ƒ¹@224771
141
251
135
141
229
150
267
395
1263
234
209
435
183
220
373
107
172
176
165
1100
329
384
1424
547
431
270
123
280
268
357
182
102
298
237
384
130
442
275
275
203
710
1027
904
753
597
474
574
574
351
351
1161
1161
307
293
306
306
221
221
140
140
190
190
223
223
327
1303
1303
281
281
261
263
264
518
273
185
247
179
247
254
261
162
656
268
434
328
886
474
302
433
368
749
274
757
182
195
546
244
1748
208
146
294
171
212
149
2665
424
514
656
141
253
694
276
356
178
120
130
434
182
125
1094
150
311
311
118
152
151
139
124
254
168
209
721
1205
204
213
411
213
204
293
267
294
394

K 10
svn:author
V 6
emaste
K 8
svn:date
V 27
2011-08-11T02:50:18.915511Z
K 7
svn:log
V 155
MFC r224650:
  Don't try to free() an address returned by mmap().  This appears to be
  triggered by the same .o being included twice on the command line.

END
K 10
svn:author
V 2
ru
K 8
svn:date
V 27
2011-08-11T04:58:20.542811Z
K 7
svn:log
V 44
Vendor import of bwk's 10-Aug-2011 release.

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-11T06:45:12.458647Z
K 7
svn:log
V 46
Add the LOR and recursive lock error details.

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-11T08:51:53.011574Z
K 7
svn:log
V 133
Fix the TXQ recursive locking bug that sometimes occurs when an interface
is destroyed.

Only lock the TXQ if it isn't locked by us.

END
K 10
svn:author
V 2
ru
K 8
svn:date
V 27
2011-08-11T10:24:09.125073Z
K 7
svn:log
V 59
Corrected description of the bugfix; distfile was updated.

END
K 10
svn:author
V 2
ru
K 8
svn:date
V 27
2011-08-11T10:29:10.509727Z
K 7
svn:log
V 175
- Merged awk upstream that includes a fix for a bug exposed by kmod_syms.mk.
- Provide a build aid for those who already have a buggy awk(1) installed.

Approved by:	re (kib)

END
K 10
svn:author
V 8
hselasky
K 8
svn:date
V 27
2011-08-11T11:30:21.569866Z
K 7
svn:log
V 297
Use synchronous device destruction instead of asynchronous, so that a new
device having the same name like a previous one is not created before the old
one is gone. This fixes some panics due to asserts in the devfs code which
were added recently.

Approved by:    re (kib)
MFC after:      1 week

END
K 10
svn:author
V 7
rwatson
K 8
svn:date
V 27
2011-08-11T12:30:23.307369Z
K 7
svn:log
V 1165
Second-to-last commit implementing Capsicum capabilities in the FreeBSD
kernel for FreeBSD 9.0:

Add a new capability mask argument to fget(9) and friends, allowing system
call code to declare what capabilities are required when an integer file
descriptor is converted into an in-kernel struct file *.  With options
CAPABILITIES compiled into the kernel, this enforces capability
protection; without, this change is effectively a no-op.

Some cases require special handling, such as mmap(2), which must preserve
information about the maximum rights at the time of mapping in the memory
map so that they can later be enforced in mprotect(2) -- this is done by
narrowing the rights in the existing max_protection field used for similar
purposes with file permissions.

In namei(9), we assert that the code is not reached from within capability
mode, as we're not yet ready to enforce namespace capabilities there.
This will follow in a later commit.

Update two capability names: CAP_EVENT and CAP_KEVENT become
CAP_POST_KEVENT and CAP_POLL_KEVENT to more accurately indicate what they
represent.

Approved by:	re (bz)
Submitted by:	jonathan
Sponsored by:	Google Inc

END
K 10
svn:author
V 4
nyan
K 8
svn:date
V 27
2011-08-11T13:13:08.264392Z
K 7
svn:log
V 140
Fix building a release with old release scheme.  This catches up with
renaming from Makefile to Makefile.sysinstall.

Approved by:	re (hrs)

END
K 10
svn:author
V 4
nyan
K 8
svn:date
V 27
2011-08-11T13:15:11.635839Z
K 7
svn:log
V 115
Allow to build a release for stable/[78] on a current box and vise versa.

Approved by:	re (hrs)
MFC after:	1 week

END
K 10
svn:author
V 8
jonathan
K 8
svn:date
V 27
2011-08-11T13:29:59.713175Z
K 7
svn:log
V 337
Only call fdclose() on successfully-opened FDs.

Since kern_openat() now uses falloc_noinstall() and finstall() separately,
there are cases where we could get to cleanup code without ever creating
a file descriptor. In those cases, we should not call fdclose() on FD -1.

Approved by: re (kib), mentor (rwatson)
Sponsored by: Google Inc

END
K 10
svn:author
V 6
brooks
K 8
svn:date
V 27
2011-08-11T14:17:04.323419Z
K 7
svn:log
V 88
Take Kirill's commit bit into safe keeping per his request to core@.

Approved by:	core

END
K 10
svn:author
V 2
jh
K 8
svn:date
V 27
2011-08-11T15:22:46.116095Z
K 7
svn:log
V 128
MFC r216047 by kevlo:

Fix fd leak in get_font

MFC r222568:

Don't try to close the stream if fopen(3) fails.

PR:		bin/155349

END
K 10
svn:author
V 8
jonathan
K 8
svn:date
V 27
2011-08-11T15:52:06.860578Z
K 7
svn:log
V 275
Use the right printf() format string without a cast to maxint_t.

As per kib's suggestion, we also change test_count from a size_t to an int;
its value at the moment is 4, and we only expect it to go up to 7.

Approved by: re (kib), mentor (rwatson)
Sponsored by: Google Inc

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-11T19:21:48.259970Z
K 7
svn:log
V 12
Another LOR

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-11T19:31:33.940555Z
K 7
svn:log
V 77
Update the BAW when packets from an aggregate TID session are being
deleted.

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-11T20:29:38.279212Z
K 7
svn:log
V 81
Add two new statistics fields, for software retries and software retry
failures.

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-11T20:30:22.961549Z
K 7
svn:log
V 70
Add another TXQ method, this time to prepend packets to the TXQ head.

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-11T20:34:57.722411Z
K 7
svn:log
V 1003
Implement a very basic software re-transmit algorithm.

If an ADDBA session frame fails, and it's not during some kind
of queue cleanup, attempt to retransmit the thing.

The retransmit method is very simple - set some bits to indicate
its a retry, prepend to the front of the queue, and schedule the
TID for activity in the future. The frame will be rescheduled
and resubmitted to the hardware for us by the normal TX path.

This doesn't attempt a rate lookup; it re-uses the same rate
information as the original frame. This means that if TX is
failing because the rate selection isn't optimal, it's
likely going to keep failing. I'll investigate that in the
future.

Use a (currently) hard-coded max of 10 retransmits. This is 10
retransmits on top of whatever the rate control module attempts,
which may be 4 or 10 frames. This means a frame could be retried
from 40 to 100 times.

Finally, a BAR isn't sent if TX does finally fail. I'll worry
about BAR handling and session teardown handling soon.

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-12T04:06:43.421985Z
K 7
svn:log
V 233
Begin fleshing out the cleanup code needed when transitioning from
aggregation to non-aggregation on a TID.

This will pause the TID scheduling and let hardware-queued packets
complete, before restarting the (now non-aggregate) TID.

END
K 10
svn:author
V 3
pjd
K 8
svn:date
V 27
2011-08-12T07:04:16.637919Z
K 7
svn:log
V 291
Eliminate the zfsdev_state_lock entirely and replace it with the
spa_namespace_lock. This fixes LOR between the spa_namespace_lock and
spa_config lock. LOR can cause deadlock on vdevs removal/insertion.

