]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
14 years agoMFC: r204008: realloc() with a proper amount of memory.
ru [Mon, 22 Feb 2010 15:58:10 +0000 (15:58 +0000)]
MFC: r204008: realloc() with a proper amount of memory.

14 years agoMFC r203551:
jh [Sun, 21 Feb 2010 13:17:35 +0000 (13:17 +0000)]
MFC r203551:

- Cast intptr_t, pid_t and time_t values to intmax_t and use %jd with
  printf.
- Cast the system call return value to long and use %ld in a printf in
  ktrsysret().

PR: bin/123774

14 years agoMerge r200413 and r202755 from head to stable/8:
antoine [Sun, 21 Feb 2010 11:22:01 +0000 (11:22 +0000)]
Merge r200413 and r202755 from head to stable/8:
reduces white space diff between head and stable/8

Suggested by: ru@

14 years agoMFC r203866:
kib [Sun, 21 Feb 2010 11:13:15 +0000 (11:13 +0000)]
MFC r203866:
Invalid filesystem might cause the bp to be never read.

14 years agoMFC r203085
alc [Sun, 21 Feb 2010 04:10:25 +0000 (04:10 +0000)]
MFC r203085
  Optimize pmap_demote_pde() by using the new KPTmap to access a kernel
  page table page instead of creating a temporary mapping to it.

  Set the PG_G bit on the page table entries that implement the KPTmap.

  Locore initializes the unused portions of the NKPT kernel page table
  pages that it allocates to zero.  So, pmap_bootstrap() needn't zero
  the page table entries referenced by CMAP1 and CMAP3.

  Simplify pmap_set_pg().

14 years agoMFC rev 203883:
marcel [Sat, 20 Feb 2010 22:43:12 +0000 (22:43 +0000)]
MFC rev 203883:
o   Eliminate IA64_PHYS_TO_RR6 by calling bus_space_map() or pmap_mapdev().
o   Implement bus_space_map() in terms of pmap_mapdev().
o   Have ia64_pib hold the uncached virtual address of the PIB.

14 years agoMFC r203157, r203816:
jh [Sat, 20 Feb 2010 13:35:05 +0000 (13:35 +0000)]
MFC r203157, r203816:

Handle short reads when the -P option is used and remove some dead code.

PR: bin/121502

14 years agoMFC from head to stable/8:
antoine [Sat, 20 Feb 2010 12:48:44 +0000 (12:48 +0000)]
MFC from head to stable/8:
r202896:
  Unbreak world:
  - WITHOUT_OPENSSH (and WITH_KERBEROS)
  - WITHOUT_KERBEROS and WITH_GSSAPI

  PR: 137483
  Submitted by: bf
r203016 by ru@:
  Regen the list of prebuild libraries using tools/make_libdeps.sh.

Reviewed by: ru@

14 years agoMFC r203826:
kib [Sat, 20 Feb 2010 11:58:19 +0000 (11:58 +0000)]
MFC r203826:
Use M_ZERO instead of calling bzero().
Fix function name in the comment.

MFC r203828:
Fix function name in the comment in the second location too.

14 years agoMFC r203822:
kib [Sat, 20 Feb 2010 11:54:18 +0000 (11:54 +0000)]
MFC r203822:
Remove unused macros.

14 years agoMFC r203818:
kib [Sat, 20 Feb 2010 11:50:50 +0000 (11:50 +0000)]
MFC r203818:
Clear the bp pointer when buffer is already brelse()d.

14 years agoMFC r200291, r203529:
mav [Fri, 19 Feb 2010 18:15:45 +0000 (18:15 +0000)]
MFC r200291, r203529:
Increase Max Read Request Size for PCIe chips from 512 to 1024 bytes.
It gives those beasts additional 10% of write bandwidth.
Use new helper functions to do it.

14 years agoMFC r203528:
mav [Fri, 19 Feb 2010 18:07:51 +0000 (18:07 +0000)]
MFC r203528:
Add pci_get|set_max_read_req() helper functions to control maximum PCIe
read request size.

14 years agoMFC r203873:
mav [Fri, 19 Feb 2010 17:54:03 +0000 (17:54 +0000)]
MFC r203873:
With FBS enabled, we have no idea what command caused timeout.
Implement same logic as in siis(4) - wait for other commands
complete or timeout and then give some more time.

