[Lucid][pull request] Update to 2.6.32.12 stable kernel

Tue May 25 13:36:14 UTC 2010

On Thu, May 06, 2010 at 02:32:02PM +0200, Stefan Bader wrote:
> Hi All,
> 
> http://bugs.launchpad.net/bugs/575853
> 
> The 2.6.32.12 stable kernel has recently been released.
> 
> http://kernel.org/pub/linux/kernel/v2.6/ChangeLog-2.6.32.12
> http://kernel.org/pub/linux/kernel/v2.6/patch-2.6.32.12.bz2
> 
> The 2.6.32.12 upstream stable kernel brings in 182 patches (DRM excluded). As we
> carry the DRM stack from 2.6.33, all patches that affect DRM have been dropped
> and replaced by DRM patches from 2.6.33.3 (19 patches).
> 
> The upstream process for stable tree updates is quite similar in scope to the
> Ubuntu SRU process. e.g: each patch demonstrably fixes a bug and is vetted by
> upstream either by originating directly from Linus' tree or by being minimally
> backported from some patch in Linus' tree.

Ok, below is my review of the heap added by 2.6.32.12 and 2.6.33.3.
Overall I think there are some big patches in there, but being in sync
is pretty much a requirement to continuing to take stable at all.
Overall the XFS stuff is the most worrying, but a relativly sparsly used
feature (which appears to be broken before this patch set).  Overall I
am for applying:

Acked-by: Andy Whitcroft <apw at canonical.com>

Overall comments:

About 10% of these fixes are to XFS some of which are large, they all sound
and look reasonable, but due to the complexity are hard to be 100% sure of.
They are however in a non-commonly deployed filesystem so overall risk to
the common user is zero, and most XFS users will be server type users and
highly likely to test well before deployment.

ARM: 6031/1: fix Thumb-2 decompressor -- this patch may well be already
on some of our arm branches, it seems familiar.  Somthing to watch out for
when rebasing.

There is a heap of patches which rewrite the rw-semaphores for amd64
which make them have a larger range.  This is a risky looking patch set
as it modifies the assembly used for locking.  That said it is likely
it would explode in our face instantly if wrong.  As this would be of
use most on servers and really only affects amd64 it is probabally
lower risk than it might otherwise be, and of signicant benefit for
larger server installations.  Overall I guess I am happy.

There seems to be a heap of changes related to b43 and DMA errors, maybe
just maybe it might start to support our wireless with these changes.
We should remember to test.

There are a heap of KVM changes, some on AMD some general, so we need
broad testing on KVM with these changes.  Need to get -server testing this
as soon as it is in pre-proposed.

Lots of DRM related changes, need to get the X guys testing this as soon
as its in pre-proposed.

Specific comments are below.

-apw

commit e96420da1bfebf6e538a99f70646f8cf12197dce
Author: Nikolaus Schulz <microschulz at web.de>

    fat: fix buffer overflow in vfat_create_shortname()

# Type: NEW
Limits buffer use to 4 characters exactly, protects against wrap.  Looks
good.

commit 18f93d9855529a249877c619ab92b2c1f4d784d1
Author: Oleg Nesterov <oleg at redhat.com>

    oom: fix the unsafe usage of badness() in proc_oom_score()

# Type: NEW
Looks to do what it says on the tin.  Protects against use of already
freed pointers.  Looks good.

commit c2437cc0c590f14b76a17c2e8e1a47c5e803ffed
Author: Christoph Hellwig <hch at infradead.org>

    xfs: simplify inode teardown

# Type: NEW
Complex patch but in a filesystem not commonly deployed by our userbase
so though it has a medium risk due to size it is low risk to Ubuntu as a
whole.

commit 5e433990be8bbefad461b84d1be11306f2c1decc
Author: Christoph Hellwig <hch at infradead.org>

    xfs: fix mmap_sem/iolock inversion in xfs_free_eofblocks

# Type: NEW
Again in a filesystem not commonly deployed.  Prevents a deadlock due to
lock inversion on the mmap_sem.  Looks good.

commit c75a31a4236cac10b35bdcffe984be8fbc3059ac
Author: Christoph Hellwig <hch at infradead.org>

    xfs: I/O completion handlers must use NOFS allocations

# Type: NEW
Ensures that we pass KM_NOFS during allocations to prevent memory pressure
reaping the inode caches and recursing into the same code, deadlocking
on the XFS locks.

