]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
11 years agoAllow 8259 registers to be read. This is a transient condition
grehan [Fri, 28 Jun 2013 06:25:04 +0000 (06:25 +0000)]
Allow 8259 registers to be read. This is a transient condition
during Linux boot.

Submitted by: tycho nightingale at pluribusnetworks com
Reviewed by: neel

11 years agoMake sure all CPUID values are handled, instead of exiting the
grehan [Fri, 28 Jun 2013 06:05:33 +0000 (06:05 +0000)]
Make sure all CPUID values are handled, instead of exiting the
bhyve process when an unhandled one is encountered.

Hide some additional capabilities from the guest (e.g. debug store).

This fixes the issue with FreeBSD 9.1 MP guests exiting the VM on
AP spinup (where CPUID is used when sync'ing the TSCs) and the
issue with the Java build where CPUIDs are issued from a guest
userspace.

Submitted by: tycho nightingale at pluribusnetworks com
Reviewed by: neel
Reported by: many

11 years agoFix build.
hrs [Fri, 28 Jun 2013 05:09:01 +0000 (05:09 +0000)]
Fix build.

Spotted by: gjb

11 years agoAllow the PCI config address register to be read. The Linux
grehan [Fri, 28 Jun 2013 05:01:25 +0000 (05:01 +0000)]
Allow the PCI config address register to be read. The Linux
kernel does this. Also remove an unused header file.

Submitted by: tycho nightingale at pluribusnetworks com
Reviewed by: neel

11 years ago - Add a general purpose resource allocator, vmem, from NetBSD. It was
jeff [Fri, 28 Jun 2013 03:51:20 +0000 (03:51 +0000)]
 - Add a general purpose resource allocator, vmem, from NetBSD.  It was
   originally inspired by the Solaris vmem detailed in the proceedings
   of usenix 2001.  The NetBSD version was heavily refactored for bugs
   and simplicity.
 - Use this resource allocator to allocate the buffer and transient maps.
   Buffer cache defrags are reduced by 25% when used by filesystems with
   mixed block sizes.  Ultimately this may permit dynamic buffer cache
   sizing on low KVA machines.

Discussed with: alc, kib, attilio
Tested by: pho
Sponsored by: EMC / Isilon Storage Division

11 years agoThe dtmalloc provider uses the short description of a malloc type as the
markj [Fri, 28 Jun 2013 03:14:40 +0000 (03:14 +0000)]
The dtmalloc provider uses the short description of a malloc type as the
function name of its corresponding DTrace probes. These descriptions may
contain whitespace, but probe names cannot, so just replace any whitespace
with underscores when creating probes.

MFC after: 1 week

11 years agoSupport reading registers r0-r3 when unwinding. There is a seperate
andrew [Thu, 27 Jun 2013 22:26:56 +0000 (22:26 +0000)]
Support reading registers r0-r3 when unwinding. There is a seperate
instruction to load these. We only hit it when unwinding past an trap frame
as in C r0-r3 would never have been saved onto the stack.

11 years agoMake detaching drivers from PCI devices more robust. While here, fix a
jhb [Thu, 27 Jun 2013 20:21:54 +0000 (20:21 +0000)]
Make detaching drivers from PCI devices more robust.  While here, fix a
bug where a PCI device would be powered down if it failed to probe, but
not when its driver was detached (e.g. via kldunload).
- Add a new helper method resource_list_release_active() which forcefully
  releases any active resources of a specified type from a resource list.
- Add a bus_child_detached method for the PCI bus driver which forces any
  active resources to be released (and whines to the console if it finds
  any) and then powers the device down.
- Call pci_child_detached() if we fail to probe a device when a driver
  is kldloaded.  This isn't perfect but can avoid leaking resources
  from a probe() routine in the kldload case.

Reviewed by: imp, brooks
MFC after: 1 month

11 years agoAdd "amdfam10" to the supported CPU types for amd64. This complements
jkim [Thu, 27 Jun 2013 19:47:58 +0000 (19:47 +0000)]
Add "amdfam10" to the supported CPU types for amd64.  This complements
r251691 and 251692.

11 years agoTo avoid LOR, always drop the filedesc lock before exporting fd to sbuf.
trociny [Thu, 27 Jun 2013 19:14:03 +0000 (19:14 +0000)]
To avoid LOR, always drop the filedesc lock before exporting fd to sbuf.

Reviewed by: kib
MFC after: 3 days

11 years agoUpdate T5 register ranges. This is so that regdump skips over registers
np [Thu, 27 Jun 2013 18:59:07 +0000 (18:59 +0000)]
Update T5 register ranges.  This is so that regdump skips over registers
with read side-effects.

MFC after: 3 days

11 years agoAdd UNWINDSVCFRAME to provide the unwind pseudo ops to allow us to unwind
andrew [Thu, 27 Jun 2013 18:54:18 +0000 (18:54 +0000)]
Add UNWINDSVCFRAME to provide the unwind pseudo ops to allow us to unwind
past a trapframe.

