ƒ¸b222801
147
534
286
194
670
430
528
288
567
108
132
117
99
3066
144
905
150
224
99
354
326
167
304
410
137
222
696
225
554
199
205
151
394
330
165
231
186
529
529
956
453
156
371
111
164
179
584
149
239
178
226
150
379
1016
110
167
375
181
116
281
313
633
116
157
138
379
765
508
179
422
159
255
359
359
219
903
202
259
116
289
260
132
136
190
320
99
376
140
231
128
125
174
137
137
126
198
300
448
256
365
134
127
124
131
124
115
142
738
226
133
172
99
328
269
485
662
568
1086
913
139
274
478
389
175
245
214
183
435
370
201
199
823
529
265
1580
1580
654
654
681
681
235
235
2181
2181
292
292
363
307

K 10
svn:author
V 6
marcel
K 8
svn:date
V 27
2011-06-07T01:28:12.609539Z
K 7
svn:log
V 438
Fix making kernel dumps from the debugger by creating a command
for it. Do not not expect a developer to call doadump(). Calling
doadump does not necessarily work when it's declared static. Nor
does it necessarily do what was intended in the context of text
dumps. The dump command always creates a core dump.

Move printing of error messages from doadump to the dump command,
now that we don't have to worry about being called from DDB.

END
K 10
svn:author
V 7
davidxu
K 8
svn:date
V 27
2011-06-07T02:50:14.026914Z
K 7
svn:log
V 189
Use p4prio_to_tsprio to calculate TS priority instead of using
p4prio_to_rtpprio which is for RT priority.

PR:	kern/157657
Submitted by:	krivenok.dmitry at gmail dot com
MFC after:	3 days

END
K 10
svn:author
V 6
marcel
K 8
svn:date
V 27
2011-06-07T04:38:33.156742Z
K 7
svn:log
V 99
Lower WARNS level to 3 to eliminate alignment warnings related to
casting inherent in CMSG_DATA().

END
K 10
svn:author
V 3
ken
K 8
svn:date
V 27
2011-06-07T05:04:37.642645Z
K 7
svn:log
V 577
Set pca.p_bufr to NULL when we haven't allocated a buffer.

Otherwise, p_bufr is set to garbage on the stack, and if that garbage
happens to be non-NULL, and the TOLOG or TOCONS flag is set, putbuf()
will get called and attempt to fill the non-existent buffer.

This is really only relevant for tprintf() (and only when the priority is
not -1), but set it in uprintf() and ttyprintf() for completeness.

The next step, to avoid log buffer scrambling, would be to add the
PRINTF_BUFR_SIZE code to tprintf(), but this should prevent panics.

Submitted by:	rmacklem
Found by:	pho

END
K 10
svn:author
V 3
avg
K 8
svn:date
V 27
2011-06-07T06:18:02.022653Z
K 7
svn:log
V 337
amdsbwd: update to support SB8xx southbridges

Many thanks to Tino <tinotom@gmail.com> for drawing my attention to
this, for doing a lot of testing and providing great feedback.
Many thanks to AMD for continuing to release public specifications for
their chipsets.

PR:		kern/157568
Tested by:	Tino <tinotom@gmail.com>
MFC after:	1 week

END
K 10
svn:author
V 2
ae
K 8
svn:date
V 27
2011-06-07T06:42:29.018540Z
K 7
svn:log
V 436
Make a behaviour of the libalias based in-kernel NAT a bit closer to
how natd(8) does work. natd(8) drops packets only when libalias returns
PKT_ALIAS_IGNORED and "deny_incoming" option is set, but ipfw_nat
always did drop packets that were not aliased, even if they should
not be aliased and just are going through.

PR:		kern/122109, kern/129093, kern/157379
Submitted by:	Alexander V. Chernikov (previous version)
MFC after:	1 month

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-06-07T06:45:51.430206Z
K 7
svn:log
V 192
Break out ath_hx_handoff() into hardware and software varieties.
The only software path at the moment is the mcastq which is then
added to the CAB queue (HW) before the beacon is transmitted.

END
K 10
svn:author
V 2
ae
K 8
svn:date
V 27
2011-06-07T06:48:42.085743Z
K 7
svn:log
V 475
Sync ng_nat with recent (r222806) ipfw_nat changes:

  Make a behaviour of the libalias based in-kernel NAT a bit closer to
  how natd(8) does work. natd(8) drops packets only when libalias returns
  PKT_ALIAS_IGNORED and "deny_incoming" option is set, but ipfw_nat
  always did drop packets that were not aliased, even if they should
  not be aliased and just are going through.

Also add SCTP support: mark response packets to skip firewall processing.

MFC after:	1 month

END
K 10
svn:author
V 2
ae
K 8
svn:date
V 27
2011-06-07T06:57:22.330560Z
K 7
svn:log
V 17
Fix indentation.

END
K 10
svn:author
V 7
attilio
K 8
svn:date
V 27
2011-06-07T08:05:23.410962Z
K 7
svn:log
V 36
Bring back the number of CPU to 32.

END
K 10
svn:author
V 7
attilio
K 8
svn:date
V 27
2011-06-07T08:07:41.612096Z
K 7
svn:log
V 21
Add an UPDATE entry.

END
K 10
svn:author
V 7
attilio
K 8
svn:date
V 27
2011-06-07T08:24:29.218050Z
K 7
svn:log
V 4
MFC

END
K 10
svn:author
V 7
attilio
K 8
svn:date
V 27
2011-06-07T08:46:13.422082Z
K 7
svn:log
V 2968
etire the cpumask_t type and replace it with cpuset_t usage.

This is intended to fix the bug where cpu mask objects are
capped to 32.  MAXCPU, then, can now arbitrarely bumped to whatever
value.  Anyway, as long as several structures in the kernel are
statically allocated and sized as MAXCPU, it is suggested to keep it
as low as possible for the time being.

Technical notes on this commit itself:
- More functions to handle with cpuset_t objects are introduced.
  The most notable are cpusetobj_ffs() (which calculates a ffs(3)
  for a cpuset_t object), cpusetobj_strprint() (which prepares a string
  representing a cpuset_t object) and cpusetobj_strscan() (which
  creates a valid cpuset_t starting from a string representation).
- pc_cpumask and pc_other_cpus are target to be removed soon.
  With the moving from cpumask_t to cpuset_t they are now inefficient
  and not really useful.  Anyway, for the time being, please note that
  access to pcpu datas is protected by sched_pin() in order to avoid
  migrating the CPU while reading more than one (possible) word
- Please note that size of cpuset_t objects may differ between kernel
  and userland.  While this is not directly related to the patch itself,
  it is good to understand that concept and possibly use the patch
  as a reference on how to deal with cpuset_t objects in userland, when
  accessing kernland members.
- KTR_CPUMASK is changed and now is represented through a string, to be
  set as the example reported in NOTES.

Please additively note that no MAXCPU is bumped in this patch, but
private testing has been done until to MAXCPU=128 on a real 8x8x2(htt)
machine (amd64).

Please note that the FreeBSD version is not yet bumped because of
the upcoming pcpu changes.  However, note that this patch is not
targeted for MFC.

People to thank for the time spent on this patch:
- sbruno, pluknet and Nicholas Esborn (nick AT desert DOT net) tested
  several revision of the patches and really helped in improving
  stability of this work.
- marius fixed several bugs in the sparc64 implementation and reviewed
  patches related to ktr.
- jeff and jhb discussed the basic approach followed.
- kib and marcel made targeted review on some specific part of the
  patch.
- marius, art, nwhitehorn and andreast reviewed MD specific part of
  the patch.
- marius, andreast, gonzo, nwhitehorn and jceel tested MD specific
  implementations of the patch.
- Other people have made contributions on other patches that have been
  already committed and have been listed separately.

