]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
10 years agoWhen detaching a [USB] keyboard, keys might still be pressed. Ensure
hselasky [Tue, 28 Jan 2014 09:55:07 +0000 (09:55 +0000)]
When detaching a [USB] keyboard, keys might still be pressed. Ensure
that all pressed keys are released before completing the USB keyboard
detach. This will prevent so-called "ghost-keys" from appearing after
that the USB device generating the key event(s) has been detached.

MFC after: 1 week

10 years agoRemove STARTUP_PAGETABLE_ADDR from the ARM configs and replace it with
andrew [Tue, 28 Jan 2014 09:12:04 +0000 (09:12 +0000)]
Remove STARTUP_PAGETABLE_ADDR from the ARM configs and replace it with
memory at the end of the kernel.

This helps reduce the SoC and board specific configuration required.

Reviewed by: bsdimp
Tested by: jmg (armeb), br

10 years agoRevert r260622:
hselasky [Tue, 28 Jan 2014 08:49:00 +0000 (08:49 +0000)]
Revert r260622:
To be implemented a bit differently.

10 years agoComply to the official LibUSB v1.0 API:
hselasky [Tue, 28 Jan 2014 07:21:46 +0000 (07:21 +0000)]
Comply to the official LibUSB v1.0 API:
"It is legal to attempt to claim an already-claimed interface."

MFC after: 1 week

10 years agoAllow sigwait(2) in capabilities mode.
csjp [Tue, 28 Jan 2014 01:49:49 +0000 (01:49 +0000)]
Allow sigwait(2) in capabilities mode.

It's common for multi-threaded processes to create a thread for
the purpose of synchronously processing signals. Allow such processes to
utilize a capabilities sandbox.

Discussed with: rwatson, pjd
MFC after: 2 weeks

10 years agoRemove support for legacy PCI devices. These haven't been needed since
jhb [Mon, 27 Jan 2014 22:26:15 +0000 (22:26 +0000)]
Remove support for legacy PCI devices.  These haven't been needed since
support for LPC uart devices was added and it conflicts with upcoming
patches to add PCI INTx support.

Reviewed by: neel

10 years agoExplicitly enable I/O and memory decoding in the bridge's command register
jhb [Mon, 27 Jan 2014 19:49:52 +0000 (19:49 +0000)]
Explicitly enable I/O and memory decoding in the bridge's command register
when activating an I/O or memory window on the CardBus bridge.

Tested by: Olivier Cochard-Labbe <olivier@cochard.me>
Reviewed by: imp
MFC after: 3 days

10 years agoMerge from vendor branch importing dtc git rev 6a15eb2350426d285130e4c9d84c0bdb6575547a
imp [Mon, 27 Jan 2014 19:37:35 +0000 (19:37 +0000)]
Merge from vendor branch importing dtc git rev 6a15eb2350426d285130e4c9d84c0bdb6575547a

10 years agoRemove extra parens to silence clang warning.
imp [Mon, 27 Jan 2014 19:31:18 +0000 (19:31 +0000)]
Remove extra parens to silence clang warning.

10 years agoAdd support for 'clac' and 'stac' to DDB's disassembler on amd64.
jhb [Mon, 27 Jan 2014 18:53:18 +0000 (18:53 +0000)]
Add support for 'clac' and 'stac' to DDB's disassembler on amd64.

10 years agoMerge bmake-20140101 from vendor
sjg [Mon, 27 Jan 2014 17:47:21 +0000 (17:47 +0000)]
Merge bmake-20140101 from vendor

10 years agofix args to mtx_init
jmg [Mon, 27 Jan 2014 17:31:21 +0000 (17:31 +0000)]
fix args to mtx_init

Note that this commit hasn't been compile tested because these files
are not hooked up to the build...

