Changelog
linux (4.1.5-1) unstable; urgency=medium
* New upstream stable update:
https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.1.4
- Bluetooth: btbcm: allow btbcm_read_verbose_config to fail on Apple
(regression in 4.1)
- ath9k_htc: memory corruption calling set_bit()
- ieee802154: Fix sockaddr_ieee802154 implicit padding information leak.
- staging: vt6656: check ieee80211_bss_conf bssid not NULL
- staging: vt6655: check ieee80211_bss_conf bssid not NULL
- staging: vt6655: device_rx_srv check sk_buff is NULL
- staging: rtl8712: prevent buffer overrun in recvbuf2recvframe
- hid-sensor: Fix suspend/resume delay (regression in 4.0)
- ext4: fix race between truncate and __ext4_journalled_writepage()
- ext4: call sync_blockdev() before invalidate_bdev() in put_super()
- ext4: don't retry file block mapping on bigalloc fs with non-extent file
- ext4: fix fencepost error in lazytime optimization
- ext4: avoid deadlocks in the writeback path by using sb_getblk_gfp
- ext4: fix reservation release on invalidatepage for delalloc fs
- ext4: be more strict when migrating to non-extent based file
- ext4: correctly migrate a file with a hole at the beginning
- jbd2: use GFP_NOFS in jbd2_cleanup_journal_tail()
- jbd2: fix ocfs2 corrupt when updating journal superblock fails
- [armhf] rtc: snvs: fix wakealarm by call enable_irq_wake earlier
(regression in 3.19)
- i2c: mux: Use __i2c_transfer() instead of calling parent's master_xfer()
(regression in 4.1)
- i2c: use parent adapter quirks in mux (regression in 4.1)
- vb2: Don't WARN when v4l2_buffer.bytesused is 0 for multiplanar buffers
(regression in 4.1)
- media: Fix regression in some more dib0700 based devices
(regression in 3.17)
- rc-core: fix dib0700 scancode generation for RC5 (regression in 3.17)
- cx24117: fix a buffer overflow when checking userspace params
- af9013: Don't accept invalid bandwidth
- s5h1420: fix a buffer overflow when checking userspace params
- cx24116: fix a buffer overflow when checking userspace params
- libata: Fall back to unqueued READ LOG EXT if the DMA variant fails
(regression in 4.1)
- libata: Expose TRIM capability in sysfs
- libata: add ATA_HORKAGE_NOTRIM
- libata: add ATA_HORKAGE_MAX_SEC_1024 to revert back to previous
max_sectors limit (regression in 3.19)
- libata: force disable trim for SuperSSpeed S238
- [armhf] usb: dwc3: gadget: don't clear EP_BUSY too early
(regression in 3.15)
- USB: devio: fix a condition in async_completed()
- [armhf] phy: twl4030-usb: remove incorrect pm_runtime_get_sync() in
probe function. (regression in 3.17)
- usb: gadget: composite: Fix NULL pointer dereference (regression in 4.1)
- usb: gadget: f_fs: do not set cancel function on synchronous {read,write}
(regression in 4.0)
- USB: OHCI: Fix race between ED unlink and URB submission
(regression in 3.17)
- usb: core: lpm: set lpm_capable for root hub device (regression in 3.15)
- usb: xhci: Bugfix for NULL pointer deference in xhci_endpoint_init()
function
- dm cache: fix race when issuing a POLICY_REPLACE operation
- dm btree remove: fix bug in redistribute3
- block: loop: convert to per-device workqueue (regression in 4.0)
- block: loop: avoiding too many pending per work I/O (regression in 4.0)
- block: Do a full clone when splitting discard bios (Closes: #793326)
(regression in 3.14)
- [armhf] drm/tegra: dpaux: Fix transfers larger than 4 bytes
(regression in 3.15)
- drm/qxl: Do not cause spice-server to clean our objects
- drm/atomic: fix out of bounds read in for_each_*_in_state helpers
(regression in 4.1)
- drm/radeon: take the mode_config mutex when dealing with hpds (v2)
- [x86] drm/i915/ppgtt: Break loop in gen8_ppgtt_clear_range failure path
(regression in 4.1)
- [x86] drm/i915: fix backlight after resume on 855gm (regression in 3.19)
- [x86] drm/i915: Snapshot seqno of most recently submitted request.
