Changelog
linux (3.16.5-1) unstable; urgency=medium
* New upstream stable update:
https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.16.4
- module: Clean up ro/nx after early module load failures
(regression in 3.16)
- [armhf] cpufreq: OPP: Avoid sleeping while atomic
- [armhf] drm/tilcdc: Fix various bugs in removal path
- drm/ttm: Fix possible stack overflow by recursive shrinker calls.
- [x86] drm/i915: Fix crash when failing to parse MIPI VBT
(regression in 3.16)
- [x86] drm/i915: read HEAD register back in init_ring_common() to enforce
ordering (Closes: #763583)
- libata: widen Crucial M550 blacklist matching
- pata_scc: propagate return value of scc_wait_after_reset
- pwm: Fix period and polarity in pwm_get() for non-perfect matches
- aio: add missing smp_rmb() in read_events_ring
- [arm64] flush TLS registers during exec
- [arm64] use irq_set_affinity with force=false when migrating irqs
(regression in 3.15)
- [arm*] KVM: Nuke Hyp-mode tlbs before enabling MMU
- [x86] i2c: ismt: use correct length when copy buffer
- ftrace: Use current addr when converting to nop in
__ftrace_replace_code() (regression in 3.16)
- ALSA: core: fix buffer overflow in snd_info_get_line()
- ALSA: firewire-lib/dice: add arrangements of PCM pointer and interrupts
for Dice quirk (regression in 3.16)
- HID: picolcd: sanity check report size in raw_event() callback
(CVE-2014-3186)
- HID: magicmouse: sanity check report size in raw_event() callback
(CVE-2014-3181)
- HID: logitech-dj: prevent false errors to be shown (regression in 3.16.2)
- [x86] drm/i915: Skip load detect when intel_crtc->new_enable==true
(regression in 3.16)
- [x86] drm/i915: fix plane/cursor handling when runtime suspended
(regression in 3.14)
- [x86] drm/i915: Ignore VBT backlight presence check on Acer C720 (4005U)
(regression in 3.15)
- [x86] drm/i915: Wait for vblank before enabling the TV encoder
(regression in 3.16)
- [x86] drm/i915/hdmi: fix hdmi audio state readout (regression in 3.16)
- drm/radeon: Add ability to get and change dpm state when radeon PX card
is turned off (regression in 3.15)
- locks: pass correct "before" pointer to locks_unlink_lock in
generic_add_lease
- ufs: fix deadlocks introduced by sb mutex merge (regression in 3.16)
- USB: serial: fix potential stack buffer overflow
- USB: serial: fix potential heap buffer overflow
- USB: option: reduce interrupt-urb logging verbosity (regression in 3.16)
- [armhf] usb: phy: twl4030-usb: Fix lost interrupts after ID pin goes down
(regression in 3.13)
- [armhf] usb: phy: twl4030-usb: Fix regressions to runtime PM on omaps
(regressions in 3.14, 3.15)
- uwb: init beacon cache entry before registering uwb device
- usb: hub: take hub->hdev reference when processing from eventlist
- USB: EHCI: unlink QHs even after the controller has stopped
- Revert "ACPI / battery: fix wrong value of capacity_now reported when
fully charged" (regression in 3.16)
- [x86] iommu/vt-d: Check return value of acpi_bus_get_device()
(regression in 3.15)
- [armhf/armmp-lpae] iommu/arm-smmu: fix programming of SMMU_CBn_TCR for
stage 1
- cgroup: check cgroup liveliness before unbreaking kernfs
(regression in 3.15)
- NFSv4: Fix another bug in the close/open_downgrade code
(regression in 3.16.2)
- nfsd4: fix corruption of NFSv4 read data (regression in 3.16)
- nfs: check wait_on_bit_lock err in page_group_lock
- nfs: clear_request_commit while holding i_lock
- nfs: fix nonblocking calls to nfs_page_group_lock
- nfs: use blocking page_group_lock in add_request
- nfs: fix error handling in lock_and_join_requests
- nfs: don't sleep with inode lock in lock_and_join_requests
- nfs: disallow duplicate pages in pgio page vectors
- nfs: can_coalesce_requests must enforce contiguity
- [armhf] 8129/1: errata: work around Cortex-A15 erratum 830321 using dummy
strex
- [armhf] 8133/1: use irq_set_affinity with force=false when migrating irqs
(regression in 3.15)
- [armel,armhf] 8148/1: flush TLS and thumbee register state during exec
- [armel,armhf] 8149/1: perf: Don't sleep while atomic when enabling
per-cpu interrupts (regression in 3.15)
- [armhf] imx: fix .is_enabled() of shared gate clock (regression in 3.16)
- [armhf] 8165/1: alignment: don't break misaligned NEON load/store
- [mips*] Fix MFC1 & MFHC1 emulation for 64-bit MIPS systems
(regression in 3.15)
- ACPICA: Update to GPIO region handler interface.