PR: 186129
Submitted by: Takanori Sawada
Approved by: rpaulo

10 years agoAdd myself as a developer
rodrigo [Mon, 27 Jan 2014 11:10:06 +0000 (11:10 +0000)]
Add myself as a developer
Add bapt@ and kwm@ as mentors

Approved by: kwm@ (co-mentor)

10 years agoNote merge to head for sendmail 8.14.8.
gshapiro [Sun, 26 Jan 2014 23:55:35 +0000 (23:55 +0000)]
Note merge to head for sendmail 8.14.8.

MFC after: 5 days

10 years agoMinor changes to force commit these files so new freebsd*.cf files are
gshapiro [Sun, 26 Jan 2014 23:51:16 +0000 (23:51 +0000)]
Minor changes to force commit these files so new freebsd*.cf files are
built to use the new sendmail-8.14.8/cf tree.

MFC after: 5 days

10 years agoAdd missing svn:keywords property to new files
gshapiro [Sun, 26 Jan 2014 23:47:03 +0000 (23:47 +0000)]
Add missing svn:keywords property to new files

MFC after: 5 days

10 years agoUpdate for sendmail 8.14.8 import
gshapiro [Sun, 26 Jan 2014 23:44:12 +0000 (23:44 +0000)]
Update for sendmail 8.14.8 import

MFC after: 5 days

10 years agoRemove local FreeBSD workaround now that upstream project has a better fix.
gshapiro [Sun, 26 Jan 2014 23:42:40 +0000 (23:42 +0000)]
Remove local FreeBSD workaround now that upstream project has a better fix.

MFC after: 5 days

10 years agoAdd new sendmail 8.14.8 file
gshapiro [Sun, 26 Jan 2014 23:40:31 +0000 (23:40 +0000)]
Add new sendmail 8.14.8 file

MFC after: 5 days

10 years agoMerge sendmail 8.14.8 to HEAD
gshapiro [Sun, 26 Jan 2014 23:39:11 +0000 (23:39 +0000)]
Merge sendmail 8.14.8 to HEAD

MFC after: 5 days

10 years agologin: Clean up PAM and audit, then exit, on SIGHUP and SIGTERM.
jilles [Sun, 26 Jan 2014 22:49:24 +0000 (22:49 +0000)]
login: Clean up PAM and audit, then exit, on SIGHUP and SIGTERM.

This avoids leaving stale entries in utmpx after the connection is closed on
an open login session. It also allows a clean way (SIGTERM) to forcibly
terminate a user's terminal session.

This does not affect the situation for "hung" processes after the connection
is closed. The foreground process group receives SIGHUP and the tty becomes
inaccessible.

Also replace all use of the obsolete signal() function with sigaction() (not
only the part where it is actually required: SIGHUP and SIGTERM must mask
the other as well when caught).

PR: misc/183495
Reviewed by: ed

10 years agosh: Allow aliases to force alias substitution on the following word.
jilles [Sun, 26 Jan 2014 21:19:33 +0000 (21:19 +0000)]
sh: Allow aliases to force alias substitution on the following word.

If an alias's value ends with a space or tab, the next word is also
checked for aliases.

This is a POSIX feature. It is useful with utilities like command and
nohup (alias them to themselves followed by a space).

10 years agoImport sendmail 8.14.8
gshapiro [Sun, 26 Jan 2014 20:46:55 +0000 (20:46 +0000)]
Import sendmail 8.14.8

10 years agoUpdate link to vendor import instructions
gshapiro [Sun, 26 Jan 2014 20:13:28 +0000 (20:13 +0000)]
Update link to vendor import instructions

10 years agoStop the options string leak if it is not attached into the options
kevlo [Sun, 26 Jan 2014 07:20:32 +0000 (07:20 +0000)]
Stop the options string leak if it is not attached into the options
linked list.

Obtained from: OpenBSD

10 years agoBefore resetting the USART, delay a bit to allow the transmitter to
imp [Sun, 26 Jan 2014 04:57:58 +0000 (04:57 +0000)]
Before resetting the USART, delay a bit to allow the transmitter to
finish the current character to drain to avoid glitching. Also,
simplify the code a smidge.