Use this macro in exception_exit as it is the function the unwinder enters
as the functions that store the frame setting lr to point to it.

11 years ago- Add vnode-backed swap space specification support. This is enabled when
hrs [Thu, 27 Jun 2013 18:28:45 +0000 (18:28 +0000)]
- Add vnode-backed swap space specification support.  This is enabled when
  device names "md" or "md[0-9]*" and a "file" option are specified in
  /etc/fstab like this:

  md    none    swap    sw,file=/swap.bin       0       0

- Add GBDE/GELI encrypted swap space specification support, which
  rc.d/encswap supported.  The /etc/fstab lines are like the following:

  /dev/ada1p1.bde       none    swap    sw      0       0
  /dev/ada1p2.eli       none    swap    sw      0       0

  .eli devices accepts aalgo, ealgo, keylen, and sectorsize as options.

  swapctl(8) can understand an encrypted device in the command line
  like this:

  # swapctl -a /dev/ada2p1.bde

- "-L" flag is added to support "late" option to defer swapon until
  rc.d/mountlate runs.

- rc.d script change:

    rc.d/encswap -> removed
    rc.d/addswap -> just display a warning message if $swapfile is defined
    rc.d/swap1 -> renamed to rc.d/swap
    rc.d/swaplate -> newly added to support "late" option

These changes alleviate a race condition between device creation/removal
and swapon/swapoff.

MFC after: 1 week
Reviewed by: wblock (manual page)

11 years agoFix build.
glebius [Thu, 27 Jun 2013 10:42:09 +0000 (10:42 +0000)]
Fix build.

11 years agoAdd support for the NTT Docomo L-02C card.
remko [Thu, 27 Jun 2013 09:08:07 +0000 (09:08 +0000)]
Add support for the NTT Docomo L-02C card.
(This file didn't get along in the previous commit)

PR: 180017
Submitted by: Masaharu FUJITA
Glanced at by: imp
MFC after: 1 week

11 years agoAdd support for the NTT Docomo L-02C card.
remko [Thu, 27 Jun 2013 09:06:00 +0000 (09:06 +0000)]
Add support for the NTT Docomo L-02C card.

PR: 180017
Submitted by: Masaharu FUJITA
Glanced at by: imp
MFC after: 1 week

11 years ago- Request non-cached memory for framebuffer
gonzo [Thu, 27 Jun 2013 00:33:08 +0000 (00:33 +0000)]
- Request non-cached memory for framebuffer
- Properly probe/initialize syscons

11 years agoMove definitions required by userland applications out of acpica_machdep.h.
jkim [Thu, 27 Jun 2013 00:22:40 +0000 (00:22 +0000)]
Move definitions required by userland applications out of acpica_machdep.h.

11 years agoMerge ACPICA 20130626.
jkim [Thu, 27 Jun 2013 00:16:54 +0000 (00:16 +0000)]
Merge ACPICA 20130626.

11 years agoAdd firmware replacement and activation support to nvmecontrol(8) through
jimharris [Thu, 27 Jun 2013 00:08:25 +0000 (00:08 +0000)]
Add firmware replacement and activation support to nvmecontrol(8) through
a new firmware command.

NVMe controllers may support up to 7 firmware slots for storing of
different firmware revisions.  This new firmware command supports
firmware replacement (i.e. firmware download) with or without immediate
activation, or activation of a previously stored firmware image.  It
also supports selection of the firmware slot during replacement
operations, using IDENTIFY information from the controller to
check that the specified slot is valid.

Newly activated firmware does not take effect until the new controller
reset, either via a reboot or separate 'nvmecontrol reset' command to the
same controller.

Submitted by: Joe Golio <joseph.golio@emc.com>
Obtained from: EMC / Isilon Storage Division
MFC after: 3 days

11 years agoAdd log page support to nvmecontrol(8) through a new logpage command.
jimharris [Wed, 26 Jun 2013 23:53:54 +0000 (23:53 +0000)]
Add log page support to nvmecontrol(8) through a new logpage command.

This includes pretty printers for all of the standard NVMe log pages
(Error, SMART/Health, Firmware), as well as hex output for non-standard
or vendor-specific log pages.

Submitted by: Joe Golio <joseph.golio@emc.com>
Obtained from: EMC / Isilon Storage Division
MFC after: 3 days

11 years agoConsistently cast ACPICA 64-bit integer types when we print them.
jkim [Wed, 26 Jun 2013 23:52:10 +0000 (23:52 +0000)]
Consistently cast ACPICA 64-bit integer types when we print them.

11 years agoAdd generalized helper functions for printing hex data as part of
jimharris [Wed, 26 Jun 2013 23:46:31 +0000 (23:46 +0000)]
Add generalized helper functions for printing hex data as part of
nvmecontrol commands.