- gpio / ACPI: Use pin index and bit length
- ACPI / platform / LPSS: disable async suspend/resume of LPSS devices
(regression in 3.16)
- ACPI / hotplug: Generate online uevents for ACPI containers
(regression in 3.14)
- ACPI / video: disable native backlight for ThinkPad X201s
(regression in 3.16)
- regmap: Fix regcache debugfs initialization (regression in 3.15)
- regmap: Fix handling of volatile registers for format_write() chips
- regmap: Don't attempt block writes when syncing cache on single_rw
devices
- cgroup: reject cgroup names with '\n'
- cgroup: delay the clearing of cgrp->kn->priv
- cgroup: fix unbalanced locking (regression in 3.14)
- [s390*] KVM: Fix user triggerable bug in dead code
- [s390*] KVM: mm: try a cow on read only pages for key ops
- [s390*] KVM: mm: Fix storage key corruption during swapping
- [s390*] KVM: mm: Fix guest storage key corruption in
ptep_set_access_flags
- [x86] xen: don't copy bogus duplicate entries into kernel page tables
- [x86] early_ioremap: Increase FIX_BTMAPS_SLOTS to 8 (regression in 3.16)
- shmem: fix nlink for rename overwrite directory
- SMB3: Fix oops when creating symlinks on smb3
- iio: Fix indio_dev->trig assignment in several drivers
- Target/iser: Don't put isert_conn inside disconnected handler
- target: Fix inverted logic in SE_DEV_ALUA_SUPPORT_STATE_STORE
(regression in 3.13)
- iscsi-target: Fix memory corruption in iscsit_logout_post_handler_diffcid
- SCSI: libiscsi: fix potential buffer overrun in __iscsi_conn_send_pdu
- Revert "iwlwifi: dvm: don't enable CTS to self" (regression in 3.16)
- iwlwifi: mvm: fix endianity issues with Smart Fifo commands
(regression in 3.14)
- iwlwifi: mvm: set MAC_FILTER_IN_BEACON correctly for STA/P2P client
(regression in 3.16)
- workqueue: apply __WQ_ORDERED to create_singlethread_workqueue()
(regression in 3.10)
- futex: Unlock hb->lock in futex_wait_requeue_pi() error path
- block: Fix dev_t minor allocation lifetime
- dm cache: fix race causing dirty blocks to be marked as clean
- percpu: fix pcpu_alloc_pages() failure path
- percpu: perform tlb flush after pcpu_map_pages() failure
- regulatory: add NUL to alpha2
- lockd: fix rpcbind crash on lockd startup failure (regression in 3.15)
- genhd: fix leftover might_sleep() in blk_free_devt()
- eventpoll: fix uninitialized variable in epoll_ctl
- kcmp: fix standard comparison bug
- fs/notify: don't show f_handle if exportfs_encode_inode_fh failed
- nilfs2: fix data loss with mmap()
- mm, slab: initialize object alignment on cache creation
- fs/cachefiles: add missing \n to kerror conversions (regression in 3.16)
- mm: softdirty: keep bit when zapping file pte
- sched: Fix unreleased llc_shared_mask bit during CPU hotplug
- brcmfmac: handle IF event for P2P_DEVICE interface (regression in 3.12)
- ath9k_htc: fix random decryption failure (regression in 3.15)
- [powerpc,ppc*] Add smp_mb() to arch_spin_is_locked()
- [powerpc,ppc*] Add smp_mb()s to arch_spin_unlock_wait()
- [hppa] Implement new LWS CAS supporting 64 bit operations.
- alarmtimer: Return relative times in timer_gettime
- alarmtimer: Do not signal SIGEV_NONE timers
- alarmtimer: Lock k_itimer during timer callback
- GFS2: fix d_splice_alias() misuses
- IB/qib: Correct reference counting in debugfs qp_stats
- IB/mlx4: Avoid null pointer dereference in mlx4_ib_scan_netdevs()
(regression in 3.14)
- IB/mlx4: Don't duplicate the default RoCE GID (regression in 3.14)
- IB/core: When marshaling uverbs path, clear unused fields
(regression in 3.14)
- mm: Fix unbalanced mutex in dma_pool_create(). (regression in 3.16)
- PCI: Add pci_ignore_hotplug() to ignore hotplug events for a device
(regression in 3.15)
- Revert "PCI: Don't scan random busses in pci_scan_bridge()"
(regression in 3.15)
- drm/nouveau/runpm: fix module unload
- drm/radeon/px: fix module unload
- fs: Fix nasty 32-bit overflow bug in buffer i/o code.