10 years agogcc: Add support for -Wmissing-prototypes in C++
pfg [Sun, 26 Jan 2014 02:36:06 +0000 (02:36 +0000)]
gcc: Add support for -Wmissing-prototypes in C++

Support for warnings about missing prototypes in C++ was added by Apple
GCC (Radar 6261539).  Most of the code crept into r260311 so it felt
natural to make use of it.

Obtained from: Apple GCC - 5646
MFC after: 5 days

10 years agobinutils: add support for Intel SMAP-related instructions
pfg [Sun, 26 Jan 2014 00:37:21 +0000 (00:37 +0000)]
binutils: add support for Intel SMAP-related instructions

Add support for stac/clac instructions to manipulate the flag
that controls the behaviour of Intel's Supervisor Mode Access
Prevention (SMAP) feature.

Tested by: dim
Obtained from: OpenBSD
MFC after: 5 days

10 years agoMPC74xx should not fall through, to the error case.
jhibbits [Sat, 25 Jan 2014 22:50:42 +0000 (22:50 +0000)]
MPC74xx should not fall through, to the error case.

MFC after: 1 week

10 years agoMake early printf output nicer by inserting a carriage return before
imp [Sat, 25 Jan 2014 21:57:45 +0000 (21:57 +0000)]
Make early printf output nicer by inserting a carriage return before
any linefeeds that are output.

10 years agoBus space handles need to be the VA of the requested resource, not the
imp [Sat, 25 Jan 2014 21:52:19 +0000 (21:52 +0000)]
Bus space handles need to be the VA of the requested resource, not the
rounded page VA. Correct so the DBGU device can be mapped for FDT
console since it isn't on a page boundary.

10 years agoSupport level triggered interrupts with VT-x virtual interrupt delivery.
neel [Sat, 25 Jan 2014 20:58:05 +0000 (20:58 +0000)]
Support level triggered interrupts with VT-x virtual interrupt delivery.

The VMCS field EOI_bitmap[] is an array of 256 bits - one for each vector.
If a bit is set to '1' in the EOI_bitmap[] then the processor will trigger
an EOI-induced VM-exit when it is doing EOI virtualization.

The EOI-induced VM-exit results in the EOI being forwarded to the vioapic
so that level triggered interrupts can be properly handled.

Tested by: Anish Gupta (akgupt3@gmail.com)

10 years agoFix compilation with IGB_LEGACY_TX defined.
glebius [Sat, 25 Jan 2014 20:39:23 +0000 (20:39 +0000)]
Fix compilation with IGB_LEGACY_TX defined.

PR: 185909
Submitted by: Aurelien Rougemont <beorn binaries.fr>

10 years agoCheck for a full virtqueue in the multiqueue transmit path
bryanv [Sat, 25 Jan 2014 19:58:53 +0000 (19:58 +0000)]
Check for a full virtqueue in the multiqueue transmit path

With most hosts, we'll negotiate indirect descriptors, so all we
need is one available descriptor to transmit a frame.

10 years agoAvoid queue unlock followed by relock when the enable interrupt race is lost
bryanv [Sat, 25 Jan 2014 19:57:30 +0000 (19:57 +0000)]
Avoid queue unlock followed by relock when the enable interrupt race is lost

This already happens infrequently, and the hold time is still bounded since
we defer to a taskqueue after a few tries.

10 years agoMove duplicated transmit start code into a single function
bryanv [Sat, 25 Jan 2014 19:55:42 +0000 (19:55 +0000)]
Move duplicated transmit start code into a single function

10 years agoFix gcc with -Wstrict-prototypes by telling it bi_emac takes no parameters.
andrew [Sat, 25 Jan 2014 19:36:27 +0000 (19:36 +0000)]
Fix gcc with -Wstrict-prototypes by telling it bi_emac takes no parameters.