Submitted by: Joe Golio <joseph.golio@emc.com>
Obtained from: EMC / Isilon Storage Division
MFC after: 3 days

11 years agoEnsure that a device node is specified when invoking the identify or reset
jimharris [Wed, 26 Jun 2013 23:41:07 +0000 (23:41 +0000)]
Ensure that a device node is specified when invoking the identify or reset
commands.

Also improve the checking of device node names, so that better error
messages are displayed when incorrect names are specified.

Sponsored by: Intel
MFC after: 3 days

11 years agoRemove remaining uio-related code.
jimharris [Wed, 26 Jun 2013 23:37:11 +0000 (23:37 +0000)]
Remove remaining uio-related code.

The nvme_physio() function was removed quite a while ago, which was the
only user of this uio-related code.

Sponsored by: Intel
MFC after: 3 days

11 years agoFail any passthrough command whose transfer size exceeds the controller's
jimharris [Wed, 26 Jun 2013 23:32:45 +0000 (23:32 +0000)]
Fail any passthrough command whose transfer size exceeds the controller's
max transfer size.  This guards against rogue commands coming in from
userspace.

Also add KASSERTS for the virtual address and unmapped bio cases, if the
transfer size exceeds the controller's max transfer size.

Sponsored by: Intel
MFC after: 3 days

11 years agoUse MAXPHYS to specify the maximum I/O size for nvme(4).
jimharris [Wed, 26 Jun 2013 23:27:17 +0000 (23:27 +0000)]
Use MAXPHYS to specify the maximum I/O size for nvme(4).

Also allow admin commands to transfer up to this maximum I/O size, rather
than the artificial limit previously imposed.  The larger I/O size is very
beneficial for upcoming firmware download support.  This has the added
benefit of simplifying the code since both admin and I/O commands now use
the same maximum I/O size.

Sponsored by: Intel
MFC after: 3 days

11 years agoCreate #defines for NVME_CTRLR_PREFIX and NVME_NS_PREFIX for the "nvme"
jimharris [Wed, 26 Jun 2013 23:20:08 +0000 (23:20 +0000)]
Create #defines for NVME_CTRLR_PREFIX and NVME_NS_PREFIX for the "nvme"
and "ns" strings, rather than hardcoding the string values throughout the
nvmecontrol code base.

Sponsored by: Intel
MFC after: 3 days

11 years agoAdd an nvme_function structure array, defining the name, C function and
jimharris [Wed, 26 Jun 2013 23:11:20 +0000 (23:11 +0000)]
Add an nvme_function structure array, defining the name, C function and
usage message for each nvmecontrol command.  This helps reduce some code
clutter both now and for future commits which will add logpage and
firmware support to nvmecontrol(8).

Also move helper function prototypes to the end of the header file, after
the per-command functions.

Sponsored by: Intel
MFC after: 3 days

11 years agoBreak out code related to the reset command into a separate source file.
jimharris [Wed, 26 Jun 2013 23:05:48 +0000 (23:05 +0000)]
Break out code related to the reset command into a separate source file.

Sponsored by: Intel
MFC after: 3 days

11 years agoBreak out code related to the perftest command into a separate file.
jimharris [Wed, 26 Jun 2013 23:02:48 +0000 (23:02 +0000)]
Break out code related to the perftest command into a separate file.

Sponsored by: Intel
MFC after: 3 days

11 years agoBreak out code related to the identify command into a separate source file.
jimharris [Wed, 26 Jun 2013 23:00:42 +0000 (23:00 +0000)]
Break out code related to the identify command into a separate source file.

Sponsored by: Intel
MFC after: 3 days

11 years agoBreak out code related to the devlist command into a separate source file.
jimharris [Wed, 26 Jun 2013 22:58:59 +0000 (22:58 +0000)]
Break out code related to the devlist command into a separate source file.

Sponsored by: Intel
MFC after: 3 days

11 years agoAdd separate *_usage() functions for each nvmecontrol command.
jimharris [Wed, 26 Jun 2013 22:53:21 +0000 (22:53 +0000)]
Add separate *_usage() functions for each nvmecontrol command.

Sponsored by: Intel
MFC after: 3 days

11 years agoAdd an open_dev() helper function to be used in the various places
jimharris [Wed, 26 Jun 2013 22:50:49 +0000 (22:50 +0000)]
Add an open_dev() helper function to be used in the various places
where nvmecontrol needs to open a handle to a controller or namespace
device node.

Sponsored by: Intel
MFC after: 3 days

11 years agoFor ATA_PASSTHROUGH commands, pretend isci(4) supports multiword DMA
jimharris [Wed, 26 Jun 2013 22:08:45 +0000 (22:08 +0000)]
For ATA_PASSTHROUGH commands, pretend isci(4) supports multiword DMA
by treating it as UDMA.

This fixes a problem introduced in r249933/r249939, where CAM sends
ATA_DSM_TRIM to SATA devices using ATA_PASSTHROUGH_16.  scsi_ata_trim()
sets protocol as DMA (not UDMA) which is for multi-word DMA, even
though no such mode is selected for the device.  isci(4) would fail
these commands which is the correct behavior but not consistent with
other HBAs, namely LSI's.

smh@ did some further testing on an LSI controller, which rejected
ATA_PASSTHROUGH_16 commands with mode=UDMA_OUT, even though only
a UDMA mode was selected on the device.  So this precludes adding
any kind of mode detection in CAM to determine which mode to use on
a per-device basis.

