[APPLIED][Karmic] SRU: Update to 2.6.31.7 stable kernel

Stefan Bader stefan.bader at canonical.com
Wed Jan 6 16:50:31 UTC 2010


Leann Ogasawara wrote:
> Hi All,
> 
> https://bugs.edge.launchpad.net/ubuntu/karmic/+source/linux/+bug/494633
> 
> SRU Justification/Impact:
> 
> The 2.6.31.7 stable kernel has recently been released.  The 2.6.31.7
> Upstream Stable Kernel brings in 123 patches. 5 of the 123 patches have
> already been released in a security update. Also 1 patch replaces an
> existing SAUCE patch we were temporarily carrying.  Stefan has applied
> and pushed these patches to a stable branch in his tree for review.
> 
> git://kernel.ubuntu.com/smb/ubuntu-karmic.git stable
> 
> http://kernel.ubuntu.com/git?p=smb/ubuntu-karmic.git;a=shortlog;h=refs/heads/stable
> 
> The upstream process for stable tree updates is quite similar in scope
> to the Ubuntu SRU process, e.g., each patch has to demonstrably fix a
> bug, and each patch is vetted by upstream by originating either directly
> from Linus' tree or in a minimally backported form of that patch. The
> 2.6.31.7 upstream stable patch set is now available. It should be
> included in the Ubuntu kernel as well.
> 
> TEST CASE: TBD
> 
> http://kernel.org/pub/linux/kernel/v2.6/ChangeLog-2.6.31.7
> 
> http://kernel.org/pub/linux/kernel/v2.6/patch-2.6.31.7.bz2
> 
> The following bugs are noted as fixed via patches from 2.6.31.7:
> 
> https://bugs.launchpad.net/ubuntu/+bug/460664
>  * Fixed with "drm: work around EDIDs with bad htotal/vtotal values"
> 
> https://bugs.edge.launchpad.net/ubuntu/+bug/435352
>  * Fixed with "USB: work around for EHCI with quirky periodic schedules"
> 
> The 123 stable patches are as follows.  I've inlined my own review
> comments (marked by #).
> 
> Thanks,
> Leann
> 
> commit b88b724607b4e9472398c7b3ebc5f8cb3f1d98d6
> Author: Ryusuke Konishi <konishi.ryusuke at lab.ntt.co.jp>
> 
>     nilfs2: fix kernel oops in error case of nilfs_ioctl_move_blocks
> 
> # Properly moves error checking for duplication of blocks prior to list
> insertion in nilfs_ioctl_move_inode_block
> 
> commit 6804b96f48a32bca16fee60212ad8e4b201bb99f
> Author: Jeff Layton <jlayton at redhat.com>
> 
>     cifs: don't use CIFSGetSrvInodeNumber in is_path_accessible
> 
> # Appears not all servers support this so remove it in favor of just
> doing a normal CIFSSMBQPathInfo which is noted to be widely supported
> and shouldn't increase overhead significantly.
> 
> commit ad4316779dc2ca73c3d2b11d0f13ff3d3b0fa100
> Author: Jeff Layton <jlayton at redhat.com>
> 
>     cifs: clean up handling when server doesn't consistently support inode numbers
> 
> # Doesn't allow query of server inode numbers when it's not supported.
> 
> commit fb598664ca7b2b807380be5a526378b4877e5e0d
> Author: Suresh Jayaraman <sjayaraman at suse.de>
> 
>     cifs: clear server inode number flag while autodisabling
> 
> # Seems to be more of a typo fix for above commit ad43167 to actually
> clear a flag ie. s/CIFS_MOUNT_SERVER_INUM/~CIFS_MOUNT_SERVER_INUM/
> 
> commit 0fbad7ae3c0d00c5e5f5951b1de5ef536dec2a5e
> Author: Steve French <sfrench at us.ibm.com>
> 
>     CIFS: fix oops in cifs_lookup during net boot
> 
> # Fixes Null pointer dereference.  Looks sane.
> 
> commit 24fa7e7df85aae43e2ac0af24a56ca093a613460
> Author: Steve French <sfrench at us.ibm.com>
> 
>     CIFS: Duplicate data on appending to some Samba servers
> 
> # Prevents use of problematic O_APPEND flag. 
> 
> commit 174388981606378894ab74cae9467d5bbf0e5830
> Author: Dave Jones <davej at redhat.com>
> 
>     gdth: Prevent negative offsets in ioctl CVE-2009-3080
> 
> # Already included and fixed with a security update
> 
> commit 146d0c086cc8b6d580cb08b6e8ad149f91c1e03d
> Author: Larry Finger <Larry.Finger at lwfinger.net>
> 
>     rtl8187: Fix kernel oops when device is removed when LEDS enabled
> 
> # Moved cancel_delayed_work_sync() to after rtl8187_unregister_led().
> 
> commit 98bc571940095198eec1e4b9af70bf9024b5f539
> Author: NeilBrown <neilb at suse.de>
> 
>     md: don't clear endpoint for resync when resync is interrupted.
> 
> # Even if resync is interrupted, don't reset the resync endpoint until
> it's actually reached the endpoint.  
> 
> commit 8a7963564a8288a2c98d1c924d09ac0697b8f92c
> Author: NeilBrown <neilb at suse.de>
> 
>     md/raid5: make sure curr_sync_completes is uptodate when reshape starts
> 
> # Update sysfs attribute to always have the correct value.