Reported by:	gibbs, delphij
Tested by:	delphij
Approved by:	re (kib)
MFC after:	1 week

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-12T09:18:02.267715Z
K 7
svn:log
V 1327
Commit some (very untested) initial code to clean up after
a DELBA.

Once a DELBA is received, there may be:

* packets in the software TX TID queue
* packets in the hardware queue (aggregate or individual frames)

The software TX queue is fine - those packets don't yet have a
personality as they haven't been "scheduled".

The packets in the hardware TX queues need to be completed or
deleted.

So the logic here is:

* packets in the SWQ which are retries are discarded;
* packets in the SWQ which aren't retries are modified to have their
  completion handler be the default, non-aggregate one
* if there are any packets left in the hardware queue,
  mark the TID as "cleanup" and keep track of how many are outstanding
* if the TID is "cleanup", pause the TID so no further traffic
  is scheduled.

Then, when the packets are completed, the "incomplete" packet count
is updated and when it hits 0 (ie, all packets on the hardware queue)
it'll be unpaused and will begin running as a non-aggregate TID.

To avoid races, pause the TID in the addba handler before disabling the
addba state and calling ath_tx_cleanup(). Since the TXQ lock is held for
the duration of queue operations, any packets that snuck in before
the TID was paused will be downgraded to non-aggregate status by
ath_tx_cleanup().

Obtained from:	Atheros

END
K 10
svn:author
V 8
jonathan
K 8
svn:date
V 27
2011-08-12T10:52:46.711054Z
K 7
svn:log
V 449
Test *at(2) calls with capability-mode lookup.

This commit adds regression testing for openat(), fstatat(), etc. with
capability scoping ("strict relative" lookup), which applies:
 - in capability mode
 - when performing any *at() lookup relative to a capability

These tests will fail until the *at() code is committed; on my local
instance, with the *at() changes, they all pass.

Approved by: re (kib), mentor (rwatson)
Sponsored by: Google Inc

END
K 10
svn:author
V 8
jonathan
K 8
svn:date
V 27
2011-08-12T11:43:56.240675Z
K 7
svn:log
V 333
Reorder and renumber capability rights.

This patch does three things:
 - puts capability rights in a more pleasing declaration order
 - changes mask values to match the new declaration order
 - declare new rights which will be used soon (e.g. CAP_LOOKUP, CAP_MKDIR)

Approved by: re (kib), mentor (rwatson)
Sponsored by: Google Inc

END
K 10
svn:author
V 5
gabor
K 8
svn:date
V 27
2011-08-12T14:07:18.167137Z
K 7
svn:log
V 175
- Use malloc() instead of alloca()
- Introduce REG_GNU for GNU extensions
- Only process GNU-specific word-boundary notation if REG_GNU is specified
- Fix a compilation error

END
K 10
svn:author
V 5
gabor
K 8
svn:date
V 27
2011-08-12T14:16:45.535576Z
K 7
svn:log
V 29
- TRE-specific style changes

END
K 10
svn:author
V 8
jonathan
K 8
svn:date
V 27
2011-08-12T14:26:47.043180Z
K 7
svn:log
V 182
Rename CAP_*_KEVENT to CAP_*_EVENT.

Change the names of a couple of capability rights to be less
FreeBSD-specific.

Approved by: re (kib), mentor (rwatson)
Sponsored by: Google Inc

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-12T14:43:37.705655Z
K 7
svn:log
V 172
Don't assign sequence numbers to null qos frames.

This creates a 1 packet hole in the receivers BAW, which causes packets to get received
but put in the reassembly queue.

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-12T14:59:21.298857Z
K 7
svn:log
V 261
Fix some missing locks:

* I didn't delete an ATH_TXQ_UNLOCK() from ath_tx_cleanup(); do that
* Since the tid flush function now updates the BAW, the TXQ lock is required.
  This may cause another LOR with the net80211 node lock. I'll investigate this
  later.

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-12T15:06:12.202108Z
K 7
svn:log
V 87
Ok, don't use seqnum=0, use the non-qos seqnum (tid 16.)
This matches what Linux does.

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-12T15:13:06.518736Z
K 7
svn:log
V 8
Updates

END
K 10
svn:author
V 5
gabor
K 8
svn:date
V 27
2011-08-12T16:17:15.386237Z
K 7
svn:log
V 203
- Introduce new flag for word-boundary matching: REG_WORD
- Partly recover broken word-boundary matching; rest is TODO
- Macroify fastcomp() and fastcomp_literal() initialization code
- Adjust a comment

END
K 10
svn:author
V 5
gabor
K 8
svn:date
V 27
2011-08-12T16:21:05.839592Z
K 7
svn:log
V 142
- REG_NOSUB is always applicable
- Avoid calculating pattern length various times
- Drop code for -w and rely on recently introduced REG_WORD

END
K 10
svn:author
V 6
mjacob
K 8
svn:date
V 27
2011-08-12T19:51:28.238175Z
K 7
svn:log
V 288
Fixes zombie device and loop down timers so that they work more than
once. Use taskqueues to do the actual work.

Fix an offset line.

Fix isp_prt so that prints from just one buffer, which makes it
appear cleanly cleanly in logs on SMP systems.

Approved by:	re (kib)
MFC after:	1 month

END
K 10
svn:author
V 5
gabor
K 8
svn:date
V 27
2011-08-12T20:02:47.329528Z
K 7
svn:log
V 36
- Add support for TRE's REG_NEWLINE

END
K 10
svn:author
V 6
mjacob
K 8
svn:date
V 27
2011-08-12T20:09:38.043435Z
K 7
svn:log
V 346
Fixes for sure bus reference miscounting and potential device and
target reference miscounts.  It also adds a helper function to get
the current reference counts for components of cam_path for debug
aid.  One minor style(9) change.

Partially Obtained from: Chuck Tuffli (Emulex)
Reviewed by:	scsi@ (ken)
Approved by:	re (kib)
MFC after:	1 month

END
K 10
svn:author
V 5
dougb
K 8
svn:date
V 27
2011-08-12T21:40:24.782456Z
K 7
svn:log
V 180
MFC r224726:

Make clear that the -U option can be dangerous.

In -p mode install the group file with the same permissions as normally
done in etc/Makefile.

Update the copyright.

END
K 10
svn:author
V 5
dougb
K 8
svn:date
V 27
2011-08-12T21:41:20.210418Z
K 7
svn:log
V 180
MFC r224726:

Make clear that the -U option can be dangerous.

In -p mode install the group file with the same permissions as normally
done in etc/Makefile.

Update the copyright.

END
K 10
svn:author
V 7
delphij
K 8
svn:date
V 27
2011-08-13T00:56:42.051368Z
K 7
svn:log
V 106
MFC r224001:

Use prototype.  While I'm there, add a pair of parenthesis to mark an if
statment's border.

END
K 10
svn:author
V 8
jonathan
K 8
svn:date
V 27
2011-08-13T09:21:16.590995Z
K 7
svn:log
V 612
Allow Capsicum capabilities to delegate constrained
access to file system subtrees to sandboxed processes.

- Use of absolute paths and '..' are limited in capability mode.
- Use of absolute paths and '..' are limited when looking up relative
  to a capability.
- When a name lookup is performed, identify what operation is to be
  performed (such as CAP_MKDIR) as well as check for CAP_LOOKUP.

With these constraints, openat() and friends are now safe in capability
mode, and can then be used by code such as the capability-mode runtime
linker.