commit 4adc60848ad8dfc603967af08e256eef90a418d3
Author: Andy Poling <andy at realbig.com>

    xfs: Wrapped journal record corruption on read at recovery

# Type: NEW
Handles wrapped journal record loading.  Looks sane.

commit 4e47d01f7cd3cac21fc30760f23879aa11fb6f3d
Author: Jason Gunthorpe <jgunthorpe at obsidianresearch.com>

    xfs: Fix error return for fallocate() on XFS

# Type: NEW
Ensures errors are mapped into syscall errors by negating.  Looks ok.

commit ea853e6aa95f63bfaa56ad73d4eff11318f518d3
Author: Christoph Hellwig <hch at infradead.org>

    xfs: check for not fully initialized inodes in xfs_ireclaim

# Type: NEW
Adds a sensible and obvious new sanity check.  Looks good.

commit 50db1fa67c09f625a09878f1fbeed584ad033876
Author: Christoph Hellwig <hch at infradead.org>

    xfs: fix timestamp handling in xfs_setattr

# Type: NEW
Cleans up stamp handling, code looks correct to my eye and cleaner and
simpler.  Looks fine.

commit c6b84f4202638981d6eb0960c2dc46195c3e945c
Author: Dave Chinner <david at fromorbit.com>

    xfs: Don't flush stale inodes

# Type: NEW
Prevents stale inodes from being written.  Seems ok.

commit dface51c06914af0ffd4dc65a3d5c97d7088e859
Author: Dave Chinner <david at fromorbit.com>

    xfs: Ensure we force all busy extents in range to disk

# Type: NEW
New code finds the highest transaction number of all ranges in the
extent.  Looks good.

commit 2da6ffee05c0fa320e6023373e99a4eb976c9758
Author: Dave Chinner <david at fromorbit.com>

    xfs: reclaim inodes under a write lock

# Type: NEW
Appears to do what is claimed, again in a less commonly deployed filesystem
so general risk is low.  Looks ok.

commit be64d1c631ca8dd58040dc472d5cec7b03dae6dd
Author: Dave Chinner <david at fromorbit.com>

    xfs: Avoid inodes in reclaim when flushing from inode cache

# Type: NEW
Looks like it does what is claimed.  Looks ok.

commit ad93ba560be9034b10154deb8d432b3de4c84ca4
Author: Dave Chinner <david at fromorbit.com>

    xfs: reclaim all inodes by background tree walks

# Type: NEW
Switches all reclaim to background.  Seems ok.

commit d2a7471956dc5e8f52ee69600204c01a6286db56
Author: Dave Chinner <david at fromorbit.com>

    xfs: fix stale inode flush avoidance

# Type: NEW
A fix for the above patches, seems to also be sane.

commit 48f75a72446b6f5051d2285cf051835ba903cea4
Author: Dave Chinner <david at fromorbit.com>

    xfs: xfs_swap_extents needs to handle dynamic fork offsets

# Type: NEW
Seems to do something sensible rejecting incompatible changes requested
in defragmentation.

commit 94994e6ceb30e89a17d7351d09b1497085f17ec3
Author: Christoph Hellwig <hch at infradead.org>

    xfs: quota limit statvfs available blocks

# Type: NEW
Limits available space and total space to quota, meaning that the available
is always <= to total.  Looks sane.

commit 15d4f435c0f2fa1afcba132b406b89b1a5d6ed71
Author: Dave Chinner <david at fromorbit.com>

    xfs: don't hold onto reserved blocks on remount, ro

# Type: NEW
Frees any reserve on switch to read-only and reaquires reserve on move
to read-write.  Seems sane.

commit a37947edf8eeacfd777e96e53a04a43316fd4cec
Author: Christoph Hellwig <hch at lst.de>

    xfs: remove invalid barrier optimization from xfs_fsync

# Type: NEW
Removes an invalid optimisation.  Ok.

commit 47c6ccd092f769dffa78484d4798f839bb4ce6e5
Author: Dave Chinner <david at fromorbit.com>

    xfs: Non-blocking inode locking in IO completion

# Type: NEW
Allows io completion to be blocked by requeueing later.  Seems to fix
an issue where XFS is loopback mounted on a file on an XFS filesystem.
Looks ok.