(regression in 4.0)
- Revert "drm/radeon: dont switch vt on suspend" (regression in 4.1)
- drm/radeon: unpin cursor BOs on suspend and pin them again on resume (v2)
(regression in 4.1)
- drm/radeon: fix user ptr race condition
- drm/radeon/ci: silence a harmless PCC warning
- drm: add a check for x/y in drm_mode_setcrtc
- drm: Stop resetting connector state to unknown (regression in 4.1)
- libata: Fix regression when the NCQ Send and Receive log page is absent
- xfs: fix remote symlinks on V5/CRC filesystems
- xfs: don't truncate attribute extents if no extents exist
(regression in 4.1)
- tpm: Fix initialization of the cdev (regression in 4.0)
- tpm, tpm_crb: fail when TPM2 ACPI table contents look corrupted
(regression in 4.1)
- [armel,armhf] dmaengine: mv_xor: bug fix for racing condition in
descriptors cleanup
- md: clear mddev->private when it has been freed. (regression in 4.0)
- md: unlock mddev_lock on an error path. (regression in 4.0)
- md: Skip cluster setup for dm-raid (regression in 4.1)
- Btrfs: don't invalidate root dentry when subvolume deletion fails
(regression in 3.18)
- Btrfs: use kmem_cache_free when freeing entry in inode cache
- Btrfs: fix race between caching kthread and returning inode to inode
cache (regression in 3.15)
- Btrfs: fix fsync data loss after append write
- Btrfs: fix list transaction->pending_ordered corruption
(regression in 3.19)
- Btrfs: fix file corruption after cloning inline extents
- selinux: fix mprotect PROT_EXEC regression caused by mm change
(regression in 4.1)
- ceph/crush: fix a bug in tree bucket decode
- ACPI / PNP: Reserve ACPI resources at the fs_initcall_sync stage
(regression in 4.1)
- ACPI / LPSS: Fix up acpi_lpss_create_device() (regression in 4.0)
- ACPICA: Tables: Enable both 32-bit and 64-bit FACS (regression in 3.14)
- ACPICA: Tables: Enable default 64-bit FADT addresses favor
- ACPI / PCI: Fix regressions caused by resource_size_t overflow with
32-bit kernel (regression in 4.0)
- mmc: sdhci: Restore behavior while creating OCR mask (regression in 3.17)
- PM / clk: Fix clock error check in __pm_clk_add() (regression in 3.19)
- RDMA/ocrdma: fix double free on pd (regression in 4.0)
- mm/hugetlb: introduce minimum hugepage order (regression in 3.12)
- PM / sleep: Increase default DPM watchdog timeout to 60
(regression in 3.13)
- firmware: dmi_scan: Only honor end-of-table for 64-bit tables
(regression in 3.19)
- mac80211: fix the beacon csa counter for mesh and ibss
(regression in 3.17)
- iwlwifi: mvm: fix ROC reference accounting (regression in 3.17)
- cfg80211: ignore netif running state when changing iftype
- e1000e: Cleanup handling of VLAN_HLEN as a part of max frame size
(regression in 3.15)
- ovl: lookup whiteouts outside iterate_dir() (regression in 4.0)
- pNFS: Fix a memory leak when attempted pnfs fails
- pNFS/flexfiles: Fix the reset of struct pgio_header when resending
- nfs: fixing infinite OPEN loop in 4.0 stateid recovery
- NFS: Ensure we set NFS_CONTEXT_RESEND_WRITES when requeuing writes
- nfs: fix potential credential leak in ff_layout_update_mirror_cred
- nfs: always update creds in mirror, even when we have an already
connected ds
- 9p: forgetting to cancel request on interrupted zero-copy RPC
- 9p: don't leave a half-initialized inode sitting around
- p9_client_write(): avoid double p9_free_req()
- [arm64] bpf: fix out-of-bounds read in bpf2a64_offset()
- [arm64] bpf: fix endianness conversion bugs
- [arm64] Don't report clear pmds and puds as huge
- mm: avoid setting up anonymous pages into file mapping
https://www.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.1.5
- [powerpc*] powernv: Fix race in updating core_idle_state
(regression in 3.19)
- Revert "Input: synaptics - allocate 3 slots to keep stability in image
sensors" (regression in 4.1)
- [hppa] Fix some PTE/TLB race conditions and optimize __flush_tlb_range
based on timing results
- [hppa] mm: Fix a memory leak related to pmd not attached to the pgd
(regression in 4.0)
- [armel,armhf] 8404/1: dma-mapping: fix off-by-one error in bitmap size
check (regression in 3.15)
- [mips*] c-r4k: Fix cache flushing for MT cores
- can: replace timestamp as unique skb attribute (regression in 4.1)
- vfs: freeing unlinked file indefinitely delayed
- [armhf] crypto: omap-des - Fix unmapping of dma channels
- [s390x] nmi: fix vector register corruption
- [s390x] bpf: clear correct BPF accumulator register
- genirq: Prevent resend to interrupts marked IRQ_NESTED_THREAD
- [x86] perf: Fix static_key bug in load_mm_cr4() (regression in 4.0)
- Revert "dm: only run the queue on completion if congested or no requests
pending" (regression in 4.1)
- scsi: fix host max depth checking for the 'queue_depth' sysfs interface
(regression in 3.19)
- scsi: fix memory leak with scsi-mq (regression in 3.17)
- mnt: Clarify and correct the disconnect logic in umount_tree
(regressions in 4.0.2, 4.1)
- mnt: In detach_mounts detach the appropriate unmounted mount
(regression in 4.1)
- ftrace: Fix breakage of set_ftrace_pid (regression in 3.16)
- [x86] iommu/vt-d: Fix VM domain ID leak (regression in 3.17)
- mmc: sdhci check parameters before call dma_free_coherent
(regression in 3.16)
- iwlwifi: mvm: fix antenna selection when BT is active
(regression in 3.17)
- md/raid1: fix test for 'was read error from last working device'.