> 
> commit 57a0aa351bff86bd529c8638a376cf0a18b60eae
> Author: NeilBrown <neilb at suse.de>
> 
>     md/raid1/raid10: add a cond_resched
> 
> # Does what is says and only adds cond_resched()
> 
> commit 44cf344afe9976e27b64c2bac861f5d45009f3fc
> Author: Julian Anastasov <ja at ssi.bg>
> 
>     ALSA: usb-audio: fix combine_word problem
> 
> # Add missing parentheses which fixes some cases of input control names
> having output control names.
> 
> commit b381ea627f10c081d4f45ce601463732bda0e765
> Author: Daniel J Blueman <daniel.blueman at gmail.com>
> 
>     ALSA: hda - Dell Studio 1557 hd-audio quirk
> 
> # Add stac92hd73xx audio quirk for Dell Studio 1557
> 
> commit c20be9b482978de39a489dfc2e2a4ccb8ee9ea56
> Author: Russell King <rmk+kernel at arm.linux.org.uk>
> 
>     ALSA: AACI: fix AC97 multiple-open bug
> 
> # Call snd_ac97_pcm_close() only if aacirun->pcm_open
> 
> commit 2a2c59e64de72e390d427e165d937020c7f417fe
> Author: Russell King <rmk+kernel at arm.linux.org.uk>
> 
>     ALSA: AACI: fix recording bug
> 
> # use aacirun->pcm->r[0].slots (the capture slot info) instead of
> aacirun->pcm->r[1].slots (the double rate slot info)
> 
> commit ea4cf642637ddf61ef992568ea3e960aaa9b609a
> Author: David Woodhouse <dwmw2 at infradead.org>
> 
>     jffs2: Fix memory corruption in jffs2_read_inode_range()
> 
> # prevent corruption by ensuring we stay within the proper range
> 
> commit e38dcb2b06e60459054478d79e22ef179f8ae798
> Author: Clemens Ladisch <clemens at ladisch.de>
> 
>     sound: rawmidi: disable active-sensing-on-close by default
> 
> # Does what it says and disables this by default
> 
> commit 1a65ef117b0bcb58f5e8b97dc477728e98d3a795
> Author: Clemens Ladisch <clemens at ladisch.de>
> 
>     sound: rawmidi: fix checking of O_APPEND when opening MIDI device
> 
> # Check that the substream was opened with O_APPEND if opening it for a
> second time
> 
> commit dfe0b47c1e940dbe2f59133c08c485dee78668c5
> Author: Clemens Ladisch <clemens at ladisch.de>
> 
>     sound: rawmidi: fix double init when opening MIDI device with O_APPEND
> 
> # Substream initialization was being done every time a substream was
> opened.  If the substream was opened multiple times it would get
> re-initialized which resulted in data being dropped.  Now, only
> initialize if substream->use_count is 0.
> 
> commit 80bc5c1318442367dbcdccfa737ec952548088a3
> Author: Clemens Ladisch <clemens at ladisch.de>
> 
>     sound: rawmidi: fix MIDI device O_APPEND error handling
> 
> # If opening the substream fails, make sure to error out and free the
> substream.
> 
> commit 52f9034cfc63fad0aa06eecb1fd943432b6dc9f1
> Author: Soeren Sandmann <sandmann at daimi.au.dk>
> 
>     highmem: Fix race in debug_kmap_atomic() which could cause warn_count to underflow
> 
> # Properly check for warn_count underflow.  Also prevents flood of
> messages.
> 
> commit 8eed84d8062c4b41ac9722b6d121ccdffb508d05
> Author: Soeren Sandmann <sandmann at daimi.au.dk>
> 
>     highmem: Fix debug_kmap_atomic() to also handle KM_IRQ_PTE, KM_NMI, and KM_NMI_PTE
> 
> # Prevents spurious warnings
> 
> commit 22e633d1ba54044a66de913d164987f1fa946eea
> Author: Becky Bruce <beckyb at kernel.crashing.org>
> 
>     powerpc: Fix DEBUG_HIGHMEM build break from d4515646699
> 
> # We don't have this enabled which explains why we haven't seen the
> build issue.  Since we don't have this enabled, this fix won't affect
> us.  If we did enable it, all is does is define dummy definitions.
> 
> commit ee83348407722a76b2cabc6c6628c55876fd51e3
> Author: Russell King <rmk+kernel at arm.linux.org.uk>
> 
>     kmap: fix build errors with DEBUG_HIGHMEM enabled
> 
> # Again, same comment as above for commit 22e633d
> 
> commit d6581525739470f4dfaadcb071011315ea2724d7
> Author: Mike Isely <isely at pobox.com>
> 
>     V4L/DVB (13169): bttv: Fix potential out-of-order field processing
> 
> # Very verbose explanation in commit.  Basically, only conditionally set
> the "bottom" field the prevent out of order processing. 
> 
> commit 21b6d2edf649e48e884c0b1d6414982a5e3d070b
> Author: Mike Isely <isely at pobox.com>
> 
>     V4L/DVB (13170): bttv: Fix reversed polarity error when switching video standard
> 
> # memcmp return value was being mis-interpreted.  Update check to
> proceed if memcmp returns non zero value.