14 years agoMFC r203931:
mav [Fri, 19 Feb 2010 17:45:47 +0000 (17:45 +0000)]
MFC r203931:
Make CD driver a bit more robust and predictable to unreported errors.

14 years agoMFC: r203849
rmacklem [Fri, 19 Feb 2010 16:36:08 +0000 (16:36 +0000)]
MFC: r203849
Change the default value for vfs.newnfs.enable_locallocks to 0 for
the experimental NFS server, since local locking is known to be
broken and the patch to fix it is still a work in progress.

14 years agoMFC: r203918: Unbreak makefiles by removing mentions of DEFAULT_THREAD_LIB.
ru [Thu, 18 Feb 2010 16:52:49 +0000 (16:52 +0000)]
MFC: r203918: Unbreak makefiles by removing mentions of DEFAULT_THREAD_LIB.

14 years agoMFC: r203848
rmacklem [Thu, 18 Feb 2010 16:23:13 +0000 (16:23 +0000)]
MFC: r203848
This fixes the experimental NFS server so that it won't crash in the
caching code for IPv6 by fixing a typo that used the incorrect variable.
It also fixes the indentation of the statement above it.

Reported by: simon AT comsys.ntu-kpi.kiev.ua

14 years agoMFC r203685:
brucec [Thu, 18 Feb 2010 11:07:36 +0000 (11:07 +0000)]
MFC r203685:

Document the usfs driver and the NO_SYSCTL_DESCR option, and update the comment
for umass.
Don't include the sysctl description variables in aic7xxx when NO_SYSCTL_DESCR
is used.

Approved by: rrs (mentor)

14 years agoMFC r203679:
brucec [Thu, 18 Feb 2010 11:03:35 +0000 (11:03 +0000)]
MFC r203679:

Remove the usb2_input_kbd directive that was missed during the renaming of the
drivers in the usb2 stack.

Approved by: rrs (mentor)

14 years agoMFC r203678:
brucec [Thu, 18 Feb 2010 10:59:32 +0000 (10:59 +0000)]
MFC r203678:

Rename usb2_ structures and variables to usb_.

Approved by: rrs (mentor)

14 years agoMFC r203323:
brucec [Thu, 18 Feb 2010 10:55:42 +0000 (10:55 +0000)]
MFC r203323:

Remove extra semicolon.

Approved by: rrs (mentor)

14 years agoMFC r203322:
brucec [Thu, 18 Feb 2010 10:51:31 +0000 (10:51 +0000)]
MFC r203322:

Fix typo of ENOTCONN.
Add missing RETURN VALUES section in sctp_opt_info(3).

Approved by: rrs (mentor)

14 years agoMFC r203690:
brucec [Thu, 18 Feb 2010 10:48:37 +0000 (10:48 +0000)]
MFC r203690:

Xorg isn't treated as a distribution, so /usr/X11R6/lib shouldn't be
configured when running ldconfig.

PR: bin/138945
Approved by: rrs (mentor)

14 years agoMFC r203688:
brucec [Thu, 18 Feb 2010 10:46:25 +0000 (10:46 +0000)]
MFC r203688:

Initialize the execfile argument to NULL instead of _PATH_DEVNULL. This allows
the -M option to be used without specifying -N.

PR: bin/138146
Approved by: rrs (mentor)

14 years agoMFC r203681:
brucec [Thu, 18 Feb 2010 10:39:53 +0000 (10:39 +0000)]
MFC r203681:

Fix typo and remove extra spaces.

Approved by: rrs (mentor)

14 years agoAdd the long missing "destroy" option [mfc of 203505]
mjacob [Thu, 18 Feb 2010 01:15:58 +0000 (01:15 +0000)]
Add the long missing "destroy" option [mfc of 203505]

14 years agoMFC r203441:
kib [Wed, 17 Feb 2010 09:09:12 +0000 (09:09 +0000)]
MFC r203441:
Placate new binutils, by using 16-bit %ax instead of 32-bit %eax as an
argument for fnstsw. Explicitely specify sizes for the XMM control and
status word and X87 control and status words.

