Inotify instance is destroyed when all references to it are dropped.
That not only means that the corresponding file descriptor needs to be
closed but also that all corresponding instance marks are freed (as each
mark holds a reference to the inotify instance).  However marks are
freed only after SRCU period ends which can take some time and thus if
user rapidly creates and frees inotify instances, number of existing
inotify instances can exceed max_user_instances limit although from user
point of view there is always at most one existing instance.  Thus
inotify_init() returns EMFILE error which is hard to justify from user
point of view.  This problem is exposed by LTP inotify06 testcase on
some machines.

We fix the problem by making sure all group marks are properly freed
while destroying inotify instance.  We wait for SRCU period to end in
that path anyway since we have to make sure there is no event being
added to the instance while we are tearing down the instance.  So it
takes only some plumbing to allow for marks to be destroyed in that path
as well and not from a dedicated work item.

[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: NJan Kara <jack@suse.cz>
Reported-by: NXiaoguang Wang <wangxg.fnst@cn.fujitsu.com>
Tested-by: NXiaoguang Wang <wangxg.fnst@cn.fujitsu.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Pull tracing updates from Steven Rostedt:
 "This includes two new updates for the ftrace infrastructure.

   - With the changing of the code for filtering events by pid, from a
     list of pids to a bitmask, we can now easily implement following
     forks.  With a new tracing option "event-fork" which, when set,
     will have tasks with pids in set_event_pid, when they fork, to have
     their child pids added to set_event_pid and the child will be
     traced as well.

     Note, if "event-fork" is set and a task with its pid in
     set_event_pid exits, its pid will be removed from set_event_pid

   - The addition of Tom Zanussi's hist triggers.  This includes a very
     thorough documentatino on how to use the hist triggers with events.
     This introduces a quick and easy way to get histogram data from
     events and their fields.

  Some other cleanups and updates were added as well.  Like Masami
  Hiramatsu added test cases for the event trigger and hist triggers.
  Also I added a speed up of filtering by using a temp buffer when
  filters are set"

* tag 'trace-v4.7' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace: (45 commits)
  tracing: Use temp buffer when filtering events
  tracing: Remove TRACE_EVENT_FL_USE_CALL_FILTER logic
  tracing: Remove unused function trace_current_buffer_lock_reserve()
  tracing: Remove one use of trace_current_buffer_lock_reserve()
  tracing: Have trace_buffer_unlock_commit() call the _regs version with NULL
  tracing: Remove unused function trace_current_buffer_discard_commit()
  tracing: Move trace_buffer_unlock_commit{_regs}() to local header
  tracing: Fold filter_check_discard() into its only user
  tracing: Make filter_check_discard() local
  tracing: Move event_trigger_unlock_commit{_regs}() to local header
  tracing: Don't use the address of the buffer array name in copy_from_user
  tracing: Handle tracing_map_alloc_elts() error path correctly
  tracing: Add check for NULL event field when creating hist field
  tracing: checking for NULL instead of IS_ERR()
  tracing: Do not inherit event-fork option for instances
  tracing: Fix unsigned comparison to zero in hist trigger code
  kselftests/ftrace: Add a test for log2 modifier of hist trigger
  tracing: Add hist trigger 'log2' modifier
  kselftests/ftrace: Add hist trigger testcases
  kselftests/ftrace : Add event trigger testcases
  ...

Pull audit updates from Paul Moore:
 "Four small audit patches for 4.7.

  Two are simple cleanups around the audit thread management code, one
  adds a tty field to AUDIT_LOGIN events, and the final patch makes
  tty_name() usable regardless of CONFIG_TTY.

