]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
11 years agoAdd the ability to have a 'fallback' search for memory ranges.
Peter Grehan [Fri, 22 Feb 2013 00:46:32 +0000 (00:46 +0000)]
Add the ability to have a 'fallback' search for memory ranges.
These set of ranges will be looked at if a standard memory
range isn't found, and won't be installed in the cache.
Use this to implement the memory behaviour of the PCI hole on
x86 systems, where writes are ignored and reads always return -1.
This allows breakpoints to be set when issuing a 'boot -d', which
has the side effect of accessing the PCI hole when changing the
PTE protection on kernel code, since the pmap layer hasn't been
initialized (a bug, but present in existing FreeBSD releases so
has to be handled).

Reviewed by: neel
Obtained from: NetApp

11 years agoDocument VFCF_SBDRY.
Sergey Kandaurov [Thu, 21 Feb 2013 22:21:45 +0000 (22:21 +0000)]
Document VFCF_SBDRY.

Reviewed by: jhb
X-MFC with: r247116

11 years agoDisable debugging entries about BAW issues. I haven't seen any issues
Adrian Chadd [Thu, 21 Feb 2013 21:47:35 +0000 (21:47 +0000)]
Disable debugging entries about BAW issues.  I haven't seen any issues
to do with BAW tracking in the last 9 months or so.

11 years agoReflect CARP media-type departure.
Sergey Kandaurov [Thu, 21 Feb 2013 21:35:09 +0000 (21:35 +0000)]
Reflect CARP media-type departure.

X-MFC after: never

11 years agoImprove mxge's receive performance for IPv6:
Andrew Gallatin [Thu, 21 Feb 2013 21:28:33 +0000 (21:28 +0000)]
Improve mxge's receive performance for IPv6:

- Add support for IPv6 rx csum offload
- Finally switch mxge from using its own driver lro, to
using tcp_lro

MFC after: 7 days
Sponsored by: Myricom Inc.

11 years agoSeparate items in the list of System values with .Pp to ease readability.
Sergey Kandaurov [Thu, 21 Feb 2013 20:45:19 +0000 (20:45 +0000)]
Separate items in the list of System values with .Pp to ease readability.

11 years agocxgbe(4): Add sysctls to extract debug information from the chip:
Navdeep Parhar [Thu, 21 Feb 2013 20:13:15 +0000 (20:13 +0000)]
cxgbe(4): Add sysctls to extract debug information from the chip:

dev.t4nex.X.misc.cim_la         logic analyzer dump
dev.t4nex.X.misc.cim_qcfg       queue configuration
dev.t4nex.X.misc.cim_ibq_xxx    inbound queues
dev.t4nex.X.misc.cim_obq_xxx    outbound queues

Obtained from: Chelsio
MFC after: 1 week

11 years agoreorder so all the flags are together and make the PCLMUL flag unique..
John-Mark Gurney [Thu, 21 Feb 2013 19:13:19 +0000 (19:13 +0000)]
reorder so all the flags are together and make the PCLMUL flag unique..

This fixes the problem on amd64 miscompiling mpboot.s causing boot
issues...  We are still using gas for a few files in the kernel...

Submitted by: kib
MFC after: 1 month

11 years agoFurther refine the handling of stop signals in the NFS client. The
John Baldwin [Thu, 21 Feb 2013 19:02:50 +0000 (19:02 +0000)]
Further refine the handling of stop signals in the NFS client.  The
changes in r246417 were incomplete as they did not add explicit calls to
sigdeferstop() around all the places that previously passed SBDRY to
_sleep().  In addition, nfs_getcacheblk() could trigger a write RPC from
getblk() resulting in sigdeferstop() recursing.  Rather than manually
deferring stop signals in specific places, change the VFS_*() and VOP_*()
methods to defer stop signals for filesystems which request this behavior
via a new VFCF_SBDRY flag.  Note that this has to be a VFC flag rather than
a MNTK flag so that it works properly with VFS_MOUNT() when the mount is
not yet fully constructed.  For now, only the NFS clients are set this new
flag in VFS_SET().

A few other related changes:
- Add an assertion to ensure that TDF_SBDRY doesn't leak to userland.
- When a lookup request uses VOP_READLINK() to follow a symlink, mark
  the request as being on behalf of the thread performing the lookup
  (cnp_thread) rather than using a NULL thread pointer.  This causes
  NFS to properly handle signals during this VOP on an interruptible
  mount.