Companies that should be mentioned for having participated at several
degrees:
- Yahoo! for having offered the machines used for testing on big
  count of CPUs.
- The FreeBSD Foundation for having sponsored my devsummit attendance,
  which has been instrumental.
- Sandvine for having offered offices and infrastructure during
  development.

(I really hope I didn't forget anyone, if it happened I apologize in
advance).

END
K 10
svn:author
V 7
attilio
K 8
svn:date
V 27
2011-06-07T09:02:25.931916Z
K 7
svn:log
V 48
After reintegration the branch must be removed.

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-06-07T09:03:28.061739Z
K 7
svn:log
V 809
Flesh out a new HAL method to fetch the radar PHY error frame information.

For the AR5211/AR5212, this is apparently a one byte pulse duration
counter value. It is only coded up here for the AR5212 as I don't have
any AR5211-series hardware to test it on.

This information was extracted from the Madwifi DFS branch along with
some local additions.

Please note - all this does is extract out the radar event duration,
it in no way reflects the presence of a radar. Further code is needed
to take a set of radar events and filter them to extract out correct
radar pulse trains (and ignore other events.)

For further information, please see:

http://wiki.freebsd.org/dev/ath_hal%284%29/RadarDetection

This includes references to the relevant patents which describe what
is going on.

Obtained from:	Madwifi

END
K 10
svn:author
V 7
attilio
K 8
svn:date
V 27
2011-06-07T09:03:41.308113Z
K 7
svn:log
V 54
Regenerate the largeSMP branch for finishing the job.

END
K 10
svn:author
V 8
jlaffaye
K 8
svn:date
V 27
2011-06-07T09:06:15.485815Z
K 7
svn:log
V 126
- Add myself to calendar.freebsd
- Add the forgotten relation to mentors in committers-ports.dot

Approved by:	bapt@ (mentor)

END
K 10
svn:author
V 7
attilio
K 8
svn:date
V 27
2011-06-07T09:37:08.415919Z
K 7
svn:log
V 4
MFC

END
K 10
svn:author
V 2
ae
K 8
svn:date
V 27
2011-06-07T11:11:11.677505Z
K 7
svn:log
V 262
Do not use LCM from stripesize and user specified alignment value.
When user wants have specific alignment - do what user wants.
Use stripesize as alignment value in case, when some of gpart's
arguments are ommitted for automatic calculation.

Suggested by:	mav

END
K 10
svn:author
V 3
hrs
K 8
svn:date
V 27
2011-06-07T12:00:29.455913Z
K 7
svn:log
V 233
- Use ELM_MALLOC() for struct rainfo.

- Fix a missing back pointer assignment in struct prefix to struct rainfo
  when addr="" is specified.  This caused SIGSEGV.

- Insert a prefix element to a tail queue after setting parameters.

END
K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-06-07T14:00:47.185896Z
K 7
svn:log
V 72
Since HAL_PHYERR_* is used in the radar code, always include ah_desc.h.

END
K 10
svn:author
V 10
nwhitehorn
K 8
svn:date
V 27
2011-06-07T14:35:18.357022Z
K 7
svn:log
V 203
MFC r222430:
Require an error instead of a timeout to decide the new-style fan
commands won't work. This prevents a busy system from making smu(4)
suddenly decide its fans use the old-style command set.

END
K 10
svn:author
V 10
nwhitehorn
K 8
svn:date
V 27
2011-06-07T14:36:13.407553Z
K 7
svn:log
V 309
MFC r222666:
Fix a typo derived from a mismerge from mmu_oea that would cause
pmap_sync_icache() to sync random (possibly uncached or nonexisting!)
memory, causing kernel page faults or machine checks, most easily
triggered by using GDB. While here, add an additional safeguard to only
sync cacheable memory.

END
K 10
svn:author
V 3
hrs
K 8
svn:date
V 27
2011-06-07T15:40:17.973473Z
K 7
svn:log
V 45
Fix a bug that prevents tc=xxx from working.

END
K 10
svn:author
V 2
jh
K 8
svn:date
V 27
2011-06-07T16:23:27.193466Z
K 7
svn:log
V 130
Remove a gratuitous newline. This should be a non-functional change.

PR:		bin/153667
Submitted by:	Craig Leres
MFC after:	1 week

END
K 10
svn:author
V 3
mav
K 8
svn:date
V 27
2011-06-07T17:01:52.115521Z
K 7
svn:log
V 603
Make automatic hw.snd.default_unit choice a bit more intelligent. Instead
of just setting it to the first registered device, reevaluate it for each
device registered, trying to choose best candidate, unless one was forced.
For now use such preference order: play&rec, play, rec.

As side effect, this should workaround the situation when HDMI audio output
of the video card, usually not connected to anything, becomes default, that
requires manual user intervention to make sound working. If at some point
this won't be enough, we can try to fetch some additional priority flags
from the device driver.

END
K 10
svn:author
V 6
marius
K 8
svn:date
V 27
2011-06-07T17:19:14.066151Z
K 7
svn:log
V 129
Fix a problem with r222813; given that we may only operate on interrupt
globals here but clobber %y save and restore the latter.

END
K 10
svn:author
V 6
marius
K 8
svn:date
V 27
2011-06-07T17:33:39.294254Z
K 7
svn:log
V 458
Adapt CATR() to r222813. This is somewhat tricky as we can't afford using
more than three temporary register in several places CATR() is used so
this code trades instructions in for registers. Actually, this still isn't
sufficient and CATR() has the side-effect of clobbering %y. Luckily, with
the current uses of CATR() this either doesn't matter or we are able to
(save and) restore it.
Now that there's only one use of AND() and TEST() left inline these.

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-07T17:40:33.659493Z
K 7
svn:log
V 106
Log the socket address passed as the destination to sendto() and sendmsg()
via ktrace.

MFC after:	1 week

END
K 10
svn:author
V 6
grehan
K 8
svn:date
V 27
2011-06-07T18:35:45.353786Z
K 7
svn:log
V 109
Allow access to the device's config area with any size i/o access at any
offset. This is now spec-compliant.

END
K 10
svn:author
V 7
dchagin
K 8
svn:date
V 27
2011-06-07T18:43:02.608899Z
K 7
svn:log
V 55
MFC r219242:

Print out shared flag for debug purpose.

END
K 10
svn:author
V 7
delphij
K 8
svn:date
V 27
2011-06-07T18:48:49.465451Z
K 7
svn:log
V 297
Add a special mount option "failok" to indicate that the administrator wants
the system to proceed to boot without bailing out into single user mode,
even when the file system can not be successfully mounted.

This option is implemented in mount(8) and not passed into kernel.

MFC after:	1 month

END
K 10
svn:author
V 2
bz
K 8
svn:date
V 27
2011-06-07T19:39:34.143525Z
K 7
svn:log
V 238
For the moment document the possible problem introduced with dynamic address
family detection in world, mostly noticed by ifconfig(8), when running with
an old kernel.

Reported by:	Andrzej Tobola (ato iem.pw.edu.pl)
Reported by:	gcooper

END
K 10
svn:author
V 3
zec
K 8
svn:date
V 27
2011-06-07T20:46:03.187830Z
K 7
svn:log
V 73
Set curvnet context in a callout-trigerred code path.