14 years agoMFC r203875:
kib [Wed, 17 Feb 2010 09:03:38 +0000 (09:03 +0000)]
MFC r203875:
Do not leak process lock when current thread is not allowed to see target.

14 years agoMerge 203689 from head:
gavin [Tue, 16 Feb 2010 22:23:33 +0000 (22:23 +0000)]
Merge 203689 from head:

  Install the padlock(4) man page on amd64 as well as i386, to match the
  platforms where the driver itself is compiled and installed.

PR: docs/130895
Reported by: George Hartzell <hartzell alerce.com>

14 years agoMerge r202461 from head:
gavin [Tue, 16 Feb 2010 22:19:55 +0000 (22:19 +0000)]
Merge r202461 from head:

  Implement an "-x" option to cp(1), for compatibility with Linux and
  feature parity with du(1) and similar:  When set, cp(1) will not traverse
  mount points.

PR: bin/88056
Initial patch by: Graham J Lee   leeg teaching.physics.ox.ac.uk

14 years agoMFC: r203676
emax [Tue, 16 Feb 2010 19:00:47 +0000 (19:00 +0000)]
MFC: r203676

Introduce new rc.conf variable firewall_coscripts. It can be used to
specify list of executables and/or rc scripts that should be executed
after firewall starts/stops.

Submitted by: Yuri Kurenkov <y dot kurenkov at init dot ru>
Reviewed by: rhodes, rc@

14 years agoMFC r203408:
delphij [Tue, 16 Feb 2010 06:34:44 +0000 (06:34 +0000)]
MFC r203408:

Prevent NULL deference by checking return value of
gctl_get_asciiparam.

14 years agoMFC r203124:
mav [Mon, 15 Feb 2010 16:41:30 +0000 (16:41 +0000)]
MFC r203124:
Note added FIS-based switching support.

14 years agoMFC r203637:
raj [Mon, 15 Feb 2010 15:00:40 +0000 (15:00 +0000)]
MFC r203637:

Improve checking whether an ARM VA has a valid mapping before performing cache
sync.

VIPT/PIPT caches need valid VA-PA mapping in PTE for a cache operation to
succeed (unlike VIVT). Prior to this fix pmap was using l2pte_valid() for that
check, but this is not sufficient as the function merely checks if a PTE
exists (there can be existing but _invalid_ entries in the table).

A new pmap_has_valid_mapping() routine is introduced to do this job right by
checking proper PTE flags.

Among other potential problems this cures coherency issues with L2 caches on
MV-78100.

Submitted by: Grzegorz Bernacki, Piotr Ziecik
Reviewed, tested by: marcel
Obtained from: Semihalf

14 years agoMFC: r202969: Actualize.
ru [Mon, 15 Feb 2010 11:29:27 +0000 (11:29 +0000)]
MFC: r202969: Actualize.

14 years agoMFC r203426:
mav [Sun, 14 Feb 2010 20:00:21 +0000 (20:00 +0000)]
MFC r203426:
Disable PHY of unconnected ports when interface power management enabled.
It allows to save a bit more power (about 0.5W on 2 unused ports of ICH8M).

14 years agoMFC r203165:
mav [Sun, 14 Feb 2010 19:59:06 +0000 (19:59 +0000)]
MFC r203165:
Reset port on disconnect event, to abort any running requests.

14 years agoMFC r203123:
mav [Sun, 14 Feb 2010 19:57:54 +0000 (19:57 +0000)]
MFC r203123:
Add FIS-based switching support. If controller supports FBS, it allows
several devices beyond Port Multiplier to work simultaneously, substantially
increasing performance.

14 years agoMFC r203449:
mav [Sun, 14 Feb 2010 19:56:42 +0000 (19:56 +0000)]
MFC r203449:
Implement poll method for atapicam.
It is not perfect, but better then nothing.

14 years agoMFC r203347:
mav [Sun, 14 Feb 2010 19:55:41 +0000 (19:55 +0000)]
MFC r203347:
NetCell is a PCI hardware RAID without cable and mode setting.

14 years agoMFC r203525:
mav [Sun, 14 Feb 2010 19:54:16 +0000 (19:54 +0000)]
MFC r203525:
Report SATA300 chips also as SATA.