PR: kern/176179
Reported by: Russell Cattelan (sigdeferstop() recursion)
Reviewed by: kib
MFC after: 1 month

11 years agosplsoftvm() is no longer in the tree. gc.
Warner Losh [Thu, 21 Feb 2013 17:54:14 +0000 (17:54 +0000)]
splsoftvm() is no longer in the tree. gc.

11 years agobus_dmamap_load_ccb doesn't exist on earlier versions of FreeBSD or
Warner Losh [Thu, 21 Feb 2013 17:50:21 +0000 (17:50 +0000)]
bus_dmamap_load_ccb doesn't exist on earlier versions of FreeBSD or
DragonFlyBSD, so it certainly doesn't need splsoftvm(). Remove it.

# I doubt this driver will now compile on older FreeBSD versions or DFBSD
# We should consider unifdefing it since that code seems unmaintained.

11 years agoRemove splsoftclock() since it is now gone.
Warner Losh [Thu, 21 Feb 2013 17:40:59 +0000 (17:40 +0000)]
Remove splsoftclock() since it is now gone.

11 years agoFix tcp_lro_rx_ipv4() for drivers that do not set CSUM_IP_CHECKED.
Andrew Gallatin [Thu, 21 Feb 2013 17:00:35 +0000 (17:00 +0000)]
Fix tcp_lro_rx_ipv4() for drivers that do not set CSUM_IP_CHECKED.
Specifcially, in_cksum_hdr() returns 0 (not 0xffff) when the IPv4
checksum is correct. Without this fix, the tcp_lro code will reject
good IPv4 traffic from drivers that do not implement IPv4 header
harder csum offload.

Sponsored by: Myricom Inc.

MFC after: 7 days

11 years agoIn the EXAMPLES section: append the subsystem regular expression with a
Sergey Kandaurov [Thu, 21 Feb 2013 15:41:09 +0000 (15:41 +0000)]
In the EXAMPLES section: append the subsystem regular expression with a
typical interface name pattern as devd(8) uses full string match.

Discussed with: avg, glebius

11 years agoUse critical_enter/critical_exit around the time sensitive part of
Warner Losh [Thu, 21 Feb 2013 15:35:48 +0000 (15:35 +0000)]
Use critical_enter/critical_exit around the time sensitive part of
this code to depessimize the worst case we've lived with silently and
uneventfully for the past 12 years. Add a comment about a refinement
for those needing more assurance of accuracy.

Fix ddb's show rtc command deadlock potential when debugging rtc code
by not taking the lock if we're in the debugger. If you need a thumb
to count the number of people that have encountered this, I'd be
surprised.

Submitted by: bde

11 years agoFix build.
Gleb Smirnoff [Thu, 21 Feb 2013 12:52:18 +0000 (12:52 +0000)]
Fix build.

11 years agoFix build.
Gleb Smirnoff [Thu, 21 Feb 2013 12:40:52 +0000 (12:40 +0000)]
Fix build.

11 years agoBe slightly more paranoid with the TX DMA buffer maximum threshold.
Adrian Chadd [Thu, 21 Feb 2013 08:42:40 +0000 (08:42 +0000)]
Be slightly more paranoid with the TX DMA buffer maximum threshold.

Specifically - never jack the TX FIFO threshold up to the absolute
maximum; always leave enough space for two DMA transactions to
appear.

This is a paranoia from the Linux ath9k driver.  It can't hurt.

Obtained from: Linux ath9k

11 years agoRemove incorrect comment about splsoftclock.
Warner Losh [Thu, 21 Feb 2013 07:19:50 +0000 (07:19 +0000)]
Remove incorrect comment about splsoftclock.

11 years agoLocking for todr got pushed down into inittodr and the client
Warner Losh [Thu, 21 Feb 2013 07:16:40 +0000 (07:16 +0000)]
Locking for todr got pushed down into inittodr and the client
libraries it calls (although some might not be doing it right). We are
serialized right now by giant as well. This means the splsoftclock are
now an anachronism that has no benefit, even marking where locking
needs to happen. Remove them.

11 years agoAdd an option to allow the minimum number of delimiters to be tweaked.
Adrian Chadd [Thu, 21 Feb 2013 06:38:49 +0000 (06:38 +0000)]
Add an option to allow the minimum number of delimiters to be tweaked.

This is primarily for debugging purposes.

Tested:

* AR5416, STA mode

11 years agoCorrect comment about use of pmtimer, and the real reason it isn't
Warner Losh [Thu, 21 Feb 2013 06:38:24 +0000 (06:38 +0000)]
Correct comment about use of pmtimer, and the real reason it isn't
used or desirable for amd64.