Sponsored by: Intel
Discussed with: scottl, smh
Reported by: scottl
Tested by: scottl
MFC after: 3 days

11 years agoIn the Xen block front driver, take advantage of backends that
gibbs [Wed, 26 Jun 2013 20:39:07 +0000 (20:39 +0000)]
In the Xen block front driver, take advantage of backends that
support cache flush and write barrier commands.

sys/dev/xen/blkfront/block.h:
Add per-command flag that specifies that the I/O queue must
be frozen after this command is dispatched.  This is used
to implement "single-stepping".

Remove the unused per-command flag that indicates a polled
command.

Add block device instance flags to record backend features.

Add a block device instance flag to indicate the I/O queue
is frozen until all outstanding I/O completes.

Enhance the queue API to allow the number of elements in a
queue to be interrogated.

Prefer "inline" to "__inline".

sys/dev/xen/blkfront/blkfront.c:
Formalize queue freeze semantics by adding methods for both
global and command-associated queue freezing.

Provide mechanism to freeze the I/O queue until all outstanding
I/O completes.  Use this to implement barrier semantics
(BIO_ORDERED) when the backend does not support
BLKIF_OP_WRITE_BARRIER commands.

Implement BIO_FLUSH as either a BLKIF_OP_FLUSH_DISKCACHE
command or a 0 byte write barrier.  Currently, all publicly
available backends perform a diskcache flush when processing
barrier commands, and this frontend behavior matches what
is done in Linux.

Simplify code by using new queue length API.

Report backend features during device attach and via sysctl.

Submitted by: Roger Pau MonnĂ©
Submitted by: gibbs (Merge with new driver queue API, sysctl support)

11 years agoMinor sorting.
pfg [Wed, 26 Jun 2013 19:43:22 +0000 (19:43 +0000)]
Minor sorting.

MFC after: 3 days

11 years agoFix some UTF-8 chars slipped into r252204 via copy/paste.
mav [Wed, 26 Jun 2013 09:56:33 +0000 (09:56 +0000)]
Fix some UTF-8 chars slipped into r252204 via copy/paste.

11 years agoAdd the EEPROM start offset and size in so ath_ahb will use the
adrian [Wed, 26 Jun 2013 05:03:47 +0000 (05:03 +0000)]
Add the EEPROM start offset and size in so ath_ahb will use the
supplied EEPROM data.

This needs to die in a fire and replaced with the firmware API
to store the calibration data early ..

11 years agoAdd in the vendor / device id so the ath_ahb glue works.
adrian [Wed, 26 Jun 2013 05:02:47 +0000 (05:02 +0000)]
Add in the vendor / device id so the ath_ahb glue works.

11 years agoAdd the wireless support to the AR933x base kernel, as the support
adrian [Wed, 26 Jun 2013 05:02:30 +0000 (05:02 +0000)]
Add the wireless support to the AR933x base kernel, as the support
actually works now.

11 years agoAdd in two new configuration options for the current generation SoCs.
adrian [Wed, 26 Jun 2013 05:01:50 +0000 (05:01 +0000)]
Add in two new configuration options for the current generation SoCs.

11 years agoExtend the AHB code to work on chips besides the AR9130.
adrian [Wed, 26 Jun 2013 04:58:25 +0000 (04:58 +0000)]
Extend the AHB code to work on chips besides the AR9130.

The AHB code:

* hard coded the AR9130 device id;
* assumes a 4k flash calibration space.

This code now extends this:

* hint.ath.X.eepromsize now overrides the eeprom range, instead of 4k
* hint.ath.X.device_id and hint.ath.X.vendor_id can now be overridden.

Tested:

* AR9330 board (Carambola 2)

11 years agoIf EEPROM data is provided by the AHB layer, copy it in.
adrian [Wed, 26 Jun 2013 04:53:33 +0000 (04:53 +0000)]
If EEPROM data is provided by the AHB layer, copy it in.

This is required for the flash layer code to correctly work.

Tested:

* AR9330 SoC (carambola 2)

11 years agoRe-enable the channel set code for the AR933x.
adrian [Wed, 26 Jun 2013 04:46:43 +0000 (04:46 +0000)]
Re-enable the channel set code for the AR933x.

This required a HAL change to map the 2GHz frequency back to an IEEE
channel number in order to fetch some value(s) to program in.

11 years agoAdd a HAL local routine to map the 2GHz channel frequency to an IEEE
adrian [Wed, 26 Jun 2013 04:46:03 +0000 (04:46 +0000)]
Add a HAL local routine to map the 2GHz channel frequency to an IEEE
channel.

There's some HAL code in the AR9300 HAL that requires a back-mapping
and using the net80211 code isn't appropriate here.