10 years agoRemove stray space
bryanv [Sat, 25 Jan 2014 18:34:57 +0000 (18:34 +0000)]
Remove stray space

10 years ago * Mark static inline functions with __fenv_static.
andrew [Sat, 25 Jan 2014 18:13:43 +0000 (18:13 +0000)]
 * Mark static inline functions with __fenv_static.
 * Correctly shift the mask when masking/unmasking exceptions.

10 years agoPull in r195679 from upstream llvm trunk:
dim [Sat, 25 Jan 2014 16:35:56 +0000 (16:35 +0000)]
Pull in r195679 from upstream llvm trunk:

  Don't use nopl in cpus that don't support it.

  Patch by Mikulas Patocka. I added the test. I checked that for cpu names that
  gas knows about, it also doesn't generate nopl.

  The modified cpus:
  i686 - there are i686-class CPUs that don't have nopl: Via c3, Transmeta
         Crusoe, Microsoft VirtualBox - see
         https://bbs.archlinux.org/viewtopic.php?pid=775414
  k6, k6-2, k6-3, winchip-c6, winchip2 - these are 586-class CPUs
  via c3 c3-2 - see https://bugs.archlinux.org/task/19733 as a proof that
         Via c3 and c3-Nehemiah don't have nopl

PR: bin/185777
MFC after: 3 days

10 years agoUse __fenv_static for all static inline functions.
andrew [Sat, 25 Jan 2014 16:03:08 +0000 (16:03 +0000)]
Use __fenv_static for all static inline functions.

10 years agosh: Add tests for alias names after another alias.
jilles [Sat, 25 Jan 2014 14:59:08 +0000 (14:59 +0000)]
sh: Add tests for alias names after another alias.

Since the first alias's value does not end with a blank, the next word
should not be checked for aliases.

10 years agoAlso include the mbuf's csum_flags in an assert message
bryanv [Sat, 25 Jan 2014 07:35:09 +0000 (07:35 +0000)]
Also include the mbuf's csum_flags in an assert message

10 years agoRead and write the MAC address in the config space byte by byte
bryanv [Sat, 25 Jan 2014 07:13:47 +0000 (07:13 +0000)]
Read and write the MAC address in the config space byte by byte

10 years agoRead each field of the configuration individually
bryanv [Sat, 25 Jan 2014 07:01:51 +0000 (07:01 +0000)]
Read each field of the configuration individually

In the forthcoming VirtIO spec, the device configuration is
always in little endian instead of guest edian. This is a
noop change for now.

10 years agoChange RWX to XWR in comments to match intent and bit patterns
grehan [Sat, 25 Jan 2014 06:58:41 +0000 (06:58 +0000)]
Change RWX to XWR in comments to match intent and bit patterns
in discussion of valid EPT pte protections.

Discussed with: neel
MFC after: 3 days

10 years agoRemove spaces before tabs in the function prototype list
bryanv [Sat, 25 Jan 2014 06:54:04 +0000 (06:54 +0000)]
Remove spaces before tabs in the function prototype list

10 years agoBad timespec_subtract() calculations produce negative tv_nsec on i386
ache [Sat, 25 Jan 2014 02:16:09 +0000 (02:16 +0000)]
Bad timespec_subtract() calculations produce negative tv_nsec on i386
which cause EINVAL returned from nanosleep() which cause loop in
cron_sleep() and making all cron jobs to start about 30 seconds earlier
(which cause f.e. logfiles rotation by newsyslog delayed by 1 hour).

Use simple and proved calculations from kernel's timespecsub() instead.

MFC after:      3 days

10 years agoTemporarily turn off the services.db generation during distributeworld.
gjb [Sat, 25 Jan 2014 00:03:33 +0000 (00:03 +0000)]
Temporarily turn off the services.db generation during distributeworld.
I'll work on an update to test for endianness-compatibility.

Submitted by: many
Sponsored by: The FreeBSD Foundation

10 years agosh: Do not depend on parse/execute split in new alias test.
jilles [Fri, 24 Jan 2014 23:00:35 +0000 (23:00 +0000)]
sh: Do not depend on parse/execute split in new alias test.