11 years agoAdd a new option to limit the maximum size of aggregates.
Adrian Chadd [Thu, 21 Feb 2013 06:18:40 +0000 (06:18 +0000)]
Add a new option to limit the maximum size of aggregates.
The default is to limit them to what the hardware is capable of.

Add sysctl twiddles for both the non-RTS and RTS protected aggregate
generation.

Whilst here, add some comments about stuff that I've discovered during
my exploration of the TX aggregate / delimiter setup path from the
reference driver.

11 years agoRemove this unneeded printf(), sorry!
Adrian Chadd [Thu, 21 Feb 2013 02:52:13 +0000 (02:52 +0000)]
Remove this unneeded printf(), sorry!

11 years agoThe request queue is already locked, so we don't need the splsofclock/splx
Warner Losh [Thu, 21 Feb 2013 02:43:44 +0000 (02:43 +0000)]
The request queue is already locked, so we don't need the splsofclock/splx
here to note future work.

11 years agoDon't worry if a module is already loaded when looking for a fstype to mount
Jamie Gritton [Thu, 21 Feb 2013 02:41:37 +0000 (02:41 +0000)]
Don't worry if a module is already loaded when looking for a fstype to mount
(possible in a race condition).

Reviewed by: kib
MFC after: 1 week

11 years agoMost other giant locked storage drivers in the tree don't use
Warner Losh [Thu, 21 Feb 2013 02:40:20 +0000 (02:40 +0000)]
Most other giant locked storage drivers in the tree don't use
splsoftclock to note the need for future locking, so remove it from
here.

11 years agoThe other giant locked storage drivers have removed splbio(), for the
Warner Losh [Thu, 21 Feb 2013 02:34:04 +0000 (02:34 +0000)]
The other giant locked storage drivers have removed splbio(), for the
most part, so remove it here too. Anybody locking this driver will need
far more than locks where splbio() were, so remove these nops.

11 years agoFix broken usage of splhigh() by removing it.
Warner Losh [Thu, 21 Feb 2013 00:40:08 +0000 (00:40 +0000)]
Fix broken usage of splhigh() by removing it.

11 years agoKill now-bogus splhigh() -- it is a nop
Warner Losh [Thu, 21 Feb 2013 00:36:12 +0000 (00:36 +0000)]
Kill now-bogus splhigh() -- it is a nop

11 years agoReplace splhigh() with critical_enter()/leave() to ensure we write the
Warner Losh [Thu, 21 Feb 2013 00:27:51 +0000 (00:27 +0000)]
Replace splhigh() with critical_enter()/leave() to ensure we write the
config mode unlock sequence quickly enough. This likely isn't too critical,
since splhigh() has been a noop for a decade...

11 years agoNo longer need splhigh() since locking was done, delete it and
Warner Losh [Thu, 21 Feb 2013 00:26:31 +0000 (00:26 +0000)]
No longer need splhigh() since locking was done, delete it and
comments about it.

11 years agoRefresh on the shared code for the E1000 drivers.
Jack F Vogel [Thu, 21 Feb 2013 00:25:45 +0000 (00:25 +0000)]
Refresh on the shared code for the E1000 drivers.
  - bear with me, there are lots of white space changes, I would not
    do them, but I am a mere consumer of this stuff and if these drivers
    are to stay in shape they need to be taken.

em driver changes: support for the new i217/i218 interfaces

igb driver changes:
  - TX mq start has a quick turnaround to the stack
  - Link/media handling improvement
  - When link status changes happen the current flow control state
    will now be displayed.
  - A few white space/style changes.

lem driver changes:
  - the shared code uncovered a bogus write to the RLPML register
    (which does not exist in this hardware) in the vlan code,this
    is removed.

11 years agocxgbe(4): Assume that CSUM_TSO in the transmit path implies CSUM_IP and
Navdeep Parhar [Wed, 20 Feb 2013 23:15:40 +0000 (23:15 +0000)]
cxgbe(4): Assume that CSUM_TSO in the transmit path implies CSUM_IP and
CSUM_TCP too.  They are all set explicitly by the kernel usually.

While here, fix an unrelated bug where hardware L4 checksum calculation
was accidentally disabled for some IPv6 packets.

Reported by: alfred@
MFC after: 3 days

11 years agoWhen porting XTS-related code from OpenBSD I forgot to update copyright (only
Pawel Jakub Dawidek [Wed, 20 Feb 2013 22:59:53 +0000 (22:59 +0000)]
When porting XTS-related code from OpenBSD I forgot to update copyright (only
OpenBSD was credited in one of two commits). Fix it.