> 
> commit 7e1209105d7bc831bc7aed3e07f780e8e75c0ad3
> Author: Seth Barry <seth at cyberseth.com>
> 
>     V4L/DVB (13109): tda18271: fix signedness issue in tda18271_rf_tracking_filters_init
> 
> # Avoid overflows by using signed calculations
> 
> commit 53017a39386cce1b7b6da3013176364424793115
> Author: Michael Krufky <mkrufky at kernellabs.com>
> 
>     V4L/DVB (13107): tda18271: fix overflow in FM radio frequency calculation
> 
> # Avoid 32bit multiplication overflow by scaling the frequency up front
> 
> commit 27bff6b55ad0fe861530315bb75eb3401fa1b660
> Author: Devin Heitmueller <dheitmueller at kernellabs.com>
> 
>     V4L/DVB (13190): em28xx: fix panic that can occur when starting audio streaming
> 
> # Prevent panic by making sure to initialize variables.
> 
> commit c43d7819934b796560ffc14371dcc69089d1fb81
> Author: Martin Samek <martin at marsark.sytes.net>
> 
>     V4L/DVB (13079): dib0700: fixed xc2028 firmware loading kernel oops
> 
> # This is a temporary workaround to make sure the parent is initialized
> thus preventing the oops
> 
> commit 7b9acdf264761c1a8fdd3696a04e3a47d3a44e23
> Author: Mike Isely <isely at pobox.com>
> 
>     V4L/DVB (13230): s2255drv: Don't conditionalize video buffer completion on waiting processes
> 
> # Don't abort if !waitqueue_active() as some apps might be running in an
> asynchronous manner.
> 
> commit 02caa6be73dbeacafbfa98ecaa39b1dab7d84eb6
> Author: Thomas Gleixner <tglx at linutronix.de>
> 
>     uids: Prevent tear down race
> 
> # Prevents race by moving the init/schedule_delayed work to inside the
> lock protected region
> 
> commit f1d8f4d0822df5c5f06021d221658fe2ad25c0cb
> Author: Rodolfo Giometti <giometti at linux.it>
> 
>     pps: events reporting fix up
> 
> # Only report pps events which were requested.
> 
> commit 073493a20768a109ed1e6e744a2fe88e8bbd2fab
> Author: Rodolfo Giometti <giometti at linux.it>
> 
>     pps: locking scheme fix up for PPS_GETPARAMS
> 
> # Add spinlock around params userspace has r/w access to
> 
> commit 4549526e54461f6be8b229703d3c9488da1eb131
> Author: Scott Valentine <svalentine at concentris-systems.com>
> 
>     rtc: v3020: fix v3020_mmio_read_bit()
> 
> # Add double negation to ensure a correct return value
> 
> commit bd7d2351462306b3244c3572c48670c77aa21528
> Author: Heiko Carstens <heiko.carstens at de.ibm.com>
> 
>     fs: add missing compat_ptr handling for FS_IOC_RESVSP ioctl
> 
> # Add proper compat_ptr conversion for arg
> 
> commit 0c98d0c6dcfb9f3f8608053b41192731630b41cd
> Author: KAMEZAWA Hiroyuki <kamezawa.hiroyu at jp.fujitsu.com>
> 
>     memcg: fix wrong pointer initialization at page migration when memcg is disabled.
> 
> # mem never gets initialized if mem_cgroup_disabled() resulting in a
> panic.  Ensure mem is initialize to Null.
> 
> commit 280292d5ed25f6799d4f301dea43d1c0a3c6ad68
> Author: Sukadev Bhattiprolu <sukadev at linux.vnet.ibm.com>
> 
>     pidns: fix a leak in /proc dentries and inodes with pid namespaces.
> 
> # It's no longer necessary to check for PF_EXITING flag prior to calling
> shrink_dcache_parent() since shrink_dcache_parent() is no longer
> responsible for flushing dentries for other fs's.  Remove the check thus
> avoiding possible dentries not being flushed.
> 
> commit 88b117ce4ae1a68c5be558c182b1bea1967cac73
> Author: Mel Gorman <mel at csn.ul.ie>
> 
>     page allocator: Do not allow interrupts to use ALLOC_HARDER
> 
> # Exiting logic was slightly changed and allowing interrupts to set
> ALLOC_HARDER.  Add a check to prevent this so the logic is now reverted
> back to its original intent and reduces the number of allocation
> failures.
> 
> commit ffb83b9e7d0e935ccc665ed09f2f7273828c9cfb
> Author: Mel Gorman <mel at csn.ul.ie>
> 
>     page allocator: always wake kswapd when restarting an allocation attempt after direct reclaim failed
> 
> # moves the restart: goto a few lines up to ensure kswapd is re-woken
> after direct reclaim fails.  It was noted that kswapd was always
> re-woken in the past for direct reclaim fails so this is really
> returning the code to it's original functionality.
> 
> commit 8f34cea85b340d1057d9251e38d6b000898c659f
> Author: Alan Cox <alan at linux.intel.com>
> 
>     tty_port: If we are opened non blocking we still need to raise the carrier
> 
> # If the tty is opened O_NONBLOCK, still raise the carrier.  This is a
> common bug that was inherited from other drivers.
> 
> commit 7fcbd9ff6078339a4aed470b2ddde43960291e98
> Author: Alan Cox <alan at linux.intel.com>
> 
>     tty: cp210x: Fix carrier handling
> 
> # Add dtr_rts and carrier_raised methods to properly set the modem lines
> and correct blocking semantics.