11 years agoIn preparation to support AR933x SoC builds, allow the AH_SUPPORT_AR93(xx)
adrian [Wed, 26 Jun 2013 04:34:45 +0000 (04:34 +0000)]
In preparation to support AR933x SoC builds, allow the AH_SUPPORT_AR93(xx)
options to set the AR9300 HAL defines.

This enables the Wasp, Hornet (and later Scorpion) SoC code.

11 years agoAdd a stub AR93xx RF module just to keep the linker happy.
adrian [Wed, 26 Jun 2013 04:33:57 +0000 (04:33 +0000)]
Add a stub AR93xx RF module just to keep the linker happy.

When building AR933x test images, I'd like to only build only the ar9300
HAL.  To do this, it needs to supply an RF linker entry or it won't compile.

Tested:

* AR933x test builds

11 years agoDocument ifconfig_wlanX="HOSTAP"
rpaulo [Wed, 26 Jun 2013 04:33:32 +0000 (04:33 +0000)]
Document ifconfig_wlanX="HOSTAP"

11 years agosed: use getline() instead of fgetln().
pfg [Wed, 26 Jun 2013 04:14:19 +0000 (04:14 +0000)]
sed: use getline() instead of fgetln().

In BSD, fgetln() available in libc but in Illumos the Solaris port had to
include it internally. It also seems to have caused problems [1].

Aid portability by using getline() instead.

Reference:
https://www.illumos.org/issues/3820 [1]

Submitted by: Johann 'Myrkraverk' Oskarsson <johann@myrkraverk.com>
Reviewed by: dds
MFC after: 2 weeks

11 years agoImplement ifconfig_wlanX="HOSTAP".
rpaulo [Wed, 26 Jun 2013 04:00:52 +0000 (04:00 +0000)]
Implement ifconfig_wlanX="HOSTAP".

Not only this is a bit cleaner, it allows multiple instances of hostapd to be
running on the system host, useful for simultaneous dual-band WiFi.
This is similar to ifconfig_wlanX="WPA" but it uses /etc/hostapd-wlanX.conf.
Compatibility with hostapd_enable=YES/NO was kept.

Reviewed by: adrian

11 years agoPrint the 'setting internal ...' message only with bootverbose.
rpaulo [Wed, 26 Jun 2013 02:56:54 +0000 (02:56 +0000)]
Print the 'setting internal ...' message only with bootverbose.

11 years ago- Adding myself to the committers-ports.dot
feld [Wed, 26 Jun 2013 02:21:37 +0000 (02:21 +0000)]
- Adding myself to the committers-ports.dot
- Adding myself to the freebsd calendar

Approved by: swills (mentor)

11 years agoDon't blidly clear GPIOs configuration. Just use firmware configured
yongari [Wed, 26 Jun 2013 01:15:40 +0000 (01:15 +0000)]
Don't blidly clear GPIOs configuration. Just use firmware configured
one.  This change also fixes non-working traffic LED on BCM57780.

Submitted by: Masanobu SAITOH <msaitoh@NetBSD.org>
Tested by: Alexander Milanov <a@amilanov.com>

11 years ago - Resolve bucket recursion issues by passing a cookie with zone flags
jeff [Wed, 26 Jun 2013 00:57:38 +0000 (00:57 +0000)]
 - Resolve bucket recursion issues by passing a cookie with zone flags
   through bucket_alloc() to uma_zalloc_arg() and uma_zfree_arg().
 - Make some smaller buckets for large zones to further reduce memory
   waste.
 - Implement uma_zone_reserve().  This holds aside a number of items only
   for callers who specify M_USE_RESERVE.  buckets will never be filled
   from reserve allocations.

Sponsored by: EMC / Isilon Storage Division

11 years agoMFV r252215:
delphij [Tue, 25 Jun 2013 22:14:32 +0000 (22:14 +0000)]
MFV r252215:

Restore a previous behavior before r251646, where when destructing
ZFS snapshot, the ioctl would return ENOENT when it hit any of
them in the errlist (the new behavior was only return ENOENT when
all returns error).

Illumos ZFS issues:
  3829 fix for 3740 changed behavior of zfs destroy/hold/release ioctl

MFC after: 1 week

11 years agoDiff reduction against Illumos, no real change to code itself.
delphij [Tue, 25 Jun 2013 21:51:52 +0000 (21:51 +0000)]
Diff reduction against Illumos, no real change to code itself.

This marks vendor branch revision 252213 as merged, the actual code was
committed in r245479.

MFC after: 1 week

11 years agoA few mostly cosmetic nits to aid in debugging:
jhb [Tue, 25 Jun 2013 20:23:08 +0000 (20:23 +0000)]
A few mostly cosmetic nits to aid in debugging:
- Call lock_init() first before setting any lock_object fields in
  lock init routines.  This way if the machine panics due to a duplicate
  init the lock's original state is preserved.
- Somewhat similarly, don't decrement td_locks and td_slocks until after
  an unlock operation has completed successfully.