Reported by: Theo de Raadt <deraadt@cvs.openbsd.org>
Reviewed by: Damien Miller <djm@mindrot.org>

11 years agoDo not use deprecated functions from libarchive
Baptiste Daroussin [Wed, 20 Feb 2013 22:51:42 +0000 (22:51 +0000)]
Do not use deprecated functions from libarchive

11 years agoRemove the unused spl functions: spl0, splsoftcam, splsofttty,
Warner Losh [Wed, 20 Feb 2013 22:20:49 +0000 (22:20 +0000)]
Remove the unused spl functions: spl0, splsoftcam, splsofttty,
splsofttq and splstatclock.

Other used spl functions to follow.

11 years agorevert 247035
Dag-Erling Smørgrav [Wed, 20 Feb 2013 21:16:50 +0000 (21:16 +0000)]
revert 247035

11 years agoFix parallel build race with DEBUG_FLAGS
Ed Maste [Wed, 20 Feb 2013 20:42:56 +0000 (20:42 +0000)]
Fix parallel build race with DEBUG_FLAGS

Reported by: Jan Beich on freebsd-current
Tested by: markj

11 years agoSupport hardlinks in manifest files by the same logic as the treewalk
Brooks Davis [Wed, 20 Feb 2013 19:32:31 +0000 (19:32 +0000)]
Support hardlinks in manifest files by the same logic as the treewalk
code.

Reviewed by: marcel
Sponsored by: DARPA, AFRL

11 years agoSort sections.
Joel Dahl [Wed, 20 Feb 2013 19:05:13 +0000 (19:05 +0000)]
Sort sections.

11 years agoVarious improvements to the qsort(3) usage example:
Giorgos Keramidas [Wed, 20 Feb 2013 18:31:55 +0000 (18:31 +0000)]
Various improvements to the qsort(3) usage example:

- Remove unused #include.
- Do not cast away const.
- Use the canonical idiom to compare two numbers.
- Use proper type for sizes, i.e. size_t instead of int.
- Correct indentation.
- Simplify printf("\n") to puts("").
- Use return instead of exit() in main().

Submitted by: Christoph Mallon, christoph.mallon at gmx.de
Approved by: gjb (mentor)
Reviewed by: stefanf
MFC after: 1 week

11 years agoAvoid panic when tearing down the DTrace pid provider for a
Justin T. Gibbs [Wed, 20 Feb 2013 17:55:17 +0000 (17:55 +0000)]
Avoid panic when tearing down the DTrace pid provider for a
process that has crashed.

sys/cddl/contrib/opensolaris/uts/common/dtrace/fasttrap.c:
In fasttrap_pid_disable(), we cannot PHOLD the proc
structure for a process that no longer exists, but
we still have other, fasttrap specific, state that
must be cleaned up for probes that existed in the
dead process.  Instead of returning early if the
process related to our probes isn't found,
conditionalize the locking and carry on with a NULL
proc pointer.  The rest of the fasttrap code already
understands that a NULL proc is possible and does
the right things in this case.

Sponsored by: Spectra Logic Corporation
Reviewed by: rpaulo, gnn
MFC after: 1 week

11 years agoOrphaned processes that are being traced are killed by the
Justin T. Gibbs [Wed, 20 Feb 2013 17:46:38 +0000 (17:46 +0000)]
Orphaned processes that are being traced are killed by the
kernel.  Properly restore, continue, and detach from processes
being DTraced when DTrace exits with an error so the program
being inspected is not terminated.

cddl/contrib/opensolaris/cmd/dtrace/dtrace.c:
In fatal(), the generic error handler, close the DTrace
handle as is done in the "probe/script" error handler
dfatal().  fatal() can be invoked after DTrace attaches
to processes (e.g. a script specified by command line
argument can't be found) and closing the handle will
release them.

Submitted by: Spectra Logic Corporation
Reviewed by: rpaulo, gnn

11 years agoConvert machine/elf.h, machine/frame.h, machine/sigframe.h,
Konstantin Belousov [Wed, 20 Feb 2013 17:39:52 +0000 (17:39 +0000)]
Convert machine/elf.h, machine/frame.h, machine/sigframe.h,
machine/signal.h and machine/ucontext.h into common x86 includes,
copying from amd64 and merging with i386.

Kernel-only compat definitions are kept in the i386/include/sigframe.h
and i386/include/signal.h, to reduce amd64 kernel namespace pollution.
The amd64 compat uses its own definitions so far.

