]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
13 years agoo) Tear down receive interrupt on detach.
Juli Mallett [Wed, 16 Mar 2011 08:56:22 +0000 (08:56 +0000)]
o) Tear down receive interrupt on detach.

Submitted by: Bhanu Prakash

13 years agoo) Clean up FPA pools on module unload.
Juli Mallett [Wed, 16 Mar 2011 08:51:36 +0000 (08:51 +0000)]
o) Clean up FPA pools on module unload.
o) Allocate output buffer pool based on available output queues.

Submitted by: Bhanu Prakash (with modifications)

13 years agoo) Properly size caches and TLB on Octeon.
Juli Mallett [Wed, 16 Mar 2011 08:22:29 +0000 (08:22 +0000)]
o) Properly size caches and TLB on Octeon.
o) Make COP0_SYNC do nothing on Octeon, which is fully interlocked.

Submitted by: Bhanu Prakash (with modifications)

13 years agoMFaltix:
Marcel Moolenaar [Wed, 16 Mar 2011 03:53:18 +0000 (03:53 +0000)]
MFaltix:
Add support for Pre-Boot Virtual Memory (PBVM) to the loader.

PBVM allows us to link the kernel at a fixed virtual address without
having to make any assumptions about the physical memory layout. On
the SGI Altix 350 for example, there's no usuable physical memory
below 192GB. Also, the PBVM allows us to control better where we're
going to physically load the kernel and its modules so that we can
make sure we load the kernel in memory that's close to the BSP.

The PBVM is managed by a simple page table. The minimum size of the
page table is 4KB (EFI page size) and the maximum is currently set
to 1MB. A page in the PBVM is 64KB, as that's the maximum alignment
one can specify in a linker script. The bottom line is that PBVM is
between 64KB and 8GB in size.

The loader maps the PBVM page table at a fixed virtual address and
using a single translations. The PBVM itself is also mapped using a
single translation for a maximum of 32MB.

While here, increase the heap in the EFI loader from 512KB to 2MB
and set the stage for supporting relocatable modules.

13 years agoMFV of r219685, tzdata2011d
Edwin Groothuis [Wed, 16 Mar 2011 03:28:42 +0000 (03:28 +0000)]
MFV of r219685, tzdata2011d

- Pacific/Apia will have DST on 2 April instead of 3 April in 2011.
- Turkye will go to DST on 28 April instead of 27 April in 2011.
- Cuba will go to DST on 20 April in 2011.

Obtained from:  ftp://elsie.nci.nih.gov/pub/

13 years agoVendor import of tzdata2011d:
Edwin Groothuis [Wed, 16 Mar 2011 03:07:08 +0000 (03:07 +0000)]
Vendor import of tzdata2011d:

- Pacific/Apia will have DST on 2 April instead of 3 April in 2011.
- Turkye will go to DST on 28 April instead of 27 April in 2011.
- Cuba will go to DST on 20 April in 2011.

Obtained from: ftp://elsie.nci.nih.gov/pub/

13 years agoFix mv_gpio_in() for pin numbers that occupy bits 8-31 in GPIO registers.
Marcel Moolenaar [Wed, 16 Mar 2011 00:42:15 +0000 (00:42 +0000)]
Fix mv_gpio_in() for pin numbers that occupy bits 8-31 in GPIO registers.
The compiler will truncate the 32-bit return value of mv_gpio_value_get()
to match the 8-bit return value of mv_gpio_in(). A conditional expression
is used to have mv_gpio_in() always return 0 or 1 instead.

13 years agoRevert previous commit: EFI_STATUS is a 64-bit integral on ia64. Fix the
Marcel Moolenaar [Wed, 16 Mar 2011 00:08:10 +0000 (00:08 +0000)]
Revert previous commit: EFI_STATUS is a 64-bit integral on ia64. Fix the
compile warning on i386 (where EFI_STATUS is a 32-bit integral) by casting
the status argument to u_long instead.

Pointy hat: brucec
MFC after: 3 days

13 years agorfork(2): Discourage rfork_thread-like approaches.
Jilles Tjoelker [Tue, 15 Mar 2011 23:51:47 +0000 (23:51 +0000)]
rfork(2): Discourage rfork_thread-like approaches.

Calling rfork_thread(3) does not interoperate with pthreads and global state
is not properly protected.

Remove the BUGS section suggesting LinuxThreads entirely. With the current
pthread library libthr, all threads are kernel-level entities so there seems
little reason to use LinuxThreads.

13 years agoMerge svn+ssh://svn.freebsd.org/base/head@219680
Marcel Moolenaar [Tue, 15 Mar 2011 23:28:15 +0000 (23:28 +0000)]
Merge svn+ssh://svn.freebsd.org/base/head@219680

13 years agobin: Prefer strrchr() to rindex().
Jilles Tjoelker [Tue, 15 Mar 2011 22:22:11 +0000 (22:22 +0000)]
bin: Prefer strrchr() to rindex().