> 
> commit 49cb656ba98a71f5dbc0f7c5dd2b1e35d32e673d
> Author: Libin Yang <libin.yang at amd.com>
> 
>     USB: ohci: quirk AMD prefetch for USB 1.1 ISO transfer
> 
> # Avoids USB 1.1 device failures for SB800
> 
> commit 5f1e43635144fd054218d3423ab4c86c2ea16755
> Author: Alan Stern <stern at rowland.harvard.edu>
> 
>     USB: usbmon: fix bug in mon_buff_area_shrink
> 
> # Fix bug in error path by recalculating the size of the area reserved.
> 
> commit 25838ae5be73ec1b27866b49961b50153bef9fd8
> Author: Zhang Le <r0bertz at gentoo.org>
> 
>     USB: option.c: add support for D-Link DWM-162-U5
> 
> # Add new product id for existing vendor
> 
> commit c859382e8e6c1347d0b34dc72b2c096b07113fd1
> Author: Henry Gebhardt <gebhardt at astro.uni-tuebingen.de>
> 
>     USB: cdc_acm: Fix race condition when opening tty
> 
> # Call tasklet_schedule() after tty_port_block_til_ready(), not before.
> 
> commit 4d10d9eb04a6fdbb9cd9579835f7006751e558d1
> Author: Sarah Sharp <sarah.a.sharp at linux.intel.com>
> 
>     USB: xhci: Fix bug memory free after failed initialization.
> 
> # Check that pointer is valid before de-referencing.
> 
> commit 03a3cf4a7ac175a511a4d9d89a6b2e9eefbb73fe
> Author: Sarah Sharp <sarah.a.sharp at linux.intel.com>
> 
>     USB: xhci: Fix TRB physical to virtual address translation.
> 
> # Avoid endless looping by doing the following: 1) if the entire ring
> has been traversed then break 2) if one of the ring segments is Null
> then break 3) implement a cur_seg pointer to traverse through the ring
> 
> commit ee003b24971b5d4015f01eb5f90cb204ed42afb4
> Author: Sarah Sharp <sarah.a.sharp at linux.intel.com>
> 
>     USB: xhci: Fix scratchpad deallocation.
> 
> # Prevent oops by calling scratchpad_free() prior to resetting page_size
> to 0
> 
> commit 5a68dad11fc3c75ba97113e5bbd431d20dd36786
> Author: Wey-Yi Guy <wey-yi.w.guy at intel.com>
> 
>     iwlwifi: Use RTS/CTS as the preferred protection mechanism for 6000 series
> 
> # Does what it says and adds rts/cts protection for the iwl-6000 series of cards.
> 
> commit 2b41cc435f74811a0fe4894b9d0a3e2e476cfa53
> Author: Wey-Yi Guy <wey-yi.w.guy at intel.com>
> 
>     iwlwifi: Fix issue on file transfer stalled in HT mode
> 
> # Adds rts/cts protection for iwl1000_bgn_cfg.  pre-requisite patch is
> the above commit 5a68dad
> 
> commit 8bc4be6e44a5b9931b9bf0b9b267e2264f71075d
> Author: Mimi Zohar <zohar at linux.vnet.ibm.com>
> 
>     ima: replace GFP_KERNEL with GFP_NOFS
> 
> # Avoids a potential deadlock by changing from GFP_KERNEL to GFP_NOFS
> 
> commit aa7c7f8c1b47d415f3cca42f0a2aa22d8539860e
> Author: Trond Myklebust <Trond.Myklebust at netapp.com>
> 
>     NFSv4: Fix a cache validation bug which causes getcwd() to return ENOENT
> 
> # Revert an incorrect bitmap change that resulted in a failure to return
> a fileid/inode number.
> 
> commit b79250a273007d1d7f3884c4bc50860c7910b347
> Author: Csaba Henk <csaba at gluster.com>
> 
>     fuse: reject O_DIRECT flag also in fuse_create
> 
> # Error on fuse_create if O_DIRECT flag
> 
> commit f624cb3a3de1c02ae8d7dd87cbe19c2aa376421d
> Author: Mark Brown <broonie at opensource.wolfsonmicro.com>
> 
>     ASoC: Fix suspend with active audio streams
> 
> # Force power down if suspend event received, otherwise stream would
> remain marked as active.
> 
> commit 57ee10d0308f0ae7699fef496a5f9924a82b9903
> Author: Anuj Aggarwal <anuj.aggarwal at ti.com>
> 
>     ASoC: AIC23: Fixing infinite loop in resume path
> 
> # Use the actual reference counter being incremented in the loop
> comparison.
> 
> commit a9f5433f6f88aaad161d23a595a86b3d1ae739fb
> Author: Johannes Berg <johannes at sipsolutions.net>
> 
>     mac80211: fix two remote exploits
> 
> # Already included in security update
> 
> commit b62b52ae0de1bb97f8c9dfe4609270493d77c7c4
> Author: Johannes Berg <johannes at sipsolutions.net>
> 
>     mac80211: fix spurious delBA handling
> 
> # Already included in security update
> 
> commit 8d0d5e22d59bfd6d58df4f4111bea6dd6b9f9920
> Author: Larry Finger <Larry.Finger at lwfinger.net>
> 
>     b43: Work around mac80211 race condition
> 
> # There's a bug in the mac80211 stack which allows a call to TX routine
> after the TX queues have been stopped.  Workaround this by checking if
> the queue has been stopped and refuse to transmit the extra packet.