commit 83e49e6f1f6c0331b357b9f18a17e9f603061981
Author: Christoph Hellwig <hch at infradead.org>

    xfs: fix locking for inode cache radix tree tag updates

# Type: NEW
Moves to write locks for radix tree tag updates.  Looks ok.

commit ece1e83da1bab05175e82a4dd18d55a8ee1a8f56
Author: Matt Fleming <matt at console-pimps.org>

    sh: Enable the mmu in start_secondary()

# Type: NEW
Non core architecture.  Zero risk.

commit f0c49dee20bbae7d9bdf4bf53a5d275ed061d73e
Author: Andrew Stubbs <ams at codesourcery.com>

    sh: Fix FDPIC binary loader

# Type: NEW
Non core architecture.  Zero risk.

commit d0965e49dc14c0c921a3e67a6a7ec08355c1a7eb
Author: Mike Christie <michaelc at cs.wisc.edu>

    libiscsi: Fix recovery slowdown regression

# Type: NEW
Ensures we mark a scsi path bad as soon as possible to allow, particularly
important for the dm-multipath use.  Looks ok.

commit 80d621b4397758b2e894fe873951e20cba7c9ede
Author: Matt Helsley <matthltc at us.ibm.com>

    Freezer: Fix buggy resume test for tasks frozen with cgroup freezer

# Type: NEW
Ensures that a cgroup which is in the process of being frozen and some
which are actually frozen but not yet marked as such, are not resumed
during regular resume from suspend.  Looks ok.

commit 9e403ab83a93ce06194fbb4c9becf7b513d39f26
Author: Wey-Yi Guy <wey-yi.w.guy at intel.com>

    iwlwifi: counting number of tfds can be free for 4965

# Type: NEW
Pulls in a missing hunk of a previous fix.  Ok.

commit bc0a5cf007bc9412ba836f7596e55f2057995750
Author: Stefan Bader <stefan.bader at canonical.com>

    Revert "(pre-stable) iwlwifi: fix nfreed--"

# Type: NEW
Revert in favour of upstream version.

commit 8932b3e13099cd8769deac50f239de20b8714122
Author: Stanislaw Gruszka <sgruszka at redhat.com>

    iwlwifi: fix nfreed--

# Type: NEW
Already applied, upstream version.

commit bdb35a60272470dee61643aa1c1cda65afdb495b
Author: Dan Carpenter <error27 at gmail.com>

    iwlwifi: range checking issue

# Type: NEW
Ensures we do not step out of an allocated array.  Looks right.

commit aee447476bc73f804b4a1237b48e91127bfa7719
Author: Valentin Longchamp <valentin.longchamp at epfl.ch>

    setup correct int pipe type in ar9170_usb_exec_cmd

# Type: NEW
Fills in the correct USB type for command packets.  Looks ok.

commit ce01db5186dc6d17b1587bbc7dd79abe77eedb04
Author: Johannes Berg <johannes.berg at intel.com>

    mac80211: move netdev queue enabling to correct spot

# Type: NEW
Ensure we push out any pending packets to the real device before we
allow packets from above to queue.

commit 06bdb0534d42b2d56016ad14d291c3e1e633a0ee
Author: Wey-Yi Guy <wey-yi.w.guy at intel.com>

    mac80211: tear down all agg queues when restart/reconfig hw

# Type: NEW
Ensure we clear down the aggregation queues between stoping and restarting
the hw queues.

commit ad9686e25035f3d2dded17d2b0b5538ea7ed9cce
Author: Thomas Mingarelli <Thomas.Mingarelli at hp.com>

    WATCHDOG: hpwdt - fix lower timeout limit

# Type: NEW
Lowers the timeout limit, seems arbitrary at either level, but 5s is
pretty long 30s an eternity.

commit dfbf3c433cab63eff07297f6916eabbd39bcf57d
Author: Seth Heasley <seth.heasley at intel.com>

    WATCHDOG: iTCO_wdt: TCO Watchdog patch for additional Intel Cougar Point DeviceIDs