This removes the last index/rindex usage from /bin.

13 years agoPartially revert r219672. After r198295, kernel need to seed randomness as
Jung-uk Kim [Tue, 15 Mar 2011 21:45:10 +0000 (21:45 +0000)]
Partially revert r219672.  After r198295, kernel need to seed randomness as
soon as possible for stack protector.  However, dummy timecounter does not
have enough entropy and we don't need to sacrifice Pentium class and later.

Pointed out by: Maxim Dounin (mdounin at mdounin dot ru)

13 years agoDo not let machdep.tsc_freq modify tsc_freq itself. It is bad for i386 as
Jung-uk Kim [Tue, 15 Mar 2011 19:47:20 +0000 (19:47 +0000)]
Do not let machdep.tsc_freq modify tsc_freq itself.  It is bad for i386 as
it does not operate atomically.  Actually, it serves no purpose.

Noticed by: bde

13 years agoRemove tsc_present from this file, really.
Jung-uk Kim [Tue, 15 Mar 2011 18:09:29 +0000 (18:09 +0000)]
Remove tsc_present from this file, really.

13 years agoFix a typo in the previous commit. AMD64 and Intel 64 are two brand names.
Jung-uk Kim [Tue, 15 Mar 2011 17:59:30 +0000 (17:59 +0000)]
Fix a typo in the previous commit.  AMD64 and Intel 64 are two brand names.

13 years agoDeprecate tsc_present as the last of its real consumers finally disappeared.
Jung-uk Kim [Tue, 15 Mar 2011 17:19:52 +0000 (17:19 +0000)]
Deprecate tsc_present as the last of its real consumers finally disappeared.

13 years agoUnconditionally use binuptime(9) for get_cyclecount(9) on i386. Since this
Jung-uk Kim [Tue, 15 Mar 2011 17:14:26 +0000 (17:14 +0000)]
Unconditionally use binuptime(9) for get_cyclecount(9) on i386.  Since this
function is almost exclusively used for random harvesting, there is no need
for micro-optimization.  Adjust the manual page accordingly.

13 years agoReserve 24KB for the static kernel stack. We use this stack to call into
Marcel Moolenaar [Tue, 15 Mar 2011 16:50:17 +0000 (16:50 +0000)]
Reserve 24KB for the static kernel stack. We use this stack to call into
the firmware for physical mode calls and the Altix doesn't work when it's
16KB.

13 years agoRemove #include needed for debugging.
Pawel Jakub Dawidek [Tue, 15 Mar 2011 13:53:39 +0000 (13:53 +0000)]
Remove #include needed for debugging.

MFC after: 1 week

13 years agoStaticize functions which are not used somewhere else, move the
Alexander Leidinger [Tue, 15 Mar 2011 13:40:47 +0000 (13:40 +0000)]
Staticize functions which are not used somewhere else, move the
corresponding prototypes from the header to the code file.

13 years agoTurn on softupdates by default. We need a UI to set filesystem parameters.
Nathan Whitehorn [Tue, 15 Mar 2011 13:27:34 +0000 (13:27 +0000)]
Turn on softupdates by default. We need a UI to set filesystem parameters.

13 years agoForgot to commit this change along with r219563: pam_group(8) now issues
Dag-Erling Smørgrav [Tue, 15 Mar 2011 10:13:35 +0000 (10:13 +0000)]
Forgot to commit this change along with r219563: pam_group(8) now issues
a warning if neither luser nor ruser is specified.  The correct option
for su(1) is ruser.

MFC after: 1 month

13 years agoRe-order the TRs so that there're less gaps. This could have a performance
Marcel Moolenaar [Tue, 15 Mar 2011 06:07:02 +0000 (06:07 +0000)]
Re-order the TRs so that there're less gaps. This could have a performance
impact.

13 years agoRound the size, not the top address. This makes the code less
Marcel Moolenaar [Tue, 15 Mar 2011 06:04:13 +0000 (06:04 +0000)]
Round the size, not the top address. This makes the code less
sensitive to compiler bugs (32-bit truncation).

13 years ago- Inadvertently committed files with +x attribute, fixed.
David Christensen [Tue, 15 Mar 2011 01:06:27 +0000 (01:06 +0000)]
- Inadvertently committed files with +x attribute, fixed.
- Minor change to info output string.

13 years agoDon't define IA64_PBVM_PAGE_SIZE as 1U shifted to the left by some amount.
Marcel Moolenaar [Mon, 14 Mar 2011 23:49:41 +0000 (23:49 +0000)]
Don't define IA64_PBVM_PAGE_SIZE as 1U shifted to the left by some amount.
The compiler seems to assume it's a 32-bit integral and rounding to the
page size using the standard expression (((u_long)(x) + mask) & ~mask),
results in a 32-bit value. Dropping the 'U' suffix is enough to have the
compiler treat the expression as a 64-bit integral.