MFC after:	3 days

END
K 10
svn:author
V 2
mm
K 8
svn:date
V 27
2011-06-07T20:46:31.316935Z
K 7
svn:log
V 139
Silence notice on pool creation, import and access.

Suggested by:	Jeremy Chadwick (freebsd-stable@)
Discussed with:	pjd
MFC after:	1 week

END
K 10
svn:author
V 5
gibbs
K 8
svn:date
V 27
2011-06-07T20:49:29.064634Z
K 7
svn:log
V 92
Create a project branch for intergration of the ZFS fault monitoring and
management daemon.

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-07T21:15:43.657301Z
K 7
svn:log
V 436
MFC 221138:
Only align MSI message groups based on the number of messages being
allocated, not the maximum number of messages the device supports.  The
spec only requires the former, and I believe I implemented the latter due
to misunderstanding an e-mail.  In particular, this fixes an issue where
having several devices that all support 16 messages can run out of
IDT vectors on x86 even though the driver only uses a single message.

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-07T21:16:02.664126Z
K 7
svn:log
V 436
MFC 221138:
Only align MSI message groups based on the number of messages being
allocated, not the maximum number of messages the device supports.  The
spec only requires the former, and I believe I implemented the latter due
to misunderstanding an e-mail.  In particular, this fixes an issue where
having several devices that all support 16 messages can run out of
IDT vectors on x86 even though the driver only uses a single message.

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-07T21:48:36.760361Z
K 7
svn:log
V 863
MFC 219968:
Fix some locking nits with the p_state field of struct proc:
- Hold the proc lock while changing the state from PRS_NEW to PRS_NORMAL
  in fork to honor the locking requirements.  While here, expand the scope
  of the PROC_LOCK() on the new process (p2) to avoid some LORs.  Previously
  the code was locking the new child process (p2) after it had locked the
  parent process (p1).  However, when locking two processes, the safe order
  is to lock the child first, then the parent.
- Fix various places that were checking p_state against PRS_NEW without
  having the process locked to use PROC_LOCK().  Every place was already
  locking the process, just after the PRS_NEW check.
- Remove or reduce the use of PROC_SLOCK() for places that were checking
  p_state against PRS_NEW.  The PROC_LOCK() alone is sufficient for reading
  the current state.

END
K 10
svn:author
V 6
marius
K 8
svn:date
V 27
2011-06-07T23:15:21.858340Z
K 7
svn:log
V 357
- For the case when tl1_align(_trap) is used to call rsf_fatal via
  RSF_FATAL we need to switch to alternate globals for KSTACK_CHECK just
  like tl1_data_excptn(_trap) does. This is more or less cosmetic because
  in case RSF_FATAL is called we're already heading south.
- Correct an END().
- Read the window state from the correct register for a CATR().

END
K 10
svn:author
V 7
delphij
K 8
svn:date
V 27
2011-06-08T00:31:10.502897Z
K 7
svn:log
V 60
Use .Dq instead of hardcoding `` and ''.

Suggested by:	gjb

END
K 10
svn:author
V 8
rmacklem
K 8
svn:date
V 27
2011-06-08T01:14:31.727892Z
K 7
svn:log
V 273
MFC: r222289
Fix the new NFS client so that it correctly sets the "must_commit"
argument for a write RPC when it succeeds for the first one and
fails for a subsequent RPC within the same call to the function.
This makes it compatible with the old NFS client for this case.

END
K 10
svn:author
V 7
delphij
K 8
svn:date
V 27
2011-06-08T01:18:47.455240Z
K 7
svn:log
V 15
MFC less v443.

END
K 10
svn:author
V 8
rmacklem
K 8
svn:date
V 27
2011-06-08T01:28:46.493527Z
K 7
svn:log
V 67
MFC: r222291
Add some missing mutex locking to the new NFS client.

END
K 10
svn:author
V 2
bz
K 8
svn:date
V 27
2011-06-08T03:02:11.567657Z
K 7
svn:log
V 88
Correct comments and debug logging in ipsec to better match reality.

MFC after:	3 days

END
K 10
svn:author
V 7
dchagin
K 8
svn:date
V 27
2011-06-08T03:45:40.727216Z
K 7
svn:log
V 487
MFC r219405:
Extend struct sysvec with new method sv_schedtail, which is used for an
explicit process at fork trampoline path instead of eventhadler(schedtail)
invocation for each child process.

Remove eventhandler(schedtail) code and change linux ABI to use newly added
sysvec method.

While here replace explicit comparing of module sysentvec structure with the
newly created process sysentvec to detect the linux ABI.

MFC r219421:
Remove bogus since r219405 check of the Linux ABI.

END
K 10
svn:author
V 7
dchagin
K 8
svn:date
V 27
2011-06-08T03:55:58.749411Z
K 7
svn:log
V 53
MFC r219558:

Style(9) fixes. No functional changes.

END
K 10
svn:author
V 3
hrs
K 8
svn:date
V 27
2011-06-08T04:06:56.348296Z
K 7
svn:log
V 146
- Disable "resolvconf -d" temporarily to avoid extra invocations of the script
  under a certain condition.
- Fix argv handling.

Spotted by:	ume

END
K 10
svn:author
V 2
ae
K 8
svn:date
V 27
2011-06-08T04:50:23.243379Z
K 7
svn:log
V 87
MFC r222559:
  Hide useless warning under debug macro.

  PR:		kern/69963, kern/139226

END
K 10
svn:author
V 2
ae
K 8
svn:date
V 27
2011-06-08T04:54:22.306349Z
K 7
svn:log
V 134
MFC r222559:
  Hide useless warning under debug macro.

  PR:           kern/69963, kern/139226

  This is direct commit to stable/7.

END
K 10
svn:author
V 2
ae
K 8
svn:date
V 27
2011-06-08T04:57:14.200237Z
K 7
svn:log
V 59
MFC r222560:
  Hide some debug messages under debug macro.

END
K 10
svn:author
V 2
ru
K 8
svn:date
V 27
2011-06-08T08:08:42.478904Z
K 7
svn:log
V 287
Pull up all vendor changes to mdoc(7).

This also replaces the local fix in r219209 that made .Ac emit
ASCII angle quotes with an official fix.  In the official fix,
ASCII quotes are output when using the .Aq, .Ao and .Ac calls,
but only when nested into the .An macro.

PR:		gnu/154822

END
K 10
svn:author
V 3
avg
K 8
svn:date
V 27
2011-06-08T08:12:15.246138Z
K 7
svn:log
V 923
remove code for dynamic offlining/onlining of CPUs on x86

The code has definitely been broken for SCHED_ULE, which is a default
scheduler.  It may have been broken for SCHED_4BSD in more subtle ways,
e.g. with manually configured CPU affinities and for interrupt devilery
purposes.
We still provide a way to disable individual CPUs or all hyperthreading
"twin" CPUs before SMP startup.  See the UPDATING entry for details.

Interaction between building CPU topology and disabling CPUs still
remains fuzzy: topology is first built using all availble CPUs and then
the disabled CPUs should be "subtracted" from it.  That doesn't work
well if the resulting topology becomes non-uniform.

This work is done in cooperation with Attilio Rao who in addition to
reviewing also provided parts of code.

PR:		kern/145385
Discussed with:	gcooper, ambrisko, mdf, sbruno
Reviewed by:	attilio
Tested by:	pho, pluknet
X-MFC after:	never

END
K 10
svn:author
V 2
ru
K 8
svn:date
V 27
2011-06-08T08:16:32.178474Z
K 7
svn:log
V 19
Added FreeBSD 8.3.

