Commit Graph

145953 Commits

Author SHA1 Message Date
Robert Watson
10263f0832 Add DTrace probes to the NFS access and attribute caches. Access cache
events are:

  nfsclient:accesscache:flush:done
  nfsclient:accesscache:get:hit
  nfsclient:accesscache:get:miss
  nfsclient:accesscache:load:done

They pass the vnode, uid, and requested or loaded access mode (if any);
the load event may also report a load error if the RPC fails.

The attribute cache events are:

  nfsclient:attrcache:flush:done
  nfsclient:attrcache:get:hit
  nfsclient:attrcache:get:miss
  nfsclient:attrcache:load:done

They pass the vnode, optionally the vattr if one is present (hit or load),
and in the case of a load event, also a possible RPC error.

MFC after:	1 month
Sponsored by:	Google, Inc.
2009-03-24 17:14:34 +00:00
Sam Leffler
a390848f98 require wlan for ancillary modules; they are required to link 2009-03-24 17:10:42 +00:00
Luigi Rizzo
babec003ec remove unused "package" entry.
default to use devfs (we do that anyways)

Many small changes in comments
2009-03-24 16:03:46 +00:00
VANHULLEBUS Yvan
ff1fdd77ff Fixed indentation for LINKTYPE_ENC
Approved by:	gnn(mentor)
2009-03-24 15:57:35 +00:00
Marius Strobl
b82183f68f Revert unintended part of r190375. 2009-03-24 15:56:15 +00:00
Marius Strobl
e145f329c1 Given that zfs.ko depends on opensolaris.ko also build the latter by
default on sparc64.
2009-03-24 15:52:45 +00:00
Marius Strobl
f58fac9df4 As with ZFS use real atomic operations for sparc64. 2009-03-24 15:48:35 +00:00
Rui Paulo
cd39fead00 Bump __FreeBSD_version to reflect tcpdump 4.0.0 and libpcap 1.0.0
import.
2009-03-24 13:11:16 +00:00
Edwin Groothuis
a7c248498e MFV of tzdata2009d:
- Morocco will observe DST from 2009-06-01 00:00 to 2009-08-21 00:00
- Tunisia will not observe DST this year.
- Syria will start DST on 2009-03-27 00:00 this year
- the Province of San Luis will go to utc-04:00

MFC after:	1 week
2009-03-24 10:50:30 +00:00
Edwin Groothuis
b7e4449445 Vendor import of tzdata2009d
- Morocco will observe DST from 2009-06-01 00:00 to 2009-08-21 00:00
- Tunisia will not observe DST this year.
- Syria will start DST on 2009-03-27 00:00 this year
- the Province of San Luis will go to utc-04:00
2009-03-24 10:47:53 +00:00
Weongyo Jeong
c4f82f895d set NULL after free to avoid duplicate free.
Tested by:	Ganbold <ganbold_at_micom.mng.net>
2009-03-24 04:20:17 +00:00
Bruce M Simpson
5b35d05538 Don't call m_freem() after ip_output(), as it always consumes
the mbuf chain provided to it.

Found by:	Pierre Guinoiseau
2009-03-24 01:22:12 +00:00
Sam Leffler
aa28501a67 fix build w/ AH_DEBUG 2009-03-24 00:09:35 +00:00
Juli Mallett
1b344bf7d7 Remove gratuitous unlock in error case.
Reviewed by:	sam
2009-03-24 00:08:58 +00:00
Xin LI
6c0af71bdc Save errno before calling _close(), which may clear it.
Obtained from:	OpenBSD
2009-03-23 23:43:07 +00:00
Xin LI
7efabbb915 use more proper format string.
Obtained from:	NetBSD via OpenBSD
2009-03-23 23:22:09 +00:00
Jung-uk Kim
d2b227cd49 - Clean up suspend/resume code for amd64.
- Call acpi_resync_clock() to reset system time before hardclock is ready
to tick.  Note we assume the current timecounter hardware and RTC are
already available for read operation.