Approved by: re (bz), mentor (rwatson)
Sponsored by: Google Inc

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-13T10:23:52.622681Z
K 7
svn:log
V 931
Some fixes - fix BAW checking/update (which isn't entirely fixed,
I'll be fiddling with this shortly!) and simplify the swq TX routine.

This is unfortunately one of those aggregate commits which I dislike
doing.

* Remove an unneeded mbuf argument to ath_tx_swq(). ath_buf should already
  be setup by now and bf->bf_m is the first mbuf in the chain.

* Now that NULL data QOS frames are not receiving sequence numbers from
  the TID, we can't check them against the BAW window when we're scheduling
  packets. Otherwise it's likely it'll be outside the BAW and will remain
  so, so the TX will hang until the watchdog timer kicks in.

The code duplication involved in checking whether a frame is going to be
given a sequence number and treated as part of the BAW is .. well, horrible.

Unfortunately, I didn't think about needing to also not call the BAW update
upon packet completion. That's going to have to be the next check.

END
K 10
svn:author
V 8
jonathan
K 8
svn:date
V 27
2011-08-13T10:43:21.555558Z
K 7
svn:log
V 806
Allow openat(2), fstatat(2), etc. in capability mode.

namei() and lookup() can now perform "strictly relative" lookups.
Such lookups, performed when in capability mode or when looking up
relative to a directory capability, enforce two policies:
 - absolute paths are disallowed (including symlinks to absolute paths)
 - paths containing '..' components are disallowed

These constraints make it safe to enable openat() and friends.
These system calls are instrumental in supporting Capsicum
components such as the capability-mode-aware runtime linker.

Finally, adjust comments in capabilities.conf to reflect the actual state
of the world (e.g. shm_open(2) already has the appropriate constraints,
getdents(2) already requires CAP_SEEK).

Approved by: re (bz), mentor (rwatson)
Sponsored by: Google Inc.

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-13T10:43:56.398485Z
K 7
svn:log
V 657
This deviates a bit from the reference code; I may end up backing this out
at a later stage.

This unifies the "is this packet supposed to be considered as part of the
BAW?" logic. I've added a bit in the ath_buf.bf_state struct which indicates
whether it's going to be considered for inclusion or not.

Packets that aren't part of the BAW calculation are thus not passed to the
BAW add or update code.

Eventually, the aggregate forming code will also not include this packet
as part of an aggregate and will be sent as a non-aggregate packet.

This fixes the hangs, interface resets and general unhappiness that has been
happening with my recent changes.

END
K 10
svn:author
V 2
mm
K 8
svn:date
V 27
2011-08-13T10:58:53.212518Z
K 7
svn:log
V 505
Fix race between dmu_objset_prefetch() invoked from
zfs_ioc_dataset_list_next() and dsl_dir_destroy_check() indirectly
invoked from dmu_recv_existing_end() via dsl_dataset_destroy() by not
prefetching temporary clones, as these count as always inconsistent.
In addition, do not prefetch hidden datasets at all as we are not
going to process these later.

Filed as Illumos Bug #1346

PR:		kern/157728
Tested by:	Borja Marcos <borjam@sarenet.es>, mm
Reviewed by:	pjd
Approved by:	re (kib)
MFC after:	1 week

END
K 10
svn:author
V 7
rwatson
K 8
svn:date
V 27
2011-08-13T12:14:40.706546Z
K 7
svn:log
V 377
Regenerate system call files following r224812 changes to capabilities.conf.
A no-op for non-Capsicum kernels; for Capsicum kernels, completes the
enabling of fooat(2) system calls using capabilities.  With this change,
and subject to bug fixes, Capsicum capability support is now complete for
9.0.

Approved by:    re (kib)
Submitted by:   jonathan
Sponsored by:   Google Inc

END
K 10
svn:author
V 6
marius
K 8
svn:date
V 27
2011-08-13T12:28:58.529185Z
K 7
svn:log
V 478
MFC: r223985

- For SAS but neither FC nor SPI controllers default to using MSI/MSI-X
  (still allowing their use to be disabled via device hints though). This
  matches what the corresponding Linux driver provided by LSI does. Tested
  with SAS1064.
- There's no need to keep track of the RIDs used.
- Don't allocate MSI/MSI-X as RF_SHAREABLE.
- Remove a comment which no longer applies since r209599 (MFC'ed to stable/8
  in r210376).
- Assign NULL rather than 0 to pointers.

END
K 10
svn:author
V 6
marius
K 8
svn:date
V 27
2011-08-13T12:29:50.564003Z
K 7
svn:log
V 478
MFC: r223985

- For SAS but neither FC nor SPI controllers default to using MSI/MSI-X
  (still allowing their use to be disabled via device hints though). This
  matches what the corresponding Linux driver provided by LSI does. Tested
  with SAS1064.
- There's no need to keep track of the RIDs used.
- Don't allocate MSI/MSI-X as RF_SHAREABLE.
- Remove a comment which no longer applies since r209599 (MFC'ed to stable/7
  in r224332).
- Assign NULL rather than 0 to pointers.

END
K 10
svn:author
V 6
marius
K 8
svn:date
V 27
2011-08-13T12:33:06.321193Z
K 7
svn:log
V 255
MFC: r224493

- Staticize functions as appropriate and comment out unused ones.
- Sprinkle some const where appropriate.
- Consistently use target_id_t for the target parameter of mpt_map_physdisk()
  and mpt_is_raid_volume().
- Fix some whitespace bugs.

END
K 10
svn:author
V 6
marius
K 8
svn:date
V 27
2011-08-13T12:33:41.723909Z
K 7
svn:log
V 255
MFC: r224493

- Staticize functions as appropriate and comment out unused ones.
- Sprinkle some const where appropriate.
- Consistently use target_id_t for the target parameter of mpt_map_physdisk()
  and mpt_is_raid_volume().
- Fix some whitespace bugs.

END
K 10
svn:author
V 6
marius
K 8
svn:date
V 27
2011-08-13T12:37:22.785778Z
K 7
svn:log
V 1064
MFC: r224494, r224761

- Send the RELSIM_ADJUST_OPENINGS in response to a MPI_EVENT_QUEUE_FULL using
  the right SIM in case the HBA is RAID-capable but the target in question is
  not a hot spare or member of a RAID volume.
- Report the loss and addition of SAS and SATA targets detected via PHY link
  status changes and signalled by MPI_EVENT_SAS_DEVICE_STATUS_CHANGE to cam(4)
  as lost devices and trigger rescans as appropriate. Without this it can take
  quite some time until a lost device actually is no longer tried to be used,
  if it ever stops. [1]
- Handle MPI_EVENT_IR2, MPI_EVENT_LOG_ENTRY_ADDED, MPI_EVENT_SAS_DISCOVERY
  and MPI_EVENT_SAS_PHY_LINK_STATUS silently as these serve no additional
  purpose beyond adding cryptic entries to logs.
- Add a warning for MPI_EVENT_SAS_DISCOVERY_ERROR events, which can help
  identifying broken disks. [2]

Thanks to Hans-Joerg Sirtl for providing one of the HBAs these changes were
developed with and RIP to the mainboard that didn't survive testing them.

