Commit Graph

135818 Commits

Author SHA1 Message Date
Jason Evans
26b5e3a18e Implement lazy deallocation of small objects. For each arena, maintain a
vector of slots for lazily freed objects.  For each deallocation, before
doing the hard work of locking the arena and deallocating, try several times
to randomly insert the object into the vector using atomic operations.

This approach is particularly effective at reducing contention for
multi-threaded applications that use the producer-consumer model, wherein
one producer thread allocates objects, then multiple consumer threads
deallocate those objects.
2007-11-27 03:13:15 +00:00
Jason Evans
bcd3523138 Avoid re-zeroing memory in calloc() when possible. 2007-11-27 03:12:15 +00:00
Jason Evans
1bbd1b8613 Fix stats printing of the amount of memory currently consumed by huge
allocations. [1]

Fix calculation of the number of arenas when 'n' is specified via
MALLOC_OPTIONS.

Clean up various style inconsistencies.

Obtained from:	[1] NetBSD
2007-11-27 03:09:23 +00:00
Bruce A. Mah
66288637d1 Use the most recent Xorg manpage set that we have. 2007-11-26 23:21:24 +00:00
Bruce A. Mah
3ca6e6e828 New release note: wpi(4). 2007-11-26 23:19:15 +00:00
Attilio Rao
49aead8a10 Simplify the adaptive spinning algorithm in rwlock and mutex:
currently, before to spin the turnstile spinlock is acquired and the
waiters flag is set.
This is not strictly necessary, so just spin before to acquire the
spinlock and to set the flags.
This will simplify a lot other functions too, as now we have the waiters
flag set only if there are actually waiters.
This should make wakeup/sleeping couplet faster under intensive mutex
workload.
This also fixes a bug in rw_try_upgrade() in the adaptive case, where
turnstile_lookup() will recurse on the ts_lock lock that will never be
really released [1].

[1] Reported by: jeff with Nokia help
Tested by: pho, kris (earlier, bugged version of rwlock part)
Discussed with: jhb [2], jeff
MFC after: 1 week

[2] John had a similar patch about 6.x and/or 7.x about mutexes probabilly
2007-11-26 22:37:35 +00:00
John Birrell
221a97c194 Although the entire src tree builds cleanly now without -fno-strict-aliasing
in the default CFLAGS, we're in the middle of a ports freeze, so we can't
really go making the corresponding change to the ports mk files.

I'll take -fno-strict-aliasing out again when the ports freeze ends.
2007-11-26 21:46:21 +00:00
John Baldwin
8ec770db87 Add a note to indicate that these files do borrow in part from mbr.s and
boot1.S

Requested by:	rnordier
2007-11-26 21:29:59 +00:00
Sam Leffler
1a4ae5c45b Be more careful handling off-channel frames: if the driver (wrongly)
sends frames up the stack after changing the current channel then
the lookup by ieee channel number may fail leaving a null ptr in
se_chan; if this happens fallback to the channel recorded when the
frame is processed (curchan).  Since the frame doesn't contribute
to scan results for the sta this is acceptable.

Reviewed by:	thompsa
MFC after:	3 days
2007-11-26 21:28:18 +00:00
Remko Lodder
1b72425ab3 Add the Intel 82801FB (ICH6) SMBus controller and the
Intel 82801GB (ICH7) SMBus controller.

PR:		85106 [1]
PR:		99663 [2]
Approved by:	imp (mentor), jhb
Submitted by:	Oliver Fromme <olli at lurza.secnetix dot de>, [1]
		Arthur Hartwig <arthur dot hartwig at nokia dot com>, [1]
		Lowell Gilbert <lgfbsd at be-well dot ilk dot org> [2]@
MFC After:	3 days
2007-11-26 19:54:54 +00:00
Jack F Vogel
abcc80c87a Fix for a reported panic in certain circumstances. When
calling em_stop() now make sure the TX lock is held as
well as CORE.
2007-11-26 19:47:03 +00:00
Søren Schmidt
91b4ff52c7 Use device_set_desc_copy in the generic ident as its used on a temp buffer. 2007-11-26 19:08:08 +00:00
Remko Lodder
0d985fab86 Add the FNW3603TX Planex NIC.
PR:		76081
Approved by:	imp (mentor)
Submitted by:	umi at pocke dot org
MFC After:	3 days
2007-11-26 18:25:07 +00:00
Remko Lodder
64d0afa7f8 Add the 845M GMCH controller.
PR:		114802
Approved by:	imp (mentor), anholt (private mail)
Submitted by:	Alex Goncharov <algo1 at comcast dot net>
MFC After:	3 days
2007-11-26 18:17:07 +00:00
Christian Brueffer
23ae24e09a Bump .Dd for previous revision. 2007-11-26 15:21:21 +00:00
Konstantin Belousov
6259969d36 Implement read_default_ldt in linux_modify_ldt(). It copies out zeroed
descriptor, like real Linux does.