- [armhf] spi: imx: Fix small DMA transfers (regression in 3.18)
- [armhf] regulator: s2mps11: Fix GPIO suspend enable shift wrapping bug
- usb-storage: ignore ZTE MF 823 card reader in mode 0x1225
- [armhf] Revert "serial: imx: initialized DMA w/o HW flow enabled"
(regression in 4.0)
- serial: core: Fix crashes while echoing when closing (regression in 3.19)
- efi: Handle memory error structures produced based on old versions of
standard (regression in 3.13)
- [x86] efi: Use all 64 bit of efi_memmap in setup_e820()
- rds: rds_ib_device.refcount overflow
- n_tty: signal and flush atomically (regression in 4.0)
- [x86] perf/intel/cqm: Return cached counter value from IRQ context
(regression in 4.1)f
- NFS: Don't revalidate the mapping if both size and change attr are up to
date (regression in 3.16)
- NFSv4: We must set NFS_OPEN_STATE flag in nfs_resync_open_stateid_locked
(regression in 4.0)
- NFS: Fix a memory leak in nfs_do_recoalesce (regression in 4.0)
- iscsi-target: Fix use-after-free during TPG session shutdown
- iscsi-target: Fix iscsit_start_kthreads failure OOPs (regression in 4.0)
- iscsi-target: Fix iser explicit logout TX kthread leak
(regression in 4.0)
- qla2xxx: Fix hardware lock/unlock issue causing kernel panic.
(regression in 3.18)
- qla2xxx: release request queue reservation. (regression in 3.18)
- qla2xxx: Remove msleep in qlt_send_term_exchange (regression in 3.18)
- qla2xxx: fix command initialization in target mode. (regression in 3.18)
- qla2xxx: kill sessions/log out initiator on RSCN and port down efvents
(regression in 3.18)
- drm/nouveau/fbcon/nv11-: correctly account for ring space usage
- drm/nouveau/kms/nv50-: guard against enabling cursor on disabled heads
- drm/nouveau: hold mutex when calling nouveau_abi16_fini()
- drm/nouveau/drm/nv04-nv40/instmem: protect access to priv->heap by mutex
- xfs: remote attribute headers contain an invalid LSN
- xfs: remote attributes need to be considered data
[ Ian Campbell ]
* [armhf] Enable cpufreq on some sunxi platforms (Closes: #793185)
- Enable autoloading of the cpufreq-dt driver.
- Enable autoloading of AXP20x regulator driver.
[ Ben Hutchings ]
* [x86] iio: Enable drivers for ACPI-discoverable devices as modules:
AK8975, BMC150_ACCEL, BMG160, BMP280, INV_MPU6050_IIO, JSA1212, KMX61,
KXCJK1013, MMA9551, MMA9553, SX9500
* Adjust for migration to git:
- Update .gitignore files
- debian/control: Update Vcs-* fields
- README.Debian, README.source: Update references to svn
* Bump ABI to 2
* virtio-net: drop NETIF_F_FRAGLIST (CVE-2015-5156)
* ovl: Add support for NFS as lower filesystem (Closes: #786925)
* perf: Fix double-free of the AUX buffer
* linux-source: Fix compression of the source tarball (Closes: #795199)
* udeb: Add newly enabled drivers:
- [i386] Add gpio_keys_polled, leds-gpio to input-modules
- Add virtio-input to input-modules
- Add toshsd to mmc-modules
- Add wd719x to scsi-extra-modules
- Move eeprom_93cx6 to core-modules; make scsi-extra-modules depend on it
-- Ben Hutchings <email address hidden> Sat, 15 Aug 2015 23:37:20 +0200