10 years agoAdd MLINK for fdclosedir.3 to directory.3 and sort fdopendir(3) entry.
bdrewery [Fri, 24 Jan 2014 22:37:35 +0000 (22:37 +0000)]
Add MLINK for fdclosedir.3 to directory.3 and sort fdopendir(3) entry.

Reported by: bde
Approved by: bapt (mentor)

10 years agoAccept O_CLOEXEC in shm_open().
rmh [Fri, 24 Jan 2014 21:05:07 +0000 (21:05 +0000)]
Accept O_CLOEXEC in shm_open().

Reviewed by: jilles, jhb
MFC after: 1 week

10 years agoCorrect the alignment of sp through functions that use UNWINDSVCFRAME. We
andrew [Fri, 24 Jan 2014 20:51:56 +0000 (20:51 +0000)]
Correct the alignment of sp through functions that use UNWINDSVCFRAME. We
were incorrectly adding the trap frame padding to the stack pointer after
reading it's value and unaligning it.

10 years agoext2fs: Re-enable reallocblk.
pfg [Fri, 24 Jan 2014 20:26:00 +0000 (20:26 +0000)]
ext2fs: Re-enable reallocblk.

The major corruption issues affecting this code have been fixed
a while ago.

MFC after: 1 week

10 years agoIn an effort to diagnose possible corruption of struct vm_page on some
alc [Fri, 24 Jan 2014 19:08:42 +0000 (19:08 +0000)]
In an effort to diagnose possible corruption of struct vm_page on some
sparc64 machines make the page queue assert in vm_page_dequeue() more
precise.  While I'm here switch the page lock assert to the newer style.

10 years agoAdd more USB quirks.
hselasky [Fri, 24 Jan 2014 18:52:04 +0000 (18:52 +0000)]
Add more USB quirks.

Submitted by: Gerrit Kuehn <gerrit.kuehn@aei.mpg.de>
MFC after: 1 week

10 years agoFix the Makefiles so that the tests I submitted in r261081 will actually
asomers [Fri, 24 Jan 2014 18:04:16 +0000 (18:04 +0000)]
Fix the Makefiles so that the tests I submitted in r261081 will actually
get built.

Sponsored by: Spectra Logic
MFC after: 13 days
X-MFC-With: r261081

10 years agoFix the build so -DNO_TESTS is passed in various phases that don't
asomers [Fri, 24 Jan 2014 18:01:46 +0000 (18:01 +0000)]
Fix the build so -DNO_TESTS is passed in various phases that don't
require tests in order to build or install.  Crucially, don't try to
install tests during the lib32 install phase.  This commit supersedes
r261081, which fixed the lib32 install phase problem, but didn't fix
other phases.

Submitted by: Garrett Cooper
Reviewed by: sjg
MFC after: 13 days

10 years agoRemove obsolete option
imp [Fri, 24 Jan 2014 17:00:27 +0000 (17:00 +0000)]
Remove obsolete option

10 years agoFix comment to match option
imp [Fri, 24 Jan 2014 16:58:50 +0000 (16:58 +0000)]
Fix comment to match option

10 years agoFix comment to match option
imp [Fri, 24 Jan 2014 16:58:21 +0000 (16:58 +0000)]
Fix comment to match option

10 years agoFix comment to match option
imp [Fri, 24 Jan 2014 16:57:42 +0000 (16:57 +0000)]
Fix comment to match option

10 years agoAdd NAND support
imp [Fri, 24 Jan 2014 16:56:27 +0000 (16:56 +0000)]
Add NAND support
Fix comment
This board has 4 wire support

10 years agoRemove redundant declaration for uart devclass.
imp [Fri, 24 Jan 2014 16:50:15 +0000 (16:50 +0000)]
Remove redundant declaration for uart devclass.
Commint some unrelated, but harmless, FDT ifdefs.