Tested by: Yuriy Tsibizov <yuriy.tsibizov at gmail com>
Submitted by:	rdivacky
MFC after:	1 week
2007-11-26 11:06:19 +00:00
Xin LI
91f5fdd5fd Mention less upgrade to v416. 2007-11-26 08:58:52 +00:00
Xin LI
3951d17b6c Resolve conflicts. 2007-11-26 08:58:07 +00:00
Xin LI
ddccfc1d45 This commit was generated by cvs2svn to compensate for changes in r173932,
which included commits to RCS files with non-trunk default branches.
2007-11-26 08:57:30 +00:00
Xin LI
33a9f6ab73 Virgin import of less v416. 2007-11-26 08:57:30 +00:00
Darren Reed
b351a3c4b3 Fix 3 issues relating to the use of "auth" rules in IPFilter, from sourceforge:
1837014 Kernel panics after authentication of an outgoing packet
1836992 Potential bugs in packet auth code (w/patches)
1836967 Kernel panic when using auth rule with keep state
and another reported only to FreeBSD by Andiry (see PR)

PR:		kern/118251
Submitted by:	Andriy Syrovenko <andriys@gmail.com>
Reviewed by:	darrenr
MFC after:	5 days
2007-11-26 08:18:19 +00:00
John Birrell
ba90c265b0 Implement the _long functions using u_long rather than trying to
cast as uint32_t which is defined as unsigned int. gcc doesn't want to
consider that there might not be much difference between an int and
a long on a 32 bit architecture.
2007-11-26 05:52:45 +00:00
John Birrell
1b655ceb37 Add Dell's Photo AIO Printer 926. 2007-11-26 05:47:15 +00:00
Sam Leffler
901e004d2a silence compiler complaints 2007-11-26 05:03:55 +00:00
Joseph Koshy
742fe6de5d - Allow source descriptors with no data to be used as arguments to the
elf{32,64}_xlateto[fm]() translation functions.  This change makes our
  libelf compatible with other ELF(3) implementations. [1]
- Update manual page to reflect this change.
- Style fixes: wrap a long line.

Submitted by:	jb [1]
2007-11-26 03:09:33 +00:00
David Xu
eb326f14a2 Move the added code in revision 1.26 into function pthread_key_create,
it should be there.
2007-11-26 02:35:17 +00:00
Daniel Gerzo
89396d2512 Polish this manual page a bit:
- refer to the dummynet(4) man page only once, later use rather
  the .Nm macro.
- use .Va macro when refering to the sysctl variables
- grammar and markup fixes

Reviewed by:	keramida, trhodes, ru (roughly)
MFC-after:	1 week
2007-11-26 00:36:40 +00:00
Max Khon
cb0ed6eefb Fix -jX when makefiles are remade. 2007-11-25 20:43:27 +00:00
Alan Cox
da31e3aa04 Make contigmalloc(9)'s page laundering more robust. Specifically, use
vm_pageout_fallback_object_lock() in vm_contig_launder_page() to better
handle a lock-ordering problem.  Consequently, trylock's failure on the
page's containing object no longer implies that the page cannot be
laundered.

MFC after: 6 weeks
2007-11-25 20:37:29 +00:00
Christian Brueffer
4faa4b080e Clean up after previous revision. 2007-11-25 16:14:32 +00:00
Max Laier
1030a1a9cb pfil(9) locking take 3: Switch to rmlock(9)
This has the benefit that rmlocks have proper support for reader recursion
(in contrast to rwlock(9) which could potential lead to writer stravation).
It also means a significant performance gain, eventhough only visible in
microbenchmarks at the moment.

Discussed on:	-arch, -net
2007-11-25 12:41:47 +00:00
Alan Cox
9c5ce94257 Tidy up: Add comments. Eliminate the pointless
malloc_type_allocated(..., 0) calls that occur when contigmalloc() has
failed.  Eliminate the acquisition and release of the page queues lock
from vm_page_release_contig().  Rename contigmalloc2() to
contigmapping(), reflecting what it does.
2007-11-25 07:42:34 +00:00
Joseph Koshy
c89d1cf588 - Add an overview of the pmc(3) API.
- Defer detailed descriptions of individual functions in the API to
  function-specific manual pages.
2007-11-25 06:50:31 +00:00
Joseph Koshy
677d4530c2 Move the following functions into their own manual pages:
- pmc_attach(), pmc_detach().
- pmc_capabilities(), pmc_cpuinfo(), pmc_ncpu(), pmc_npmc(), pmc_pmcinfo(),
  pmc_width().
- pmc_get_driver_stats().
- pmc_get_msr().
- pmc_read(), pmc_rw(), pmc_write().
- pmc_set().
- pmc_start(), pmc_stop().
2007-11-25 06:38:55 +00:00
Andrew Thompson
80ddfb40e4 Have the lagg interface generate link up/down events, the interface is marked
as up if at least one of its ports also has a link up. This fixes using
carp+lagg together and any other system that relies on linkstate events.