> Also remove B43_WARN_ON(ring->stopped) noise as it's recognized this is
> not the fault of the b43 driver and only WARN_ON in the event the queue
> is not stopped but is full.
> 
> commit 899da70ca5dc094f506e82c60ef09ce5bd09bb94
> Author: Johannes Berg <johannes at sipsolutions.net>
> 
>     rfkill: fix miscdev ops
> 
> # set rfkill_fops->owner to THIS_MODULE
> 
> commit 57ce46ef790bc67038322ed59c26f6d8eed9ee61
> Author: Roel Kluin <roel.kluin at gmail.com>
> 
>     thinkpad-acpi: fix sign of ERESTARTSYS return
> 
> # Fix return value to be negative
> 
> commit 37ed9bf8e31377284ca719b823e6794386ee0d5b
> Author: Harald Welte <HaraldWelte at viatech.com>
> 
>     Enable ACPI PDC handshake for VIA/Centaur CPUs
> 
> # Call init_intel_pdc for VENDOR_INTEL or VENDOR_CENTAUR
> 
> commit 975894c3493caf60061d228262107d2636863125
> Author: Robert Lowery <rglowery at exemail.com.au>
> 
>     V4L/DVB (13436): cxusb: Fix hang on DViCO FusionHDTV DVB-T Dual Digital 4 (rev 1)
> 
> # Fixes regression after introduction of disable_i2c_gate field
> 
> commit 7087c84b4239187f0196a35dcc577aecf7eca77c
> Author: Hans Verkuil <hverkuil at xs4all.nl>
> 
>     V4L/DVB (13321): radio-gemtek-pci: fix double mutex_lock
> 
> # prevent double lock since gemtek_pci_setfrequency() locks the mutex
> 
> commit a2ddf6aff5cd6f0a6a00e226720900a473480278
> Author: Hans Verkuil <hverkuil at xs4all.nl>
> 
>     V4L/DVB (12948): v4l1-compat: fix VIDIOC_G_STD handling
> 
> # VIDIOC_G_STD ioctl might not be present for radio receivers so ignore
> errors.
> 
> commit 0c8953efd523d634df7a65e7e15f269af46bd384
> Author: Jean-Francois Moine <moinejf at free.fr>
> 
>     V4L/DVB (12280): gspca - sonixj: Remove auto gain/wb/expo for the ov7660 sensor.
> 
> #  Updates ov7660_sensor_init
> 
> commit 8e6666b224aeac899eccbf29020b1556b9f8f151
> Author: Denis Loginov <dinvlad at gmail.com>
> 
>     V4L/DVB (12356): gspca - sonixj: Webcam 0c45:6148 added
> 
> # Adds device
> 
> commit ab5b96818a7ad40cab085ce978040811fb857aa0
> Author: Jean-Francois Moine <moinejf at free.fr>
> 
>     V4L/DVB (12501): gspca - sonixj: Do the ov7660 sensor work again.
> 
> # Only touches the ov7660 sensor to fix sensor power, edge
> gain/threshold, auto gain, light frequency filter
> 
> commit 338d606407cbb3e80555a513347bdd2359f3c48a
> Author: Jean-Francois Moine <moinejf at free.fr>
> 
>     V4L/DVB (12691): gspca - sonixj: Don't use mdelay().
> 
> # Use msleep() instead
> 
> commit 1519b64ba60656727f91962e65ca2ae667e81f8a
> Author: Jean-Francois Moine <moinejf at free.fr>
> 
>     V4L/DVB (12696): gspca - sonixj / sn9c102: Two drivers for 0c45:60fc and 0c45:613e.
> 
> # Only claim device 0c45:60fc if !defined CONFIG_USB_SN9C102 && !defined CONFIG_USB_SN9C102_MODULE
> Only claim device 0c45:613e if !defined CONFIG_USB_GSPCA && !defined CONFIG_USB_GSPCA_MODULE
> 
> commit 2ca6ea590d90073119d1471c79cc6f0df28f4747
> Author: Chris Wilson <chris at chris-wilson.co.uk>
> 
>     drm/i915: Select CONFIG_SHMEM
> 
> # This seems more like a workaround to prevent a panic.  The patch
> claims the driver requires shmfs as the backing filesystem so ensures it
> is selected if the i915 driver is enabled.  However, reading the
> upstream bug report (bugzilla.kernel.org bug 14662) it seems i915
> actually requires TMPFS which happens to depends on SHMEM in the
> upstream kernel fs/Kconfig.  For Karmic, TMPFS doesn't show it depends
> on SHMEM.  Regardless, it appears we already enable CONFIG_TMPFS=y.  I
> don't think merging this patch gives us a fix but rather just adds more
> bloat.  Seems we already have the needed fix of CONFIG_TMPFS=y so I
> suspect we don't even see bugs with this panic (and I've searched
> Launchpad just to be sure).  The only reason I can see to take this it
> to keep us in sync with upstream.
> 
> commit a75c6447907a732d612fd399e8765f661cca77a7
> Author: Jesse Barnes <jbarnes at virtuousgeek.org>
> 
>     drm: work around EDIDs with bad htotal/vtotal values
> 
> # Tested and confirmed to fix LP Bug 460664.