The _MACHINE_ELF_WANT_32BIT definition is to allow the
sys/boot/userboot/userboot/elf32_freebsd.c to use i386 ELF definitions
on the amd64 compile host.  The same hack could be usefully abused by
other code too.

11 years agoInitialize vm_max_kernel_address on non-FDT platforms. (This should have
Alan Cox [Wed, 20 Feb 2013 16:48:52 +0000 (16:48 +0000)]
Initialize vm_max_kernel_address on non-FDT platforms.  (This should have
been included in r246926.)

The second parameter to pmap_bootstrap() is redundant.  Eliminate it.

Reviewed by: andrew

11 years agoFix "fdt addr" to accept literal addresses rather than va offsets.
Tim Kientzle [Wed, 20 Feb 2013 16:32:38 +0000 (16:32 +0000)]
Fix "fdt addr" to accept literal addresses rather than va offsets.
When initializing the fdt, query U-Boot as well.

With this change, it is now feasible to have U-Boot load
the FDT, ubldr will pull it from U-Boot and hand it to the
kernel.

11 years agoip_savecontrol() style fixes. No functional changes.
Sergey Kandaurov [Wed, 20 Feb 2013 15:44:40 +0000 (15:44 +0000)]
ip_savecontrol() style fixes. No functional changes.
- fix indentation
- put the operator at the end of the line for long statements
- remove spaces between the type and the variable in a cast
- remove excessive parentheses

Tested by: md5

11 years agoAllow '.' components in manifest paths. They are always the first
Brooks Davis [Wed, 20 Feb 2013 15:28:40 +0000 (15:28 +0000)]
Allow '.' components in manifest paths.  They are always the first
component of mtree -C and install -M output and are easily skipped.

Reviewed by: marcel
Sponsored by: DARPA, AFRL

11 years agoFix the -N option in manifest mode by using pwcache(3). This also
Brooks Davis [Wed, 20 Feb 2013 15:25:40 +0000 (15:25 +0000)]
Fix the -N option in manifest mode by using pwcache(3).  This also
speeds up image creation appreciably.

Reviewed by: marcel
Sponsored by: DARPA, AFRL

11 years agoAdd a -D flag that causes duplicate entries in an mtree manifest to be
Brooks Davis [Wed, 20 Feb 2013 15:18:42 +0000 (15:18 +0000)]
Add a -D flag that causes duplicate entries in an mtree manifest to be
treated as warnings rather than errors.

Reviewed by: marcel
Sponsored by: DARPA, AFRL

11 years agoFix several new & old style issues.
Alexander V. Chernikov [Wed, 20 Feb 2013 14:19:09 +0000 (14:19 +0000)]
Fix several new & old style issues.

Pointed by: ae, pluknet, zont
MFC with: r247036

11 years agoAdd interface name filtering via 'match' cmd.
Alexander V. Chernikov [Wed, 20 Feb 2013 13:47:05 +0000 (13:47 +0000)]
Add interface name filtering via 'match' cmd.
Add 'pps' cmd for switching beetween interface packets/bytes statistics.

Submitted by: vsevolod
MFC after: 2 weeks

11 years agoReduce excessive nesting.
Dag-Erling Smørgrav [Wed, 20 Feb 2013 12:59:21 +0000 (12:59 +0000)]
Reduce excessive nesting.

11 years agoCheck if the -sec option is given without an argument.
Sergey Kandaurov [Wed, 20 Feb 2013 12:40:26 +0000 (12:40 +0000)]
Check if the -sec option is given without an argument.

PR: bin/170413
Submitted by: Andrey Simonenko <simon@comsys.ntu-kpi.kiev.ua>
MFC after: 1 week

11 years agoConfigure larger TX FIFO default and maximum level values.
Adrian Chadd [Wed, 20 Feb 2013 12:14:49 +0000 (12:14 +0000)]
Configure larger TX FIFO default and maximum level values.

This has reduced the number of TX delimiter and data underruns when
doing large UDP transfers (>100mbit).

This stops any HAL_INT_TXURN interrupts from occuring, which is a good
sign!

Obtained from: Qualcomm Atheros

11 years agoIf any of the TX queues have underrun reporting enabled, enable
Adrian Chadd [Wed, 20 Feb 2013 11:24:11 +0000 (11:24 +0000)]
If any of the TX queues have underrun reporting enabled, enable
HAL_INT_TXURN in the interrupt mask register.