  Nothing controversial, and it all passes our regression test"

* 'stable-4.7' of git://git.infradead.org/users/pcmoore/audit:
  tty: provide tty_name() even without CONFIG_TTY
  audit: add tty field to LOGIN event
  audit: we don't need to __set_current_state(TASK_RUNNING)
  audit: cleanup prune_tree_thread

Pull remoteproc updates from Bjorn Andersson:
 "Introduce a synchronization point between the async firmware loading
  and clients requesting the remote processor to boot, as well as
  support for remote processors that are not interested in the resource
  table information"

* tag 'rproc-v4.7' of git://github.com/andersson/remoteproc:
  remoteproc: Add additional crash reasons
  remoteproc: core: Make the loaded resource table optional
  remoteproc: core: Task sync during rproc_fw_boot()

Pull rpmsg updates from Bjorn Andersson:
 "Refactor rpmsg module registration to follow other subsystems; by
  introduction of module_rpmsg_driver and hiding of THIS_MODULE from
  clients"

* tag 'rpmsg-v4.7' of git://github.com/andersson/remoteproc:
  rpmsg: use module_rpmsg_driver in existing drivers and examples
  rpmsg: add helper macro module_rpmsg_driver
  rpmsg: drop owner assignment from rpmsg_drivers
  rpmsg: add THIS_MODULE to rpmsg_driver in rpmsg core

Pull input updates from Dmitry Torokhov:
 "First round of updates for the input subsystem.  No new drivers here,
  just some driver fixes"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input:
  Input: rotary-encoder - fix bare use of 'unsigned'
  Input: cm109 - spin_lock in complete() cleanup
  Input: cm109 - fix handling of volume and mute buttons
  Input: byd - don't wipe dynamically allocated memory twice
  Input: twl4030 - fix unsafe macro definition
  Input: twl6040-vibra - remove mutex
  Input: bcm_iproc_tsc - DT spelling s/clock-name/clock-names/
  Input: bcm_iproc_tsc - use syscon to access shared registers
  Input: ti_am335x_tsc - use SIMPLE_DEV_PM_OPS
  Input: omap-keypad - remove set_col_gpio_val() and get_row_gpio_val()
  Input: omap-keypad - drop empty PM stubs
  Input: omap-keypad - remove adjusting of scan delay
  Input: gpio-keys - clean up device tree binding example
  Input: kbtab - stop saving struct usb_device
  Input: gtco - stop saving struct usb_device
  Input: aiptek - stop saving struct usb_device
  Input: acecad - stop saving struct usb_device

Pull media updates from Mauro Carvalho Chehab:
 - added support for Intersil/Techwell TW686x-based video capture cards
 - v4l PCI skeleton driver moved to samples directory
 - Documentation cleanups and improvements
 - RC: reduced the memory footprint for IR raw events
 - tpg: Export the tpg code from vivid as a module
 - adv7180: Add device tree binding documentation
 - lots of driver improvements and fixes

* tag 'media/v4.7-1' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media: (173 commits)
  [media] exynos-gsc: avoid build warning without CONFIG_OF
  [media] samples: v4l: from Documentation to samples directory
  [media] dib0700: add USB ID for another STK8096-PVR ref design based card
  [media] tvp5150: propagate I2C write error in .s_register callback
  [media] tvp5150: return I2C write operation failure to callers
  [media] em28xx: add support for Hauppauge WinTV-dualHD DVB tuner
  [media] em28xx: add missing USB IDs
  [media] update cx23885 and em28xx cardlists
  [media] media: au0828 fix au0828_v4l2_device_register() to not unlock and free
  [media] c8sectpfe: Rework firmware loading mechanism
  [media] c8sectpfe: Demote print to dev_dbg
  [media] c8sectpfe: Fix broken circular buffer wp management
  [media] media-device: Simplify compat32 logic
  [media] media: i2c: ths7303: remove redundant assignment on bt
  [media] dvb-usb: hide unused functions
  [media] xilinx-vipp: remove unnecessary of_node_put
  [media] drivers/media/media-devnode: clear private_data before put_device()
  [media] drivers/media/media-device: move debug log before _devnode_unregister()
  [media] drivers/media/rc: postpone kfree(rc_dev)
  [media] media/dvb-core: forward media_create_pad_links() return value
  ...