# Type: NEW
Cougar Point enablement, adds ids.  Looks ok.

commit 42e89c0d18e3cc38c35c45f70e3298e33d6bb8b0
Author: Thomas Gleixner <tglx at linutronix.de>

    genirq: Force MSI irq handlers to run with interrupts disabled

# Type: NEW
Disables interrupt nesting for MSI interrupts.  Reasoning sounds correct.
Looks ok.

commit f960388fee3411bbebc6c6257da373f157c43611
Author: Oleg Nesterov <oleg at redhat.com>

    tty: release_one_tty() forgets to put pids

# Type: NEW
Closes memory leak, leaking pid structures.  Looks ok.

commit 5a1124791db180621c2aa117e4f420ab474a405d
Author: Éric Piel <eric.piel at tremplin-utc.net>

    lis3: fix show rate for 8 bits chips

# Type: NEW
Corrects 12 bit reads.  Looks ok.

commit 0ac561eb6c9ad453a1578ec1c20fe2a3b819ec10
Author: Alan Cox <alan at linux.intel.com>

    pata_ali: Fix regression with old devices

# Type: NEW
Fixes support for older devices, device id specific.  Looks ok.

commit 7d1e052358c8edc39c7e4a7210d97d4ad1277250
Author: Jiri Kosina <jkosina at suse.cz>

    HID: fix oops in gyration_event()

# Type: NEW
Prevents OOPS for HID device which the hid subsystem has not claimed.  Looks
safe.

commit 2510e1597366893c6b8a7d2d348b7d8f9f258254
Author: Anton Blanchard <anton at samba.org>

    raw: fsync method is now required

# Type: NEW
Fixes sync on raw devices.  Looks ok.

commit 55badc470af8de7090e37fe8947d508ab1ea399e
Author: Wu Fengguang <fengguang.wu at intel.com>

    readahead: fix NULL filp dereference

# Type: NEW
Fixes oops from btrfs.  Protects against null filp.  Looks sane.

commit 424199a228881098d2e5e0ec50e11fa23371f55f
Author: Dan Carpenter <error27 at gmail.com>

    ALSA: mixart: range checking proc file

# Type: NEW
Looks ok.

commit 65782addea84d0d514f22f775881ce484c2ff1b5
Author: Daniel T Chen <crimsun at ubuntu.com>

    ALSA: hda: Fix 0 dB offset for Lenovo Thinkpad models using AD1981

# Type: NEW
Fixes upper volume to 0dB which matches where distortion starts.  Tested
by Jane Silber :).  Looks ok.

commit 6c1f7de8f15b83780a2816b294485e390fdad969
Author: Andreas Herrmann <herrmann.der.user at googlemail.com>

    x86, amd: Get multi-node CPU info from NodeId MSR instead of PCI config space

# Type: NEW
Moves to using CPU MSRs to work out cpu/numa node topology.  Looks ok.

commit 930f8302477b2cfffcfe6b309c299e6dec146f1d
Author: Jiri Slaby <jslaby at suse.cz>

    resource: move kernel function inside __KERNEL__

# Type: NEW
Hides a kernel only function from userspace headers.  Looks sane.

commit 83dae423bff8cf84902cfa1f4325b981a09c2ebf
Author: Stefan Bader <stefan.bader at canonical.com>

    Revert "backlight: mbp_nvidia_bl - add five more MacBook variants"

# Type: NEW
Revert in favour of upstream patch.

commit 2d61077cc57d656ed3a71ce9b4a50fcd73a546cd
Author: Evan McClain <evan.mcclain at gatech.edu>

    backlight: mbp_nvidia_bl - add five more MacBook variants

# Type: NEW
Upstream patch already reviewed.

commit 80cc2a92546074a6031bda3fa7eb1e5dff618b21
Author: Stefan Bader <stefan.bader at canonical.com>

    Revert "(pre-stable) pata_via: Add VIA VX900 support"

# Type: NEW
Revert in favour of upstream patch.

commit 08697e77775c9f5973143fb64089f510e03c4cba
Author: JosephChan at via.com.tw <JosephChan at via.com.tw>

    pata_via: Add VIA VX900 support

# Type: NEW
Upstream patch already reviewed.

commit bab16589de54c8b90ac0d4fbd90694214b422c4d
Author: Eric Sandeen <sandeen at redhat.com>

    ext3: Don't update the superblock in ext3_statfs()

# Type: NEW
Do not update the on-disk superblock when stating, this could lead to
out of sync checksums and fsck errors.  Looks ok.