This should now allow for TXURN interrupts to be posted.

11 years agoA couple of quick tidyups:
Adrian Chadd [Wed, 20 Feb 2013 11:22:44 +0000 (11:22 +0000)]
A couple of quick tidyups:

* Delete this debugging print - I used it when debugging the initial
  TX descriptor chaining code.  It now works, so let's toss it.
  It just confuses people if they enable TX descriptor debugging as they
  get two slightly different versions of the same descriptor.

* Indenting.

11 years agoEnable TX FIFO underrun interrupts. This allows the TX FIFO threshold
Adrian Chadd [Wed, 20 Feb 2013 11:20:51 +0000 (11:20 +0000)]
Enable TX FIFO underrun interrupts.  This allows the TX FIFO threshold
adjustment code to now run.

Tested:

* AR5416, STA

TODO:

* Much more thorough testing on the other chips, AR5210 -> AR9287

11 years agooops, tab!
Adrian Chadd [Wed, 20 Feb 2013 11:17:29 +0000 (11:17 +0000)]
oops, tab!

11 years agoPost interrupts in the ath alq trace.
Adrian Chadd [Wed, 20 Feb 2013 11:17:03 +0000 (11:17 +0000)]
Post interrupts in the ath alq trace.

11 years agoCFG_ERR, DATA_UNDERRUN and DELIM_UNDERRUN are all flags, rather than
Adrian Chadd [Wed, 20 Feb 2013 11:14:55 +0000 (11:14 +0000)]
CFG_ERR, DATA_UNDERRUN and DELIM_UNDERRUN are all flags, rather than
part of ts_status. Thus:

* make sure we decode them from ts_flags, rather than ts_status;
* make sure we decode them regardless of whether there's an error or not.

This correctly exposes descriptor configuration errors, TX delimiter
underruns and TX data underruns.

11 years agoAdd a sample program that shows how a custom comparison function and
Giorgos Keramidas [Tue, 19 Feb 2013 23:57:39 +0000 (23:57 +0000)]
Add a sample program that shows how a custom comparison function and
qsort(3) can work together to sort an array of integers.

PR:             docs/176197
Submitted by:   Fernando, fapesteguia at opensistemas.com
Approved by:    gjb (mentor)
MFC after:      1 week

11 years agosh: Fix a crash with the stackmark code.
Jilles Tjoelker [Tue, 19 Feb 2013 23:46:51 +0000 (23:46 +0000)]
sh: Fix a crash with the stackmark code.

If a stack mark is set while the current stack block is empty, the stack
block may move later on (because of realloc()) and the stack mark needs to
be updated. This updating does not happen after popstackmark() has been
called; therefore, call setstackmark() again if the stack mark is still
being used.

For some reason, this only affects a few users. I cannot reproduce it. The
situation seems quite rare as well because an empty stack block would
usually be freed (by popstackmark()) before execution reaches a
setstackmark() call.

PR: 175922
Tested by: KT Sin

11 years agoadd support for AES and PCLMULQDQ instructions to binutils...
John-Mark Gurney [Tue, 19 Feb 2013 21:35:17 +0000 (21:35 +0000)]
add support for AES and PCLMULQDQ instructions to binutils...

Thanks to Mike Belopuhov for the pointer to the OpenBSD patch, though
OpenBSD's gcc is very different that it only helped w/ where to modify,
not how...  Thanks to jhb for some early reviews...

Reviewed by: imp, kib
MFC after: 1 month

11 years agoAdd support to mxge for IPv6 TX csum offload & IPv6 TSO.
Andrew Gallatin [Tue, 19 Feb 2013 21:33:21 +0000 (21:33 +0000)]
Add support to mxge for IPv6 TX csum offload & IPv6 TSO.

Sponsored by: Myricom, Inc.
MFC after: 7 days

11 years agoSpelling fixes
Oleksandr Tymoshenko [Tue, 19 Feb 2013 21:24:52 +0000 (21:24 +0000)]
Spelling fixes

Spotted by: N. J. Mann

11 years agoRoll back change of frequency for initialization sequence since it
Oleksandr Tymoshenko [Tue, 19 Feb 2013 20:33:55 +0000 (20:33 +0000)]
Roll back change of frequency for initialization sequence since it
seems to cause more problems then previous behavior: it either breaks
initilization sequence in other places or uncovers problems with
high-speed mode timing for SDHCI 3.0

11 years agoMFcalloutng:
Alexander Motin [Tue, 19 Feb 2013 19:25:50 +0000 (19:25 +0000)]
MFcalloutng:
Make led(4) fire callouts at 10Hz only when there is at least one LED that
is requested to blink.  Do not fire if all LEDs are static (usual case).