Pull SCSI updates from James Bottomley:
 "First round of SCSI updates for the 4.6+ merge window.

  This batch includes the usual quota of driver updates (bnx2fc, mp3sas,
  hpsa, ncr5380, lpfc, hisi_sas, snic, aacraid, megaraid_sas).  There's
  also a multiqueue update for scsi_debug, assorted bug fixes and a few
  other minor updates (refactor of scsi_sg_pools into generic code, alua
  and VPD updates, and struct timeval conversions)"

* tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi: (138 commits)
  mpt3sas: Used "synchronize_irq()"API to synchronize timed-out IO & TMs
  mpt3sas: Set maximum transfer length per IO to 4MB for VDs
  mpt3sas: Updating mpt3sas driver version to 13.100.00.00
  mpt3sas: Fix initial Reference tag field for 4K PI drives.
  mpt3sas: Handle active cable exception event
  mpt3sas: Update MPI header to 2.00.42
  Revert "lpfc: Delete unnecessary checks before the function call mempool_destroy"
  eata_pio: missing break statement
  hpsa: Fix type ZBC conditional checks
  scsi_lib: Decode T10 vendor IDs
  scsi_dh_alua: do not fail for unknown VPD identification
  scsi_debug: use locally assigned naa
  scsi_debug: uuid for lu name
  scsi_debug: vpd and mode page work
  scsi_debug: add multiple queue support
  bfa: fix bfa_fcb_itnim_alloc() error handling
  megaraid_sas: Downgrade two success messages to info
  cxlflash: Fix to resolve dead-lock during EEH recovery
  scsi_debug: rework resp_report_luns
  scsi_debug: use pdt constants
  ...

Pull iscsi_ibft updates from Konrad Rzeszutek Wilk:
 "The pull has two features - both of them expand the SysFS entries:

   - 'prefix-len' - which is subnet_mask_prefix of the iBFT header.

   - 'acpi_header' dir with: 'iBFT', OEM-ID (whatever it extracts from
     the iBFT header) and OEM_TABLE_ID (also whatever it extracts from
     the iBFT header).  This is to help NIC drivers to figure out during
     bootup how to deal with BIOS created iBFT tables (like by TianoCore
     UEFI implemenation)"

* 'stable/for-linus-4.7' of git://git.kernel.org/pub/scm/linux/kernel/git/konrad/ibft:
  ibft: Expose iBFT acpi header via sysfs
  iscsi_ibft: Add prefix-len attr and display netmask

Pull ARM SoC driver updates from Arnd Bergmann:
 "Driver updates for ARM SoCs, these contain various things that touch
  the drivers/ directory but got merged through arm-soc for practical
  reasons.

  For the most part, this is now related to power management
  controllers, which have not yet been abstracted into a separate
  subsystem, and typically require some code in drivers/soc or arch/arm
  to control the power domains.

  Another large chunk here is a rework of the NVIDIA Tegra USB3.0
  support, which was surprisingly tricky and took a long time to get
  done.

  Finally, reset controller handling as always gets merged through here
  as well"

* tag 'armsoc-drivers' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (97 commits)
  arm-ccn: Enable building as module
  soc/tegra: pmc: Add generic PM domain support
  usb: xhci: tegra: Add Tegra210 support
  usb: xhci: Add NVIDIA Tegra XUSB controller driver
  dt-bindings: usb: xhci-tegra: Add Tegra210 XUSB controller support
  dt-bindings: usb: Add NVIDIA Tegra XUSB controller binding
  PCI: tegra: Support per-lane PHYs
  dt-bindings: pci: tegra: Update for per-lane PHYs
  phy: tegra: Add Tegra210 support
  phy: Add Tegra XUSB pad controller support
  dt-bindings: phy: tegra-xusb-padctl: Add Tegra210 support
  dt-bindings: phy: Add NVIDIA Tegra XUSB pad controller binding
  phy: core: Allow children node to be overridden
  clk: tegra: Add interface to enable hardware control of SATA/XUSB PLLs
  drivers: firmware: psci: make two helper functions inline
  soc: renesas: rcar-sysc: Add support for R-Car H3 power areas
  soc: renesas: rcar-sysc: Add support for R-Car E2 power areas
  soc: renesas: rcar-sysc: Add support for R-Car M2-N power areas
  soc: renesas: rcar-sysc: Add support for R-Car M2-W power areas
  soc: renesas: rcar-sysc: Add support for R-Car H2 power areas
  ...