13 years agoMake get_cyclecount(9) little bit more useful where binuptime(9) is used.
Jung-uk Kim [Mon, 14 Mar 2011 23:30:14 +0000 (23:30 +0000)]
Make get_cyclecount(9) little bit more useful where binuptime(9) is used.

13 years ago- Initial release of bxe(4) to support Broadcom NetXtreme II 10GbE.
David Christensen [Mon, 14 Mar 2011 22:42:41 +0000 (22:42 +0000)]
- Initial release of bxe(4) to support Broadcom NetXtreme II 10GbE.
  (BCM57710, BCM57711, BCM57711E)

MFC after: One month

13 years agoWhen TSC is unavailable, broken or disabled and the current timecounter has
Jung-uk Kim [Mon, 14 Mar 2011 22:05:59 +0000 (22:05 +0000)]
When TSC is unavailable, broken or disabled and the current timecounter has
better quality than i8254 timer, use it for DELAY(9).

13 years agoThanks to sysinstall for its 15 years of service. And now to sleep,
Nathan Whitehorn [Mon, 14 Mar 2011 13:47:17 +0000 (13:47 +0000)]
Thanks to sysinstall for its 15 years of service. And now to sleep,
perchance to dream...

13 years agoAdd ssse3 capability for CPUTYPE=core2 to MACHINE_CPU in bsd.cpu.mk
Martin Matuska [Mon, 14 Mar 2011 13:36:51 +0000 (13:36 +0000)]
Add ssse3 capability for CPUTYPE=core2 to MACHINE_CPU in bsd.cpu.mk

MFC after: 2 weeks

13 years agoBackport SSSE3 instruction set support to base gcc.
Martin Matuska [Mon, 14 Mar 2011 13:31:34 +0000 (13:31 +0000)]
Backport SSSE3 instruction set support to base gcc.
Enabled by default for -march=core2

Obtained from: gcc 4.3 (rev. 117958, 121687, 121726, 123639; GPLv2)
MFC after: 2 weeks

13 years agoAdd my birthday to the freebsd calendar.
Dmitry Chagin [Mon, 14 Mar 2011 13:02:12 +0000 (13:02 +0000)]
Add my birthday to the freebsd calendar.

13 years agoFix potential panic in dbuf_sync_list() relate to spill blocks handling.
Pawel Jakub Dawidek [Mon, 14 Mar 2011 11:07:12 +0000 (11:07 +0000)]
Fix potential panic in dbuf_sync_list() relate to spill blocks handling.

Obtained from: IllumOS
MFC after: 1 month

13 years agoo Deal with unmapped PBVM in the alternate instruction and data TLB fault
Marcel Moolenaar [Mon, 14 Mar 2011 05:29:45 +0000 (05:29 +0000)]
o   Deal with unmapped PBVM in the alternate instruction and data TLB fault
    handlers.
o   Put the IVT in its own section and keep the supporting code close.
o   Make sure the VHPT is sized so that it can be mapped using a single
    translation.
o   Map the PAL code and VHPT with a translation that has the right size.
    Assume the platform has a PAL code size that can be mapped with a
    single translations.
o   Pass the pointer to the bootinfo structure as an argument to ia64_init().
o   Get rid of LOG2_ID_PAGE_SIZE and IA64_ID_PAGE_SIZE. It was used to map
    the regions 6 & 7 and was as large as possible. The problem is that we
    can't support memory attributes easily if the granuratity is not a page.
    We need to support memory attributes because the new USB stack violates
    the BUS_DMA(9) interface.
o   Update some comments...

NOTE: this is broken for SMP kernels, because the AP startup code hasn't
been updated yet.

13 years agoExpose the PBVM constants to assembly.
Marcel Moolenaar [Mon, 14 Mar 2011 05:18:27 +0000 (05:18 +0000)]
Expose the PBVM constants to assembly.

13 years agoMove the actual code to the front of the segment and make sure the IVT
Marcel Moolenaar [Mon, 14 Mar 2011 05:16:12 +0000 (05:16 +0000)]
Move the actual code to the front of the segment and make sure the IVT
can be put right at the front when put in its own section. This makes
sure that when the loader can only map a relatively small part of the
PBVM, the IVT and the startup code is wired.

13 years agoChange the sz argumnt to mmu_wire() from u_int to vm_size_t. It's a
Marcel Moolenaar [Mon, 14 Mar 2011 05:13:52 +0000 (05:13 +0000)]
Change the sz argumnt to mmu_wire() from u_int to vm_size_t. It's a
size in bytes...