10 years agosh: Solve the alias recursion problem in a less hackish way.
jilles [Fri, 24 Jan 2014 16:40:51 +0000 (16:40 +0000)]
sh: Solve the alias recursion problem in a less hackish way.

Add the space to avoid alias recursion when the alias is expanded, not when
it is added.

As a result, displaying an alias via command -v, command -V or type no
longer erroneously appends a space. Adjust the tests so they now require
this bug to be absent.

10 years agoFix comment.
kevlo [Fri, 24 Jan 2014 15:34:22 +0000 (15:34 +0000)]
Fix comment.

10 years agoHide now harmless warning from dmesg.
hselasky [Fri, 24 Jan 2014 15:05:28 +0000 (15:05 +0000)]
Hide now harmless warning from dmesg.

MFC after: 1 week

10 years agodtrace: remove unexplained 16MB limitation from dt_alloc/dt_zalloc
avg [Fri, 24 Jan 2014 15:04:02 +0000 (15:04 +0000)]
dtrace: remove unexplained 16MB limitation from dt_alloc/dt_zalloc

The limitation was introduced in r178556 without any note or comment.
It seems pretty artificial and now it leads to problems like the following:
  $ dtrace -x bufsize=17m -n ...
  dtrace: processing aborted: Memory allocation failure
OpenSolaris and illumos never had this limitation.

Sponsored by: HybridCluster

10 years agosh: Add test for nested alias.
jilles [Fri, 24 Jan 2014 15:03:56 +0000 (15:03 +0000)]
sh: Add test for nested alias.

10 years agoext2fs: fix a bug in dirindex and re-enable.
pfg [Fri, 24 Jan 2014 13:51:38 +0000 (13:51 +0000)]
ext2fs: fix a bug in dirindex and re-enable.

The IN_* flags should be set in i_flag instead of corrupting
i_flags [1].

Re-enable HTree dirindex as the last series of bug fixes
seems to have fixed the issues.

Reported by: bde [1]
Tested by: kevlo
MFC after: 1 week

10 years agoGet rid of memmove(). It's not portable.
kevlo [Fri, 24 Jan 2014 09:17:29 +0000 (09:17 +0000)]
Get rid of memmove().  It's not portable.

Tested on RT3071, RT3573, RT3570, RT3572, and RT5572.

Reviewed by: hselasky

10 years agoReorder struct ip_fw_chain:
melifaro [Fri, 24 Jan 2014 09:13:30 +0000 (09:13 +0000)]
Reorder struct ip_fw_chain:
* move rarely-used fields down
* move uh_lock to different cacheline
* remove some usused fields

Sponsored by: Yandex LLC

10 years agoFix 32-bit signal handling on ppc64. This was broken when the PSL_USERSTATIC
jhibbits [Fri, 24 Jan 2014 02:57:00 +0000 (02:57 +0000)]
Fix 32-bit signal handling on ppc64.  This was broken when the PSL_USERSTATIC
macro was changed.  Since copying 64-bit srr1 into 32-bit srr1 drops the upper
32 bits, any bits set in the context were dropped, meaning the context check
fails.  Since 32-bit set_context can't change those bits anyway, copy the ones
from the current context (td->td_frame) before calling set_context().

MFC after: 3 weeks

10 years agoBe more robust with malformed interrupt config data. Instead of crashing
ian [Fri, 24 Jan 2014 00:42:18 +0000 (00:42 +0000)]
Be more robust with malformed interrupt config data.  Instead of crashing
or going into a near-infinite loop, warn and make potentially-reasonable
assumptions.

Reviewed by: brooks, nwhitehorn

10 years agoclock-frequency is a FreeBSD-specific extention. Make it optional and
imp [Fri, 24 Jan 2014 00:14:14 +0000 (00:14 +0000)]
clock-frequency is a FreeBSD-specific extention. Make it optional and
allow the client uart drivers to decide if a frequency is required.