Pull ARM SoC defconfig updates from Arnd Bergmann:
 "As usual, a bunch of commits, mostly adding drivers and other options
  to defconfigs.

  We are adding three new defconfig files for the newly added 32-bit
  machines (aspeed and mps2), the rest is mainly housekeeping.

  The changes outside of arch/arm/config/ are for a Kconfig symbol that
  got renamed"

* tag 'armsoc-defconfig' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (63 commits)
  ARM: aspeed: adapt defconfigs for new CONFIG_PRINTK_TIME
  ARM: u8500_defconfig: update sensor config
  ARM: u8500_defconfig: remove staging from defconfig
  ARM: multi_v7_defconfig: Remove unused Kconfig option MACH_UX500_DT
  ARM: at91/defconfig: sama5: add CONFIG_FHANDLE
  arm/configs: Add Aspeed defconfig
  arm/configs/multi_v5: Add Aspeed ast2400
  ARM: at91: sama5: Update defconfig
  ARM: imx_v6_v7_defconfig: add CONFIG_MICREL_PHY
  ARM: imx_v6_v7_defconfig: add CONFIG_I2C_GPIO
  ARM: multi_v7: Enable Tegra XUSB controller in defconfig
  ARM: tegra: Enable XUSB controller in defconfig
  ARM: omap2plus_defconfig: Enable PWM and ir-rx51 as loadable modules
  ARM: multi_v7_defconfig: add the Atmel sama5d2-compatible ADC driver
  ARM: multi_v7_defconfig: add the Atmel Audio microphone interface PDMIC
  ARM: multi_v7_defconfig: add Atmel ISI (Image Sensor Interface) driver
  ARM: multi_v7_defconfig: add Atmel watchdog timers
  ARM: multi_v7_defconfig: add HLCDC drivers as modules
  ARM: at91/defconfig: add PDMIC driver to sama5_defconfig
  ARM: at91/defconfig: add HLCDC driver to sama5_defconfig
  ...

Pull ARM 64-bit DT updates from Arnd Bergmann:
 "We continue ramping up platform support for 64-bit ARM machines, with
  111 individual non-merge changesets touching 21 platforms.

  The LG1312 platform is completely new and is the first ARM platform by
  LG that we support in the mainline kernel.  Two other SoCs got added
  that are updated versions of existing SoC families, so the port mainly
  consists of new dts files:

   - The Hisilicon Hip06/D03 is the latest server platform from
     Huawei/Hisilicon, and follows the Hip05/D02 platform.

   - Rockchip RK3399 follows the 32-bit RK3288 that is popular in
     low-end Chromebooks and the 64-bit RK3368 that is mainly found in
     chinese Android TV boxes.

  The 96Boards HiKey based on the Hisilicon Hi6220 (Kirin 620) gets a
  long-awaited overhaul with a lot of devices enabled in the DT, so it
  should be much more usable with a mainline kernel now.  See also

     https://plus.google.com/111524780435806926688/posts/PeGb2VsNhJd

  A lot of work went into enabling new device drivers on existing
  machines, but we also have a couple of new commercially available
  machines:

   - Google Pixel C laptop based on Tegra210
   - Hardkernel Odroid C2 Based on Amlogic Meson GXBB (S905)
   - Geekbuying GeekBox based on Rockchip RK3368