13 years agoFix typo that snuck in.
Adrian Chadd [Mon, 14 Mar 2011 02:32:10 +0000 (02:32 +0000)]
Fix typo that snuck in.

13 years agoBring over the AR9285 board update code from ath9k.
Adrian Chadd [Mon, 14 Mar 2011 00:42:48 +0000 (00:42 +0000)]
Bring over the AR9285 board update code from ath9k.

This does a few things in particular:

* Abstracts out the gain control settings into a separate function;
* Configure antenna diversity, LNA and antenna gain parameters;
* Configure ob/db entries - the later v4k EEPROM modal revisions have
  multiple OB/DB parameters which are used for some form of
  calibration. Although the radio does have defaults for each,
  the EEPROM can override them.

This resolves the AR2427 related issues I've been seeing and makes
it stable at all 11g rates for both TX and RX.

13 years agoo Make sure the page table has a size that is mappable. Certain page
Marcel Moolenaar [Sun, 13 Mar 2011 22:04:19 +0000 (22:04 +0000)]
o   Make sure the page table has a size that is mappable. Certain page
    sizes are not supported.
o   Map the PBVM page table.
o   Map the PBVM using the largest possible power of 2 that is less than
    the amount of PBVM used and round down to a valid page size. Note
    that the current kernel is between 8MB and 16MB in size, which would
    mean that 8MB would be the typical size of the mapping, if only 8MB
    wasn't an invalid page size. In practice, we end up mapping the first
    4MB of PBVM in most cases.

13 years agoDon't sleep while setting the clock. This can cause panics when
Nathan Whitehorn [Sun, 13 Mar 2011 21:23:25 +0000 (21:23 +0000)]
Don't sleep while setting the clock. This can cause panics when
periodic_resettodr() calls CLOCK_SETTIME() and smu tries to sleep while
running from a callout.

Reported by: Torfinn Ingolfsen

13 years agosh: Fix some parameter expansion variants ${#...}.
Jilles Tjoelker [Sun, 13 Mar 2011 20:02:39 +0000 (20:02 +0000)]
sh: Fix some parameter expansion variants ${#...}.

These already worked: $# ${#} ${##} ${#-} ${#?}
These now work as well: ${#+word} ${#-word} ${##word} ${#%word}

There is an ambiguity in the standard with ${#?}: it could be the length of
$? or it could be $# giving an error in the (impossible) case that it is not
set. We continue to use the former interpretation as it seems more useful.

13 years agoPOSIX accepts only ELOOP if O_NOFOLLOW is specified and target is a symlink.
Pawel Jakub Dawidek [Sun, 13 Mar 2011 19:35:13 +0000 (19:35 +0000)]
POSIX accepts only ELOOP if O_NOFOLLOW is specified and target is a symlink.

13 years agoIn command line options allow size to be specified using k/M/G/T
Mikolaj Golub [Sun, 13 Mar 2011 19:23:32 +0000 (19:23 +0000)]
In command line options allow size to be specified using k/M/G/T
suffixes.

Approved by: pjd (mentor)
MFC after: 1 week

13 years agoRewind manifest file in case distributions are in a different order there.
Nathan Whitehorn [Sun, 13 Mar 2011 18:26:16 +0000 (18:26 +0000)]
Rewind manifest file in case distributions are in a different order there.

13 years agoUse the manifest file count, if available, to determine the number of files
Nathan Whitehorn [Sun, 13 Mar 2011 18:23:47 +0000 (18:23 +0000)]
Use the manifest file count, if available, to determine the number of files
in the archive instead of spinning through it to get a file count for the
progress bar. This speeds up installation a lot on systems with slow
CD drives.

13 years agoFix some typos due to unremoved commented-out lines. A great big pointy
Nathan Whitehorn [Sun, 13 Mar 2011 18:02:28 +0000 (18:02 +0000)]
Fix some typos due to unremoved commented-out lines. A great big pointy
hat goes to me for this one.

13 years agoAdd generation of an installation manifest containing SHA256 checksums as
Nathan Whitehorn [Sun, 13 Mar 2011 17:15:31 +0000 (17:15 +0000)]
Add generation of an installation manifest containing SHA256 checksums as
well as package descriptions and add code in the installer to check the
checksums.

13 years agoadd myself
Pawel Pekala [Sun, 13 Mar 2011 17:03:19 +0000 (17:03 +0000)]
add myself

Approved by:     miwi (mentor)

13 years agoFix typo.
Rebecca Cran [Sun, 13 Mar 2011 16:47:21 +0000 (16:47 +0000)]
Fix typo.

13 years agoRevert r219578
Doug Barton [Sun, 13 Mar 2011 16:27:36 +0000 (16:27 +0000)]
Revert r219578

13 years agosh: Add some tests for ${#parameter}.
Jilles Tjoelker [Sun, 13 Mar 2011 16:20:38 +0000 (16:20 +0000)]
sh: Add some tests for ${#parameter}.