11 years agoSeveral improvements to rmlock(9). Many of these are based on patches
jhb [Tue, 25 Jun 2013 18:44:15 +0000 (18:44 +0000)]
Several improvements to rmlock(9).  Many of these are based on patches
provided by Isilon.
- Add an rm_assert() supporting various lock assertions similar to other
  locking primitives.  Because rmlocks track readers the assertions are
  always fully accurate unlike rw_assert() and sx_assert().
- Flesh out the lock class methods for rmlocks to support sleeping via
  condvars and rm_sleep() (but only while holding write locks), rmlock
  details in 'show lock' in DDB, and the lc_owner method used by
  dtrace.
- Add an internal destroyed cookie so that API functions can assert
  that an rmlock is not destroyed.
- Make use of rm_assert() to add various assertions to the API (e.g.
  to assert locks are held when an unlock routine is called).
- Give RM_SLEEPABLE locks their own lock class and always use the
  rmlock's own lock_object with WITNESS.
- Use THREAD_NO_SLEEPING() / THREAD_SLEEPING_OK() to disallow sleeping
  while holding a read lock on an rmlock.

Submitted by: andre
Obtained from: EMC/Isilon

11 years agoAdd include path to the ar9300 HAL.
adrian [Tue, 25 Jun 2013 17:50:48 +0000 (17:50 +0000)]
Add include path to the ar9300 HAL.

11 years agoAdd Qlogic 10Gb Ethernet Driver for Qlogic 8100 Series CNA Adapter
davidcs [Tue, 25 Jun 2013 17:50:22 +0000 (17:50 +0000)]
Add Qlogic 10Gb Ethernet Driver for Qlogic 8100 Series CNA Adapter
Driver version (v2.0.0)

Submitted by: David C Somayajulu (davidcs@freebsd.org) QLogic Corporation
Approved by: George Neville-Neil (gnn@freebsd.org)

11 years agoIf daily_status_security_inline is set, the rc value needs to be
jhb [Tue, 25 Jun 2013 15:45:31 +0000 (15:45 +0000)]
If daily_status_security_inline is set, the rc value needs to be
forced to 3 so that the output of this script is always displayed.
In fact, setting this flag is identical to setting
daily_status_security_output to an empty string.  To make the logic
less confusing, change the behavior of daily_status_security_inline
such that it just forces daily_status_security_output to an empty
string and then applies the normal logic.

PR: conf/178611
Submitted by: Jason Unovitch <jason.unovitch@gmail.com>
MFC after: 3 days

11 years agoAdd bunch of names for Seagate and HGST vennor-specififc ASC/ASCQ codes.
mav [Tue, 25 Jun 2013 10:50:17 +0000 (10:50 +0000)]
Add bunch of names for Seagate and HGST vennor-specififc ASC/ASCQ codes.

11 years agoAdd test for SATA registers writability and skip using them if it failed.
mav [Tue, 25 Jun 2013 09:15:49 +0000 (09:15 +0000)]
Add test for SATA registers writability and skip using them if it failed.

There are some systems reported, where PCI BAR(5), used for SATA registers
access, is present, but not functional.  Attempt to use it brakes devices
detection logic.  Try to detect those cases on attach by setting and testing
some bits in SControl register.  If bits are unsettable, fallback to legacy
ATA without hot-plug detection, speed control/reporting, etc.

MFC after: 2 weeks

11 years agoRegenerate usb.conf after r252196.
kevlo [Tue, 25 Jun 2013 07:17:30 +0000 (07:17 +0000)]
Regenerate usb.conf after r252196.

Spotted by: rpaulo

11 years agoAdd support for D-Link DWA-131.
kevlo [Tue, 25 Jun 2013 06:43:04 +0000 (06:43 +0000)]
Add support for D-Link DWA-131.

11 years agoAdd Lenovo USB 2.0 Ethernet adapter to the list of supported
yongari [Tue, 25 Jun 2013 00:29:13 +0000 (00:29 +0000)]
Add Lenovo USB 2.0 Ethernet adapter to the list of supported
devices.

11 years agoAdd Lenovo USB 2.0 Ethernet adapter.
yongari [Tue, 25 Jun 2013 00:26:30 +0000 (00:26 +0000)]
Add Lenovo USB 2.0 Ethernet adapter.

PR: usb/179920
MFC After: 1 week

11 years agoDue to the routing related networking kernel redesign work
qingli [Tue, 25 Jun 2013 00:10:49 +0000 (00:10 +0000)]
Due to the routing related networking kernel redesign work
in FBSD 8.0, interface routes have been returened to the
applications without the RTF_GATEWAY bit. This incompatibility
has caused some issues with Zebra, Qugga and the like.
This patch provides the RTF_GATEWAY flag bit in returned interface
routes so to behave similarly to pre 8.0 systems.