  And finally, a couple of reference or development platforms that are
  not end-user platforms but are used for trying out the respective SoC
  platforms:

   - Amlogic Meson GXBB P200 and P201 development systems
   - NXP Layerscape 1043A QDS development board
   - Hisilicon Hip06 D03 server board, as mentioned above
   - LG1312 Reference Design
   - RK3399 Evaluation Board"

* tag 'armsoc-dt64' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (104 commits)
  arm64: dts: marvell: add XOR node for Armada 3700 SoC
  dt-bindings: document rockchip rk3399-evb board
  arm64: dts: rockchip: add dts file for RK3399 evaluation board
  arm64: dts: rockchip: add core dtsi file for RK3399 SoCs
  dt-bindings: rockchip-dw-mshc: add description for rk3399
  arm64: dts: marvell: Use a SoC-specific compatible for xHCI on Armada37xx
  arm64: dts: marvell: Rename armada-37xx USB node
  arm64: dts: marvell: Clean up armada-3720-db
  Documentation: arm64: Add Hisilicon Hip06 D03 dts binding
  arm64: dts: Add initial dts for Hisilicon Hip06 D03 board
  arm64: dts: hip05: Add nor flash support
  arm64: dts: hip05: fix its node without msi-cells
  arm64: dts: r8a7795: Don't disable referenced optional clocks
  arm64: dts: salvator-x: populate EXTALR
  arm64: dts: r8a7795: enable PCIe on Salvator-X
  arm64: dts: r8a7795: Add PCIe nodes
  arm64: tegra: Add IOMMU node to GM20B on Tegra210
  arm64: tegra: Add reference clock to GM20B on Tegra210
  dt-bindings: Add documentation for GM20B GPU
  dt-bindings: gk20a: Document iommus property
  ...

Pull ARM DT updates from Arnd Bergmann:
 "These are all the updates to device tree files for 32-bit platforms,
  which as usual makes up the bulk of the ARM SoC changes: 462 non-merge
  changesets, 450 files changed, 23340 insertions, 5216 deletions.

  The three platforms that are added with the "soc" branch are here as
  well, and we add some related machine files:

   - For Aspeed AST2400/AST2500, we get the evaluation platform and the
     Tyan Palmetto POWER8 mainboard that uses the AST2400 BMC
   - For Oxnas 810SE, the Western Digital "My Book World Edition" is
     added as the only platform at the moment.
   - For ARM MPS2, the AN385 (Cortex-M3) and AN399 (Cortex-M7) are
     supported

  On the ARM Realview development platform, we now support all machines
  with device tree, previously only the board files were supported,
  which in turn will likely be removed soon.

  Qualcomm IPQ4019 is the second generation ARM based "Internet
  Processor", following the IPQ806x that is used in many high-end WiFi
  routers.  This one integrates two ath10k wifi radios that were
  previously on separate chips.

  Other boards that got added for existing chips are:

  Ti OMAP family:
     - Amazon Kindle Fire, first generation, tablet and ebook reader
     - OnRISC Baltos iR 2110 and 3220 embedded industrial PCs
     - TI AM5728 IDK, TI AM3359 ICE-V2, and TI DRA722 Rev C EVM
       development systems

  Samsung EXYNOS platform:
     - Samsung ARTIK5 evaluation board, see

        https://www.artik.io/modules/overview/artik-5/

  NXP i.MX platforms:
     - Ka-Ro electronics TX6S-8034, TX6S-8035, TX6U-8033, TX6U-81xx,
       TX6Q-1036, TX6Q-1110/-1130, TXUL-0010 and TXUL-0011 industrial
       SoM modules
     - Embest MarS Board i.MX6Dual DIY platform
     - Boundary Devices i.MX6 Quad Plus Nitrogen6_MAX and SoloX
       Nitrogen6sx embedded boards
     - Technexion Pico i.MX6UL compute module
     - ZII VF610 Development Board

