Commit Graph

113715 Commits

Author SHA1 Message Date
David Xu
c5fa3778bb Add locking code for tls routines. 2005-03-20 23:28:25 +00:00
Ian Dowse
a3a574f503 Remove trailing whitespace as per NetBSD's revision 1.91. Also
update the TODO comments to more closely match current reality.
2005-03-20 22:22:18 +00:00
Ian Dowse
2bcccc224c Don't skip the initialisation of tl->len when we hit an EOF immediately
after allocating a new buffer. This bug caused `tail -r < /dev/null'
to core dump when the `J' malloc option is set, and also affected
any other input that was an exact multiple of 128k.
2005-03-20 22:08:52 +00:00
Gleb Smirnoff
c1249c6338 - Don't lose TCP flags of the first packet in a flow.
- Don't account length of the first packet in a flow twice.
2005-03-20 21:03:43 +00:00
Tim Kientzle
1ca479296c Deorbit gtar, step 1: Remove WITH_GTAR knob.
"tar" is now always a synonym for "bsdtar".
  People who need gtar can access it as "gtar."

Approved by:	No objections on arch@
2005-03-20 20:38:46 +00:00
Murray Stokely
ddc5ac8b73 Return BUS_PROBE_DEFAULT in preference to 0.
Pointed out by:	Nate Lawson <nate@root.org>
2005-03-20 20:13:21 +00:00
Bruce A. Mah
52957bbfd8 When building release documentation for release builds or for
release documentation snapshots, force URLS_ABSOLUTE.  This
allows us to relocate HTML renderings anywhere without worrying
about breaking relative links to pages on the Web site.  This can
happen when we use &base; as part of a URL.

I wonder if we should do this for all release documentation builds,
but I'm not quite sure what all the cases are.
2005-03-20 17:59:17 +00:00
Simon L. B. Nielsen
4bced63f6d Add byteorder(9) to SEE ALSO.
MFC after:	3 days
2005-03-20 17:27:57 +00:00
Ian Dowse
189dcd7f0f Interrupt transfers and split transactions work now, so remove the
part of the BUGS section that says they don't. Leave the "quite
buggy" comment for now though, as it's still partially true.
2005-03-20 17:13:43 +00:00
Robert Watson
6220dcba84 Add a read-only kern.sched.preemption sysctl so that user space can tell
if "options PREEMPTION" is compiled into the kernel.
2005-03-20 17:05:12 +00:00
Philip Paeps
a75ae58445 Document acpi_fujitsu(4).
Reviewed by:	brueffer
2005-03-20 15:27:38 +00:00
Christian Brueffer
3b166e9624 Add a missing word. 2005-03-20 14:57:46 +00:00
Gleb Smirnoff
d4d2297060 ifma_protospec is a pointer. Use NULL when assigning or compating it. 2005-03-20 14:31:45 +00:00
Pawel Jakub Dawidek
1897c44125 Add regression tests for pkill/pgrep.
MFC after:	3 days
2005-03-20 12:38:08 +00:00
Pawel Jakub Dawidek
d0cf19a161 Whitespace fixes. 2005-03-20 11:47:44 +00:00
Pawel Jakub Dawidek
0a5f326cd2 - Introduce an '-o' option which allows to match oldest of the matching
processes.
  This option can be also found in Solaris and Linux.
- Use timercmp(9) macro for timeval comparsion.
- Include time.h directly, don't depend on stat.h doing it for us.

Reviewed by:	gad (first point)
MFC after:	3 days
2005-03-20 11:42:26 +00:00
Pawel Jakub Dawidek
d054f5d7cd - Introduce '-S' option which allows to match system processes (pgrep only).
- Rename IS_KERNPROC() macro to PSKIP() and extend its functionality.
  Now it'll skip calling process and system processes when -S is not given.
  As a side effect it fixes '-n' option. Before it was always matching
  calling process (because of missing 'if (kp->ki_pid == mypid)' check)
  and after that, calling process was ignored.
- When '-l' option is given and there are no arguments, use p_comm as an
  arguments list (this is helpful for kernel threads matching).

Reviewed by:	gad
MFC after:	3 days
2005-03-20 11:30:26 +00:00
Pawel Jakub Dawidek
b94b78867e Sort options properly.
Reviewed by:	gad
MFC after:	3 days
2005-03-20 11:17:06 +00:00
Pawel Jakub Dawidek
46a4b46cd7 Add a -i option to ignore case in the process match.
Obtained from:	NetBSD
Reviewed by:	gad
MFC after:	3 days
2005-03-20 11:12:13 +00:00
Pawel Jakub Dawidek
87a8221aeb Add and document '-F' option which allows to use file where PID is stored
for matching.

Reviewed by:	gad
MFC after:	3 days
2005-03-20 11:03:11 +00:00
Pawel Jakub Dawidek
2b0a135497 Add and document '-j' option which allows to match processes based on its
jail ID.

Reviewed by:	gad
MFC after:	3 days
2005-03-20 10:53:54 +00:00
Pawel Jakub Dawidek
a990344f7e Add and document the 'jid' keyword for the '-o' option.
Reviewed by:	gad
MFC after:	3 days
2005-03-20 10:40:36 +00:00
Pawel Jakub Dawidek
f8197bf090 Make kvm(3) aware of ki_jid field.
Reviewed by:	gad
MFC after:	3 days
2005-03-20 10:37:56 +00:00
Pawel Jakub Dawidek
c78941e69e Add ki_jid field to the kinfo_proc structure and store jail ID there.
Reviewed by:	gad
MFC after:	3 days
2005-03-20 10:35:23 +00:00
Pawel Jakub Dawidek
e69f1fa2ef Make prison structure visible from userland if _WANT_PRISON is defined
(simlar to _WANT_UCRED).