PR:		157534 [1]
Submitted by:	Andrew Boyer [2]

END
K 10
svn:author
V 6
marius
K 8
svn:date
V 27
2011-08-13T12:37:23.612937Z
K 7
svn:log
V 1064
MFC: r224494, r224761

- Send the RELSIM_ADJUST_OPENINGS in response to a MPI_EVENT_QUEUE_FULL using
  the right SIM in case the HBA is RAID-capable but the target in question is
  not a hot spare or member of a RAID volume.
- Report the loss and addition of SAS and SATA targets detected via PHY link
  status changes and signalled by MPI_EVENT_SAS_DEVICE_STATUS_CHANGE to cam(4)
  as lost devices and trigger rescans as appropriate. Without this it can take
  quite some time until a lost device actually is no longer tried to be used,
  if it ever stops. [1]
- Handle MPI_EVENT_IR2, MPI_EVENT_LOG_ENTRY_ADDED, MPI_EVENT_SAS_DISCOVERY
  and MPI_EVENT_SAS_PHY_LINK_STATUS silently as these serve no additional
  purpose beyond adding cryptic entries to logs.
- Add a warning for MPI_EVENT_SAS_DISCOVERY_ERROR events, which can help
  identifying broken disks. [2]

Thanks to Hans-Joerg Sirtl for providing one of the HBAs these changes were
developed with and RIP to the mainboard that didn't survive testing them.

PR:		157534 [1]
Submitted by:	Andrew Boyer [2]

END
K 10
svn:author
V 6
marius
K 8
svn:date
V 27
2011-08-13T12:43:33.078353Z
K 7
svn:log
V 211
MFC: r224682

- Merge from r147740:
  When the last, possibly partially filled buffer is flushed, we didn't
  reset fragsz to 0 and as such would stop reflecting reality.
- Use __FBSDID.
- Wrap a too long line.

END
K 10
svn:author
V 6
marius
K 8
svn:date
V 27
2011-08-13T12:43:35.165696Z
K 7
svn:log
V 197
MFC: r224682 (partial)

- Merge from r147740:
  When the last, possibly partially filled buffer is flushed, we didn't
  reset fragsz to 0 and as such would stop reflecting reality.
- Use __FBSDID.

END
K 10
svn:author
V 6
marius
K 8
svn:date
V 27
2011-08-13T12:47:09.076044Z
K 7
svn:log
V 210
MFC: r224683, r224693

The tid member of struct pthread actually is long so read it as such.
Accessing it as an int causes failure on big-endian LP64, i.e. mips64be,
powerpc64 and sparc64.

Reviewed by:	marcel

END
K 10
svn:author
V 6
marius
K 8
svn:date
V 27
2011-08-13T12:47:10.867036Z
K 7
svn:log
V 210
MFC: r224683, r224693

The tid member of struct pthread actually is long so read it as such.
Accessing it as an int causes failure on big-endian LP64, i.e. mips64be,
powerpc64 and sparc64.

Reviewed by:	marcel

END
K 10
svn:author
V 6
marius
K 8
svn:date
V 27
2011-08-13T12:49:00.822645Z
K 7
svn:log
V 125
MFC: r224684

Use the size of struct fpreg rather than of the pointer to it when copying
the FPU state.

Reviewed by:	marcel

END
K 10
svn:author
V 6
marius
K 8
svn:date
V 27
2011-08-13T12:49:02.503853Z
K 7
svn:log
V 125
MFC: r224684

Use the size of struct fpreg rather than of the pointer to it when copying
the FPU state.

Reviewed by:	marcel

END
K 10
svn:author
V 6
marius
K 8
svn:date
V 27
2011-08-13T12:50:38.568082Z
K 7
svn:log
V 45
MFC: r224685

Implement

Reviewed by:	marcel

END
K 10
svn:author
V 6
marius
K 8
svn:date
V 27
2011-08-13T12:50:39.359918Z
K 7
svn:log
V 45
MFC: r224685

Implement

Reviewed by:	marcel

END
K 10
svn:author
V 6
marius
K 8
svn:date
V 27
2011-08-13T12:57:48.387294Z
K 7
svn:log
V 95
MFC: r224686

Implement functions necessary for compiling fbsd-threads.c.

Reviewed by:	marcel

END
K 10
svn:author
V 6
marius
K 8
svn:date
V 27
2011-08-13T12:57:51.871886Z
K 7
svn:log
V 95
MFC: r224686

Implement functions necessary for compiling fbsd-threads.c.

Reviewed by:	marcel

END
K 10
svn:author
V 6
marius
K 8
svn:date
V 27
2011-08-13T13:09:43.863889Z
K 7
svn:log
V 127
MFC: r224687

Compile fbsd-threads.c. Amongst others this is necessary for a working
kgdb(1).

PR:		158719
Reviewed by:	marcel

END
K 10
svn:author
V 6
marius
K 8
svn:date
V 27
2011-08-13T13:09:46.544674Z
K 7
svn:log
V 127
MFC: r224687

Compile fbsd-threads.c. Amongst others this is necessary for a working
kgdb(1).

PR:		158719
Reviewed by:	marcel

END
K 10
svn:author
V 7
rwatson
K 8
svn:date
V 27
2011-08-13T13:11:28.762474Z
K 7
svn:log
V 230
Bump __FreeBSD_version to reflect the availability of capabilities, but
also capability-related changes to fget(9).  This is likely not part of
a formal KPI, but the nvidia driver (at least) uses it.

Approved by:    re (bz, kib)

END
K 10
svn:author
V 6
marius
K 8
svn:date
V 27
2011-08-13T13:15:27.607847Z
K 7
svn:log
V 1206
MFC: r224762

Sync makefs(8) ISO 9660 support with NetBSD:
o cd9960 -> cd9660
o Move inclusion of sys/endian.h from cd9660_eltorito.c to cd9660.h
  since actual user is not cd9660_eltorito.c but iso.h and
  cd9660_eltorito.h.

  Actually, include order/place of sys/endian.h doesn't matter on
  netbsd since it is always included by sys/types.h but it's not
  true on other system.  This should fix cross build breakage on
  freebsd introduced by rev. 1.16 of cd9660_eltorito.c.
  Problem reported and fix suggested on twitter.
o Fix fd leaks in error cases. Found by cppcheck.
o RRIP RE length should be 4, not 0
o Apply fixes for PR bin/44114 (makefs(8) -t cd9660 -o rockridge creates
  corrupted cd9660fs), iso9660_rrip.c part:
  - cd9660_rrip_finalize_node() should check rr_real_parent in node->parent,
    not in node itself in RRIP_PL case
  - cd9660_rrip_initialize_node() should update only node passed as arg
    so handle RRIP_PL in DOTDOT case

  Fixes malformed dotdot entries in deep (more than 8 level) directories
  moved into .rr_moved dir.

  Should be pulled up to netbsd-5.
  (no official ISO has such deep dirs, but cobalt restorecd is affected)

Reviewed by:	mm
Obtained from:	NetBSD

END
K 10
svn:author
V 6
marius
K 8
svn:date
V 27
2011-08-13T13:15:38.549412Z
K 7
svn:log
V 1206
MFC: r224762

Sync makefs(8) ISO 9660 support with NetBSD:
o cd9960 -> cd9660
o Move inclusion of sys/endian.h from cd9660_eltorito.c to cd9660.h
  since actual user is not cd9660_eltorito.c but iso.h and
  cd9660_eltorito.h.

  Actually, include order/place of sys/endian.h doesn't matter on
  netbsd since it is always included by sys/types.h but it's not
  true on other system.  This should fix cross build breakage on
  freebsd introduced by rev. 1.16 of cd9660_eltorito.c.
  Problem reported and fix suggested on twitter.
o Fix fd leaks in error cases. Found by cppcheck.
o RRIP RE length should be 4, not 0
o Apply fixes for PR bin/44114 (makefs(8) -t cd9660 -o rockridge creates
  corrupted cd9660fs), iso9660_rrip.c part:
  - cd9660_rrip_finalize_node() should check rr_real_parent in node->parent,
    not in node itself in RRIP_PL case
  - cd9660_rrip_initialize_node() should update only node passed as arg
    so handle RRIP_PL in DOTDOT case

  Fixes malformed dotdot entries in deep (more than 8 level) directories
  moved into .rr_moved dir.