14 years agoMFC r203445:
mav [Sun, 14 Feb 2010 19:53:09 +0000 (19:53 +0000)]
MFC r203445:
Report to CAM unit number of underlying ataX bus device, instead of this
atapicamX. It is more apropriate and useful together with "ata" name.

14 years agoMFC r203499:
mav [Sun, 14 Feb 2010 19:51:39 +0000 (19:51 +0000)]
MFC r203499:
On probe error, if restart requested, skip any retries and recovery.
Just restart probe from the beginning immediately.

14 years agoMFC r203421:
mav [Sun, 14 Feb 2010 19:50:33 +0000 (19:50 +0000)]
MFC r203421:
Add Power Up In Stand-by feature support. Device with PUIS enabled
require explicit command to do initial spin-up. Mark that command
with CAM_HIGH_POWER flag, to allow CAM manage staggered spin-up.

14 years agoMFC r203386:
mav [Sun, 14 Feb 2010 19:48:53 +0000 (19:48 +0000)]
MFC r203386:
Change the way in which fake async events generated. Do not use
taskqueue for lock decoupling, as it causes unwanted races.

14 years agoMFC r203385:
mav [Sun, 14 Feb 2010 19:46:54 +0000 (19:46 +0000)]
MFC r203385:
- Use separate buffer for identify data fetching. We can't use main buffer
here if device already running, as data need to be formatted before use.
- Remove some saved_ccb variables. They are unused now.

14 years agoMFC r203376, r203384:
mav [Sun, 14 Feb 2010 19:44:48 +0000 (19:44 +0000)]
MFC r203376, r203384:
- Give ATA/SATA SIMs info about ATAPI packet size, supported by device.
- Make ATA XPT to reject longer SCSI CDBs then supported by device, or
any SCSI CDBs, if device doesn't support ATAPI.

14 years agoMFC r203108:
mav [Sun, 14 Feb 2010 19:38:27 +0000 (19:38 +0000)]
MFC r203108:
Large set of CAM improvements:
- Unify bus reset/probe sequence. Whenever bus attached at boot or later,
CAM will automatically reset and scan it. It allows to remove duplicate
code from many drivers.
- Any bus, attached before CAM completed it's boot-time initialization,
will equally join to the process, delaying boot if needed.
- New kern.cam.boot_delay loader tunable should help controllers that
are still unable to register their buses in time (such as slow USB/
PCCard/ CardBus devices), by adding one more event to wait on boot.
- To allow synchronization between different CAM levels, concept of
requests priorities was extended. Priorities now split between several
"run levels". Device can be freezed at specified level, allowing higher
priority requests to pass. For example, no payload requests allowed,
until PMP driver enable port. ATA XPT negotiate transfer parameters,
periph driver configure caching and so on.
- Frozen requests are no more counted by request allocation scheduler.
It fixes deadlocks, when frozen low priority payload requests occupying
slots, required by higher levels to manage theit execution.
- Two last changes were holding proper ATA reinitialization and error
recovery implementation. Now it is done: SATA controllers and Port
Multipliers now implement automatic hot-plug and should correctly
recover from timeouts and bus resets.
- Improve SCSI error recovery for devices on buses without automatic sense
reporting, such as ATAPI or USB. For example, it allows CAM to wait, while
CD drive loads disk, instead of immediately return error status.
- Decapitalize diagnostic messages and make them more readable and sensible.
- Teach PMP driver to limit maximum speed on fan-out ports.
- Make boot wait for PMP scan completes, and make rescan more reliable.
- Fix pass driver, to return CCB to user level in case of error.
- Increase number of retries in cd driver, as device may return several UAs.

14 years agoMFC r203030:
mav [Sun, 14 Feb 2010 19:28:45 +0000 (19:28 +0000)]
MFC r203030:
Add support for SATA part of Marvell 88SE912x controllers to ahci(4).
Limit early revisions from 6Gb/s to 3Gb/s by default, or they negotiate
only 1.5Gb/s, when 3Gb/s devices connected.

Add dummy driver for PATA part of these controllers, preventing generic
driver attach them. It causes system freeze when SATA controller used after
PATA was touched.