Reviewed by:	gad
MFC after:	3 days
2005-03-20 10:33:27 +00:00
Gleb Smirnoff
50bb170471 Remove a workaround from previos revision. It proved to be incorrect.
Add two another workarounds for carp(4) interfaces:
- do not add connected route when address is assigned to carp(4) interface
- do not add connected route when other interface goes down

Embrace workarounds with #ifdef DEV_CARP
2005-03-20 10:27:17 +00:00
Nate Lawson
ac5f2dab74 s/SLIST/STAILQ to catch up with changes to resource lists.
Missed by:	imp
2005-03-20 06:55:49 +00:00
Yoshihiro Takahashi
f227591555 s/SLIST/STAILQ/ 2005-03-20 05:22:57 +00:00
Nate Lawson
6638c61aa6 Attach acpi_perf early. Especially when it is being used to provide info
to other devices, it needs to be attached first.  (Multi-pass newbus
probes would be a better solution.)
2005-03-20 03:51:18 +00:00
John-Mark Gurney
ca9ac06c99 I forgot to commit this fix that is now kgdb aware... 2005-03-20 02:08:21 +00:00
Sam Leffler
eb9883565a fix braino introduced when converting from madwifi 2005-03-20 01:55:02 +00:00
Sam Leffler
57a4a75853 eliminate mid-block variable decls 2005-03-20 01:27:33 +00:00
Nate Lawson
e33bea8d9e Add the acpi_ec_read and write methods. This allows an external driver
(like an EC/SMbus controller) to access the EC address space.  Access
is synchronized by the EcLock/Unlock routines in EcSpaceHandler().

Tested by:	Hans Petter Selasky
2005-03-20 01:27:27 +00:00
Nate Lawson
823d404a6d Only activate ICH speedstep if we're going to use it. No bugs were observed
due to this but it's cleaner this way.
2005-03-20 01:25:21 +00:00
Olivier Houchard
2b2714935b Use the correct values for softfloat, in both the little endian and the big
endian cases.
2005-03-20 00:53:52 +00:00
Olivier Houchard
7fc53c7b12 Bring in a version of float.h more correct for softfloat. 2005-03-20 00:34:24 +00:00
Sam Leffler
b276305606 version 1.1 (with cleanups)
Submitted by:	John Bicket
2005-03-19 21:04:53 +00:00
Bruce A. Mah
20c54b0f87 New release notes: clkbrd driver, nve, re(4) ALTQ support.
MFCs noted:  tcpdrop(8).

Updated release notes:  Mention ath_rate_sample rate control module,
libthr stack size update, Xorg 6.8.2.

Fixed a minor grammo in introduction, also adjust wording in pointers
to snapshots.
2005-03-19 19:54:09 +00:00
Bruce A. Mah
1cf444c14e Rather than pointing users at jp.freebsd.org for CURRENT
snapshots, instead point them at a Web page on the main FreeBSD
site that gives more information on the monthly (RE-created) snapshots
plus pointers to the jp.freebsd.org and se.freebsd.org snapshots.
2005-03-19 19:52:30 +00:00
Ian Dowse
4c11f583a3 Now that all architectures allow hooks to be inserted before
configure_final(), assert that "cold" is true in usb_cold_explore()
when there are busses to explore. When USB is kldloaded after boot,
usb_cold_explore() will still get invoked but the list of busses
to explore in that case should always be empty.
2005-03-19 19:27:38 +00:00
Joe Marcus Clarke
d19a25b5b3 Change the PATH for pkg_add to be the same used for pointyhat package builds.
This will give our users and porters a more consistent experience between
installing ports and installing packages.

Discussed with:	portmgr
2005-03-19 19:20:19 +00:00
Maxim Konovalov
c017b3b664 s/SLIST/STAILQ/
Spotted by:	clive
2005-03-19 19:17:17 +00:00
Ian Dowse
a3d327674a It was possible to have two threads concurrently aborting the same
transfer, which lead to panics or page faults. For example if a
transfer timed out, another thread could come along and attempt to
abort the same transfer while the timeout task was sleeping in
the *_abort_xfer() function.

Add an "aborting" flag to the private transfer state in each host
controller driver and use this to ensure that the abort is only
executed once. Also prioritise normal abort requests over timeouts
so that the callback is always given a status of USB_CANCELLED even
if the timeout-initiated abort began first.

The crashes caused by this bug were mainly reported in connection
with lpd printing to a USB printer.

PR:	usb/78208, usb/78986
2005-03-19 19:08:46 +00:00
Christian Brueffer
4029efa504 Xref carp(4) and polling(4).
MFC after:	3 days
2005-03-19 16:14:31 +00:00
Poul-Henning Kamp
9ca3d6d5b7 Instead of a rather useless generation number, use a sample number to
keep track of things.
2005-03-19 12:55:46 +00:00
Simon L. B. Nielsen
b0e79c15cf Bump documentation date for last commit before ru reminds me. 2005-03-19 08:57:00 +00:00
Simon L. B. Nielsen
4067affa6c Add nForce4 to the list of supported audio devices.
MFC after:	1 week
2005-03-19 08:54:06 +00:00
Poul-Henning Kamp
7f661c6ba1 Use subr_unit 2005-03-19 08:22:36 +00:00
Poul-Henning Kamp
773eff9d97 Sleeping is not allowed in uma->fini 2005-03-19 08:22:13 +00:00
Scott Long
e3f3b2ed9f Don't free the SIM object twice.
Submitted by: Coverity Prevent analysis tool
2005-03-19 06:32:57 +00:00