  Should be pulled up to netbsd-5.
  (no official ISO has such deep dirs, but cobalt restorecd is affected)

Reviewed by:	mm
Obtained from:	NetBSD

END
K 10
svn:author
V 6
marius
K 8
svn:date
V 27
2011-08-13T13:18:39.937822Z
K 7
svn:log
V 185
MFC: r221466, r221494 (partial), r224652

Change several archs to use makefs (from base) instead of mkisofs (from
ports) which makes release CD generation both faster and self-hosting.

END
K 10
svn:author
V 6
marius
K 8
svn:date
V 27
2011-08-13T13:18:54.562637Z
K 7
svn:log
V 185
MFC: r221466, r221494 (partial), r224652

Change several archs to use makefs (from base) instead of mkisofs (from
ports) which makes release CD generation both faster and self-hosting.

END
K 10
svn:author
V 7
rwatson
K 8
svn:date
V 27
2011-08-13T13:26:40.943830Z
K 7
svn:log
V 164
Now that capability support has been committed, update and expand the
comment at the type of sys_capability.c to describe its new contents.

Approved by:  re (xxx)

END
K 10
svn:author
V 7
rwatson
K 8
svn:date
V 27
2011-08-13T13:34:01.875402Z
K 7
svn:log
V 166
Update use of the FEATURE() macro in sys_capability.c to reflect the move
to two different kernel options for capability mode vs. capabilities.

Approved by:	re (bz)

END
K 10
svn:author
V 2
bz
K 8
svn:date
V 27
2011-08-13T13:49:11.855973Z
K 7
svn:log
V 172
Fix jls backward compat mode broken in r222465, correctly
displaying addresses in verbose mode (jls -v) again.

Submitted by:	jamie
MFC after:	3 days
Approved by:	re (kib)

END
K 10
svn:author
V 7
rwatson
K 8
svn:date
V 27
2011-08-13T16:03:40.492177Z
K 7
svn:log
V 421
When falloc() was broken into separate falloc_noinstall() and finstall(),
a bug was introduced in kern_openat() such that the error from the vnode
open operation was overwritten before it was passed as an argument to
dupfdopen().  This broke operations on /dev/{stdin,stdout,stderr}.  Fix
by preserving the original error number across finstall() so that it is
still available.

Approved by:	re (kib)
Reported by:	cognet

END
K 10
svn:author
V 4
zack
K 8
svn:date
V 27
2011-08-13T17:08:57.595979Z
K 7
svn:log
V 179
MFC: r224077
  Small acl patch to return the aclerror that comes back from
  nfsrv_dissectacl().  This fixes a problem where ATTRNOTSUPP was being
  returned instead of BADOWNER.

END
K 10
svn:author
V 4
zack
K 8
svn:date
V 27
2011-08-13T17:16:06.763712Z
K 7
svn:log
V 92
MFC: 224078
  Move nfsvno_pathconf to be accessible to sys/fs/nfs; no functionality change.

END
K 10
svn:author
V 7
attilio
K 8
svn:date
V 27
2011-08-13T17:17:04.997457Z
K 7
svn:log
V 150
Fix a typo in cb_dumpdata() about wrongly calling wdog_kern_pat().

Submitted by:	Andrew Boyer <aboyer at averesystems dot com>
Approved by:	re (kib)

END
K 10
svn:author
V 4
zack
K 8
svn:date
V 27
2011-08-13T17:17:06.459981Z
K 7
svn:log
V 86
MFC: 224080

Remove unnecessary thread pointer from VOPLOCK macros and current users.

END
K 10
svn:author
V 4
zack
K 8
svn:date
V 27
2011-08-13T17:17:59.085817Z
K 7
svn:log
V 153
MFC: 224081

Simple find/replace of vn_lock -> NFSVOPLOCK. This is done so that NFSVOPLOCK
can be modified later to add enhanced logging and assertions.

END
K 10
svn:author
V 4
zack
K 8
svn:date
V 27
2011-08-13T17:18:45.456401Z
K 7
svn:log
V 160
MFC: 224082

Simple find/replace of VOP_UNLOCK -> NFSVOPUNLOCK. This is done so that
NFSVOPUNLOCK can be modified later to add enhanced logging and assertions.

END
K 10
svn:author
V 4
zack
K 8
svn:date
V 27
2011-08-13T17:19:24.318329Z
K 7
svn:log
V 167
MFC: r224083

Simple find/replace of VOP_ISLOCKED -> NFSVOPISLOCKED. This is done so that
NFSVOPISLOCKED can be modified later to add enhanced logging and assertions.

END
K 10
svn:author
V 4
zack
K 8
svn:date
V 27
2011-08-13T17:20:00.851372Z
K 7
svn:log
V 69
MFC: 224084

Add a small comment about unloading the kgsappi module.

END
K 10
svn:author
V 4
zack
K 8
svn:date
V 27
2011-08-13T17:21:50.455882Z
K 7
svn:log
V 562
MFC r224086:

Add DEXITCODE plumbing to NFS.

Isilon has the concept of an in-memory exit-code ring that saves the last exit
code of a function and allows for stack tracing. This is very helpful when
debugging tough issues.

This patch is essentially a no-op for BSD at this point, until we upstream
the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS
function that returns an errno error code. A number of code paths were also
reorganized to have single exit paths, to reduce code duplication.

Submitted by:   David Kwan <dkwan@isilon.com>

END
K 10
svn:author
V 7
rwatson
K 8
svn:date
V 27
2011-08-13T17:22:16.057458Z
K 7
svn:log
V 171
Trim some warnings and notes from capabilities.conf -- these are left over
from Capsicum development, and no longer apply.

Approved by:	re (kib)
Sponsored by:	Google Inc

END
K 10
svn:author
V 4
zack
K 8
svn:date
V 27
2011-08-13T17:29:37.656559Z
K 7
svn:log
V 340
MFC r224079 and r224121:

r224121 was meant to revert r224079, except I accidentally checked in an
additional ACL patch in r224121. Committing both of these ends up with just the
ACL fix. Here is the real commit note for the ACL patch:

Running pynfs verify/nverify tests produces a panic in nfsrv_compareacl(). This
patch fixes the panic.

END
K 10
svn:author
V 6
jilles
K 8
svn:date
V 27
2011-08-13T21:33:50.070146Z
K 7
svn:log
V 232
MFC r222716: sh: Fix $? in heredocs on simple commands.

The patch is slightly different because redirected compound commands are
handled differently in stable/8 (no separate function, redirection errors
are fatal).

PR:		bin/41410

END
K 10
svn:author
V 2
mm
K 8
svn:date
V 27
2011-08-13T21:35:22.196481Z
K 7
svn:log
V 794
zfs_ioctl.c: improve code readability in zfs_ioc_dataset_list_next()

zvol.c: fix calling of dmu_objset_prefetch() in zvol_create_minors()
by passing full instead of relative dataset name and prefetching all
visible datasets to be processed later instead of just the pool name

Reviewed by:	pjd
Approved by:	re (kib)
MFC after:	1 week
> Reviewed by:   If someone else reviewed your modification.
> Approved by:   If you needed approval for this commit.
> Obtained from: If the change is from a third party.
> MFC after:     N [day[s]|week[s]|month[s]].  Request a reminder email.
> Security:      Vulnerability reference (one per line) or description.
> Empty fields above will be automatically removed.