  Marvell embedded (mvebu, orion, kirkwood) platforms:
     - Linksys Viper (E4200v2 / EA4500) WiFi router
     - Buffalo Kurobox Pro NAS

  Qualcomm Snapdragon:
     - Arrow DragonBoard 600c (96boards) with APQ8064 Snapdragon 600

  Rockchips platform:
     - mqmaker MiQi single-board computer

  Altera SoCFPGA:
     - samtec VIN|ING 1000 vehicle communication interface

  Allwinner Sunxi platforms:
     - Dserve DSRV9703C tablet
     - Difrnce DIT4350 tablet
     - Colorfly E708 Q1 tablet
     - Polaroid MID2809PXE04 tablet
     - Olimex A20 OLinuXino LIME2 single board computer
     - Xunlong Orange Pi 2, Orange Pi One, and Orange Pi PC single board
       computers

  Across many platforms, bug fixes went in to address warnings that dtc
  now emits with 'make dtbs W=1'.  Further changes for device enablement
  went into Ti OMAP, bcm283x (Raspberry Pi), bcm47xx (wifi router), Ti
  Davinci, Samsung EXYNOS, Marvell mvebu/kirkwood/orion, NXP i.MX/Vybrid
  NXP LPC18xx, NXP LPC32xx, Renesas shmobile/r-mobile/r-car, Rockchips
  rk3xxx, ST Ux500, ST STi, Atmel AT91/SAMA5, Altera SoCFPGA, Allwinner
  Sunxi, Sigma Designs Tango, NVIDIA Tegra, Socionext Uniphier and ARM
  Versatile Express"

* tag 'armsoc-dt' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (458 commits)
  ARM: dts: tango4: Import watchdog node
  ARM: dts: tango4: Update cpus node for cpufreq
  ARM: dts: tango4: Update DT to match clk driver
  ARM: dts: tango4: Initial thermal support
  arm/dst: Add Aspeed ast2500 device tree
  arm/dts: Add Aspeed ast2400 device tree
  ARM: sun7i: dt: Add pll3 and pll7 clocks
  ARM: dts: sunxi: Add a olinuxino-lime2-emmc
  ARM: dts: at91: sama5d4: add trng node
  ARM: dts: at91: sama5d3: add trng node
  ARM: dts: at91: sama5d2: add trng node
  ARM: dts: at91: at91sam9g45 family: reduce the trng register map size
  ARM: sun4i: dt: Add pll3 and pll7 clocks
  ARM: sun5i: chip: Enable the TV Encoder
  ARM: sun5i: r8: Add display blocks to the DTSI
  ARM: sun5i: a13: Add display and TCON clocks
  ARM: dts: ux500: configure the accelerometers open drain
  ARM: mx5: dts: Enable USB OTG on M53EVK
  ARM: dts: imx6ul-14x14-evk: Add audio support
  ARM: dts: imx6qdl: Remove unneeded unit-addresses
  ...

Pull ARM SoC 64-bit changes from Arnd Bergmann:
 "One new platform gets added this time: The Cortex-A53 based LG
  Electronics LG1K platform used in digital TVs.

  The other changes are mostly smaller updates to the defconfig files,
  to enable additional platform specific drivers, as they get merged
  through the subsystem trees"

* tag 'armsoc-arm64' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc:
  arm64: configs: add options useful for Armada 7K/8K support
  arm64: defconfig: Add Juno SATA controller
  arm64: defconfig: enable freescale/nxp config options
  arm64: defconfig: enable 48-bit virtual addresses
  arm64: defconfig: cleanup the defconfig
  MAINTAINERS: update entry for Marvell ARM platform maintainers
  arm64: marvell: enable AP806 and CP110 syscon driver
  arm64: Kconfig: select sp804 timer for ARCH_HISI
  arm64: defconfig: enable configs for WLAN and TI WL1835 as modules
  arm64: defconfig: enable several common USB network adapters
  arm64: defconfig: add CONFIG_SPI_SPIDEV as module
  arm64: defconfig: Enable the PMIC and regulator for Hi6220 and 96boards HiKey
  arm64: defconfig: Add Renesas R-Car USB 3.0 driver support
  MAINTAINERS: add Chanho Min as ARM/LG1K maintainer
  arm64: defconfig: enable ARCH_LG1K
  arm64: add Kconfig entry for LG1K SoC family
  arm64: defconfig: Enable PL330 DMA controller
  arm64: defconfig: enable basic boot for Amlogic meson