14 years agoMFC r202699:
mav [Sun, 14 Feb 2010 19:23:05 +0000 (19:23 +0000)]
MFC r202699:
Make ata_getrev() an optional method by implementing ata_null_getrev().
This fixes a bogus '???' boot message on Cambria boards with a CompactFlash
card.

14 years agoMFp4:
mav [Sun, 14 Feb 2010 12:24:12 +0000 (12:24 +0000)]
MFp4:
After last running command completed, give commands in timeout state
second time.

14 years agoMFC r203524:
mav [Sun, 14 Feb 2010 12:04:25 +0000 (12:04 +0000)]
MFC r203524:
When hacking INQUIRY result, make sure that it is right INQUIRY and there
is enough of result to hack.

14 years agoMFC r203489:
mav [Sun, 14 Feb 2010 12:03:04 +0000 (12:03 +0000)]
MFC r203489:
Return CAM_RELEASE_SIMQ flag only when it is needed, when SIM really
was frozen before and should be released.

14 years agoMFC r203484:
mav [Sun, 14 Feb 2010 12:00:23 +0000 (12:00 +0000)]
MFC r203484:
Do not release device, when changing number of openings.

14 years agoMFC r202150, r202170:
mav [Sun, 14 Feb 2010 11:53:51 +0000 (11:53 +0000)]
MFC r202150, r202170:
Make OSS_GETVERSION ioctl really work. It has 'M' group, not 'P',
as different nearby ones, and was grabbed by MIXER_xxx() handler.

While there, replace '(cmd & MIXER_xxx(0)) == MIXER_xxx(0)' expressions
with more correct '(cmd & ~0xff) == MIXER_xxx(0)'. Use of bit operations
to compare numeric fields doesn't looks sane.

14 years agoMerge r203310,203547,203717 from head:
gavin [Sun, 14 Feb 2010 09:48:53 +0000 (09:48 +0000)]
Merge r203310,203547,203717 from head:

  Implement the "-i" option to sysctl(8), to ignore failures while
  retrieving individual OIDs.  This allows the same list of OIDs to be
  passed to sysctl(8) across different systems where particular OIDs may
  not exist, and still get as much information as possible from them.

PR: bin/123644
Submitted by: dhw

14 years agoMerge r203620,203621 from head:
gavin [Sun, 14 Feb 2010 09:40:58 +0000 (09:40 +0000)]
Merge r203620,203621 from head:

  Document support for the D-Link DFE520-TX card (supported with the vr(4)
  driver)

PR: kern/135989
Submitted by: "Rashid N. Achilov"  citycat4 ngs.ru

14 years agoMerge r203636 from head:
gavin [Sun, 14 Feb 2010 09:37:13 +0000 (09:37 +0000)]
Merge r203636 from head:

  Correct arguments to free_unr(), "item" was missing.

14 years agoMerge r200530 from head:
gavin [Sun, 14 Feb 2010 09:34:27 +0000 (09:34 +0000)]
Merge r200530 from head:

  Don't panic on failure to attach if we fail before or during the
  if_alloc() of ifp.

  This fixes the panic reported in the PRs, but not the attach failure.

PR: kern/139079, kern/143874
Tested by: Steven Noonan <steven uplinklabs.net>
Reviewed by: thompsa

14 years agoMFC: r203303
rmacklem [Sun, 14 Feb 2010 00:43:42 +0000 (00:43 +0000)]
MFC: r203303
Patch the experimental NFS client so that there is a timeout
for negative name cache entries in a manner analogous to
r202767 for the regular NFS client. Also, make the code in
nfs_lookup() compatible with that of the regular client
and replace the sysctl variable that enabled negative name
caching with the mount point option.

14 years agoMerge r201920 to stable/8:
antoine [Sat, 13 Feb 2010 17:48:52 +0000 (17:48 +0000)]
Merge r201920 to stable/8:
 libusb20 was renamed libusb several months ago.

14 years agoMerge r201919 to stable/8:
antoine [Sat, 13 Feb 2010 17:41:22 +0000 (17:41 +0000)]
Merge r201919 to stable/8:
 Fix a typo.

14 years agoMerge r201918 to stable/8:
antoine [Sat, 13 Feb 2010 17:35:29 +0000 (17:35 +0000)]
Merge r201918 to stable/8:
  Fix a typo.