END
K 10
svn:author
V 2
ru
K 8
svn:date
V 27
2011-06-08T08:22:54.280442Z
K 7
svn:log
V 76
Sync Groff with what's in HEAD except for not re-enabling ANSI SGR support.

END
K 10
svn:author
V 2
bz
K 8
svn:date
V 27
2011-06-08T10:59:36.184686Z
K 7
svn:log
V 283
Add the missing call to ip6_ipsec_filtertunnel() to be able to control
whether decapsulated IPsec packets will be passed to pfil again depending
on the setting of the net.ip6.ipsec6.filtertunnel sysctl.

PR:		kern/157670
Submitted by:	Manuel Kasper (mk neon1.net)
MFC after:	2 weeks

END
K 10
svn:author
V 8
netchild
K 8
svn:date
V 27
2011-06-08T11:36:06.533701Z
K 7
svn:log
V 84
MFC r221430:
  Use proper return codes (valuable output, invalid config, problems).

END
K 10
svn:author
V 3
hrs
K 8
svn:date
V 27
2011-06-08T12:03:34.365635Z
K 7
svn:log
V 24
Merge from HEAD@222848.

END
K 10
svn:author
V 10
nwhitehorn
K 8
svn:date
V 27
2011-06-08T13:23:35.632615Z
K 7
svn:log
V 180
Compile RTLD with global dot symbols on 64-bit PowerPC, as a crutch for
GDB's ability to locate r_debug_state (which is actually the only function
that need be compiled this way).

END
K 10
svn:author
V 8
andreast
K 8
svn:date
V 27
2011-06-08T16:00:30.975419Z
K 7
svn:log
V 215
- Improve error handling.
- Add retry loops in the i2c read/write functions.
- Combied the ADC channel selection and readout of the value into
  one iicbus_transfer to avoid possible races.

Reviewed by: nwhitehorn

END
K 10
svn:author
V 3
hrs
K 8
svn:date
V 27
2011-06-08T16:03:29.408011Z
K 7
svn:log
V 540
- Accumulate RA options instead of replacing old ones when a new RA arrived.
  RFC 4861 6.3.4 clearly defines handling multiple RAs in this way.

- RDNSS/DNSSL options from multiple RAs on a single link will be
  gathered and sent to resolvconf(8).

- Call "resolvconf -d" only after at least one RDNSS or DNSSL option is
  received and then all of them are expired.

- The rtsold.dump output now supports displaying a list of the RA options.

- Use more human-readable expression for logging values of struct timeval.

Discussed with:	ume

END
K 10
svn:author
V 3
hrs
K 8
svn:date
V 27
2011-06-08T16:06:27.453765Z
K 7
svn:log
V 24
Merge from HEAD@222861.

END
K 10
svn:author
V 3
hrs
K 8
svn:date
V 27
2011-06-08T16:07:44.202357Z
K 7
svn:log
V 65
Revert bogus changes accidentally merged in the previous commit.

END
K 10
svn:author
V 4
jkim
K 8
svn:date
V 27
2011-06-08T18:52:42.168656Z
K 7
svn:log
V 45
Remove a redundant assignment since r221703.

END
K 10
svn:author
V 7
attilio
K 8
svn:date
V 27
2011-06-08T19:28:59.370287Z
K 7
svn:log
V 282
In the current code, a double panic condition may lead to dumps
interleaving.
Signal dumping to happen only for the first panic which should be the
most important.

Sponsored by:	Sandvine Incorporated
Submitted by:	Nima Misaghian (nmisaghian AT sandvine DOT com)
MFC after:	2 weeks

END
K 10
svn:author
V 4
jkim
K 8
svn:date
V 27
2011-06-08T19:38:31.975813Z
K 7
svn:log
V 671
Introduce low-resolution TSC timecounter "TSC-low".  It replaces the normal
TSC timecounter if TSC frequency is higher than ~4.29 MHz (or 2^32-1 Hz) or
multiple CPUs are present.  The "TSC-low" frequency is always lower than a
preset maximum value and derived from TSC frequency (by being halved until
it becomes lower than the maximum).  Note the maximum value for SMP case is
significantly lower than UP case because we want to reduce (rare but known)
"temporal anomalies" caused by non-serialized RDTSC instruction.  Normally,
it is still higher than "ACPI-fast" timecounter frequency (which was default
timecounter hardware for long time until r222222) to be useful.

END
K 10
svn:author
V 2
mm
K 8
svn:date
V 27
2011-06-08T19:45:47.085232Z
K 7
svn:log
V 416
MFC r218056:

Fix bug in the netfront driver that caused excessive packet drops during
receive processing.

Remove unnecessary restrictions on the mbuf chain length built during an
LRO receive.  This restriction was copied from the Linux netfront driver
where the LRO implementation cannot handle more than 18 discontinuities.
The FreeBSD implementation has no such restriction.

Approved by:	gibbs (private e-mail)

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-08T19:58:20.275193Z
K 7
svn:log
V 87
MFC 220390:
Fix several places to ignore processes that are not yet fully constructed.

END
K 10
svn:author
V 4
jkim
K 8
svn:date
V 27
2011-06-08T20:08:06.207802Z
K 7
svn:log
V 328
Increase quality of TSC (or TSC-low) timecounter to 1000 if it is P-state
invariant.  For SMP case (TSC-low), it also has to pass SMP synchronization
test and the CPU vendor/model has to be white-listed explicitly.  Currently,
all Intel CPUs and single-socket AMD Family 15h processors are listed here.

Discussed with:	hackers

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-08T20:12:38.984684Z
K 7
svn:log
V 67
MFC 221179:
Trim an additional unnecessary <linker_set.h> include.

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-08T20:57:22.641497Z
K 7
svn:log
V 162
MFC 221182,222760:
Add support for Oxford PCI Express Expresso family devices.
For these devices, the number of supported ports is read from a register
in BAR 0.

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-08T21:10:44.795792Z
K 7
svn:log
V 266
MFC 221208:
Add a 'show progress' command that shows a summary of all in-progress
commands for a given adapter.  Specifically, it shows the status of any
drive or volume activities currently in progress similar to the
'drive process' and 'volume progress' commands.

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-08T21:10:55.458384Z
K 7
svn:log
V 266
MFC 221208:
Add a 'show progress' command that shows a summary of all in-progress
commands for a given adapter.  Specifically, it shows the status of any
drive or volume activities currently in progress similar to the
'drive process' and 'volume progress' commands.

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-08T21:14:04.394513Z
K 7
svn:log
V 126
MFC 221326:
Add an entry for the Kuroutoshikou SERIAL4P-LPPCI2 which uses an Oxford
4 port chip but with a nonstandard clock.

END
K 10
svn:author
V 7
davidch
K 8
svn:date
V 27
2011-06-08T21:18:14.736856Z
K 7
svn:log
V 806
- Major reorganization of mbuf handling throughout the driver to
  increase robustness (no more calls to panic(9)) and simplify
  code.
- Allocate RX/TX data structures as a single buffer rather than
  an array of 4KB pages to simplify code.
- Fixed LRO (aka TPA) code.  Removed kernel module parameter and
  support enabling disabling LRO through ifconfig(8) command line.
  LRO is still disabled by default but should be enabled for best
  performance on an endpoint device.
- Fixed statistcs code and removed kernel module parameter (stats
  should just work).
- Added many software counters to help identify the cause of some
  performance issues.
- Streamlined adapter internal init/stop code paths.
- Fiddled with debug code (adding some here, removing some there).
- Continued style(9) adjustments.