Tested by:	mav
2009-03-23 22:35:30 +00:00
Jung-uk Kim
0755473ba2 Add a function to reset system time after resuming, which will be used
by amd64 shortly.  It can be turned off by setting "debug.acpi.reset_clock"
tunable to zero.
2009-03-23 22:12:33 +00:00
Jung-uk Kim
d42f0ad883 Check whether devd is running before calling resume notifier and
reshuffle code to reduce unnecessary locking coverage.
2009-03-23 22:06:09 +00:00
Jung-uk Kim
eae44ae03d Clean up MI inittodr(9) and kill noop code.
It was derived from i386 version long ago but never resync'ed again.
Originally, i386 version compared the current time from realtime clock
with time_second (which was just `time' in the old days).  When this MI
version was written, it was wrongly compared against `base' AND never
used because of a bug (typo?) in the code.  This check was killed
in i386 version when home-rolled calendaric calculation was removed.
Now, we just remove the code here as well to make the code simpler.
2009-03-23 21:16:21 +00:00
John Baldwin
9b84ba1cbb Improve the description of a few sysctls.
Submitted by:	bde (partially)
MFC after:	3 days
2009-03-23 20:18:06 +00:00
George V. Neville-Neil
0c1ff9c605 Minor updates to the Chelsio driver, including removing an LOR.
Submitted by:	Navdeep Parhar at Chelsio
Reviewed by:	gnn
MFC after: 	3 weeks
2009-03-23 19:58:26 +00:00
John Baldwin
f7c701d528 Reenable 64-bit DMA for twa(4) controllers, but use a boundary of 4GB to
prevent individual transactions from crossing a 4GB address boundary.  Due
to bus_size_t type limitations, the driver uses a 2GB boundary in PAE
kernels.

Reviewed by:	scottl
MFC after:	1 week
2009-03-23 19:27:23 +00:00
Andrew Thompson
91f72cedda Fix order of debug printf items, addr and config# were swapped. 2009-03-23 19:10:38 +00:00
John Baldwin
1241692334 Release driver lock at the end of the watchdog routine instead of trying to
acquire it again.

Submitted by:	bland
MFC after:	3 days
2009-03-23 17:51:07 +00:00
Xin LI
569e2ef6a9 Support for a new environment variable, LD_ELF_HINTS_PATH for overriding
the rtld hints file.  This environment variable would be unset if the
process is considered as tainted with setuid/setgid.  This feature gives
a convenient way of using a custom set of shared library that is not
located in the default location and switch back.

Feature requested by:	iXsystems
Original patch by:	John Hixson
MFC after:		2 weeks
2009-03-23 16:49:00 +00:00
VANHULLEBUS Yvan
485cf78203 Fixed comments so it stays in 80 chars by line
with hard tabs of 8 chars....

Approved by:	gnn(mentor)
2009-03-23 16:20:39 +00:00
Doug Barton
86780e2538 Don't tempt svn to expand the example $FreeBSD strings 2009-03-23 14:42:41 +00:00
Marius Strobl
c9ffd9f058 - Ensure that INTx isn't disabled, as these chips apparently have a
quirk requiring it to be enabled even when using MSI. This makes
  the latter work again after r189285.
- Remove a comment which no longer applies since r190194.
2009-03-23 14:36:50 +00:00
Poul-Henning Kamp
0353e472df Save (empty) worklist at successful completion 2009-03-23 11:07:34 +00:00
Pyun YongHyeon
6d6588a152 Togging ALLMULTI does not require controller reset so have Rx
filtering handle this. Introduce a new function msk_rxfilter that
handles Rx filter configuration and multicast setup as well as
promiscuous mode. This simplifies code a lot.
Promiscuous mode always have preference to any other Rx
filtering so don't disable the mode when ALLMULTI is set.
2009-03-23 05:46:28 +00:00
Alexander Kabaev
62a5f39429 Do not pass uninitialized data buffer to devctl_queue_data function.
This was botched in revision 190191.
2009-03-23 01:24:17 +00:00
Alexander Kabaev
9999864a87 Add safety check that does not allow empty strings to be queued
to the devctl notification queue. Empty strings cause devctl read
call to return 0 and result in devd exiting prematurely.