> 
> commit 6a36e5969e91cd0c13ffc83e83a2b8a718efe862
> Author: Nanhai Zou <nanhai.zou at intel.com>
> 
>     drm/i915: Fix IRQ stall issue on Ironlake
> 
> # Fixes stall by disabling master irq before irq handling, and
> re-enabling master irq after irq handling.
> 
> commit f2f3a6990f1c7dd8bb8e14cc8026faaf2d9ade21
> Author: Eric Dumazet <eric.dumazet at gmail.com>
> 
>     udp: Fix udp_poll() and ioctl()
> 
> # Lock socket and drop all invalid frames (ie bad checksum frames).
> 
> commit e09fa919760c2639c6e49d2132bfdc6202cfa5e6
> Author: Ben Hutchings <ben at decadent.org.uk>
> 
>     acenic: Pass up error code from ace_load_firmware()
> 
> # Add error handling which prevents oops
> 
> commit fb97d3d0b3bab856c9419e727d543e4497da24b9
> Author: Jamal Hadi Salim <hadi at cyberus.ca>
> 
>     pkt_sched: pedit use proper struct
> 
> # the wrong stuct being used in sizeof(), use the right one
> 
> commit 5213d268ccd71d4e294650d83161cd93cfb6a21e
> Author: Eric Dumazet <eric.dumazet at gmail.com>
> 
>     net: fix sk_forward_alloc corruption
> 
> # Avoid corruption by locking socket before calling skb_free_datagram()
> 
> commit 193fe66e59fd31d6046727d6fdb9a232ad4cbd46
> Author: Jasper Spaans <spaans at fox-it.com>
> 
>     bonding: Modify hash transmit policies to use the packet's source MAC address
> 
> # Modify policy to use the packets source MAC address instead of the
> bonding device MAC address.  I'm not sure what the reasoning for this
> was, it's not explained in the commit.
> 
> commit f7f7ef6ecdc4626f6b8a6a43b2c5be0e0e0a1bc7
> Author: Ben Hutchings <bhutchings at solarflare.com>
> 
>     sfc: Set ip_summed correctly for page buffers passed to GRO
> 
> # Adds checksummed parameter to be passed to efx_rx_packet_lro().  Then
> set skb->ip_summed value based on checksummed value rather than always
> defaulting to CHECKSUM_UNNECESSARY.
> 
> commit e5ac34fd383ddae70158d220e45a25ad97b02452
> Author: Roel Kluin <roel.kluin at gmail.com>
> 
>     sparc64: replace parentheses in pmul()
> 
> # Fix parentheses to ensure operator precedence
> 
> commit 3c6f31de43c672ece8bcdd8262efd57c14c20c8f
> Author: David S. Miller <davem at davemloft.net>
> 
>     sparc: Move of_set_property_mutex acquisition outside of devtree_lock grab.
> 
> # Avoids trying to sleep when preemption is disabled by locking
> of_set_property_mutex before grabbing the devtree_lock. 
> 
> commit 8526322d0f88031465b2235c9e18a9f242dd6669
> Author: Rusty Russell <rusty at rustcorp.com.au>
> 
>     sched: Fix boot crash by zalloc()ing most of the cpu masks
> 
> # Use zalloc_cpumask_var() instead of alloc_cpumask_var() to ensure we
> zero out cpu masks.
> 
> commit b4b4c13e3bb21e3c5f193b395dbb9c1a1202f399
> Author: Rusty Russell <rusty at rustcorp.com.au>
> 
>     sched: Fix isolcpus boot option
> 
> # Fixes a regression caused by the above commit 8526322.  Use
> alloc_bootmem_cpumask_var() if isolcpus is set, otherwise zalloc.
> 
> commit 3a1e1a6cc13c5de16afe95304b7984a71c912829
> Author: Michael Krufky <mkrufky at kernellabs.com>
> 
>     V4L/DVB (13202): smsusb: add autodetection support for three additional Hauppauge USB IDs
> 
> # Adds support for 3 new devices 2040:b900, 2040:b910, 2040:c000
> 
> commit ed3876a4e0cfc5cfcbbc4b3419586a858ecf8b56
> Author: Michael Krufky <mkrufky at kernellabs.com>
> 
>     V4L/DVB (13313): saa7134: add support for FORCE_TS_VALID mode for mpeg ts input
> 
> # Some new boards require support for FORCE_TS_VALID mode to function
> properly.
> 
> commit c3f57df2bda0d830179b776166df0a100da733ac
> Author: Michael Krufky <mkrufky at kernellabs.com>
> 
>     V4L/DVB (13314): saa7134: set ts_force_val for the Hauppauge WinTV HVR-1150
> 
> # Requires the above commit ed3876a and sets ts_force_val = 1
> 
> commit be488339a245d9b235a47ad563b0719bcfd5928c
> Author: David Ford <david at blue-labs.org>
> 
>     ipv4: additional update of dev_net(dev) to struct *net in ip_fragment.c, NULL ptr OOPS
> 
> # Already included in security update.
> 
> commit 7eeea230e8e078b125436e796cf30b71098107b7
> Author: Rusty Russell <rusty at rustcorp.com.au>
> 
>     speedstep-ich: fix error caused by 394122ab144dae4b276d74644a2f11c44a60ac5c
> 
> # Need to pass in the type of processor to speedstep_get_frequency(),
> not the processor itself.