Reviewed by:     hrs
Verified by:     mackn at opendns dot com

11 years agoFlag mpt(4) as supporting unmapped I/O; all necessary conversion actually
marius [Mon, 24 Jun 2013 21:27:15 +0000 (21:27 +0000)]
Flag mpt(4) as supporting unmapped I/O; all necessary conversion actually
already has been done as part of r246713 except for a comment update.

MFC after: 3 days

11 years agogcc: configuration fix.
pfg [Mon, 24 Jun 2013 21:13:58 +0000 (21:13 +0000)]
gcc:  configuration fix.

-Fix configuration support for FreeBSD 10 and 11.

Note this change is based on GCC-SVN-131197 with permission
by gerald@ .

Reported by: jmallet

11 years agoMore whitespace.
dteske [Mon, 24 Jun 2013 20:58:54 +0000 (20:58 +0000)]
More whitespace.

11 years agoWhitespace.
dteske [Mon, 24 Jun 2013 20:39:07 +0000 (20:39 +0000)]
Whitespace.

11 years agogcc: add some configuration and references.
pfg [Mon, 24 Jun 2013 20:38:27 +0000 (20:38 +0000)]
gcc: add some configuration and references.

-Add configure support for FreeBSD 10 and 11.
-Adapt a threading fix to gnu POSIX95 (which we don't use).
-Refer to a bug fix for the disabled vrptree support.

This is all useless in our current build but it is included
for convenience in case someone may want to re-package our
older gcc.

Reviewed by: gerald (long ago)

11 years agoAdd missing Obsolete Files
eadler [Mon, 24 Jun 2013 20:36:12 +0000 (20:36 +0000)]
Add missing Obsolete Files

Submitted by: Kurt Lidl <lidl@pi-coral.com>

11 years agoRemove request to email suggestions and fixes to the author.
eadler [Mon, 24 Jun 2013 19:57:25 +0000 (19:57 +0000)]
Remove request to email suggestions and fixes to the author.
He is no longer involved with the FreeBSD project.

While here: remove "no known bugs" and related.  This isn't present in other manual pages.

PR: docs/179914

11 years agoMake the order of operations for lib/msun more clear.
eadler [Mon, 24 Jun 2013 19:12:17 +0000 (19:12 +0000)]
Make the order of operations for lib/msun more clear.

Tested with md5 sum of object code

Reported by: swildner@DragonFlyBSD.org
Submitted by: bde

11 years agoDisable hw.pci.realloc_bars by default. It wasn't needed for the original
jhb [Mon, 24 Jun 2013 18:30:44 +0000 (18:30 +0000)]
Disable hw.pci.realloc_bars by default.  It wasn't needed for the original
tester of this fix, and realloc_bars breaks some other cases as a small
BAR that is reallocated can end up grabbing space needed by a much larger
BAR in the existing window of a PCI-PCI bridge.

MFC after: 3 days

11 years agoTypo in comment.
glebius [Mon, 24 Jun 2013 13:36:16 +0000 (13:36 +0000)]
Typo in comment.

11 years agoRevert r252126, r252124, r252121.
gjb [Mon, 24 Jun 2013 09:18:41 +0000 (09:18 +0000)]
Revert r252126, r252124, r252121.

Submitted by: hrs

11 years agoWhen RX checksum offloading is active, AX88772B will prepend a
yongari [Mon, 24 Jun 2013 05:18:31 +0000 (05:18 +0000)]
When RX checksum offloading is active, AX88772B will prepend a
checksum header.  The header contains a received frame length but
the defined length for AX88772B is different with other ASIX
controllers.  When the RX checksum is off, AX88772B controller does
not prepend a checksum header so driver has to use normal header
length mask.
This change should fix RX errors when RX checksum offloading is
off.

Tested by: kevlo
MFC After: 1 week

11 years agoDelete the nd6 entries associated with an off-link prefix
qingli [Mon, 24 Jun 2013 05:01:13 +0000 (05:01 +0000)]
Delete the nd6 entries associated with an off-link prefix
if the same prefix cannot be found on an alternative
interface.

Reviewed by: hrs
MFC after: 1 week

11 years agoDocument the "-o" option added by r252138.
rmacklem [Mon, 24 Jun 2013 02:24:22 +0000 (02:24 +0000)]
Document the "-o" option added by r252138.

MFC after: 10 days

11 years agoAdd a new "-o" option to the gssd which forces gss_init_sec_context()
rmacklem [Mon, 24 Jun 2013 02:19:23 +0000 (02:19 +0000)]
Add a new "-o" option to the gssd which forces gss_init_sec_context()
to use DES and the associated old style GSS initialization token.
This appears to be required for some non-FreeBSD servers to
get a kerberized NFS mount to work. Also, ignore some signals when daemonized,
which might fix the gssd from "disappearing" without leaving a core dump.
Given the tight timeframe for the FreeBSD9.2 release, I have
committed this while waiting for code review. I will commit
changes recommended by the review in a separate commit.