10 years agoAlways free sbuf in gctl_free().
ae [Thu, 23 Jan 2014 21:30:31 +0000 (21:30 +0000)]
Always free sbuf in gctl_free().

MFC after: 1 week

10 years agoRemove another unneeded NULL check from geom_alloc_copyin().
ae [Thu, 23 Jan 2014 20:25:38 +0000 (20:25 +0000)]
Remove another unneeded NULL check from geom_alloc_copyin().
Do copyout in case of gctl version mismatch and fix sbuf leak in
g_ctl_ioctl_ctl().

MFC after: 1 week

10 years agoMove <machine/apicvar.h> to <x86/apicvar.h>.
jhb [Thu, 23 Jan 2014 20:10:22 +0000 (20:10 +0000)]
Move <machine/apicvar.h> to <x86/apicvar.h>.

10 years agoIn gctl_copyin() remove unused error variable.
ae [Thu, 23 Jan 2014 19:55:02 +0000 (19:55 +0000)]
In gctl_copyin() remove unused error variable.
geom_alloc_copyin() can't return ENOMEM, so describe its fail as bad
control request. Add check for NULL pointer in gctl_dump(), since it
can be NULL when geom_alloc_copyin() failed.

MFC after: 1 week

10 years agoFix typo in r261084.
ae [Thu, 23 Jan 2014 19:31:17 +0000 (19:31 +0000)]
Fix typo in r261084.
Add to the gctl_error() an ability to specify error description even
if numeric error code is already specified. Also by default set
error code to EINVAL.

PR: 185852
MFC after: 1 week

10 years agomalloc() with M_WAITOK doesn't return NULL.
ae [Thu, 23 Jan 2014 19:07:22 +0000 (19:07 +0000)]
malloc() with M_WAITOK doesn't return NULL.

MFC after: 1 week

10 years agoAdd Atmel serial drivers.
imp [Thu, 23 Jan 2014 18:09:10 +0000 (18:09 +0000)]
Add Atmel serial drivers.

10 years agoReplace the old unix_seqpacket and unix_seqpacket_exercise tests, which
asomers [Thu, 23 Jan 2014 17:26:28 +0000 (17:26 +0000)]
Replace the old unix_seqpacket and unix_seqpacket_exercise tests, which
were a little broken and not automatable, with unix_seqpacket_test.
It's coverage is a superset of the old tests and it uses ATF.  It
includes test cases for bugs kern/185813 and kern/185812.

PR: kern/185812
PR: kern/185813
Sponsored by: Spectra Logic
MFC after: 2 weeks

10 years agoThe posix_fallocate(2) syscall should return error number on error,
kib [Thu, 23 Jan 2014 17:24:26 +0000 (17:24 +0000)]
The posix_fallocate(2) syscall should return error number on error,
without modifying errno.

Reported and tested by: Gennady Proskurin <gpr@mail.ru>
Reviewed by: mdf
PR: standards/186028
Sponsored by: The FreeBSD Foundation
MFC after: 1 week

10 years agojemalloc has been updated to version 3.5.0.
skreuzer [Thu, 23 Jan 2014 14:13:12 +0000 (14:13 +0000)]
jemalloc has been updated to version 3.5.0.

Approved by: hrs (mentor)

10 years agoRevert r261014. Let Kevin fix it.
hselasky [Thu, 23 Jan 2014 10:18:28 +0000 (10:18 +0000)]
Revert r261014. Let Kevin fix it.

MFC after: 1 week

10 years agoUpdate EINVAL description.
pluknet [Thu, 23 Jan 2014 09:37:03 +0000 (09:37 +0000)]
Update EINVAL description.
This matches current POSIX standards and actual FreeBSD behavior.

MFC after: 1 week

10 years agoSet "Interrupt Window Exiting" in the case where there is a vector to be
neel [Thu, 23 Jan 2014 06:06:50 +0000 (06:06 +0000)]
Set "Interrupt Window Exiting" in the case where there is a vector to be
injected into the vcpu but the VM-entry interruption information field
already has the valid bit set.