commit d45049b95f192a481ed8020f078907c83a8b34e9
Author: Eric Sandeen <sandeen at redhat.com>

    ext3: journal all modifications in ext3_xattr_set_handle

# Type: NEW
Moves inode clears under journal control.  Looks sensible.

commit 7ba2d8ae50f5d8f2920ef6e0247f3be0816cda60
Author: Alan Jenkins <alan-jenkins at tuffmail.co.uk>

    eeepc-laptop: disable cpu speed control on EeePC 701

# Type: NEW
Prevent cpu speed control being use by default, it is not apparently
supported but sort of works for some people.  Sane default.

commit c56146769c440345783e097698fba24711c07323
Author: Corentin Chary <corentincj at iksaif.net>

    eeepc-laptop: dmi blacklist to disable pci hotplug code

# Type: NEW
Disable wlan hotplug as the connectors disable another device all together.
Looks sane.

commit 58c6f385a0d196718a3cc0d53f96bff578cdcd87
Author: Corentin Chary <corentincj at iksaif.net>

    eeepc-laptop: add hotplug_disable parameter

# Type: NEW
Adds a manual parameter to disabe hotplug, allowing online debugging.
Looks ok.

commit 000b9d3b65692bcfb835bab942385b589e61513d
Author: Corentin Chary <corentincj at iksaif.net>

    eeepc-laptop: disable wireless hotplug for 1201N

# Type: NEW
Blacklists wireless hotplug on another eeepc model.  Looks good.

commit 76c00f8cf8fee7695f0a9eb9e56ed4dfedbc303b
Author: Alan Jenkins <alan-jenkins at tuffmail.co.uk>

    eeepc-laptop: disable wireless hotplug for 1005PE

# Type: NEW
Blacklists wireless hotplug on another eeepc model.  Looks good.

commit 25f26d2cdf570aeb4ec9cfeb8098305c80079c47
Author: Tejun Heo <tj at kernel.org>

    libata: disable NCQ on Crucial C300 SSD

# Type: NEW
Disables command queuing on older SSDs from Crucial.  Looks ok.

commit ab13416d5ff1e9edbdf75b8fe6f85f7119649bd0
Author: Suresh Jayaraman <sjayaraman at suse.de>

    cifs: Fix a kernel BUG with remote OS/2 server (try #3)

# Type: NEW
A VILE workaround for some mad OS/2 server.  Looks like it would do
what is claimed, and exposure to these should be limited.  Ok.

commit c1bdac8322df46db2b2e547c0a22532a9a849ad5
Author: Steve French <sfrench at us.ibm.com>

    CIFS: initialize nbytes at the beginning of CIFSSMBWrite()

# Type: NEW
Early initialisation of a return pointer.  Looks sensible and safer.

commit 2aa1fa88cc9c41979f5a0900b7e25090ee20668b
Author: Wey-Yi Guy <wey-yi.w.guy at intel.com>

    iwlwifi: need check for valid qos packet before free

# Type: NEW
Ensure we have a valid packet before attempting to free based on tids.
Looks to do what is claimed.  Ok.

commit 02d36b629ad62e0de6aad6ff81ef7f4a477207f9
Author: Rabin Vincent <rabin at rab.in>

    ARM: 6031/1: fix Thumb-2 decompressor

# Type: NEW
Looks valid.  We may have this on our arm branches.  Ok.

commit a777b0379107aa9ed1407496a8af73bd9df9fa9b
Author: Stefan Bader <stefan.bader at canonical.com>

    Revert "(pre-stable) x86-32, resume: do a global tlb flush in S4 resume"

# Type: NEW
Revert in favour of upstream patch.

commit 7f4bf4674c29a1d6b553028aa3c4a7fe0761bc01
Author: Shaohua Li <shaohua.li at intel.com>

    x86-32, resume: do a global tlb flush in S4 resume

# Type: NEW
Upstream patch already reviewed.

commit 72b1df4bd564a6c6f2abf94d735d2be40e42c0e3
Author: Thomas Gleixner <tglx at linutronix.de>

    x86: hpet: Make WARN_ON understandable

# Type: NEW
Clarification of a HPET related warning during operation.  Ok.

commit 87fc22b013563b3f971effc8ea52dc968a76f0d4
Author: Pallipadi, Venkatesh <venkatesh.pallipadi at intel.com>

    x86, hpet: Erratum workaround for read after write of HPET comparator

# Type: NEW
General work around for a common HPET erratum exposed by the fix
modified above which is a work around for another ATI HPET erratum.  Yay.
Looks ok.