Reviewed by: dfr (pending)
MFC after: 10 days

11 years agoBump shlib for dialog because ABI has changed when version 1.2 was imported
bapt [Sun, 23 Jun 2013 21:59:52 +0000 (21:59 +0000)]
Bump shlib for dialog because ABI has changed when version 1.2 was imported

Reviewed by: kib

11 years agoRemove variable added by ru@ in r94940 that is no longer used by source or ports...
eadler [Sun, 23 Jun 2013 20:34:54 +0000 (20:34 +0000)]
Remove variable added by ru@ in r94940 that is no longer used by source or ports in share/mk/sys.mk

PR: conf/155737
Reviewed by: ed (two years ago!)
Exp-Run by: bdrewery

11 years agoTrim copyright years, and add the current year.
gjb [Sun, 23 Jun 2013 20:24:43 +0000 (20:24 +0000)]
Trim copyright years, and add the current year.

11 years agoRevert previous change to uark.c (restore previous rev), which was
thomas [Sun, 23 Jun 2013 20:22:49 +0000 (20:22 +0000)]
Revert previous change to uark.c (restore previous rev), which was
committed by mistake.

11 years agoRewrap after previous commit.
gjb [Sun, 23 Jun 2013 20:20:34 +0000 (20:20 +0000)]
Rewrap after previous commit.

11 years agoFix minor typo in comment
thomas [Sun, 23 Jun 2013 20:19:51 +0000 (20:19 +0000)]
Fix minor typo in comment

11 years agoAdd a new macro, release.current.release, to denote the head/ branch
gjb [Sun, 23 Jun 2013 20:19:00 +0000 (20:19 +0000)]
Add a new macro, release.current.release, to denote the head/ branch
with the -RELEASE suffix.  This fixes the incorrect text on the -CURRENT
errata page from showing '10.0-CURRENT' followed by 'until 9.1-RELEASE is
released.'

11 years agoAdd missing copyright years.
gjb [Sun, 23 Jun 2013 19:47:59 +0000 (19:47 +0000)]
Add missing copyright years.

11 years agoCache the Open Firmware CPU properties at attach time, so we don't always
jhibbits [Sun, 23 Jun 2013 14:20:54 +0000 (14:20 +0000)]
Cache the Open Firmware CPU properties at attach time, so we don't always
enter it at runtime to get static data.

11 years agoAdd a newline character to the end of the "Check URL again" error message
dteske [Sun, 23 Jun 2013 10:51:26 +0000 (10:51 +0000)]
Add a newline character to the end of the "Check URL again" error message
because long URLs do not induce extra height despite wrapping by dialog(1).
NOTE: For even longer lines, the cursor up/down keys work to scroll through

11 years agoMerge r248313 from stable/9 sysinstall(8) to head bsdconfig(8):
dteske [Sun, 23 Jun 2013 10:48:26 +0000 (10:48 +0000)]
Merge r248313 from stable/9 sysinstall(8) to head bsdconfig(8):
Add support for installation directly via HTTP.

While we're here, remove the menu-item for Passive FTP (since moving to
ftp(1) and switching FTPMODE to `auto' by default -- see r251613 -- the
single remaining FTP menu-item works for both ftp.f.o and ftp-archive.f.o;
previously each requiring separately active versus passive both work with
the `auto' setting). In scripting you still have mediaSetFTPActive and
mediaSetFTPPassive but the remaining FTP menu-item uses mediaSetFTP which
defaults to `auto' (aforementioned SVN r251613).

11 years agoImplement the $probe_only parameter (previously unimplemented).
dteske [Sun, 23 Jun 2013 10:04:23 +0000 (10:04 +0000)]
Implement the $probe_only parameter (previously unimplemented).

11 years agoFine-tune the parsing of the URL. Re-order, comment, and add debugging to
dteske [Sun, 23 Jun 2013 09:41:47 +0000 (09:41 +0000)]
Fine-tune the parsing of the URL. Re-order, comment, and add debugging to
each case of unique URL format.

11 years agoDefine and use e2fs_lbn_t in ext2fs.
pfg [Sun, 23 Jun 2013 02:44:42 +0000 (02:44 +0000)]
Define and use e2fs_lbn_t in ext2fs.

In line to what is done in UFS, define an internal type
e2fs_lbn_t for the logical block numbers.

This change is basically a no-op as the new type is unchanged
(int32_t) but it may be useful as bumping this may be required
for ext4fs.

Also, as pointed out by Bruce Evans:

-Use daddr_t for daddr in ext2_bmaparray(). This seems to
improve reliability with the reallocblks option.
- Add a cast to the fsbtodb() macro as in UFS.

Reviewed by: bde
MFC after: 3 days

11 years agoIf MAKE_CONF and/or SRC_CONF are set and not character devices,
gjb [Sat, 22 Jun 2013 22:12:24 +0000 (22:12 +0000)]
If MAKE_CONF and/or SRC_CONF are set and not character devices,
copy to chroot to use with release build.

Submitted by: Garrett Cooper (original version)
Approved by: kib (mentor)
MFC after: 1 week