14 years agoMFC r203387: Exclude the interfaces which IPv6 and/or accepting
ume [Sat, 13 Feb 2010 16:28:25 +0000 (16:28 +0000)]
MFC r203387: Exclude the interfaces which IPv6 and/or accepting
RA is disabled from the auto probed interface list.

14 years agoMFC r203378: Make -a option actually work.
ume [Sat, 13 Feb 2010 16:25:33 +0000 (16:25 +0000)]
MFC r203378: Make -a option actually work.

14 years agoMerge r203393,r203395 from head:
gavin [Sat, 13 Feb 2010 10:26:00 +0000 (10:26 +0000)]
Merge r203393,r203395 from head:

  The multiplicand a = 0x5deece66d = 25214903917, not 0xfdeece66d.
  This bug in the man page has gone unnoticed for over 15 years!

PR: docs/143461
Submitted by: Jeremy Huddleston  jeremyhu apple.com

14 years agoMFC r202130:
delphij [Sat, 13 Feb 2010 00:39:46 +0000 (00:39 +0000)]
MFC r202130:

Set svn:executable to *.

14 years agoMigrate mergeinfo which was done on wrong target back to etc/ (203163).
delphij [Sat, 13 Feb 2010 00:39:03 +0000 (00:39 +0000)]
Migrate mergeinfo which was done on wrong target back to etc/ (203163).

14 years agoMFC r203053:
delphij [Sat, 13 Feb 2010 00:30:50 +0000 (00:30 +0000)]
MFC r203053:

Add a manual page for nvram(4).

14 years agoMFC r203534: Correct two typos.
delphij [Sat, 13 Feb 2010 00:29:01 +0000 (00:29 +0000)]
MFC r203534: Correct two typos.

Reported by:    Brandon Falk <falkman gamozo org>

14 years agoPick up some changes the the MFC missed.
mjacob [Fri, 12 Feb 2010 00:47:23 +0000 (00:47 +0000)]
Pick up some changes the the MFC missed.

14 years agoMFC: r203119
rmacklem [Thu, 11 Feb 2010 21:25:48 +0000 (21:25 +0000)]
MFC: r203119
Patch the experimental NFS client in a manner analogous to
r203072 for the regular NFS client. Also, delete two fields
of struct nfsmount that are not used by the FreeBSD port of
the client.

14 years agoMFC a number of changes from head for ISP (203478,203463,203444,202418,201758,
mjacob [Thu, 11 Feb 2010 18:34:06 +0000 (18:34 +0000)]
MFC a number of changes from head for ISP (203478,203463,203444,202418,201758,
201408,201325,200089,198822,197373,197372,197214,196162). Since one of those
changes was a semicolon cleanup from somebody else, this touches a lot more.

14 years agoMFC of r201700 | mckusick | 2010-01-06
mckusick [Wed, 10 Feb 2010 20:35:20 +0000 (20:35 +0000)]
MFC of r201700 | mckusick | 2010-01-06

This corrects a bug that manifested itself as identifying the last
cylinder group of a UFS1 filesystem as bad. The error was in the check
and not in the cylinder group itself. So even though fsck fixed the
cylinder group correctly, it was still endlessly reported as bad.

This bug first appeared in 8.0 so does not apply to earlier releases.

PR:             141992
Reported by:    Dan Strick

14 years agoMFC: r203072
rmacklem [Wed, 10 Feb 2010 16:16:50 +0000 (16:16 +0000)]
MFC: r203072
Fix a race that can occur when nfs nfsiod threads are being created.
Without this patch it was possible for a different thread that calls
nfs_asyncio() to snitch a newly created nfsiod thread that was
intended for another caller of nfs_asyncio(), because the nfs_iod_mtx
mutex was unlocked while the new nfsiod thread was created. This patch
labels the newly created nfsiod, so that it is not taken by another
caller of nfs_asyncio(). This is believed to fix the problem reported
on the freebsd-stable email list under the subject:
FreeBSD NFS client/Linux NFS server issue.