M    opensolaris/uts/common/fs/zfs/zfs_ioctl.c
M    opensolaris/uts/common/fs/zfs/zvol.c

END
K 10
svn:author
V 6
mjacob
K 8
svn:date
V 27
2011-08-13T23:34:17.740720Z
K 7
svn:log
V 378
Most of these changes to isp are to allow for isp.ko unloading.
We also revive loop down freezes. We also externaliz within isp
isp_prt_endcmd so something outside the core module can print
something about a command completing. Also some work in progress to
assist in handling timed out commands better.

Partially Sponsored by: Panasas
Approved by:	re (kib)
MFC after:	1 month

END
K 10
svn:author
V 10
nwhitehorn
K 8
svn:date
V 27
2011-08-14T00:20:37.144613Z
K 7
svn:log
V 201
Add support for the Blu-Ray drive found in the Sony Playstation 3 and fix
some realted minor bugs in PS3 internal storage support.

Submitted by:	glevand <geoffrey.levand@mail.ru>
Approved by:	re (bz)

END
K 10
svn:author
V 8
rmacklem
K 8
svn:date
V 27
2011-08-14T00:32:43.970034Z
K 7
svn:log
V 335
MFC: r224532
The new NFS client failed to vput() the new vnode if a setattr
failed after the file was created in nfs_create(). This would
probably only happen during a forced dismount. The old NFS client
does have a vput() for this case. Detected by pho during recent
testing, where an open syscall returned with a vnode still locked.

END
K 10
svn:author
V 7
rwatson
K 8
svn:date
V 27
2011-08-14T00:42:09.885949Z
K 7
svn:log
V 271
Updates to libprocstat(3) and procstat(1) to allow monitoring Capsicum
capability mode and capabilities.

Right now no attempt is made to unwrap capabilities when operating on
a crashdump, so further refinement is required.

Approved by:	re (bz)
Sponsored by:	Google Inc

END
K 10
svn:author
V 8
rmacklem
K 8
svn:date
V 27
2011-08-14T00:55:18.621152Z
K 7
svn:log
V 651
MFC: r224554
Fix rename in the new NFS server so that it does not require a
recursive vnode lock on the directory for the case where the
new file name is in the same directory as the old one. The patch
handles this as a special case, recognized by the new directory
having the same file handle as the old one and just VREF()s the old
dir vnode for this case, instead of doing a second VFS_FHTOVP() to get it.
This is required so that the server will work for file systems like
msdosfs, that do not support recursive vnode locking.
This problem was discovered during recent testing by pho@
when exporting an msdosfs file system via the new NFS server.

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-14T11:47:42.324948Z
K 7
svn:log
V 178
Add the 4ms framelength maximum values.

This'll be eventually used by the aggregation logic to determine how long the
maximum frame length will be at the selected rate control.

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-14T12:21:08.253446Z
K 7
svn:log
V 661
Begin fleshing out the BAR handling.

net80211 implements BAR frame queuing and completion handofing.
The first place a BAR should be TXed is when the software retry
limit has been exceeded for a TID currently doing aggregation.
The process should be (just so I don't forget):

* pause the TID;
* send the BAR;
* unpause the TID in ath_bar_response()

If multiple BAR's are sent, paused will be incremented and decremented
appropriately until all BAR frames have been transmitted (succesfully
or otherwise.)

However, using it with the current code is likely not going to work
because the ath TX code isn't re-entrant w/ the way locking currently
is implement.

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-14T12:26:24.886308Z
K 7
svn:log
V 87
Write out some notes about BAR response processing and what we're (currently)
missing.

END
K 10
svn:author
V 7
rwatson
K 8
svn:date
V 27
2011-08-14T12:41:44.453377Z
K 7
svn:log
V 99
Cross-reference cap_new(2) from dup(2), as they have similar functionality.

Approved by:	re (kib)

END
K 10
svn:author
V 6
jilles
K 8
svn:date
V 27
2011-08-14T13:37:38.707306Z
K 7
svn:log
V 450
tail: Fix crash if -F'ed file's filesystem disappears.

If tail notices that a file it is following no longer exists (because stat()
fails), it will output any final lines and then close the file. If the read
operation also causes an error, such as when the filesystem is forcefully
unmounted, it closes the file as well, leading to fclose(NULL) and a
segmentation fault.

PR:		bin/159750
Submitted by:	swills
Approved by:	re (kib)
MFC after:	1 week

END
K 10
svn:author
V 10
nwhitehorn
K 8
svn:date
V 27
2011-08-14T14:36:32.688922Z
K 7
svn:log
V 143
Provide a kboot.conf to allow powerpc64 install media to be booted on
Playstation 3 consoles running kboot or petitboot.

Approved by:	re (bz)

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-14T16:03:26.233948Z
K 7
svn:log
V 1651
Begin to attempt to handle sending BAR frames. This doesn't work yet and will
panic the kernel.

BAR frames should be generated to resync the remote side with where the block-ack
window left edge is - eg, after some packet loss.

The problem here is that said BAR frames will be generated from:

* the TX packet completion handler; and
* via a timer callout if the BAR TX wasn't successful.

The former will occur with the hardware TXQ lock held, as that's currently
how I've implemented the locking (ie, the hardware TXQ protects the TXQ and
the TID state.) This results in lock recursion (as we're recursing back
through net80211 and then back into the driver via ic->ic_raw_xmit()) and
thus TX'ing a BAR frame will very much fail.

The problem here (among other things) is that TX'ing a BAR frame should
occur after all the hardware-queued frames for that AC, in case there's
some other packets (which may or may not be successfully transmitted
in the future), and in the same hardware queue. This means we can't just
use the normal software queue method, as the TID should be paused when
a BAR frame is TX'ed. Hence, the direct dispatch to hardware.

Anyway, this is all quite messy. I think I'm going to have to bite the
bullet and (re) introduce more fine-grained locking in a subsequent commit
before supporting this in a clean fashion is even possible.

(The other option is to modify the ieee80211_send_bar()) function to just
schedule a callback even and have the TX occur from there, rather than
recurse through the driver. I may end up doing that at a later date, but
I really need to fix the locking anyway. So, I'll just do that first.)

END
K 10
svn:author
V 3
mav
K 8
svn:date
V 27
2011-08-14T16:17:00.403273Z
K 7
svn:log
V 115
Fix integer overflow on 32bit systems when calculating media size,
reintroduced by r222475.

Approved by:	re (kib)

END
K 10
svn:author
V 5
gabor
K 8
svn:date
V 27
2011-08-14T20:54:13.585849Z
K 7
svn:log
V 52
- Fix word boundary matching for the remaining case

END
K 10
svn:author
V 6
tuexen
K 8
svn:date
V 27
2011-08-14T20:55:32.616357Z
K 7
svn:log
V 198
Add support for the spp_dscp field in the SCTP_PEER_ADDR_PARAMS
socket option. Backwards compatibility is provided by still
supporting the spp_ipv4_tos field.

Approved by: re@
MFC after: 2 months.

END
K 10
svn:author
V 5
gabor
K 8
svn:date
V 27
2011-08-14T22:53:02.755000Z
K 7
svn:log
V 77
- Implement TRE-specific REG_NOTBOL and REG_NOTEOL
- Adjust an older comment

END
K 10
svn:author
V 5
gibbs
K 8
svn:date
V 27
2011-08-14T23:26:32.803439Z
K 7
svn:log
V 117
Fix building kdump.  The new scsi_ses.h requires <stdint.h> and <stddef.h>.

Sponsored by:	Spectra Logic Corporation

END
K 10
svn:author
V 5
gabor
K 8
svn:date
V 27
2011-08-15T00:38:14.572414Z
K 7
svn:log
V 55
- Add some comments
- One TRE-specific portability nit

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-15T03:41:48.415845Z
K 7
svn:log
V 2568
Do some relatively major changes to my software TX scheduling code.