11 years agodtc: fix bootstrapping from 8.2-STABLE
Ulrich Spörlein [Tue, 19 Feb 2013 18:28:25 +0000 (18:28 +0000)]
dtc: fix bootstrapping from 8.2-STABLE

MFC after: 1 week
Approved by: theraven

11 years agoMFcalloutng:
Alexander Motin [Tue, 19 Feb 2013 18:22:25 +0000 (18:22 +0000)]
MFcalloutng:
Make dcons input polling adaptive, reducing poll rate to 1Hz after several
minutes of inactivty to reduce global interrupt rate.  Most of users never
used FireWire debugging, so it is not very useful to consume power by it.

11 years agoPull in r175360 from upstream llvm trunk:
Dimitry Andric [Tue, 19 Feb 2013 17:53:32 +0000 (17:53 +0000)]
Pull in r175360 from upstream llvm trunk:

  MCParser: Reject .balign with non-pow2 alignments.

  GNU as rejects them and there are configure scripts in the wild that
  check if the assembler rejects ".align 3" to determine whether the
  alignment is in bytes or powers of two.

MFC after: 3 days

11 years agoImport change 6d783560e4aad1e680097d11e89755647a5aba87 from upstream
Dimitry Andric [Tue, 19 Feb 2013 17:38:18 +0000 (17:38 +0000)]
Import change 6d783560e4aad1e680097d11e89755647a5aba87 from upstream
heimdal:

  fix sizeof(uuid)

Found by: clang ToT
Reviewed by: stas

11 years agoFix includes for use in libstand.
Tim Kientzle [Tue, 19 Feb 2013 17:09:23 +0000 (17:09 +0000)]
Fix includes for use in libstand.

11 years agoFixup r246916 in case gcc is used to build.
Davide Italiano [Tue, 19 Feb 2013 16:43:48 +0000 (16:43 +0000)]
Fixup r246916 in case gcc is used to build.

Reported by: attilio, simon

11 years agoFix a few typos.
John Baldwin [Tue, 19 Feb 2013 16:35:27 +0000 (16:35 +0000)]
Fix a few typos.

11 years agoWARNS=6 compliance
Philippe Charnier [Tue, 19 Feb 2013 13:32:26 +0000 (13:32 +0000)]
WARNS=6 compliance

11 years agoWARNS=6 compliance
Philippe Charnier [Tue, 19 Feb 2013 13:17:16 +0000 (13:17 +0000)]
WARNS=6 compliance

11 years agoRemove old-style function definition
Philippe Charnier [Tue, 19 Feb 2013 12:57:07 +0000 (12:57 +0000)]
Remove old-style function definition

11 years agoFix typo
Ulrich Spörlein [Tue, 19 Feb 2013 10:25:51 +0000 (10:25 +0000)]
Fix typo

Submitted by: Matt Burke <mattblists@icritical.com>

11 years agoPrint out the RTS/CTS rate in the TX descriptor.
Adrian Chadd [Tue, 19 Feb 2013 06:42:12 +0000 (06:42 +0000)]
Print out the RTS/CTS rate in the TX descriptor.

11 years ago- Make sure to set an error code when trying to obtain a data descriptor for
Mark Johnston [Tue, 19 Feb 2013 03:23:13 +0000 (03:23 +0000)]
- Make sure to set an error code when trying to obtain a data descriptor for
  a section of type SHT_NULL.
- Update the man page to reflect the fact that elf_getdata() and
  elf_rawdata() may return with an error of ELF_E_SECTION.

PR: bin/175491
Approved by: emaste (co-mentor)
MFC after: 1 week

11 years agokgdb enhancements!
Adrian Chadd [Tue, 19 Feb 2013 02:09:18 +0000 (02:09 +0000)]
kgdb enhancements!

* document the kgdb -b flag
* better verify what's valid with -b
* add more comprehensive command line help

PR: kern/175743
Submitted by: Christoph Mallon <christoph.mallon@gmx.de>

11 years agoDefine gpio constants rather than using enum.
Ganbold Tsagaankhuu [Tue, 19 Feb 2013 02:01:35 +0000 (02:01 +0000)]
Define gpio constants rather than using enum.
Fix pull-up and pull-down values of gpio.
According to A10 user manual possible pull register
values are 00 Pull-up/down disable, 01 Pull-up, 10 Pull-down.