commit f2efec3362c2a70c34c5481019e77c6c671c07a4
Author: Suresh Siddha <suresh.b.siddha at intel.com>

    x86: Fix double enable_IR_x2apic() call on SMP kernel on !SMP boards

# Type: NEW
Prevent double initialisation of the apics.  Looks OK.

commit 3ee53a94bdb51d2b8c7ba273cf773cdff2733f17
Author: KOSAKI Motohiro <kosaki.motohiro at jp.fujitsu.com>

    sched: sched_getaffinity(): Allow less than NR_CPUS length

# Type: NEW
Handle mediumly huge machines with NR_CPUS > 1024.  Looks ok to me.

commit d887415504a4b4eec876b4a9b1a4c5a0b329c6f3
Author: Anton Blanchard <anton at samba.org>

    sched: Fix sched_getaffinity()

# Type: NEW
Fixes buffer size checks to be bits.  Looks good.

commit 7d25abf313ed4527f1f4eb22c4d0e0a6b14d7d69
Author: Trond Myklebust <Trond.Myklebust at netapp.com>

    NFSv4: Fall back to ordinary lookup if nfs4_atomic_open() returns EISDIR

# Type: NEW
Does what it claims.  Looks ok.

commit 77b69e3417f1e7503e1d5e2c9d95eb96df88fe36
Author: Trond Myklebust <Trond.Myklebust at netapp.com>

    NFSv4: fix delegated locking

# Type: NEW
Switches locking to a per filesystems flag.  Seems to do what is claimed.
Looks ok.

commit 54d6c44822d253e7b0f62bed8074067086d37847
Author: Joerg Schirottke <master at kanotix.com>

    ALSA: hda - add a quirk for Clevo M570U laptop

# Type: NEW
Adds a simple quirk.  Device specific.  Looks ok.

commit 0a3dd61ab9f0931e04db4a53905c8a4072384db0
Author: Takashi Iwai <tiwai at suse.de>

    ALSA: usb - Fix Oops after usb-midi disconnection

# Type: NEW
Ensure we don't touch things when disconnected.  Seems ok.

commit 9bae40392a593c293d74856e13490abef4d17915
Author: Jerome Oufella <jerome.oufella at savoirfairelinux.com>

    hwmon: (sht15) Fix sht15_calc_temp interpolation function

# Type: NEW
Clean up temperature reporting.  Device specific.  Looks sane.

commit e195b3bf559ae0372be045d3607b06f5225423cc
Author: Jean Delvare <khali at linux-fr.org>

    hwmon: (sht15) Properly handle the case CONFIG_REGULATOR=n

# Type: NEW
Handle no regulator.  Device specific.  Looks ok.

commit 7531dea4e4d002cfc9a0b55aa2b0a0247afdec5c
Author: Joerg Roedel <joerg.roedel at amd.com>

    x86/amd-iommu: Use helper function to destroy domain

# Type: NEW
Use after free of io domains.  Use the correct free helper here.  Looks ok.

commit 1e2de1a78bccf974ae7a1490f5a2e9b826a827f7
Author: Chris Wright <chrisw at sous-sol.org>

    x86/amd-iommu: enable iommu before attaching devices

# Type: NEW
Ensure the IOMMU is enabled early enough that kdump invalidations take effect.
Looks ok.

commit 7e62b98dd7e5200c33a879dba5b318425aa92949
Author: Chris Wright <chrisw at sous-sol.org>

    Revert "x86: disable IOMMUs on kernel crash"

# Type: NEW
This prevents in progress DMAs from completing which would be very bad.
Relies on the IOMMU being setup correctly in kdump.