In the first cut, and inspired by some other code I have here which
implements this, I decided to hide the per-TID locking behind
the hardware TXQ lock. That worked, as the mapping for TID->TXQ
is constant and predictable, but it meant the hardware lock is
held a lot longer than it needs to be. This has caused all kinds
of problems.

A 'better' way (for values of 'better') would be to implement
fine-grained locking of the per-node and per-TID state. Then those
locks can be held as long (short) as they need to be.

But for now, I'm going with another temporary solution designed
to give me more breathing room whilst I port over the code.
I've separated out the TX code into a sort of split setup,
with the upper half being net80211/network code, and the lower
half being the TX scheduling and completion.

* Create a "TX sched" task, which will run the TX scheduling
  code as a task
* Remove a lot of the hardware TXQ locking and asserting
* Re-introduce the per-TID software TXQ lock
* Since the entry pathways into this code now aren't locked behind
  the hardware TXQ locks, re-lock the TID software queue access
* Re-introduce the short-held hardware TXQ locks, so top and bottom
  level fiddling of the hardware TXQs don't trample on each other.

Now, the "top" level code (ie, anything which wishes to queue packets)
will enter via ath_start / ath_raw_xmit / ath_tx_start / ath_tx_raw_start
and will either queue directly to the hardware (and that's protected
by the hardware TXQ locks) or be queued to the software queue via a call
to ath_tx_swq(). ath_tx_swq() simply queues a packet to the software
queue (protected by a software TXQ lock) and then the actual TX packet
scheduling code is invoked in a task call to ath_tx_sched_proc().

ath_tx_sched_proc() handles scheduling; ath_tx_proc() handles TX
completion and further scheduling; since neither of them run
simultaneously I can avoid a lot of the complicated locking.

This likely won't be the final solution (as I may end up introducing
fine-grained locks anyway) but it does push (almost) all of the
per-TID state and general aggregation state handling into the ath
task, rather than trying to handle concurrent accesses from TX
processes, RX/TX tasks and interrupt tasks.

Note: net80211 node cleanup and node flush may still need some further
locking. I'll look into that shortly. (as node_flush can occur during
a scan, for example, and I haven't checked to see whether that runs
within or separate to the ath taskqueue.)

END
K 10
svn:author
V 7
rwatson
K 8
svn:date
V 27
2011-08-15T07:30:48.105138Z
K 7
svn:log
V 327
Bump __FreeBSD_version to reflect the availability of capabilities, but
also capability-related changes to fget(9).  This is likely not part of
a formal KPI, but the nvidia driver (at least) uses it.

Mention /dev/{stdin,stdout,stderr} breakage that appears in certain
kernel revisions as best avoided!

Approved by:  re (xxx)

END
K 10
svn:author
V 7
rwatson
K 8
svn:date
V 27
2011-08-15T07:32:44.061720Z
K 7
svn:log
V 417
Fix two cases involving opt_capsicum.h and module builds:

(1) opt_capsicum.h is no longer required in ffs_alloc.c, so remove the
   #include.

(2) portalfs depends on opt_capsicum.h, so have the Makefile generate one
   if required.

These affect only modules built without a kernel (i.e, not buildkernel,
but yes buildworld if the dubious MODULES_WITH_WORLD is used).

Approved by:	re (bz)
Sponsored by:	Google Inc

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-15T09:07:42.787258Z
K 7
svn:log
V 560
* Flip on sending BAR frames, but don't (yet) pause the TID until that's
  actually completed.
* Fix a bug in the retry code - ATH_BUF_BUSY indicates that there may be
  frames after this one and the hardware QCU may be paused at said frame,
  waiting for the next TX opportunity to open up (when doing TDMA.)
  For now, just remove the flag. I'll revisit this before I merge this
  code into -HEAD.

This now queues BAR frames but has exposed a bug in the TID scheduling
when sending BAR frames and handling packet loss. I'll fix this in a
subsequent commit.

END
K 10
svn:author
V 3
kib
K 8
svn:date
V 27
2011-08-15T09:26:53.097179Z
K 7
svn:log
V 49
MFC r224582:
Implement the linprocfs swaps file.

END
K 10
svn:author
V 7
glebius
K 8
svn:date
V 27
2011-08-15T12:08:41.079600Z
K 7
svn:log
V 156
Describe how carp(4) status changes can be processed with
help of devd(8).

Submitted by:	"Alexander V. Chernikov" <melifaro ipfw.ru>
Approved by:	re (kib)

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-15T13:25:55.756931Z
K 7
svn:log
V 598
Fix some of the locking to correctly protect things.

Specifically, ath_tx_tid_unsched() wasn't correctly locking things,
so there was a race condition between the upper half that were
unscheduling things (eg a node flush) and the lower half (tx sched
and completion task.) Fixing it required re-doing some of the
locking.

This seems to now cope with TX failures -> BAR -> BAR failure ->
DELBA. Normal TX then continues fine. Before, the code would
get wedged whilst doing packet TX whilst doing BAR re-transmission,
and things would stop progressing until a scan or some other flush
would occur.

END
K 10
svn:author
V 5
gabor
K 8
svn:date
V 27
2011-08-15T13:27:02.392606Z
K 7
svn:log
V 181
- Use internal names for fast matching that match better TRE's internal names.
  The POSIX-like names will be used to provide an alternative interface
  for the fast matching code.

END
K 10
svn:author
V 10
nwhitehorn
K 8
svn:date
V 27
2011-08-15T13:33:14.623782Z
K 7
svn:log
V 255
Use a maximum of -O on PowerPC kernels instead of -O2 to prevent a rare
bug that could cause intermittent memory corruption on PowerPC SMP
systems using non-debug kernels. This is a temporary change until the
real problem is fixed.

Approved by:	re (kib)

END
K 10
svn:author
V 7
glebius
K 8
svn:date
V 27
2011-08-15T13:33:28.233815Z
K 7
svn:log
V 82
Fix a couple of issues in last commit.

Submitted by:	maxim
Approved by:	re (kib)

END
K 10
svn:author
V 5
gabor
K 8
svn:date
V 27
2011-08-15T13:43:43.550456Z
K 7
svn:log
V 26
- Properly free resources

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-15T14:00:55.658468Z
K 7
svn:log
V 35
Update comments to reflect reality

END
K 10
svn:author
V 3
mav
K 8
svn:date
V 27
2011-08-15T14:19:26.346193Z
K 7
svn:log
V 341
MFC r224511:
 - Fix usbhidctl and usbhidaction to handle HID devices with multiple
report ids, such as multimedia keyboards.
 - Add collection type and report id to the `usbhidctl -r` output. They
are important for proper device understanding and debugging.
 - Fix usbhidaction tool to properly handle items having report_count
more then 1.

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-15T14:59:56.862919Z
K 7
svn:log
V 87
Fix BAR handling - status == 0 means ACKed or "TX successful", non-zero
means failure.

END
K 10
svn:author
V 5
gabor
K 8
svn:date
V 27
2011-08-15T15:22:27.598694Z
K 7
svn:log
V 31
- Match macro name to filename

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-15T15:26:36.111633Z
K 7
svn:log
V 998
More BAR related stuff, which is partially locking related stuff.

* Pause/unpause the queue when BAR TX is complete.

  The problem is that we're not notified when a BAR TX times out
  and the TX won't happen again - the session is just torn down.
  So, to work around the API shortcomings for now, hard-code it
  to cause an unpause at retry == 50.

  This only works if the hardware _DOES_ call the callback upon
  frame TX failure. If the callback never gets called, this will
  all fall in a heap. Luckily, ath will.