The actual offender (ugen notes for root hubs) will be fixed
by separate commit.
2009-03-23 01:13:34 +00:00
Gregory Neil Shapiro
49d24250e2 Add the URL for RFC2142
PR:		conf/127510
2009-03-23 00:40:07 +00:00
Pyun YongHyeon
df284b3c3a Fix inversed logic. pci_find_extcap() returns 0 when it finds
specified capability.
2009-03-23 00:27:46 +00:00
Colin Percival
3f935cf342 Correctly sanity-check timer IDs. [SA-09:06]
Limit the size of malloced buffer when dumping environment
variables. [EN-09:01]

Approved by:	so (cperciva)
Approved by:	re (kensmith)
Security:	FreeBSD-SA-09:06.ktimer
Errata:		FreeBSD-EN-09:01.kenv
2009-03-23 00:00:50 +00:00
Stefan Farfeleder
acd5c42915 Test the r190298 change. 2009-03-22 23:00:52 +00:00
Stefan Farfeleder
86d8da5d5b Fix the behaviour of the read built-in when IFS is unset.
Obtained from:	NetBSD
2009-03-22 22:57:53 +00:00
Stefan Farfeleder
8b1fa08c95 Add a few tests for the read built-in. 2009-03-22 22:14:46 +00:00
Stefan Farfeleder
b6748ec20c Improve the IFS handling of the read built-in.
Obtained from:	NetBSD
Submitted by:	Jilles Tjoelker
2009-03-22 22:09:12 +00:00
Robert Watson
47294818f9 Add dtnfsclient, a first cut at an NFSv2/v3 client reuest DTrace
provider.  The NFS client exposes 'start' and 'done' probes for NFSv2
and NFSv3 RPCs when using the new RPC implementation, passing in the
vnode, mbuf chain, credential, and NFSv2 or NFSv3 procedure number.
For 'done' probes, the error number is also available.

Probes are named in the following way:

  ...
  nfsclient:nfs2:write:start
  nfsclient:nfs2:write:done
  ...
  nfsclient:nfs3:access:start
  nfsclient:nfs3:access:done
  ...

Access to the unmarshalled arguments is not easily available at this
point in the stack, but the passed probe arguments are sufficient to
to a lot of interesting things in practice.  Technically, these probes
may cover multiple RPC retransmits, and even transactions if the
transaction ID change as a result of authentication failure or a
jukebox error from the server, but usefully capture the intent of a
single NFS request, such as access, getattr, write, etc.

Typical use might involve profiling RPC latency by system call, number
of RPCs, how often a getattr leads to a call to access, when failed
access control checks occur, etc.  More detailed RPC information might
best be provided by adding a krpc provider.  It would also be useful
to add NFS client probes for events such as the access cache or
attribute cache satisfying requests without an RPC.

Sponsored by:	Google, Inc.
MFC after:	1 month
2009-03-22 22:07:52 +00:00
Stefan Farfeleder
d956d1eed5 Update test for r190284. 2009-03-22 21:12:00 +00:00
Stefan Farfeleder
d4b1e37429 - Apply the r190270 changes to printing of single aliases too.
- Sort the aliases before printing them.
2009-03-22 21:09:22 +00:00
Robert Noland
ba0b16188e Correct the documentation to reflect the fact that BUS_DMA_NOCACHE is a
flag to bus_dmamem_alloc() rather than bus_dmamap_load();

Discussed with:	kib
MFC after:	3 days
2009-03-22 21:00:58 +00:00
Robert Noland
746871c800 Fix up the flags to bus_dmamem again. The man page incorrectly showed
the BUS_DMA_NOCACHE flags as being a valid flag for load instead of alloc.

Discussed with:	kib
MFC after:	3 days
2009-03-22 20:58:29 +00:00
Alan Cox
b4862e19af Update stale comments. The alternate address space mapping was eliminated
when PAE support was added to i386.  The direct mapping exists on amd64.
2009-03-22 18:56:26 +00:00
Stefan Farfeleder
2c93c8f251 Update and extend the tests for alias. 2009-03-22 17:22:14 +00:00
Stefan Farfeleder
0de913c328 Make the output of the alias built-in POSIX-compliant: Drop the leading 'alias'
and suppress printing the trailing space which is added for internal purposes.
2009-03-22 17:20:42 +00:00
Andrew Thompson
2989a67790 Set the endpoint on the preallocated fifo so it doesnt get matched as an
endpoint fifo.

Reported by:	Pieter de Goeje
2009-03-22 17:09:56 +00:00
Konstantin Belousov
2e92f03abd 7.2 will be the first release where strndup() appears.
Submitted by:	Florian Smeets <flo kasimir com>
MFC after:	3 days
2009-03-22 15:29:12 +00:00