13 years agoConnect the full doc build (Handbook, etc.) to the build. This is not built
Nathan Whitehorn [Sun, 13 Mar 2011 15:08:06 +0000 (15:08 +0000)]
Connect the full doc build (Handbook, etc.) to the build. This is not built
if no doc checkout is found.

13 years agoEnable shared page use for amd64/linux32 and i386/linux binaries.
Dmitry Chagin [Sun, 13 Mar 2011 14:58:02 +0000 (14:58 +0000)]
Enable shared page use for amd64/linux32 and i386/linux binaries.
Move signal trampoline code from the top of the stack to the shared page.

MFC after: 2 Weeks

13 years agoRemove the advertising clause from the UCB license according to the
Marius Strobl [Sun, 13 Mar 2011 13:42:43 +0000 (13:42 +0000)]
Remove the advertising clause from the UCB license according to the
July 22, 1999 addendum.

13 years agoConstantly use MHz instead of Mhz.
Bernhard Schmidt [Sun, 13 Mar 2011 13:05:50 +0000 (13:05 +0000)]
Constantly use MHz instead of Mhz.

Pointed out by: N.J. Mann <njm at njm.me.uk>

13 years agoFix the nfarray offsets for the ar2133/ar5133 radio - (AR5416, AR9160, etc.)
Adrian Chadd [Sun, 13 Mar 2011 13:00:45 +0000 (13:00 +0000)]
Fix the nfarray offsets for the ar2133/ar5133 radio - (AR5416, AR9160, etc.)

The offsets didn't match the assumption that nfarray[] is ordered by the
chainmask bits and programmed via the register order in ar5416_cca_regs[].
This repairs that damage and ensures that chain 1 is programmed correctly.
(And extension channels will now be programmed correctly also.)

This fixes some of the stuck beacons I've been seeing on my AR9160/AR5416
setups - because Chain 1 would be programmed -80 or -85 dBm, which is
higher than the actual noise floor and thus convincing the radio that
indeed it can't ever transmit.

13 years agoWhen injecting frames a temporary node is faked, during this several
Bernhard Schmidt [Sun, 13 Mar 2011 12:56:46 +0000 (12:56 +0000)]
When injecting frames a temporary node is faked, during this several
uses of ic_curchan occur. Due to the nature of a scan, switching
channels constantly and all this happening without any kind of locks
held, it might happen that ic_curchan points to nowhere leading to
panics. Fix this by not allowing frame injections while in SCAN state.

Tested by: Paul B. Mahol <onemda at gmail.com>

13 years agoFix a cut&paste error while parsing htcap/htinfo elements. This one is
Bernhard Schmidt [Sun, 13 Mar 2011 12:21:04 +0000 (12:21 +0000)]
Fix a cut&paste error while parsing htcap/htinfo elements. This one is
reponsible for not filling ni_htrates if a pre-ht information element is
present.

13 years agoHonor device capabilities while initializing ni_htrates.
Bernhard Schmidt [Sun, 13 Mar 2011 12:16:23 +0000 (12:16 +0000)]
Honor device capabilities while initializing ni_htrates.

Instead of hardcoding MCS0-15 generate the table dynamically, also
restrict the MCS rates to our own capabilities while parsing a htcap
element.

13 years agoFix rmax calculation during BSS selection.
Bernhard Schmidt [Sun, 13 Mar 2011 11:58:40 +0000 (11:58 +0000)]
Fix rmax calculation during BSS selection.

If multiple networks are available the max bandwidth is one
condition used for selecting the "best" BSS. To achieve that
we should consider all parameters which affect the max RX rate.
This includes 20/40MHz, SGI and the of course the MCS set.

If the TX MCS parameters are available we should use those,
because an AP announcing support for receiving frames at 450Mbps
might only be able to transmit at 150Mbps (1T3R). I haven't seen
devices with support for transmitting at higher rates then
receiving, so prefering TX over RX information should be safe.

While here, remove the hardcoded assumption that MCS15 is the max
possible MCS rate, use MCS31 instead which really is the highest
rate (according to the 802.11n std). Also, fix a mismatch of an
40MHz/SGI check.

13 years agoFill hc_mcsset completely.
Bernhard Schmidt [Sun, 13 Mar 2011 11:56:33 +0000 (11:56 +0000)]
Fill hc_mcsset completely.

Contrary to the rateset information in legacy frames the MCS Set
field also contains TX capability information in cases where the
number of available TX and RX spartial streams differ. Because a
rateset doesn't contain that information we have to pull the
those directly from the hardware capabilities.

13 years agoWhile determining the maxrate for ieee80211_media_setup() honor
Bernhard Schmidt [Sun, 13 Mar 2011 11:47:43 +0000 (11:47 +0000)]
While determining the maxrate for ieee80211_media_setup() honor
the device capabilities.