* Fix locking around TXQ pause/resume calls, as they can occur from
  the upper layers. This isn't entirely complete, as I'm not
  similarly locking the pause/resume checks in the rest of the
  code. I'll do that soon.

* Only pause the TID if the ieee80211_send_bar() call was successful.

This now seems to work fine with TX packet loss in non-aggregate mode,
but I'm not yet convinced that I'm sending the correct sequence
number in the BAR frame. I'll look at that next.

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-15T16:06:28.602224Z
K 7
svn:log
V 55
This needs to be fixed too - status=0 means "success".

END
K 10
svn:author
V 7
yongari
K 8
svn:date
V 27
2011-08-15T17:08:54.954747Z
K 7
svn:log
V 214
MFC r224506:
  Add new device id of D-Link DGE-530T Rev. C controller.  DGE-503T
  Rev A1 and B1 is supported by sk(4) but the DGE-530T Rev. C
  controller is re-branded RealTek 8169 controller.

  PR:	kern/159116

END
K 10
svn:author
V 7
yongari
K 8
svn:date
V 27
2011-08-15T17:10:23.437167Z
K 7
svn:log
V 214
MFC r224506:
  Add new device id of D-Link DGE-530T Rev. C controller.  DGE-503T
  Rev A1 and B1 is supported by sk(4) but the DGE-530T Rev. C
  controller is re-branded RealTek 8169 controller.

  PR:	kern/159116

END
K 10
svn:author
V 5
gabor
K 8
svn:date
V 27
2011-08-15T17:23:06.538650Z
K 7
svn:log
V 24
- Clean up header files

END
K 10
svn:author
V 5
gabor
K 8
svn:date
V 27
2011-08-15T18:03:35.984961Z
K 7
svn:log
V 58
- Eliminate code duplication
- TRE-specific style changes

END
K 10
svn:author
V 5
gabor
K 8
svn:date
V 27
2011-08-15T18:04:42.567147Z
K 7
svn:log
V 57
- Use TRE-specific internal macro for better portability

END
K 10
svn:author
V 5
gabor
K 8
svn:date
V 27
2011-08-15T18:07:41.589919Z
K 7
svn:log
V 45
- Use TRE-specific xfree() instead of free()

END
K 10
svn:author
V 5
gabor
K 8
svn:date
V 27
2011-08-15T18:49:42.844251Z
K 7
svn:log
V 30
- Do not limit pattern length

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-08-15T19:29:25.547244Z
K 7
svn:log
V 161
Add device id for the Moxa CP-112UL dual-port serial adapters.

Submitted by:	Jan Mikkelsen  janm of transactionware com
Approved by:	re (kib)
MFC after:	1 week

END
K 10
svn:author
V 5
gibbs
K 8
svn:date
V 27
2011-08-15T20:31:27.593349Z
K 7
svn:log
V 74
Integrate FreeBSD/head into projects/zfsd/head as of SVN revision 224870.

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-16T00:43:25.375151Z
K 7
svn:log
V 113
Add a missing lock arounnd ath_tid->paused, just to be correct.

Update some comments to better reflect reality.

END
K 10
svn:author
V 5
gibbs
K 8
svn:date
V 27
2011-08-16T04:35:42.554199Z
K 7
svn:log
V 626
Correct the rendering of device aliases that reside in subdirectories
of a devfs.

devfs/devfs_vnops.c:
	In devfs_readlink(), convert the devfs root relative path
	of an alias's parent, that is recorded in the alias, into a
	fully qualified path that includes the root of the containing
	devfs.  This avoids the ugliness of generating a relative path
	by prepending "../"'s.  For a non-jailed process, the "symlink
	root" is the devfs's mount point.  For a jailed process, we
	must remove any jail prefix in the mount point so that our
	response matches the user process's world view.

Sponsored by:	Spectra Logic Corporation

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-16T06:26:51.754111Z
K 7
svn:log
V 1108
Begin to flesh out the 11n aggregation fields needed in ath_buf.

The reference driver splits out the descriptor setup and rate series
setup into functions separate from the initial TX path from net80211.

The descriptor decisions are made during the ath_tx_start() / ath_tx_raw_xmit()
calls, but the rate control selection is done at packet scheduling time.
(I'll have to handle raw frames differently, preserving the rate control
being requested by the caller.)

I may not end up using all of these fields, but the basic premise still
holds - ie, in order to do aggregation, I need to delay the descriptor
setup until the aggregates are formed, and then only set the rate
control flags on the first packet of the series. (Then I have to set
the 11n aggregation fields in the first descriptor of each subframe.)

Since I'm going for a unified TX path between legacy and non-legacy
NICs, I'll have to squirrel away 11n and non-11n copies of the
rate control decisions; and then use the correct set when setting
up the multi-rate retry fields (either ath_hal_setupxtxdesc(),
or ath_hal_set11nratescenario()).

END
K 10
svn:author
V 8
hselasky
K 8
svn:date
V 27
2011-08-16T08:34:16.877542Z
K 7
svn:log
V 106
Add missing function to get device speed to the LibUSB v1.0 API.

MFC after:	1 week
Approved by:	re (kib)

END
K 10
svn:author
V 3
kib
K 8
svn:date
V 27
2011-08-16T08:41:09.752166Z
K 7
svn:log
V 120
MFC r224743:
Do not update mountpoint generation counter to the value which was not
yet acted upon by devfs_populate().

END
K 10
svn:author
V 7
delphij
K 8
svn:date
V 27
2011-08-16T08:41:37.680692Z
K 7
svn:log
V 314
Update arcmsr(4) to 1.20.00.22 to solve recursive acquisition of buffer
mutex, which would lead to a deadlock.

Many thanks to Areca for their continued support of FreeBSD.

Submitted by:	Ching Huang <ching2048 areca com tw>
Tested by:	Willem Jan Withagen <wjw digiware nl>
MFC after:	3 days
Approved by:	re (kib)

END
K 10
svn:author
V 3
kib
K 8
svn:date
V 27
2011-08-16T09:06:44.480959Z
K 7
svn:log
V 120
MFC r224743:
Do not update mountpoint generation counter to the value which was not
yet acted upon by devfs_populate().

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-16T09:07:59.419393Z
K 7
svn:log
V 108
Fix BAR frame TX completion - successful transmission is indicated
by a status of 0.

Approved by:	re (kib)

END
K 10
svn:author
V 3
mav
K 8
svn:date
V 27
2011-08-16T10:52:29.318574Z
K 7
svn:log
V 200
MFC r224728:
Do not block zero report ID. It is correct value for devices with single
ID. This fixes USB_SET_IMMED call (synchronous operation) of the uhid(4)
driver on devices with single report ID.

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-16T13:34:34.424963Z
K 7
svn:log
V 171
Break out the descriptor setup code and stuff it into
ath_tx_setds().

The reference code also sets up the descriptor chain
in this function. I'm not doing this just yet.

END
K 10
svn:author
V 8
jonathan
K 8
svn:date
V 27
2011-08-16T14:14:56.921311Z
K 7
svn:log
V 196
poll(2) implementation for capabilities.

When calling poll(2) on a capability, unwrap first and then poll the
underlying object.

Approved by: re (kib), mentor (rwatson)
Sponsored by: Google Inc

END
K 10
svn:author
V 8
jonathan
K 8
svn:date
V 27
2011-08-16T14:23:16.404178Z
K 7
svn:log
V 296
Fix a merge conflict.

r224086 added "goto out"-style error handling to nfssvc_nfsd(), in order
to reliably call NFSEXITCODE() before returning. Our Capsicum changes,
based on the old "return (error)" model, did not merge nicely.

Approved by: re (kib), mentor (rwatson)
Sponsored by: Google Inc

END