Tested by: to DOT my DOT trociny AT gmail DOT com
Reviewed by: jhb

14 years agoMFC r203062: acpi_hpet: correctly get number of timers in a timer block
avg [Wed, 10 Feb 2010 09:46:31 +0000 (09:46 +0000)]
MFC r203062: acpi_hpet: correctly get number of timers in a timer block

14 years agoMFC r203483:
zec [Wed, 10 Feb 2010 08:50:06 +0000 (08:50 +0000)]
MFC r203483:
  Instead of spamming the console on each curvnet recursion event, print
  out each such call graph only once, along with a stack backtrace.  This
  should make kernels built with VNET_DEBUG reasonably usable again in
  busy / production environments.

  Introduce a new DDB command "show vnetrcrs" which dumps the whole log
  of distinctive curvnet recursion events.  This might be useful when
  recursion reports get burried / lost too deep in the message buffer.
  In the later case stack backtraces are not available.

  Reviewed by:  bz

14 years agoMFC r202060:
delphij [Wed, 10 Feb 2010 00:34:13 +0000 (00:34 +0000)]
MFC r202060:

Add a new option, -q howmany, which when used in conjuction with -w,
exits netstat after _howmany_ outputs.

Requested by: thomasa
Reviewed by: freebsd-net (bms, old version in early 2007)

14 years agoMFC r198593:
emaste [Tue, 9 Feb 2010 22:05:30 +0000 (22:05 +0000)]
MFC r198593:

  Rename aac_fast_intr to aac_filter to reflect its current use.  Eliminate
  the fallback of using the filter as an interrupt handler, as it is no
  longer needed.

  Discussed with:       scottl, jhb

14 years agoMFC r203401
qingli [Tue, 9 Feb 2010 19:27:54 +0000 (19:27 +0000)]
MFC r203401

Some of the existing ppp and vpn related scripts create and set
the IP addresses of the tunnel end points to the same value. In
these cases the loopback route is not installed for the local
end.

14 years agoMFC Revision: 203574
bms [Tue, 9 Feb 2010 18:43:50 +0000 (18:43 +0000)]
MFC Revision: 203574
  Add sane-port (Scanner Access Now Easy) as port 6566.

Obtained from:  http://www.iana.org/assignments/port-numbers

14 years agoMFC r203077:
emaste [Tue, 9 Feb 2010 13:07:32 +0000 (13:07 +0000)]
MFC r203077:

  Add missing return, in a rare case where we can't allocate memory in
  deallocate.

Submitted by: Ryan Stone (rysto32 at gmail dot com)
Approved by:  jasone

14 years agoMFC: r198231: Properly re-create "-s size" argument to newfs(8).
ru [Tue, 9 Feb 2010 12:20:48 +0000 (12:20 +0000)]
MFC: r198231: Properly re-create "-s size" argument to newfs(8).

14 years agoMFC r200781:
jh [Mon, 8 Feb 2010 15:55:01 +0000 (15:55 +0000)]
MFC r200781:

Cast time_t values to intmax_t and use %jd with printf.

14 years agoMFC r200780:
jh [Mon, 8 Feb 2010 15:53:28 +0000 (15:53 +0000)]
MFC r200780:

Remove non-working special case for pipe(2) from amd64-fbsd32.c and
i386-fbsd.c. Add pipe(2) to syscall table to decode it's pointer
argument properly and re-add special handling for pipe(2) return value
to print_syscall_ret().

PR: bin/120870

14 years agoMFC r200752:
jh [Mon, 8 Feb 2010 15:50:51 +0000 (15:50 +0000)]
MFC r200752:

Avoid sharing the file descriptor of the output file with traced
processes by setting the FD_CLOEXEC flag for the output file.

PR: bin/140493

14 years agoMFC r200751:
jh [Mon, 8 Feb 2010 15:48:18 +0000 (15:48 +0000)]
MFC r200751:

Add fork(2), getegid(2), geteuid(2), getgid(2), getpid(2), getpgid(2),
getpgrp(2), getppid(2), getsid(2) and getuid(2) to syscall table to
decode their arguments correctly.