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-08T21:23:18.809792Z
K 7
svn:log
V 109
MFC 221731:
Add an entry for the SIIG Quartet Serial 850 which uses an Oxford
chip with a non-default clock.

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-08T21:29:12.621736Z
K 7
svn:log
V 166
MFC 222093:
Add support for the SIIG Cyber 2S PCIe adapter.  It is based on an
Oxford Semiconductor OX16PCI954 but uses only two ports with a non-default
clock rate.

END
K 10
svn:author
V 6
obrien
K 8
svn:date
V 27
2011-06-08T21:29:33.442202Z
K 7
svn:log
V 21
Simply post r222273.

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-08T21:32:14.799195Z
K 7
svn:log
V 196
MFC 222328:
For Timedia multiport serial adapters, the first two ports use a SUN1889
which uses a non-standard clock (* 8) while any additional ports use
SUN1699 chips which use a standard clock.

END
K 10
svn:author
V 7
dchagin
K 8
svn:date
V 27
2011-06-08T21:36:36.265573Z
K 7
svn:log
V 163
MFC r215013 by mdf@:

Whitespace and other aspects of style(9).  No functional changes.

MFC r220730:

Remove malloc(9) return value checks when M_WAITOK is used.

END
K 10
svn:author
V 7
dchagin
K 8
svn:date
V 27
2011-06-08T21:39:00.173176Z
K 7
svn:log
V 36
MFC r221784:

Remove wrong comment.

END
K 10
svn:author
V 6
jilles
K 8
svn:date
V 27
2011-06-08T21:58:19.074163Z
K 7
svn:log
V 41
sh: Add simple tests for set -x and PS4.

END
K 10
svn:author
V 3
ken
K 8
svn:date
V 27
2011-06-08T21:59:07.232115Z
K 7
svn:log
V 98
Add dump.c to the rtsol build.  It is needed now that sec2str is non-static
and used in rtsold.c.

END
K 10
svn:author
V 4
jkim
K 8
svn:date
V 27
2011-06-08T23:44:59.710826Z
K 7
svn:log
V 226
Tidy up r222866.

- Re-add accidentally removed atomic op. for sysctl(9) handler.
- Remove a period(`.') at the end of a debugging message.
- Consistently spell "low" for "TSC-low" timecounter throughout.

Pointed out by:	bde

END
K 10
svn:author
V 7
attilio
K 8
svn:date
V 27
2011-06-09T00:34:56.154382Z
K 7
svn:log
V 4
MFC

END
K 10
svn:author
V 8
rmacklem
K 8
svn:date
V 27
2011-06-09T01:10:39.832863Z
K 7
svn:log
V 278
MFC: r222329
Add a check for MNTK_UNMOUNTF at the beginning of nfs_sync()
in the new NFS client so that a forced dismount doesn't
get stuck in the VFS_SYNC() call that happens before
VFS_UNMOUNT() in dounmount().
Additional changes are needed before forced dismounts will work.

END
K 10
svn:author
V 7
dchagin
K 8
svn:date
V 27
2011-06-09T05:21:08.147923Z
K 7
svn:log
V 44
MFC r219776:

Remove a now unused variable.

END
K 10
svn:author
V 7
dchagin
K 8
svn:date
V 27
2011-06-09T05:23:28.792997Z
K 7
svn:log
V 134
MFC r219783, r219786:

A bit rearranged rtalloc1_fib() code.
To avoid code duplication move the miss label to line up and jump on it.

END
K 10
svn:author
V 7
dchagin
K 8
svn:date
V 27
2011-06-09T05:24:28.336560Z
K 7
svn:log
V 32
MFC r219791:

Remove dead code.

END
K 10
svn:author
V 5
dougb
K 8
svn:date
V 27
2011-06-09T06:10:39.793026Z
K 7
svn:log
V 31
Minor typo, fix 2 attributions

END
K 10
svn:author
V 3
avg
K 8
svn:date
V 27
2011-06-09T08:25:56.987567Z
K 7
svn:log
V 82
MFC r221527: prepare code that does topology detection for amd cpus for bulldozer

END
K 10
svn:author
V 3
avg
K 8
svn:date
V 27
2011-06-09T08:30:46.076511Z
K 7
svn:log
V 45
MFC r222720,222721: corrections to amdsbwd.4

END
K 10
svn:author
V 3
avg
K 8
svn:date
V 27
2011-06-09T08:33:42.987937Z
K 7
svn:log
V 45
MFC r222720,222721: corrections to amdsbwd.4

END
K 10
svn:author
V 2
mm
K 8
svn:date
V 27
2011-06-09T10:04:11.597626Z
K 7
svn:log
V 35
MFC r222757:

Remove empty #ifndef

END
K 10
svn:author
V 2
ae
K 8
svn:date
V 27
2011-06-09T10:43:36.785405Z
K 7
svn:log
V 106
MFC r222603:
  Do not hide stripeoffset from libgeom(3), it may be useful even when
  stripesize is zero.

END
K 10
svn:author
V 6
cognet
K 8
svn:date
V 27
2011-06-09T11:54:11.902504Z
K 7
svn:log
V 204
Please welcome Grzegorz Bernacki (gber@FreeBSD.org). Grzegorz has been doing stuff on arm and ppc for quite some time, and is working on some cool stuff, including SMP support for arm.

Approved by:	core

END
K 10
svn:author
V 3
mav
K 8
svn:date
V 27
2011-06-09T16:30:13.741782Z
K 7
svn:log
V 355
Intel NM10 chipset's SATA controller has same PCI ID and revision as ICH7's,
but has only 2 SATA ports instead of 4. The worst part is that SStatus and
SError registers for missing ports are not implemented and return wrong
values (0xffffffff), that caused infinite reset loop.

Just ignore that SError value while I found no better way to identify them.

END
K 10
svn:author
V 10
nwhitehorn
K 8
svn:date
V 27
2011-06-09T19:47:30.982528Z
K 7
svn:log
V 155
Add -Wa,-many to CFLAGS on PowerPC. This aids in building a kernel using
clang, which would otherwise complain about some 64-bit bridge mode
instructions.

END
K 10
svn:author
V 2
bz
K 8
svn:date
V 27
2011-06-09T19:52:28.927668Z
K 7
svn:log
V 273
Contrary to when returning in all-good cases at the end of functions we
did not free memory (1) or close the file descriptor (2) in error cases.

Reported by:	Mark Johnston (1)
Reported by:	attilio (2)
Reviewed by:	jhb
Sponsored by:	Sandvine Incorporated
MFC after:	1 week

END
K 10
svn:author
V 2
np
K 8
svn:date
V 27
2011-06-09T20:21:45.961341Z
K 7
svn:log
V 43
cxgbetool: a tool for the cxgbe(4) driver.

END
K 10
svn:author
V 6
marcel
K 8
svn:date
V 27
2011-06-09T21:29:15.873461Z
K 7
svn:log
V 32
Bootstrap the llvm-ia64 branch.

END
K 10
svn:author
V 6
marcel
K 8
svn:date
V 27
2011-06-09T22:17:27.239428Z
K 7
svn:log
V 29
Add a skeletal IA64 backend.

END
K 10
svn:author
V 6
marcel
K 8
svn:date
V 27
2011-06-09T22:38:38.434130Z
K 7
svn:log
V 36
Add file missed in previous commit.