Pointed out by: David Reed (david.reed@tidalscale.com)

10 years agoOpenSSL has been updated to version 1.0.1.f
skreuzer [Thu, 23 Jan 2014 03:24:38 +0000 (03:24 +0000)]
OpenSSL has been updated to version 1.0.1.f

Approved by: hrs (mentor)

10 years agoRemove WITHOUT_ATF as it has been replaced by WITH_TESTS
skreuzer [Thu, 23 Jan 2014 03:14:48 +0000 (03:14 +0000)]
Remove WITHOUT_ATF as it has been replaced by WITH_TESTS

PR: conf/185326
Reported by: Igor Mozolevsky <igor@hybrid-lab.co.uk>
Approved by: hrs (mentor)

10 years agoUpdate jemalloc to version 3.5.0.
jasone [Thu, 23 Jan 2014 02:47:36 +0000 (02:47 +0000)]
Update jemalloc to version 3.5.0.

10 years agoProperly sort the arguments to mtx_init(9).
jhibbits [Thu, 23 Jan 2014 01:09:33 +0000 (01:09 +0000)]
Properly sort the arguments to mtx_init(9).

PR: misc/186020
Submitted by: alfred
MFC after: 1 week

10 years agoDisable output about duplicate services, which were not output during
gjb [Wed, 22 Jan 2014 23:00:29 +0000 (23:00 +0000)]
Disable output about duplicate services, which were not output during
testing.

MFC after: 1 month
X-MFC-After: r261031
Sponsored by: The FreeBSD Foundation

10 years agoDocument EARLY_PRINTF
imp [Wed, 22 Jan 2014 22:25:48 +0000 (22:25 +0000)]
Document EARLY_PRINTF

10 years agoMostly revert r260267 and hopefully really fix the original problem.
mav [Wed, 22 Jan 2014 22:19:53 +0000 (22:19 +0000)]
Mostly revert r260267 and hopefully really fix the original problem.

The latest draft of SBC-3 tells: "A MAXIMUM UNMAP LBA COUNT field set to
a non-zero value indicates the maximum number of LBAs that may be unmapped
by an UNMAP command."  To me it does not sound like that limit is set per
single descriptor, but rather per all command.  And I have at least one
device that behaves exactly that way.  This patch fixes the problem there.

MFC after: 1 week

10 years agoClean up HL201 config (for the hot-e 201 and 101 models from thinlinx).
imp [Wed, 22 Jan 2014 22:01:29 +0000 (22:01 +0000)]
Clean up HL201 config (for the hot-e 201 and 101 models from thinlinx).
Make comments match parameters
Add options for early printf so we get regression build testing on it.
Add preview of options for FDT support coming soon (I hope)

10 years agoImplement support for early printf. You need to define SOCDEV_{PA,VA}
imp [Wed, 22 Jan 2014 21:49:20 +0000 (21:49 +0000)]
Implement support for early printf. You need to define SOCDEV_{PA,VA}
as described in the comments for the eputc function in your config file.

10 years agoAdd support for mapping a small range of the SoC devices for debugging
imp [Wed, 22 Jan 2014 21:23:58 +0000 (21:23 +0000)]
Add support for mapping a small range of the SoC devices for debugging
purposes early in boot.

10 years agoImplement generic support for early printf. Thought I can't find the
imp [Wed, 22 Jan 2014 21:20:08 +0000 (21:20 +0000)]
Implement generic support for early printf. Thought I can't find the
paper trail now, this patch is similar to one posted for one of the
preliminary versions of a new armv6 port. I took them and made them
more generic. Option not enabled by default since each board/port has
to provide its own eputc, and possibly do other things as well...

10 years agoMerge OpenSSL 1.0.1f.
jkim [Wed, 22 Jan 2014 19:57:11 +0000 (19:57 +0000)]
Merge OpenSSL 1.0.1f.

Approved by: so (delphij), benl (silence)