- blk-mq: Avoid race condition with uninitialized requests
- [x86] crypto: ccp - Check for CCP before registering crypto algs
- nl80211: clear skb cb before passing to netlink
- Revert "PCI: Make sure bus number resources stay within their parents
bounds" (regression in 3.15)
- cpufreq: release policy->rwsem on error (regression in 3.14)
- cpufreq: fix cpufreq suspend/resume for intel_pstate (regression in 3.15)
- media: it913x: init tuner on attach (regression in 3.15)
- media: videobuf2-dma-sg: fix for wrong GFP mask to
sg_alloc_table_from_pages (regression in 3.13)
- media: vb2: fix vb2 state check when start_streaming fails
(regression in 3.16.3)
- media: vb2: fix plane index sanity check in vb2_plane_cookie()
- md/raid1: clean up request counts properly in close_sync()
(regression in 3.13)
- md/raid1: be more cautious where we read-balance during resync.
(regression in 3.13)
- md/raid1: make sure resync waits for conflicting writes to complete.
(regression in 3.13)
- md/raid1: Don't use next_resync to determine how far resync has
progressed (regression in 3.13)
- md/raid1: update next_resync under resync_lock. (regression in 3.13)
- md/raid1: count resync requests in nr_pending. (regression in 3.13)
- md/raid1: fix_read_error should act on all non-faulty devices.
- md/raid1: intialise start_next_window for READ case to avoid hang
(regression in 3.13)
- netfilter: xt_hashlimit: perform garbage collection from process context
- mmc: mmci: Reverse IRQ handling for the arm_variant (regression in 3.15)
- partitions: aix.c: off by one bug (regression in 3.11)
- cpufreq: update 'cpufreq_suspended' after stopping governors
- aio: block exit_aio() until all context requests are completed
- ext4: propagate errors up to ext4_find_entry()'s callers
- ext4: avoid trying to kfree an ERR_PTR pointer
https://www.kernel.org/pub/linux/kernel/v3.x/ChangeLog-3.16.5
- udf: Avoid infinite loop when processing indirect ICBs (CVE-2014-6410)
- ASoC: core: fix possible ZERO_SIZE_PTR pointer dereferencing error.
- perf: fix perf bug in fork()
- mm: memcontrol: do not iterate uninitialized memcgs (regression in 3.14)
- mm: migrate: Close race between migration completion and mprotect
- [x86] ACPI / i915: Update the condition to ignore firmware backlight
change request (regression in 3.16)
- [x86] cpufreq: pcc-cpufreq: Fix wait_event() under spinlock
(regression in 3.15)
- md/raid5: disable 'DISCARD' by default due to safety concerns.
- [x86] drm/i915: Flush the PTEs after updating them before suspend
(regression in 3.12)
- cifs: Fix problem recognizing symlinks (regression in 3.13)
- ring-buffer: Fix infinite spin in reading buffer (regression in 3.16.3)
- mm: numa: Do not mark PTEs pte_numa when splitting huge pages
- media: vb2: fix VBI/poll regression
[ Ian Campbell ]
* [armhf] Add Exynos5 disk/usb/nic modules to udebs.
* [armhf] Backport BananaPi device tree files. Patch from Karsten
Merker (Closes: #763897).
[ Ben Hutchings ]
* [hppa/parisc64-smp] Work around gcc 4.8 miscompilation (Closes: #762390)
* [powerpc/powerpc64,ppc64*] video/fb: Change FB_MATROX, FB_RADEON, FB_ATY,
FB_SIS, FB_3DFX, FB_VOODOO1 back to modules (Closes: #748398)
* udeb: Add pata_rdc to pata-modules (Closes: #633128)
* [s390*] 3215: fix tty output containing tabs (Closes: #758264)
* radeon: Don't check for installed firmware if driver is built-in
(Closes: #763305)
* Bump ABI to 3
* vfs: fold swapping ->d_name.hash into switch_names()
* vfs: Don't exchange "short" filenames unconditionally. (Closes: #763700)
* [hppa,m68k,mips/r4k-ip22,sparc*] bluetooth: Enable BT as module
(Closes: #764524)
[ Aurelien Jarno ]
* [arm64] Change RTC_DRV_PL031 and RTC_DRV_XGENE from modules to built-ins
as the kernel isn't able to initialize the system clock from a hardware
clock whose driver is a module, and as there is no initramfs mechanism
to do that.
* [armhf] Change RTC_DRV_DA9052, RTC_DRV_IMXDI, RTC_DRV_MC13XXX,
RTC_DRV_MV, RTC_DRV_MXC, RTC_DRV_OMAP, RTC_DRV_PL030, RTC_DRV_PL031,
RTC_DRV_S5M, RTC_DRV_SUNXI, RTC_DRV_VT8500 from modules to built-ins for
the same reason as above.
-- Ben Hutchings <email address hidden> Fri, 10 Oct 2014 09:15:17 +0100