13 years agoChange the way HT capatibilities are announced.
Bernhard Schmidt [Sun, 13 Mar 2011 11:45:58 +0000 (11:45 +0000)]
Change the way HT capatibilities are announced.

Get rid of the assumption that every device is capable of 40MHz,
SGI and 2 spartial streams. Instead of printing, in the worst case,
8 times 76 MCS rates, print logically connect ranges and the
support RX/TX streams.

A device without 40MHz and SGI support looks like:
ath0: 2T2R
ath0: 11na MCS 20Mhz
ath0: MCS 0-7: 6.5Mbps - 65Mbps
ath0: MCS 8-15: 13Mbps - 130Mbps
ath0: 11ng MCS 20Mhz
ath0: MCS 0-7: 6.5Mbps - 65Mbps
ath0: MCS 8-15: 13Mbps - 130Mbps

13 years agoCross-reference ip(4) rather than sysctl(8) for the global, shared MIB
Bjoern A. Zeeb [Sun, 13 Mar 2011 11:44:16 +0000 (11:44 +0000)]
Cross-reference ip(4) rather than sysctl(8) for the global, shared MIB
values to control the various port ranges. ip(4) describes these in detail.

MFC after: 13 days

13 years agoFirst step on removing the harcoded RX/TX stream and MCS0-15 assumptions.
Bernhard Schmidt [Sun, 13 Mar 2011 11:40:18 +0000 (11:40 +0000)]
First step on removing the harcoded RX/TX stream and MCS0-15 assumptions.

Initialize ic_rxstream/ic_txstream with 2, for compatibility reasons.
Introduce 4 new HTC flags, which are used in addition to ic_rxstream
and ic_txstream to compute the hc_mcsset content and also for initializing
ni_htrates. The number of spatial streams is enough to determine support
for MCS0-31 but not for MCS32-76 as well as some TX parameters in the
hc_mcsset field.

13 years ago* Add wireless MAC reset, in prep for bringing over AR9130 support.
Adrian Chadd [Sun, 13 Mar 2011 08:46:58 +0000 (08:46 +0000)]
* Add wireless MAC reset, in prep for bringing over AR9130 support.
* Whilst I'm here, reformat to fit inside 80 characters.

13 years agoAdd the missing AR724x DDR flush routines for if_arge0.
Adrian Chadd [Sun, 13 Mar 2011 08:36:57 +0000 (08:36 +0000)]
Add the missing AR724x DDR flush routines for if_arge0.

Submitted by: Luiz Otavio O Souza

13 years agoFix the TX underrun status reset; remove a now unused variable.
Adrian Chadd [Sun, 13 Mar 2011 08:34:14 +0000 (08:34 +0000)]
Fix the TX underrun status reset; remove a now unused variable.

Submitted by: Luiz Otavio O Souza

13 years agoCommit FIFO configuration fixes from OpenWRT. This fixes performance
Adrian Chadd [Sun, 13 Mar 2011 08:28:21 +0000 (08:28 +0000)]
Commit FIFO configuration fixes from OpenWRT. This fixes performance
issues with if_arge on the AR913x and AR724x.

Reference: https://dev.openwrt.org/ticket/6754
Submitted by: Luiz Otavio O Souza

13 years agoThe number of streams is not based on the interface stream count, but the
Adrian Chadd [Sun, 13 Mar 2011 08:23:59 +0000 (08:23 +0000)]
The number of streams is not based on the interface stream count, but the
number of streams needed for that MCS rate.

13 years agoRemove the svn:executable property on Makefile
Doug Barton [Sun, 13 Mar 2011 08:22:46 +0000 (08:22 +0000)]
Remove the svn:executable property on Makefile

Add it on hastd

13 years agoMove out some of the shared eeprom board value calculation routines into ah.c
Adrian Chadd [Sun, 13 Mar 2011 05:54:05 +0000 (05:54 +0000)]
Move out some of the shared eeprom board value calculation routines into ah.c
rather than duplicating them for the v14 (ar5416+) and v4k (ar9285) codebases.

Further chipsets (eg the AR9287) have yet another EEPROM format which will use
these routines to calculate things.

13 years ago* Add in some board settings debugging to log what's being written
Adrian Chadd [Sun, 13 Mar 2011 05:30:14 +0000 (05:30 +0000)]
* Add in some board settings debugging to log what's being written
  to the TX closed-loop power control registers.
* Modify a couple of functions to take the register chain number,
  rather than the regChainOffset value. This allows for the
  register chain to be logged.