> 
> commit 3d57f55a87be678c0e3a91842a3ca2ca4ffcdcaf
> Author: Alan Stern <stern at rowland.harvard.edu>
> 
>     USB: EHCI: don't send Clear-TT-Buffer following a STALL
> 
> # According to the commit message, this prevents ehci_clear_tt_buffer()
> following a STALL handshake (stop sending isochronous packets) as some
> hubs can't handle this correctly and it doesn't appear to hurt by not
> clearing.
> 
> commit 41e0b0605826e299a981e2da61e631efe171f4af
> Author: Sergei Shtylyov <sshtylyov at ru.mvista.com>
> 
>     USB: musb_gadget: fix STALL handling
> 
> # Rather large patch than normal with multiple fixes to fixup the STALL
> handling, remove unnecessary Tx flushing and zeroing of csr variable,
> moving/removing "done" goto's.  But seems to only touch the
> msusb_gadget.  
> 
> commit 774430b67775145d69362ca807d5f25db019919e
> Author: Thomas Dahlmann <dahlmann.thomas at arcor.de>
> 
>     usb: amd5536udc: fixed shared interrupt bug and warning oops
> 
> # Just moving chunks of code around, and enabling/disabling irq's
> 
> commit 144096993162a13f165f21aa89d7d95603b25c78
> Author: Eric W. Biederman <ebiederm at xmission.com>
> 
>     USB: ftdi_sio: Keep going when write errors are encountered.
> 
> # print message but keep going and fall through to
> usb_serial_port_softint()
> 
> commit 77d12b19a0fe01d9e81baae809903ec329f84a15
> Author: Oliver Neukum <oliver at neukum.org>
> 
>     USB: work around for EHCI with quirky periodic schedules
> 
> # Add workaround for a specific chipset.  Fixes LP Bug 435352
> 
> commit d9abf6e4f67279f0ea926e6beef7baa9661dce75
> Author: Alan Cox <alan at linux.intel.com>
> 
>     tty_port: handle the nonblocking open of a dead port corner case
> 
> # Handle corner case that some drivers allow O_NDELAY of a dead port.
> 
> commit 9530e63015d1627cb46a485f7e6b8ba83ec4dca7
> Author: Daniel Mack <daniel at caiaq.de>
> 
>     pxamci: call mmc_remove_host() before freeing resources
> 
> # Prevents oops, looks sane.
> 
> commit 96433ac605f72599f99f1300f56a57316f10d532
> Author: Rusty Russell <rusty at rustcorp.com.au>
> 
>     param: don't complain about unused module parameters.
> 
> # Remove Printk.
> 
> commit a5aeface580afa4d2daba4980cd26f53ed31787a
> Author: Helge Deller <deller at gmx.de>
> 
>     modules: don't export section names of empty sections via sysfs
> 
> # Prevents warning on parisc due to multiple .text sections (which are
> empty) for each kernel module due to the use of -ffunction-sections
> compiler flag.
> 
> commit 2a959cfd1e6eff5ce71693bb6f7e753d71f5f088
> Author: NeilBrown <neilb at suse.de>
> 
>     md: revert incorrect fix for read error handling in raid1.
> 
> # This patch was unnecessary and fixed in a much better way by commit
> dd00a99e7a.  This patch also introduced a bug.  Reverting makes sense.
> 
> commit 415cc7b7fe6fd663139da295d7bd2cde556345f0
> Author: Paul Mackerras <paulus at samba.org>
> 
>     perf_event: Adjust frequency and unthrottle for non-group-leader events
> 
> # Use list_for_each_entry_rcu() instead of list_for_each_entry() and use
> ctx->event_list rather than ctx->group_list
> 
> commit bb969fdc8c7e7675266a3697ad93945428fd89d6
> Author: Antti Kaijanmäki <antti.kaijanmaki at nomovok.com>
> 
>     hso: fix soft-lockup
> 
> # kref_put() called to early so move it.
> 
> commit ee3989464a4c4896ec29f5b1c2aae840fab8b73d
> Author: Neil Brown <neilb at suse.de>
> 
>     block: use after free bug in __blkdev_get
> 
> # Move put_disk() after module_put().
> 
> commit e3675ca0caa78b55d2c3d3bf5ef0e406c835cbd1
> Author: Jean Delvare <khali at linux-fr.org>
> 
>     hwmon: (adt7475) Fix temperature fault flags
> 
> # Again use double negation to fix up logic which was resulting in
> invalid faults.
> 
> commit 38504256170312f5fe8c097e06ba291c39ee814d
> Author: Jean Delvare <khali at linux-fr.org>
> 
>     hwmon: (adt7475) Cache limits for 60 seconds
> 
> # Update code to match comments and set cache limits to 60 secs, not 2
> secs.
> 
> commit 636addb566d468fe7159e678d4c0d195dc77707d
> Author: Zhenyu Wang <zhenyuw at linux.intel.com>
> 
>     agp/intel: new host bridge support
> 
> # Add new id.
> 
> commit 325786e848323cd6c3e658886448ceafda09bf86
> Author: Jozsef Kadlecsik <kadlec at blackhole.kfki.hu>
> 
>     netfilter: nf_nat: fix NAT issue in 2.6.30.4+
> 
> # Rather large set of changes that fixes FTP over NAT.
