12:55 +0000)]
Do not print "Label" caption for "netstat -rvA".
ok markus@

14 years agoDon't set "success" flags before error checks.
ray [Fri, 30 Jul 2010 11:02:56 +0000 (11:02 +0000)]
Don't set "success" flags before error checks.
Add check for login_getclass() failure, pointed out by tobias.

OK tobias

14 years agofix header during update
espie [Fri, 30 Jul 2010 10:06:44 +0000 (10:06 +0000)]
fix header during update

14 years agoRemove unused header file.
miod [Fri, 30 Jul 2010 07:57:09 +0000 (07:57 +0000)]
Remove unused header file.

14 years agoReplace abort(), fatal(), and errx() with return -1, since date_parse
ray [Thu, 29 Jul 2010 18:52:45 +0000 (18:52 +0000)]
Replace abort(), fatal(), and errx() with return -1, since date_parse
returns -1 on error.

OK nicm

14 years agoDon't set sc_ec until after the EC registers are mapped, but before we
kettenis [Thu, 29 Jul 2010 18:32:26 +0000 (18:32 +0000)]
Don't set sc_ec until after the EC registers are mapped, but before we
execute the _REG method to make the EC available to the AML.

ok marco@

14 years agochange URLs to point to the more recent porter's handbook;
steven [Thu, 29 Jul 2010 17:38:13 +0000 (17:38 +0000)]
change URLs to point to the more recent porter's handbook;
mention the use of packages first.

ok sthen@

14 years agoPrint a column with the routing label if "netstat -vr" or "route
bluhm [Thu, 29 Jul 2010 16:35:40 +0000 (16:35 +0000)]
Print a column with the routing label if "netstat -vr" or "route
-v show" are called with the -v switch.
ok markus@

14 years agosome error cases returned 01 when they should have been returning -1
jsg [Thu, 29 Jul 2010 14:41:21 +0000 (14:41 +0000)]
some error cases returned 01 when they should have been returning -1
spotted by Mike Belopuhov.

14 years agosort pool.h into the list of includes alphabetically
blambert [Thu, 29 Jul 2010 14:40:47 +0000 (14:40 +0000)]
sort pool.h into the list of includes alphabetically

pointed out by kettenis@, ok miod@

14 years ago#include <sys/pool.h> directly, instead of relying on mbuf.h to inlcude it
blambert [Thu, 29 Jul 2010 14:06:26 +0000 (14:06 +0000)]
#include <sys/pool.h> directly, instead of relying on mbuf.h to inlcude it

ok miod@ tedu@

14 years agoreplace two misleading ie. with e.g.; from Jan Stary
jmc [Thu, 29 Jul 2010 13:55:48 +0000 (13:55 +0000)]
replace two misleading ie. with e.g.; from Jan Stary

14 years agoDon't forget to register the i8254-based timecounter if we use the i8254
kettenis [Thu, 29 Jul 2010 13:36:30 +0000 (13:36 +0000)]
Don't forget to register the i8254-based timecounter if we use the i8254
for clock interrupts.  Unbreaks amd64 in PIC mode.

14 years agoAdd some more devices. -moj
maja [Thu, 29 Jul 2010 12:37:16 +0000 (12:37 +0000)]
Add some more devices. -moj

14 years agosync
maja [Thu, 29 Jul 2010 12:34:09 +0000 (12:34 +0000)]
sync

14 years agoAdd some more displaylink devices. -moj
maja [Thu, 29 Jul 2010 12:32:04 +0000 (12:32 +0000)]
Add some more displaylink devices. -moj

14 years agoflag if an error during attach caused us to detach our resources, and
jakemsr [Thu, 29 Jul 2010 01:39:03 +0000 (01:39 +0000)]
flag if an error during attach caused us to detach our resources, and
if so, don't do anything on suspend/resume.  should fix crash reported
by mlarkin.

14 years agoRemove stray newline, fix $OpenBSD$; ok deraadt
stsp [Thu, 29 Jul 2010 00:50:10 +0000 (00:50 +0000)]
Remove stray newline, fix $OpenBSD$; ok deraadt

14 years agoZero out the firmware waking vector on successful resume from ACPI sleep.
mlarkin [Thu, 29 Jul 2010 00:29:49 +0000 (00:29 +0000)]
Zero out the firmware waking vector on successful resume from ACPI sleep.

ok deraadt@

14 years agoTruncate output file when compressing. OK deraadt@
millert [Wed, 28 Jul 2010 23:52:01 +0000 (23:52 +0000)]
Truncate output file when compressing.  OK deraadt@

14 years agoDVD_INVALIDATE_AGID is a 'no payload' operation, so don't set
krw [Wed, 28 Jul 2010 23:47:43 +0000 (23:47 +0000)]
DVD_INVALIDATE_AGID is a 'no payload' operation, so don't set
xs->datalen and reset xs->data to NULL. This prevents a spurious
attempt to setup (dma map, etc) and possibly attempt data tranfser.

In line with what Linux and FreeBSD do as far as I can tell.

Reduces the delay before my DVD starts playing its movie.

Idea to also set xs->data to NULL from matthew@.

"No objection" matthew@

14 years agodup() the stdin fd so it isn't closed twice (once for stdin, once for tty).
nicm [Wed, 28 Jul 2010 22:15:15 +0000 (22:15 +0000)]
dup() the stdin fd so it isn't closed twice (once for stdin, once for tty).