13 years agoAdd generation of memstick images to the bsdinstall release makefile for
Nathan Whitehorn [Sun, 13 Mar 2011 01:38:24 +0000 (01:38 +0000)]
Add generation of memstick images to the bsdinstall release makefile for
i386 and amd64. This involved moving the memstick generation script to
the arch directories from scripts/, in analogy to mkisoimages.sh. This
script was never called from /usr/src/release/Makefile, so that hasn't
been updated.

13 years agoFix grammar.
Andrey V. Elsukov [Sat, 12 Mar 2011 21:51:56 +0000 (21:51 +0000)]
Fix grammar.

Pointed out: Ben Kaduk
MFC after: 2 weeks

13 years agoINET6 for some time needed in_pcb.c so make sure we compile it in case of
Bjoern A. Zeeb [Sat, 12 Mar 2011 21:50:48 +0000 (21:50 +0000)]
INET6 for some time needed in_pcb.c so make sure we compile it in case of
no inet.

MFC after: 2 weeks

13 years agoMerge the two identical implementations for local port selections from
Bjoern A. Zeeb [Sat, 12 Mar 2011 21:46:37 +0000 (21:46 +0000)]
Merge the two identical implementations for local port selections from
in_pcbbind_setup() and in6_pcbsetport() in a single in_pcb_lport().

MFC after: 2 weeks

13 years agoUse the allexport option in load_rc_config() in order to avoid having
Doug Barton [Sat, 12 Mar 2011 21:13:08 +0000 (21:13 +0000)]
Use the allexport option in load_rc_config() in order to avoid having
to repeatedly read the conf files. Depending on what is enabled the
files are being read anywhere from 15, 30, or more times currently.
By loading the values in the environment this is reduced to 1, with
perhaps a couple more, again depending on what is enabled.

The speed-up for boot and shutdown is negligible when rc.conf is
on local disk, noticable when accessing files over NFS, and dramatic
when pulling rc.conf values from a database.

This change also includes a minor optimization to the conditional
for $_rc_conf_loaded.

13 years agoAllocate the DMA memory shared between the host and the controller as
Marius Strobl [Sat, 12 Mar 2011 20:36:52 +0000 (20:36 +0000)]
Allocate the DMA memory shared between the host and the controller as
coherent.

Approved by: gibbs
MFC after: 2 weeks

13 years agoTake two. Add the missing file that should have been committed
Steve Kargl [Sat, 12 Mar 2011 19:37:35 +0000 (19:37 +0000)]
Take two.  Add the missing file that should have been committed
with r219571 and re-enable building of cbrtl.

Implement the long double version for the cube root function, cbrtl.
The algorithm uses Newton's iterations with a crude estimate of the
cube root to converge to a result.

Reviewed by:    bde
Approved by:    das

13 years agoRemove unneeded newline.
Glen Barber [Sat, 12 Mar 2011 19:07:19 +0000 (19:07 +0000)]
Remove unneeded newline.

MFC after: 1 week

13 years agoAdd example of using 'scan' for wireless networks, similarly to
Glen Barber [Sat, 12 Mar 2011 19:05:39 +0000 (19:05 +0000)]
Add example of using 'scan' for wireless networks, similarly to
OpenBSD's ifconfig(8).

PR: 151952
Submitted by: Jared (rhyous of yahoo com)
MFC after: 1 week

13 years agoDocument GEOM_PART_EBR_COMPAT option.
Andrey V. Elsukov [Sat, 12 Mar 2011 17:05:18 +0000 (17:05 +0000)]
Document GEOM_PART_EBR_COMPAT option.

MFC after: 2 weeks

13 years agoTemporary disable the building of cbrtl until I
Steve Kargl [Sat, 12 Mar 2011 17:03:41 +0000 (17:03 +0000)]
Temporary disable the building of cbrtl until I
can determine why svn will not allow one to commit
a new file.

Approved by: das (implicit)

13 years agoImplement the long double version for the cube root function, cbrtl.
Steve Kargl [Sat, 12 Mar 2011 16:50:39 +0000 (16:50 +0000)]
Implement the long double version for the cube root function, cbrtl.
The algorithm uses Newton's iterations with a crude estimate of the
cube root to converge to a result.

Reviewed by: bde
Approved by: das

13 years agoPush a possible "unbind" in some situation from in6_pcbsetport() to
Bjoern A. Zeeb [Sat, 12 Mar 2011 16:45:15 +0000 (16:45 +0000)]
Push a possible "unbind" in some situation from in6_pcbsetport() to
callers.  This also fixes a problem when the prison call could set
the inp->in6p_laddr (laddr) and a following priv_check_cred() call
would return an error and will allow us to merge the IPv4 and IPv6
implementation.

MFC after: 2 weeks

13 years agoFix warnings and style(9) issues.
Rebecca Cran [Sat, 12 Mar 2011 14:47:54 +0000 (14:47 +0000)]
Fix warnings and style(9) issues.
Set WARNS to 6.