Approved by: gonzo@

11 years agoMore white-space cleanups.
Pawel Jakub Dawidek [Mon, 18 Feb 2013 23:58:05 +0000 (23:58 +0000)]
More white-space cleanups.

Reported by: zont (the first one)

11 years agoAdd mem_load_raw() for loading data from another location in memory.
Tim Kientzle [Mon, 18 Feb 2013 23:13:13 +0000 (23:13 +0000)]
Add mem_load_raw() for loading data from another location in memory.

This will be used by some upcoming changes to loader(8) FDT
handling to allow it to use an FDT provided by an earlier
boot stage the same as an FDT loaded from disk.

11 years agosetbuf(3): Restore a BUGS section about setbuf().
Jilles Tjoelker [Mon, 18 Feb 2013 22:47:59 +0000 (22:47 +0000)]
setbuf(3): Restore a BUGS section about setbuf().

The brokenness of setbuf() is not specific to 4.2BSD and 4.3BSD but inherent
in the API definition.

Reported by: bde

11 years agoMark the coretemp(4) sysctls as MPSAFE, ensuring that Giant won't be held
Mark Johnston [Mon, 18 Feb 2013 22:47:16 +0000 (22:47 +0000)]
Mark the coretemp(4) sysctls as MPSAFE, ensuring that Giant won't be held
unnecessarily by a user thread waiting to run on a specific CPU after
calling sched_bind().

Reviewed by: rstone
Approved by: emaste (co-mentor)
Sponsored by: Sandvine Incorporated
MFC after: 1 week

11 years agomdoc: sort sections.
Joel Dahl [Mon, 18 Feb 2013 20:21:26 +0000 (20:21 +0000)]
mdoc: sort sections.

11 years agoClarify an error message. The addresses here are
Tim Kientzle [Mon, 18 Feb 2013 20:11:20 +0000 (20:11 +0000)]
Clarify an error message.   The addresses here are
really offsets within the primary mapped memory block
(which may not start at address zero).

11 years agoMerge sendmail 8.14.6 errata issue
Gregory Neil Shapiro [Mon, 18 Feb 2013 19:33:07 +0000 (19:33 +0000)]
Merge sendmail 8.14.6 errata issue

MFC after: 3 days

11 years agoFix regression in sendmail 8.14.6 release based on errata published at:
Gregory Neil Shapiro [Mon, 18 Feb 2013 19:23:41 +0000 (19:23 +0000)]
Fix regression in sendmail 8.14.6 release based on errata published at:

http://www.sendmail.com/sm/open_source/download/8.14.6/

11 years agoFix an incorrect sizeof()
Adrian Chadd [Mon, 18 Feb 2013 18:39:15 +0000 (18:39 +0000)]
Fix an incorrect sizeof()

PR: kern/176238
Submitted by: Christoph Mallon <christoph.mallon@gmx.de>

11 years agoFix bad EEPROM parsing code.
Hans Petter Selasky [Mon, 18 Feb 2013 17:55:27 +0000 (17:55 +0000)]
Fix bad EEPROM parsing code.

MFC after: 2 weeks

11 years ago* Add initial debugging of the interrupt status entry
Adrian Chadd [Mon, 18 Feb 2013 04:20:43 +0000 (04:20 +0000)]
* Add initial debugging of the interrupt status entry
* Add in the TDMA debugging entries that I added a while ago

11 years agoMake this output more terse.
Adrian Chadd [Mon, 18 Feb 2013 04:19:21 +0000 (04:19 +0000)]
Make this output more terse.

11 years agoAdd a new ATH KTR debug method to log the interrupt status.
Adrian Chadd [Mon, 18 Feb 2013 04:10:38 +0000 (04:10 +0000)]
Add a new ATH KTR debug method to log the interrupt status.

11 years agothe NO_PAM option has been finally removed in favor of WITHOUT_PAM.
Luigi Rizzo [Mon, 18 Feb 2013 02:08:41 +0000 (02:08 +0000)]
the NO_PAM option has been finally removed in favor of WITHOUT_PAM.

11 years agoAdd strtoul() to libstand by copying from libc and clipping out
Tim Kientzle [Mon, 18 Feb 2013 01:55:53 +0000 (01:55 +0000)]
Add strtoul() to libstand by copying from libc and clipping out
locale code.

11 years agoDisable this variable; the code using it is also disabled.
Adrian Chadd [Mon, 18 Feb 2013 01:37:55 +0000 (01:37 +0000)]
Disable this variable; the code using it is also disabled.