14 years agoMC r202889, r202940:
attilio [Mon, 8 Feb 2010 14:08:52 +0000 (14:08 +0000)]
MC r202889, r202940:
- Fix a race in sched_switch() of sched_4bsd.
  Block the td_lock when acquiring explicitly sched_lock in order to prevent
  races with other td_lock contenders.
- Merge the ULE's internal function thread_block_switch() into the global
  thread_lock_block() and make the former semantic as the default for
  thread_lock_block().
- Split out an invariant in order to have better checks.

14 years agoMFC r203414:
davidxu [Mon, 8 Feb 2010 03:11:55 +0000 (03:11 +0000)]
MFC r203414:
After busied the lock, re-read state word before checking waiters flag,
otherwise, the waiters bit may not be set and a wakeup is lost.

14 years agoMerge r202115 from head:
gavin [Sun, 7 Feb 2010 21:04:26 +0000 (21:04 +0000)]
Merge r202115 from head:

  Our standard "xterm" termcap entry supports colour, so this tip is redundant.
  As it happens, "xterm-color" has just been an alias for "xterm" since
  src/share/termcap/termcap.src 1.131 in September 2002.

PR: docs/132959

14 years agoMFC 202961:
dougb [Sun, 7 Feb 2010 20:28:24 +0000 (20:28 +0000)]
MFC 202961:

Upgrade to BIND 9.6.1-P3.

This version address the following vulnerabilities:

BIND 9 Cache Update from Additional Section
https://www.isc.org/advisories/CVE-2009-4022v6
http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2009-4022
A nameserver with DNSSEC validation enabled may incorrectly add
unauthenticated records to its cache that are received during the
resolution of a recursive client query

BIND 9 DNSSEC validation code could cause bogus NXDOMAIN responses
https://www.isc.org/advisories/CVE-2010-0097
http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2010-0097
There was an error in the DNSSEC NSEC/NSEC3 validation code that could
cause bogus NXDOMAIN responses (that is, NXDOMAIN responses for records
proven by NSEC or NSEC3 to exist) to be cached as if they had validated
correctly

These issues only affect systems with DNSSEC validation enabled.

14 years agoMFC 202960:
dougb [Sun, 7 Feb 2010 20:26:45 +0000 (20:26 +0000)]
MFC 202960:

Copyright-only changes to generated files as part of the 9.6.1-P3 update

14 years agoMFC: r202900
marius [Sun, 7 Feb 2010 11:59:55 +0000 (11:59 +0000)]
MFC: r202900

Merge r203608 from amd64/i386:

In syscall(), reread syscall number and arguments after ptracestop(),
if debugger modified anything in the process environment.

14 years agoMFC r202882:
kib [Sun, 7 Feb 2010 11:37:38 +0000 (11:37 +0000)]
MFC r202882:
For i386, amd64 and ia32 on amd64 MD syscall(), reread syscall number
and arguments after ptracestop(), if debugger modified anything in the
process environment.

14 years agoMFC r202529:
kib [Sun, 7 Feb 2010 10:51:17 +0000 (10:51 +0000)]
MFC r202529:
vunref() the vnode in vm object deallocation code for OBJT_VNODE
appropriate number of times to prevent possible vnode reference leak.

14 years agoMFC r202528:
kib [Sun, 7 Feb 2010 10:44:44 +0000 (10:44 +0000)]
MFC r202528:
Add vunref(9).

14 years agoMFC of 197952 and 198075
julian [Sun, 7 Feb 2010 09:00:22 +0000 (09:00 +0000)]
MFC of 197952 and 198075

    Virtualize the pfil hooks so that different jails may chose different
    packet filters. ALso allows ipfw to be enabled on on ejail and disabled
    on another. In 8.0 it's a global setting.
and
    Unbreak the VIMAGE build with IPSEC, broken with r197952 by
    virtualizing the pfil hooks.
    For consistency add the V_ to virtualize the pfil hooks in here as well.

14 years agoMFC: r203486: Shortening a passphrase caused wrong authentication key
ru [Sun, 7 Feb 2010 06:22:28 +0000 (06:22 +0000)]
MFC: r203486: Shortening a passphrase caused wrong authentication key
to be used.

14 years agoMFC r202916: Make strsignal(3) thread-safe.
ume [Sun, 7 Feb 2010 04:27:18 +0000 (04:27 +0000)]
MFC r202916: Make strsignal(3) thread-safe.