END
K 10
svn:author
V 7
delphij
K 8
svn:date
V 27
2011-06-09T22:50:10.097517Z
K 7
svn:log
V 28
Vendor import of less v444.

END
K 10
svn:author
V 7
delphij
K 8
svn:date
V 27
2011-06-09T22:51:05.279682Z
K 7
svn:log
V 19
Tag for less v444.

END
K 10
svn:author
V 7
delphij
K 8
svn:date
V 27
2011-06-09T22:52:57.312172Z
K 7
svn:log
V 46
MFV: Update to less v444.

MFC after:	2 weeks

END
K 10
svn:author
V 6
jilles
K 8
svn:date
V 27
2011-06-09T23:12:23.679196Z
K 7
svn:log
V 642
sh: Do parameter expansion before printing PS4 (set -x).

The function name expandstr() and the general idea of doing this kind of
expansion by treating the text as a here document without end marker is from
dash.

All variants of parameter expansion and arithmetic expansion also work (the
latter is not required by POSIX but it does not take extra code and many
other shells also allow it).

Command substitution is prevented because I think it causes too much code to
be re-entered (for example creating an unbounded recursion of trace lines).

Unfortunately, our LINENO is somewhat crude, otherwise PS4='$LINENO+ ' would
be quite useful.

END
K 10
svn:author
V 7
rodrigc
K 8
svn:date
V 27
2011-06-10T01:13:15.297170Z
K 7
svn:log
V 129
Obtain bswap64() function implementation from
version 1.3 of src/common/lib/libc/gen/bswap64.c in NetBSD.

Obtained from: NetBSD

END
K 10
svn:author
V 7
delphij
K 8
svn:date
V 27
2011-06-10T01:15:23.584813Z
K 7
svn:log
V 37
MFC r221793: nc(1) from OpenBSD 4.9.

END
K 10
svn:author
V 6
marcel
K 8
svn:date
V 27
2011-06-10T02:07:21.523344Z
K 7
svn:log
V 77
Teach Clang about ia64. Should be about right for as much as it is
complete.

END
K 10
svn:author
V 7
attilio
K 8
svn:date
V 27
2011-06-10T02:40:01.309443Z
K 7
svn:log
V 4
MFC

END
K 10
svn:author
V 6
emaste
K 8
svn:date
V 27
2011-06-10T03:27:25.162645Z
K 7
svn:log
V 232
MFC r222589:

  There are a couple of structs in mfireg.h with members named 'class'.
  These cause problems when trying to include the header in a C++ project.
  Rename them to 'evt_class', and track the change in mfi and mfiutil.

END
K 10
svn:author
V 2
ru
K 8
svn:date
V 27
2011-06-10T06:11:33.151380Z
K 7
svn:log
V 177
MFC r222647:

Don't pass -o1- to groff(1) by default.  If ms(7) formatted document
uses the .RP macro, a separate cover page is created as page 0 which
is not otherwise output.

END
K 10
svn:author
V 8
netchild
K 8
svn:date
V 27
2011-06-10T07:54:38.120294Z
K 7
svn:log
V 387
MFC r221423, r221424, r221426, r221428, r221434:

  This is v4l2 support for the linuxulator. This allows to access FreeBSD
  native devices which support the v4l2 API from processes running within
  the linuxulator, e.g. skype or flash can access the multimedia/pwcbsd
  or multimedia/webcamd supplied drivers.

MFC r221425:

  Fix typo in comment, improve comment.

MFC tested by:	nox

END
K 10
svn:author
V 3
mav
K 8
svn:date
V 27
2011-06-10T08:32:33.029182Z
K 7
svn:log
V 569
MFC r220929:
According to ATA specifications, when ATAPI master is the only device, it
should respond with all zeroes to any access to slave registers. Test with
PATA devices confirmed such behavior. Unluckily, Intel SATA controllers in
legacy emulation mode behave differently, not making any difference between
ATA and ATAPI devices. It causes false positive slave device detection and,
as result, command timeouts.

To workaround this problem, mask result of legacy-emulated soft-reset with
the device presence information received from the SATA-specific registers.

END
K 10
svn:author
V 3
mav
K 8
svn:date
V 27
2011-06-10T08:49:38.541129Z
K 7
svn:log
V 475
MFC rr222039:
Add support for "LED" enclosure management messages, defined by the AHCI.

When supported by hardware, this allows to control per-port activity, locate
and fault LEDs via the led(4) API for localization and status reporting
purposes. Supporting AHCI controllers may transmit that information to the
backplane controllers via SGPIO interface. Backplane controllers interpret
received statuses in some way (IBPI standard) to report them using present
indicators.

END
K 10
svn:author
V 3
mav
K 8
svn:date
V 27
2011-06-10T08:54:52.640170Z
K 7
svn:log
V 993
MFC r222285:
According to SATA specification, when Serial ATA Enclosure Management Bridge
(SEMB) is unable to communicate to Storage Enclosure Processor (SEP), in
response to hard and soft resets it should among other things return value
0x7F in Status register. The weird side is that it means DRQ bit set, which
tells that reset request is not completed. It would be fine if SEMB was the
only device on port. But if SEMB connected to PMP or built into it, it may
block access to other devices sharing same SATA port.

Make some tunings/fixes to soft-reset handling to workaround the issue:
 - ahci(4): request CLO on the port after soft reset to ignore DRQ bit;
 - siis(4): gracefully reinitialize port after soft reset timeout (hardware
doesn't detect reset request completion in this case);
 - mvs(4): if PMP is used, send dummy soft-reset to the PMP port to make it
clear DRQ bit for us.

For now this makes quirks in ata_pmp.c, hiding SEMB ports of SiI3726/SiI4726
PMPs, less important.

END
K 10
svn:author
V 3
mav
K 8
svn:date
V 27
2011-06-10T08:58:13.188878Z
K 7
svn:log
V 820
MFC r222304:
Marvell 88SE91xx controllers are known to report soft-reset completion
without waiting for device readiness (or at least not updating FIS receive
area in time). To workaround that, special quirk was added earlier to wait
for the FIS receive area update. But it was found that under same PCI ID
0x91231b4b and revision 0x11 there are two completely different chip
versions (firmware?): HBA and RAID. The problem is that RAID version in
some cases, such as hot-plug, does not update FIS receive area at all!

To workaround that, differentiate the chip versions by their capabilities,
and, if RAID version found, skip FIS receive area update waiting and read
device signature from the PxSIG register instead. This method doesn't work
for HBA version when PMP attached, so keep using previous workaround there.

END
K 10
svn:author
V 3
mav
K 8
svn:date
V 27
2011-06-10T08:59:30.314618Z
K 7
svn:log
V 47
MFC r222306:
Add Marvell 88SE9172 chip PCI ID.

END
K 10
svn:author
V 3
mav
K 8
svn:date
V 27
2011-06-10T09:12:09.912788Z
K 7
svn:log
V 181
MFC r221101:
Implement relaxed comparision for hardcoded provider names to make it
ignore adX/adaY difference in both directions to simplify migration to
the CAM-based ATA or back.

END
K 10
svn:author
V 3
mav
K 8
svn:date
V 27
2011-06-10T09:18:43.740003Z
K 7
svn:log
V 385
MFC r222652:
Update disk's stripesize and stripeoffset parameters on provider open.
They are media-dependent and may change in run-time, same as sectorsize
and/or mediasize.