PR:		kern/113956
MFC after:	3 days
2007-11-25 06:30:46 +00:00
Joseph Koshy
a5c506d293 - Document an EAGAIN error return from a PMC_OP_CONFIGURELOG request.
- Document some EINVAL error returns from PMC_OP_CONFIGURELOG,
  PMC_OP_WRITELOG and PMC_OP_FLUSHLOG that were missed earlier.
- Correct a typo: s/PMC_OP_ALLOCATE/PMC_OP_PMCALLOCATE/.
- Describe how log file writing errors are returned by a PMC_OP_FLUSHLOG
  request.
2007-11-25 06:05:41 +00:00
Joseph Koshy
f28834c1a2 - Describe function pmc_init() in its own manual page.
- Describe functions pmc_configure_logfile(), pmc_flush_logfile() and
  pmc_writelog() in their own manual page.
2007-11-25 03:18:42 +00:00
Attilio Rao
0e933b2f3c Update sx(9) lock manpage in order to add missing prototypes for function
sx_slock_sig() and sx_xlock_sig() and their respective explanation.
2007-11-25 01:55:53 +00:00
Attilio Rao
6442838d08 Update rwlock(9) manpage adding missing bits about the rw_init_flags()
function, its options and explaining that recursion for writers can be
selectively enabled.

Reported by: rwatson, Skip Ford <skip at menantico dot com>
2007-11-25 01:53:58 +00:00
Robert Watson
2b19cb1b87 More carefully handle various cases in sysctl_drop(), such as unlocking
the inpcb when there's an inpcb without associated timewait state, and
not unlocking when the inpcb has been freed.  This avoids a kernel panic
when tcpdrop(8) is run on a socket in the TIMEWAIT state.

MFC after:	3 days
Reported by:	Rako <rako29 at gmail dot com>
2007-11-24 18:43:59 +00:00
Robert Watson
3b6efc61a3 Respell 'Conatainer' as 'Container'.
MFC after:	3 days
Reported by:	Toomas Aas <toomas dot aas at raad dot tartu dot ee>
2007-11-24 18:39:01 +00:00
Mike Makonnen
e38d709cd6 Hmmm, must get reading glasses or else change the font on my terminal. That
comma looks like a period.

Noticed by: brueffer
2007-11-24 11:58:54 +00:00
Joseph Koshy
fcdeaa0525 Describe convenience functions `pmc_name_of_{capability,class,cputype,
disposition,event,mode,state}' in their own manual page.
2007-11-24 11:05:36 +00:00
Christian Brueffer
a4b1fac3d2 Minor language cleanup.
MFC after:	3 days
2007-11-24 09:32:45 +00:00
Attilio Rao
4a32616a77 Fix the spinlock static table adding missing spinlocks.
- rm_spinlock has turnstile chain as child
- srclock has callout and clk as child, found by witness "emulation".
  Just move it very high in our ranking
2007-11-24 04:32:32 +00:00
Attilio Rao
2c2bebfcb3 transferlockers() is a very dangerous and hack-ish function as waiters
should never be moved by one lock to another.
As, luckily, nothing in our tree is using it, axe the function.

This breaks lockmgr KPI, so interested, third-party modules should update
their source code with appropriate replacement.

Ok'ed by: ups, rwatson
MFC after: 3 days
2007-11-24 04:22:28 +00:00
Joseph Koshy
1a94fde847 - Improve clarity: use 'scope' to distinguish between process and system
PMCs and 'mode' to distinguish between counting and sampling.
- Document PMC row dispositions.
- Document the requirement for PRIV_PMC_MANAGE privilege when issuing
  PMC_OP_PMCADMIN requests.
2007-11-24 03:48:18 +00:00
John Birrell
962e1ce30f Fix strict alias warnings. 2007-11-23 23:56:03 +00:00
Ruslan Ermilov
5f3a6945d4 Also check setuid executables on ZFS. 2007-11-23 13:00:31 +00:00
Joseph Koshy
115cadfd60 Describe pmc_event_names_of_class() in its own manual page. 2007-11-23 12:30:55 +00:00