commit ccbb3421d447a9fa99432f6a183300ee67d698a7
Author: Borislav Petkov <borislav.petkov at amd.com>

    x86, lib: Add wbinvd smp helpers

# Type: NEW
Pre-requisite of the next patch, looks sane.

commit d83d45e7073f35708a1885d83d665ecb6d02b3d0
Author: Borislav Petkov <borislav.petkov at amd.com>

    x86, cacheinfo: Fix disabling of L3 cache indices

# Type: NEW
Seems to do what is claimed.  Looks scarey and ok.

commit e09dcaa163aa474352bb0779e44c55e872225944
Author: Borislav Petkov <borislav.petkov at amd.com>

    intel-agp: Switch to wbinvd_on_all_cpus

# Type: NEW
Simplify code using the new helpers above.  Looks good.

commit c7da12c9d350cdb917f093f94bd05f303849571e
Author: Borislav Petkov <borislav.petkov at amd.com>

    x86, cacheinfo: Add cache index disable sysfs attrs only to L3 caches

# Type: NEW
Looks to to change behaviour as described for different cache levels.  Ok.

commit 79c032bec526113549643c756e8099402ac9ffca
Author: Borislav Petkov <borislav.petkov at amd.com>

    x86, cacheinfo: Calculate L3 indices

# Type: NEW
Add support for looking up L3 indices at cpu start time.  Looks ok.

commit 568b9aa2cd7475c3d7a48d8f4e98c64ad835c43f
Author: Borislav Petkov <borislav.petkov at amd.com>

    x86, cacheinfo: Remove NUMA dependency, fix for AMD Fam10h rev D1

# Type: NEW
Fixes when built !NUMA.  Looks ok.

commit 3e5f81464a5827853db12a65e32335a35b439351
Author: Borislav Petkov <borislav.petkov at amd.com>

    x86, cacheinfo: Enable L3 CID only on AMD

# Type: NEW
Reorganises code to complile !AMD. Post-requisite for above  Looks ok.

commit cc3e88ba790efe8bad30edeada7628ea41a759d0
Author: Eugene Teo <eugeneteo at kernel.sg>

    vgaarb: fix "target=default" passing

# Type: NEW
Literally fixes detection of target=default.  This is actually disabled
in Lucid.  Looks ok.

commit a5a3f10e34c5459daf3325ce0309d271acda9b51
Author: Linus Torvalds <torvalds at linux-foundation.org>

    x86-32: clean up rwsem inline asm statements

# Type: NEW
Cleans up asm to be sharable between i386 and amd_64.  Looks ok, will
explode quickly if wrong.  Pre-req for below patches.

commit 4e0ea7627f042e760fa2048f35dbec1b7a2fa0e1
Author: Linus Torvalds <torvalds at linux-foundation.org>

    x86: clean up rwsem type system

# Type: NEW
Introduces a new rwsem type.  Pre-req for below patches.

commit 4c69b15924513a15e6f911f1562f950d62ba25ef
Author: H. Peter Anvin <hpa at zytor.com>

    x86-64, rwsem: 64-bit xadd rwsem implementation

# Type: NEW
Introduces a 64 bit version for amd64.  Looks ok.

commit c39bbf8a62332dab222d3405a1bda6c579085bbf
Author: Linus Torvalds <torvalds at linux-foundation.org>

    x86-64: support native xadd rwsem implementation

# Type: NEW
Introduces an assembly version of above.  Looks damned scarey.  Likely
to explode instantly and hideously if wrong.

commit eed4a4bea686d42de87c4d2d728a41d4ac696558
Author: Linus Torvalds <torvalds at linux-foundation.org>

    x86: Fix breakage of UML from the changes in the rwsem system

# Type: NEW
Cleanup patch following above patches.  Looks ok.

commit 390181e6d7e27c563e4b9150148016cd85b4c9d2
Author: Avi Kivity <avi at redhat.com>

    x86-64, rwsem: Avoid store forwarding hazard in __downgrade_write

# Type: NEW
Simplifies and optimised the above assembly.  Gah.

commit f03c333bfcc759857b4c136a6c57fbafc957473a
Author: Al Viro <viro at zeniv.linux.org.uk>

    fix NFS4 handling of mountpoint stat

# Type: NEW
Correctly follow all mountpoints in a stack as required by the spec.
Looks to do what is claimed.  Ok.

commit 85f1e35c7ec843605b0fcf0ec18d64e739800789
Author: Kiyoshi Ueda <k-ueda at ct.jp.nec.com>

    dm mpath: fix stall when requeueing io

# Type: NEW
Handles requeued ios immediatly.  Looks ok.

commit 97018fe4f841b0db8b5ca4f61afcaa68d9420243
Author: Andrew Perepechko <andrew.perepechko at sun.com>

    quota: Fix possible dq_flags corruption

# Type: NEW
Switches to atomic operations on flags everywhere.  Definatly the safe
direction.  Looks ok.

commit 1ac4302e87e936756562a6105a7916b02efe2d9b
Author: Bernd Porr <BerndPorr at f2s.com>