14 years agoAdd a dummy kqueue filter similar to seltrue and use it for anything
nicm [Wed, 28 Jul 2010 21:44:41 +0000 (21:44 +0000)]
Add a dummy kqueue filter similar to seltrue and use it for anything
using seltrue for poll. Based on code from NetBSD.

Also remove a stray duplicate lpt entry from loongson, from deraadt.

ok tedu deraadt

14 years agoSuspending drivers need to do as little as possible in here, so when the
deraadt [Wed, 28 Jul 2010 21:24:52 +0000 (21:24 +0000)]
Suspending drivers need to do as little as possible in here, so when the
transition is to INIT, try to not generate management frames if we can
avoid it.
ok damien

14 years agoMake legacy xxpower() functions call xxstop() on suspend, and simplify their
deraadt [Wed, 28 Jul 2010 21:21:38 +0000 (21:21 +0000)]
Make legacy xxpower() functions call xxstop() on suspend, and simplify their
resume paths.  For new-style suspend/resume, add a ca_activate function where
it is missing, and use a workq to resume because these drivers like to sleep.
ok damien

14 years agoSupport -U and -C to specify the context length for "cvs diff".
nicm [Wed, 28 Jul 2010 21:19:30 +0000 (21:19 +0000)]
Support -U and -C to specify the context length for "cvs diff".

This doesn't alter cvs rdiff, where -U and -C are not support by GNU
cvs.

ok sthen zinovik

14 years agothe default pciide chip needs no additional register restores; tested by jcs on one...
deraadt [Wed, 28 Jul 2010 18:31:51 +0000 (18:31 +0000)]
the default pciide chip needs no additional register restores; tested by jcs on one model

14 years agoFix fallout from the utf-8 commit (segfaults in disklabel on install media).
stsp [Wed, 28 Jul 2010 16:16:55 +0000 (16:16 +0000)]
Fix fallout from the utf-8 commit (segfaults in disklabel on install media).
Sorry about that.
ok sthen deraadt

14 years agomore #ifdef needed for the case of not pegging on a non-MP kernel
deraadt [Wed, 28 Jul 2010 16:15:25 +0000 (16:15 +0000)]
more #ifdef needed for the case of not pegging on a non-MP kernel

14 years agoremove stub code for rw locks we will never need; ok kettenis marco
deraadt [Wed, 28 Jul 2010 16:00:29 +0000 (16:00 +0000)]
remove stub code for rw locks we will never need; ok kettenis marco

14 years agoIt's pretty pointless to peg ourselves to the primary CPU if that's the
kettenis [Wed, 28 Jul 2010 15:59:38 +0000 (15:59 +0000)]
It's pretty pointless to peg ourselves to the primary CPU if that's the
only one we have.

ok krw@, deraadt@

14 years agosuspend save/restore for SIS3112; tested by mlarkin
deraadt [Wed, 28 Jul 2010 15:50:19 +0000 (15:50 +0000)]
suspend save/restore for SIS3112; tested by mlarkin

14 years agoChange back to the pre rev 1.11 behaviour of not treating unexpected
jsg [Wed, 28 Jul 2010 15:45:04 +0000 (15:45 +0000)]
Change back to the pre rev 1.11 behaviour of not treating unexpected
id payloads as errors.  Lets interop with strongSwan which sends
both IDi and IDr work again.

14 years agomore explicit message
espie [Wed, 28 Jul 2010 15:06:14 +0000 (15:06 +0000)]
more explicit message

14 years agolet fatal work, fix typo
espie [Wed, 28 Jul 2010 15:05:33 +0000 (15:05 +0000)]
let fatal work, fix typo

14 years agoMake all gasio failures look and behave the same.
marco [Wed, 28 Jul 2010 14:39:43 +0000 (14:39 +0000)]
Make all gasio failures look and behave the same.

14 years agopass state to Ustar objects, use it to display those pesky error and
espie [Wed, 28 Jul 2010 12:19:54 +0000 (12:19 +0000)]
pass state to Ustar objects, use it to display those pesky error and
fatal messages.

zap $opt_x from pkg_mklocated, do things like other commands do.

14 years agoRevert parts of previous change leading to assertion failure for
martinh [Wed, 28 Jul 2010 10:06:19 +0000 (10:06 +0000)]
Revert parts of previous change leading to assertion failure for
certain modify operations. Also fix logic error when replacing an attribute
with the empty set.

14 years agoMake BUFs autoextend by default. This was already done in OpenCVS
ray [Wed, 28 Jul 2010 09:07:11 +0000 (09:07 +0000)]
Make BUFs autoextend by default.  This was already done in OpenCVS
a while ago.

OK zinovik nicm

14 years agoSync hcreate(3) with NetBSD, adding some caveats.
ray [Wed, 28 Jul 2010 09:00:20 +0000 (09:00 +0000)]
Sync hcreate(3) with NetBSD, adding some caveats.

OK jmc

14 years agoDelay a little longer on suspend before giving up. Fixes machines that
mlarkin [Wed, 28 Jul 2010 07:32:16 +0000 (07:32 +0000)]
Delay a little longer on suspend before giving up. Fixes machines that
sometimes don't suspend because the sleep transition takes too long
(symptoms like failing to sleep but the backlight stays on).

ok deraadt@

14 years agopfsync_bulk_fail was mucking around with pfsync_softc and sending packets
dlg [Wed, 28 Jul 2010 06:52:05 +0000 (06:52 +0000)]
pfsync_bulk_fail was mucking around with pfsync_softc and sending packets
without holding splsoftnet. this adds the necessary protection.

reported by patrick coleman