13 years agoSync licenses and the corresponding RCS IDs with NetBSD, mainly switching
Marius Strobl [Sat, 12 Mar 2011 14:33:32 +0000 (14:33 +0000)]
Sync licenses and the corresponding RCS IDs with NetBSD, mainly switching
the licenses of Matthew R. Green and the TNF to 2-clause.

Obtained from: NetBSD

13 years agoAdd some missing consts.
Pawel Jakub Dawidek [Sat, 12 Mar 2011 14:09:43 +0000 (14:09 +0000)]
Add some missing consts.

13 years agoMove cap_enter(2) and cap_getmode(2) symbols from FBSD_1.1 to FBSD_1.2.
Robert Watson [Sat, 12 Mar 2011 12:10:17 +0000 (12:10 +0000)]
Move cap_enter(2) and cap_getmode(2) symbols from FBSD_1.1 to FBSD_1.2.

Suggested by: kib
Obtained from: Capsicum Project
Sponsored by: Google
MFC after: 3 months

13 years agoMention the name of the module in warning messages.
Dag-Erling Smørgrav [Sat, 12 Mar 2011 11:26:37 +0000 (11:26 +0000)]
Mention the name of the module in warning messages.

13 years agoAdd "ruser" and "luser" options. The former corresponds to the current
Dag-Erling Smørgrav [Sat, 12 Mar 2011 11:12:30 +0000 (11:12 +0000)]
Add "ruser" and "luser" options.  The former corresponds to the current
behavior, where the module checks that the supplicant is a member of the
required group.  The latter checks the target user instead.  If neither
option was specified, pam_group(8) assumes "ruser" and issues a warning.
I intend to eventually change the default to "luser" to match the
behavior of similarly-named service modules in other operating systems.

MFC after: 1 month

13 years agoMake sure the locally cached value of rt->rt_gateway stays stable,
Bjoern A. Zeeb [Sat, 12 Mar 2011 09:41:25 +0000 (09:41 +0000)]
Make sure the locally cached value of rt->rt_gateway stays stable,
even after dropping the reference and unlocking. Previously we
have dereferenced a NULL pointer (after r121765).
Simply unlocking after the block does not work either because of
lock ordering (see r121765) and in addition we would still hold
a pointer to something that might be gone by the time we access it.
Thus take a copy of the value rather than just caching the pointer.

PR: kern/151908
Submitted by: chenyl (netstar2008 126.com) (initial version)
MFC after: 2 weeks

13 years agoadd DTrace systrace support for linux32 and freebsd32 on amd64 syscalls
Andriy Gapon [Sat, 12 Mar 2011 09:09:25 +0000 (09:09 +0000)]
add DTrace systrace support for linux32 and freebsd32 on amd64 syscalls

Add systrace_linux32 and systrace_freebsd32 modules which provide
support for tracing compat system calls in addition to native system
call tracing provided by systrace module.

Provided that all the systrace modules are loaded now you can select
what syscalls to trace in the following manner:

syscall::xxx:yyy - work on all system calls that match the specification
syscall:freebsd:xxx:yyy - only native system calls
syscall:linux32:xxx:yyy - linux32 compat system calls
syscall:freebsd32:xxx:yyy - freebsd32 compat system calls on amd64

PR: kern/152822
Submitted by: Artem Belevich <fbsdlist@src.cx>
Reviewed by: jhb (earlier version)
MFC after: 3 weeks

13 years agoadd DTrace systrace support for linux32 and freebsd32 on amd64 syscalls
Andriy Gapon [Sat, 12 Mar 2011 08:58:19 +0000 (08:58 +0000)]
add DTrace systrace support for linux32 and freebsd32 on amd64 syscalls

Regenerate system call and systrace support files.

PR: kern/152822
Submitted by: Artem Belevich <fbsdlist@src.cx>
Reviewed by: jhb (earlier version)
MFC after: 3 weeks

13 years agoadd DTrace systrace support for linux32 and freebsd32 on amd64 syscalls
Andriy Gapon [Sat, 12 Mar 2011 08:51:43 +0000 (08:51 +0000)]
add DTrace systrace support for linux32 and freebsd32 on amd64 syscalls

This commits makes necessary changes in syscall/sysent generation
infrastructure.

PR: kern/152822
Submitted by: Artem Belevich <fbsdlist@src.cx>
Reviewed by: jhb (ealier version)
MFC after: 3 weeks

13 years agoStyle(9) fixes. No functional changes.
Dmitry Chagin [Sat, 12 Mar 2011 07:47:05 +0000 (07:47 +0000)]
Style(9) fixes. No functional changes.

MFC after: 2 Week

13 years agoMerge gdtoa-20110304.
David Schultz [Sat, 12 Mar 2011 07:03:06 +0000 (07:03 +0000)]
Merge gdtoa-20110304.