SCSI devices return physical sector size and offset via READ CAPACITY(16)
command and so can not report it until media inserted or at least until
probe sequence completed. UNMAP support is also reported there.

END
K 10
svn:author
V 3
mav
K 8
svn:date
V 27
2011-06-10T09:51:15.911330Z
K 7
svn:log
V 296
MFC r206112, r208928 (by mjacob):
Rearrange how things are done to avoid dereferencing stale pointers in
the case of immediate unconfigure after configure. Hold the periph an
extra count while we have the task to create sysctl context outstanding
so that the periph doesn't go away unexpectedly.

END
K 10
svn:author
V 3
mav
K 8
svn:date
V 27
2011-06-10T09:54:48.137300Z
K 7
svn:log
V 83
MFC r220644:
Make CAM report devices with ATA/SATA transport to devstat(9) as IDE.

END
K 10
svn:author
V 3
mav
K 8
svn:date
V 27
2011-06-10T09:57:48.928115Z
K 7
svn:log
V 152
MFC r220686 (by jh):
Move TUNABLE_INT_FETCH() out of the SIM lock to avoid sleeping while
holding the lock. The fix is analogous to r220618 for ada(4).

END
K 10
svn:author
V 3
mav
K 8
svn:date
V 27
2011-06-10T09:58:48.779501Z
K 7
svn:log
V 121
MFC r220690 (by jh):
Move a comment to the right place. The rearrange done in r208928 left
the comment to a wrong place.

END
K 10
svn:author
V 3
mav
K 8
svn:date
V 27
2011-06-10T10:09:48.593104Z
K 7
svn:log
V 91
MFC r220778, r220785:
Remove some unused variables.

Found with:     Clang Static Analyzer

END
K 10
svn:author
V 3
mav
K 8
svn:date
V 27
2011-06-10T10:13:40.016954Z
K 7
svn:log
V 342
MFC r222520:
Add quirks to hint 4K physical sector (Advanced Format) for ATA disks not
reporting it properly (none? of known disks now).

Hitachi and WDC AF disks seem could be identified more or less formally.
For Seagate and Samsung enumerate some found models/series.
For other disks it can be forced with kern.cam.ada.X.quirks=1 tunable.

END
K 10
svn:author
V 3
mav
K 8
svn:date
V 27
2011-06-10T10:43:49.869349Z
K 7
svn:log
V 277
MFC r208582 (by mjacob):
Add a new primitive, XPT_SCAN_TGT, to cover the range between scanning a
whole bus (XPT_SCAN_BUS) and a single lun on that bus (XPT_SCAN_LUN).

It's less resource comsumptive than scanning a whole bus when the
caller knows only one target has changes.

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-10T12:30:16.673918Z
K 7
svn:log
V 108
Implement BUS_ADJUST_RESOURCE() for the x86 drivers that sit between the
Host-PCI bridge drivers and nexus.

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-10T13:24:56.771587Z
K 7
svn:log
V 106
Fix some off-by-one errors with the ending address of candidate regions
when attempting to grow a window.

END
K 10
svn:author
V 8
rmacklem
K 8
svn:date
V 27
2011-06-10T13:28:14.102657Z
K 7
svn:log
V 725
MFC: r222389
Fix the new NFS client so that it handles NFSv4 state
correctly during a forced dismount. This required that
the exclusive and shared (refcnt) sleep lock functions check
for MNTK_UMOUNTF before sleeping, so that they won't block
while nfscl_umount() is getting rid of the state. As
such, a "struct mount *" argument was added to the locking
functions. I believe the only remaining case where a forced
dismount can get hung in the kernel is when a thread is
already attempting to do a TCP connect to a dead server
when the krpc client structure called nr_client is NULL.
This will only happen just after a "mount -u" with options
that force a new TCP connection is done, so it shouldn't
be a problem in practice.

END
K 10
svn:author
V 6
jilles
K 8
svn:date
V 27
2011-06-10T13:47:11.511939Z
K 7
svn:log
V 433
skel/.shrc: Improve commented CDPATH example for POSIX requirements.

POSIX says an empty entry in CDPATH shall not result in the new directory
being printed, while any non-empty entry shall result in the new directory
being printed, including ".". Therefore, the value of CDPATH should almost
always start with a colon, not dot and colon.

Our sh does not print the name for empty entries as well as "." entries.

MFC after:	1 week

END
K 10
svn:author
V 8
bschmidt
K 8
svn:date
V 27
2011-06-10T17:06:25.606289Z
K 7
svn:log
V 167
Make sure to drop the IEEE80211_RATE_BASIC flag even for MCS rates.

Reported by:	Maciej Milewski <milu at dat dot pl>
Tested by:	Maciej Milewski <milu at dat dot pl>

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-10T18:46:40.598337Z
K 7
svn:log
V 1486
MFC 221209:
TCP reuses t_rxtshift to determine the backoff timer used for both the
persist state and the retransmit timer.  However, the code that implements
"bad retransmit recovery" only checks t_rxtshift to see if an ACK has been
received in during the first retransmit timeout window.  As a result, if
ticks has wrapped over to a negative value and a socket is in the persist
state, it can incorrectly treat an ACK from the remote peer as a
"bad retransmit recovery" and restore saved values such as snd_ssthresh and
snd_cwnd.  However, if the socket has never had a retransmit timeout, then
these saved values will be zero, so snd_ssthresh and snd_cwnd will be set
to 0.

If the socket is in fast recovery (this can be caused by excessive
duplicate ACKs such as those fixed by 220794), then each ACK that arrives
triggers either NewReno or SACK partial ACK handling which clamps snd_cwnd
to be no larger than snd_ssthresh.  In effect, the socket's send window
is permamently stuck at 0 even though the remote peer is advertising a
much larger window and pending data is only sent via TCP window probes
(so one byte every few seconds).

Fix this by adding a new TCP pcb flag (TF_PREVVALID) that indicates that
the various snd_*_prev fields in the pcb are valid and only perform
"bad retransmit recovery" if this flag is set in the pcb.  The flag is set
on the first retransmit timeout that occurs and is cleared on subsequent
retransmit timeouts or when entering the persist state.

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-10T18:47:10.206729Z
K 7
svn:log
V 1486
MFC 221209:
TCP reuses t_rxtshift to determine the backoff timer used for both the
persist state and the retransmit timer.  However, the code that implements
"bad retransmit recovery" only checks t_rxtshift to see if an ACK has been
received in during the first retransmit timeout window.  As a result, if
ticks has wrapped over to a negative value and a socket is in the persist
state, it can incorrectly treat an ACK from the remote peer as a
"bad retransmit recovery" and restore saved values such as snd_ssthresh and
snd_cwnd.  However, if the socket has never had a retransmit timeout, then
these saved values will be zero, so snd_ssthresh and snd_cwnd will be set
to 0.

If the socket is in fast recovery (this can be caused by excessive
duplicate ACKs such as those fixed by 220794), then each ACK that arrives
triggers either NewReno or SACK partial ACK handling which clamps snd_cwnd
to be no larger than snd_ssthresh.  In effect, the socket's send window
is permamently stuck at 0 even though the remote peer is advertising a
much larger window and pending data is only sent via TCP window probes
(so one byte every few seconds).

Fix this by adding a new TCP pcb flag (TF_PREVVALID) that indicates that
the various snd_*_prev fields in the pcb are valid and only perform
"bad retransmit recovery" if this flag is set in the pcb.  The flag is set
on the first retransmit timeout that occurs and is cleared on subsequent
retransmit timeouts or when entering the persist state.