> 
> commit 380cf591fc6682b1297415c7e90c17e578f3bf44
> Author: Jan Engelhardt <jengelh at medozas.de>
> 
>     netfilter: xt_connlimit: fix regression caused by zero family value
> 
> # Cleanup missed a few spots.  Finish cleanup by changing par->match to
> par->family and match->family to family.
> 
> commit 4a72cdf3871e086db051c70ade06c0570ac4d5b5
> Author: Michael Buesch <mb at bu3sch.de>
> 
>     b43: Fix DMA TX bounce buffer copying
> 
> # If the tx skb is in an invalid DMA memory range, make sure to copy the
> metadate to the new skb.
> 
> commit c089a8dcd8a99f8c6505a539d45a754a9f84c9dd
> Author: Chuck Ebbert <cebbert at redhat.com>
> 
>     crypto: padlock-aes - Use the correct mask when checking whether copying is required
> 
> # Mapping against PAGE_SIZE is not correct, use PAGE_MASK instead.
> 
> commit d801d0a919fd2e54ad91626a0cb7d006a9c6de84
> Author: Brandon Philips <bphilips at suse.de>
> 
>     sky2: set carrier off in probe
> 
> # Add netif_carrier_off() to probe.
> 
> commit d02b2ced79e2a22b38d6a4fdc758d070ff807d2b
> Author: Nick Kossifidis <mick at madwifi-project.org>
> 
>     ath5k: Linear PCDAC code fixes
> 
> # Set correct high/low curve indices and don't return zero min power but
> rather the proper barrier.
> 
> commit 95c2fff142fd3d44ba55cd6b3d9758c3215b52ca
> Author: Jean Delvare <khali at linux-fr.org>
> 
>     i2c: Fix userspace_device list corruption
> 
> # Remove clients before removing adapter.
> 
> commit e7ec863bb38f66855c128df211082a026a81212e
> Author: Peter Feuerer <peter at piie.net>
> 
>     acerhdf: fix fan control for AOA150 model
> 
> # Already in Karmic per LP Bug 426691.  We should just drop the SAUCE
> patch in favor of this upstream one.
> 
> commit 321cb431c5027d83257d8fbfd957f9d90b469e74
> Author: Clemens Ladisch <clemens at ladisch.de>
> 
>     drm/fb: fix FBIOGET/PUT_VSCREENINFO pixel clock handling
> 
> # pixclock is unsigned so checking if it == -1 is not correct, use 0
> instead.  This allows X.org's fbdev driver to work.
> 
> commit e689a0b49fc3ef10f50764bd7e523ae76f849a91
> Author: Michal Simek <monstr at monstr.eu>
> 
>     tty/of_serial: add missing ns16550a id
> 
> # Add new id.
> 
> commit c48bcef7d779aa1febac7050196e24052746a672
> Author: Erik Andrén <erik.andren at gmail.com>
> 
>     V4L/DVB (13255): gspca - m5602-s5k4aa: Add vflip quirk for the Bruneinit laptop
> 
> # Add hw specific quirk.
> 
> commit c7694e85ae85e7b0f2ab6784090e496734653bd3
> Author: Erik Andrén <erik.andren at gmail.com>
> 
>     V4L/DVB (13256): gspca - m5602-s5k4aa: Add another MSI GX700 vflip quirk
> 
> # Add hw specific quirk.
> 
> commit 9509e37c579356f4e0c22f00eb7f2cb783a40c05
> Author: Erik Andrén <erik.andren at gmail.com>
> 
>     V4L/DVB (13257): gspca - m5602-s5k4aa: Add vflip for Fujitsu Amilo Xi 2528
> 
> # Add hw specific quirk.
> 
> commit 0823e602ad9d28d54fa6346289ff13ae84ad2c34
> Author: Kenji Kaneshige <kaneshige.kenji at jp.fujitsu.com>
> 
>     PCI: Prevent AER driver from being loaded on non-root port PCIE devices
> 
> # Prevents dmesg spamming and unsupported PCI express errors.
> 
> commit 77540b842ef4e6cc73e6c67c59fa884c27275a0f
> Author: Peter Feuerer <peter at piie.net>
> 
>     acerhdf: additional BIOS versions
> 
> # Add new BIOS versions.
> 
> commit c0d2a80576cde5ecb6d0e0e62f7145da53a64ebd
> Author: Peter Feuerer <peter at piie.net>
> 
>     acerhdf: return temperature in milidegree instead of degree
> 
> # sysfs-api requires milidegree return value.  Looks sane.
> 
> commit 548193715cb69952c6655f8800d81a8bfe2d8bc5
> Author: Samuel Thibault <samuel.thibault at ens-lyon.org>
> 
>     Input: keyboard - fix braille keyboard keysym generation
> 
> # Add U() call translation layer.
> 
> commit b02f6a9593cbe0c3a9cf6052acf58c11980d6efd
> Author: Roel Kluin <roel.kluin at gmail.com>
> 
>     isdn: hfc_usb: Fix read buffer overflow
> 
> # Already included in security update.
> 
> commit f8ebcb2ebc49a9ce184d738ca8f9bd570ac634b1
> Author: Greg Kroah-Hartman <gregkh at suse.de>
> 
>     Linux 2.6.31.7
> 
> # Update Makefile to 2.6.31.7
> 
> 






More information about the kernel-team mailing list