    Staging: comedi: fix usbdux timeout bug

# Type: NEW
Fixes incorrectly low timesouts.  Looks sane.

commit f68bf5d0efa99e3f890a79c84712a55cabfd4213
Author: Bernd Porr <berndporr at f2s.com>

    Staging: comedi: usbdux.c: fix locking up of the driver when the comedi ringbuffer runs empty

# Type: NEW
Fixes reader wakeup on buffer empty.  Looks ok.

commit ad14faabba56ee5fe79a912ed042a49789d2106e
Author: Mark Fasheh <mfasheh at suse.com>

    ocfs2: set i_mode on disk during acl operations

# Type: NEW
Ensures we write out the i_mode when changed.  Looks sane.

commit 47723ac6d52c1b024b9a1e195a637e6c6b44b81b
Author: Tao Ma <tao.ma at oracle.com>

    ocfs2: Change bg_chain check for ocfs2_validate_gd_parent.

# Type: NEW
Does what it claims.  Looks ok.

commit 8c1735df306a7ad8113550c5343d61e1d9773025
Author: Sachin Prabhu <sprabhu at redhat.com>

    9p: Skip check for mandatory locks when unlocking

# Type: NEW
Looks ok.

commit ce2c24073272508e34b69cf0b548af9516ad2481
Author: Mike Christie <michaelc at cs.wisc.edu>

    fc class: fail fast bsg requests

# Type: NEW
Looks to do what it claims.  Ok.

commit 36fd2e2a69a27afc5ff4663728c52605db19f368
Author: Mike Christie <michaelc at cs.wisc.edu>

    SCSI: add scsi target reset support to scsi ioctl

# Type: NEW
Adds the ioctl entry for an existing path.  Looks ok.

commit a72351924607c69bce7191983f678e66a8c413be
Author: Youquan Song <youquan.song at linux.intel.com>

    PCIe AER: prevent AER injection if hardware masks error reporting

# Type: NEW
Prevents error injection when masked in hardware.  Seems sensible.

commit b184f04ffbe42af9e047de426fe1bb32ebe7a6c2
Author: Mike Travis <travis at sgi.com>

    vgaarb: Fix VGA arbiter to accept PCI domains other than 0

# Type: NEW
Allows your VGA to be in any PCI domain.  Looks ok.

commit f214daae37d44c5249f6047c15ecf5998c59cd53
Author: Harish Zunjarrao <harish.zunjarrao at qlogic.com>

    SCSI: fc-transport: Use packed modifier for fc_bsg_request structure.

# Type: NEW
Ensure the user visible structure is sized consistantly in 32 and 64 bit.

commit faad3bcbfbd0f6d0f50b14adcd7ca54b873f6e60
Author: Mike Travis <travis at sgi.com>

    pci: Update pci_set_vga_state() to call arch functions

# Type: NEW
Looks like a pre-req patch for later.  Looks sane.

commit 8618f5a8d6d5893074f7e1485ae108fbae022452
Author: Paul Mundt <lethal at linux-sh.org>

    PCI: kill off pci_register_set_vga_state() symbol export.

# Type: NEW
Clean up previous patch.  Looks ok.

commit b8f375ab08b2d23a583e240ddbd84810b88c6c1b
Author: Andrew Patterson <andrew.patterson at hp.com>

    PCI: fix nested spinlock hang in aer_inject

# Type: NEW
Fixes recusrsive locking.  Looks ok.

commit 4cc651f4e3b808633b782bf7e82d9bb0c6b993ba
Author: Eli Cohen <eli at mellanox.co.il>

    IPoIB: Fix TX queue lockup with mixed UD/CM traffic

# Type: NEW
Fixes IB lockup.  Ok.

commit d5d0b7aa4ac48b3910264f9d9d18f6b5b2a97157
Author: Seth Heasley <seth.heasley at intel.com>

    x86/PCI: irq and pci_ids patch for Intel Cougar Point DeviceIDs