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-10T18:51:22.606212Z
K 7
svn:log
V 561
MFC 220794:
When checking to see if a window update should be sent to the remote peer,
don't force a window update if the window would not actually grow due to
window scaling.  Specifically, if the window scaling factor is larger than
2 * MSS, then after the local reader has drained 2 * MSS bytes from the
socket, a window update can end up advertising the same window.  If this
happens, the supposed window update actually ends up being a duplicate ACK.
This can result in an excessive number of duplicate ACKs when using a
higher maximum socket buffer size.

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-10T18:51:33.483952Z
K 7
svn:log
V 561
MFC 220794:
When checking to see if a window update should be sent to the remote peer,
don't force a window update if the window would not actually grow due to
window scaling.  Specifically, if the window scaling factor is larger than
2 * MSS, then after the local reader has drained 2 * MSS bytes from the
socket, a window update can end up advertising the same window.  If this
happens, the supposed window update actually ends up being a duplicate ACK.
This can result in an excessive number of duplicate ACKs when using a
higher maximum socket buffer size.

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-10T18:55:58.077511Z
K 7
svn:log
V 588
MFC 222032:
Fix a race in the SMP rendezvous code.  Specifically, the write by the
last CPU to to finish the rendezvous action may become visible to
different CPUs at different times.  As a result, the CPU that initiated
the rendezvous may exit the rendezvous and drop the lock allowing another
rendezvous to be initiated on the same CPU or a different CPU.  In that
case the exit sentinel may be cleared before all CPUs have noticed causing
those CPUs to hang forever.

Workaround this by using a generation count to notice when this race
occurs and to exit the rendezvous in that case.

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-10T18:56:12.565975Z
K 7
svn:log
V 588
MFC 222032:
Fix a race in the SMP rendezvous code.  Specifically, the write by the
last CPU to to finish the rendezvous action may become visible to
different CPUs at different times.  As a result, the CPU that initiated
the rendezvous may exit the rendezvous and drop the lock allowing another
rendezvous to be initiated on the same CPU or a different CPU.  In that
case the exit sentinel may be cleared before all CPUs have noticed causing
those CPUs to hang forever.

Workaround this by using a generation count to notice when this race
occurs and to exit the rendezvous in that case.

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-10T18:58:32.475143Z
K 7
svn:log
V 142
MFC 222252:
Simplify a stale assertion.  We have not called mi_switch() from a nested
critical section during a preemption for several years.

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-10T18:58:48.492697Z
K 7
svn:log
V 142
MFC 222252:
Simplify a stale assertion.  We have not called mi_switch() from a nested
critical section during a preemption for several years.

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-10T19:03:17.523224Z
K 7
svn:log
V 2087
MFC 222254:
Fix an issue with critical sections and SMP rendezvous handlers.
Specifically, a critical_exit() call that drops the nesting level to zero
has a brief window where the pending preemption flag is set and the
nesting level is set to zero.  This is done purposefully to avoid races
where a preemption scheduled by an interrupt could be lost otherwise (see
revision 144777).  However, this does mean that if an interrupt fires
during this window and enters and exits a critical section, it may preempt
from the interrupt context.  This is generally fine as the interrupt code
is careful to arrange critical sections so that they are not exited until
it is safe to preempt (e.g. interrupts EOI'd and masked if necessary).

However, the SMP rendezvous IPI handler does not quite follow this rule,
and in general a rendezvous can never be preempted.  Rendezvous handlers
are also not permitted to schedule threads to execute, so they will not
typically trigger preemptions.  SMP rendezvous handlers may use
spinlocks (carefully) such as the rm_cleanIPI() handler used in rmlocks,
but using a spinlock also enters and exits a critical section.  If the
interrupted top-half code is in the brief window of critical_exit() where
the nesting level is zero but a preemption is pending, then releasing the
spinlock can trigger a preemption.  Because we know that SMP rendezvous
handlers can never schedule a thread, we know that a critical_exit() in
an SMP rendezvous handler will only preempt in this edge case.  We also
know that the top-half thread will happily handle the deferred preemption
once the SMP rendezvous has completed, so the preemption will not be lost.

This makes it safe to employ a workaround where we use a nested critical
section in the SMP rendezvous code itself around rendezvous action
routines to prevent any preemptions during an SMP rendezvous.  The
workaround intentionally avoids checking for a deferred preemption
when leaving the critical section on the assumption that if there is a
pending preemption it will be handled by the interrupted top-half code.

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-10T19:03:32.168764Z
K 7
svn:log
V 2087
MFC 222254:
Fix an issue with critical sections and SMP rendezvous handlers.
Specifically, a critical_exit() call that drops the nesting level to zero
has a brief window where the pending preemption flag is set and the
nesting level is set to zero.  This is done purposefully to avoid races
where a preemption scheduled by an interrupt could be lost otherwise (see
revision 144777).  However, this does mean that if an interrupt fires
during this window and enters and exits a critical section, it may preempt
from the interrupt context.  This is generally fine as the interrupt code
is careful to arrange critical sections so that they are not exited until
it is safe to preempt (e.g. interrupts EOI'd and masked if necessary).

However, the SMP rendezvous IPI handler does not quite follow this rule,
and in general a rendezvous can never be preempted.  Rendezvous handlers
are also not permitted to schedule threads to execute, so they will not
typically trigger preemptions.  SMP rendezvous handlers may use
spinlocks (carefully) such as the rm_cleanIPI() handler used in rmlocks,
but using a spinlock also enters and exits a critical section.  If the
interrupted top-half code is in the brief window of critical_exit() where
the nesting level is zero but a preemption is pending, then releasing the
spinlock can trigger a preemption.  Because we know that SMP rendezvous
handlers can never schedule a thread, we know that a critical_exit() in
an SMP rendezvous handler will only preempt in this edge case.  We also
know that the top-half thread will happily handle the deferred preemption
once the SMP rendezvous has completed, so the preemption will not be lost.

This makes it safe to employ a workaround where we use a nested critical
section in the SMP rendezvous code itself around rendezvous action
routines to prevent any preemptions during an SMP rendezvous.  The
workaround intentionally avoids checking for a deferred preemption
when leaving the critical section on the assumption that if there is a
pending preemption it will be handled by the interrupted top-half code.

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-10T19:08:07.245424Z
K 7
svn:log
V 199
MFC 222651:
Properly return an ENOBUFS error if a write to a tun(4) device fails
due to m_uiotombuf() failing.

While here, trim unneeded error handling related to tuninit() since it
can never fail.

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-10T19:08:17.686477Z
K 7
svn:log
V 199
MFC 222651:
Properly return an ENOBUFS error if a write to a tun(4) device fails
due to m_uiotombuf() failing.

While here, trim unneeded error handling related to tuninit() since it
can never fail.

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-10T19:12:00.781326Z
K 7
svn:log
V 270
MFC 222660:
- Rename the Cronyx Omega2-PCI entry to Exar XR17C158 since that is the
  real owner of the device ID.  Also rename the associated config
  function while here.
- Add support for the 2-port and 4-port Exar parts as well: Exar XR17C/D152
  and Exar XR17C154.

END
K 10
svn:author
V 3
jhb
K 8
svn:date
V 27
2011-06-10T19:13:22.929883Z
K 7
svn:log
V 214
MFC 221079:
Generate the network byte order version of the window size structure in a
temporary variable on the stack and then copy that into the output buffer
so that the htons() conversions use aligned accesses.

END