Pull ARM SoC platform updates from Arnd Bergmann:
 "We get support for three new 32-bit SoC platforms this time.

  The amount of changes in arch/arm for any of them is miniscule, as all
  the interesting code is in device driver subsystems (irqchip, clk,
  pinctrl, ...) these days.  I'm listing them here, as the addition of
  the Kconfig statement is the main relevant milestone for a new
  platform.  In each case, some drivers are are shared with existing
  platforms, while other drivers are added for v4.7 as well, or come in
  a later release.

   - The Aspeed platform is probably the most interesting one, this is
     what most whitebox servers use as their baseboard management
     controller.  We get support for the very common ast2400 and ast2500
     SoCs.  The OpenBMC project focuses on this chip, and the LWN
     article about their ELC 2016 presentation at

        https://lwn.net/Articles/683320/

     triggered the submission, but the code comes from IBM's OpenPOWER
     team rather than the team at Facebook.  There are still a lot more
     drivers that need to get added over time, and I hope both teams can
     work together on that.

   - OXNAS is an old platform for Network Attached Storage devices from
     Oxford Semiconductor.  There are models with ARM10 (!) and
     ARM11MPCore cores, but for now, we only support the original ARM9
     based versions.  The product lineup was subsequently part of PLX,
     Avago and now the new Broadcom Ltd.

        https://wiki.openwrt.org/doc/hardware/soc/soc.oxnas

     has some more information.

   - V2M-MPS2 is a prototyping platform from ARM for their Cortex-M
     cores and is related to the existing Realview / Versatile Express
     lineup, but without MMU.

     We now support various NOMMU platforms, so adding a new one is
     fairly straightforward.

        http://infocenter.arm.com/help/topic/com.arm.doc.100112_0100_03_en/

     has detailed information about the platform.

  Other noteworthy updates:

   - Work on LPC32xx has resumed, and Vladimir Zapolskiy and Sylvain
     Lemieux are now maintaining the platform.

     This is an older ARM9 based platform from NXP (not Freescale), but
     it remains in use in embedded markets.

   - Kevin Hilman is now co-maintaining the Amlogic Meson platform for
     both 32-bit and 64-bit ARM, and started contributing some patches.

   - As is often the case, work on the OMAP platforms makes up the bulk
     of the actual SoC code changes in arch/arm, but there isn't a lot
     of that either"

* tag 'armsoc-soc' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (42 commits)
  MAINTAINERS: ARM/Amlogic: add co-maintainer, misc. updates
  MAINTAINERS: add ARM/NXP LPC32XX SoC specific drivers to the section
  MAINTAINERS: add new maintainers of NXP LPC32xx SoC
  MAINTAINERS: move ARM/NXP LPC32xx record to ARM section
  arm: Add Aspeed machine
  ARM: lpc32xx: remove duplicate const on lpc32xx_auxdata_lookup
  ARM: lpc32xx: remove leftovers of legacy clock source and provider drivers
  ARM: lpc32xx: remove reboot header file
  ARM: dove: Remove CLK_IS_ROOT
  ARM: orion5x: Remove CLK_IS_ROOT
  ARM: mv78xx0: Remove CLK_IS_ROOT
  ARM: davinci: da850: use clk->set_parent for async3
  ARM: davinci: Move clock init after ioremap.
  MAINTAINERS: Update ARM Versatile Express platform entry
  ARM: vexpress/mps2: introduce MPS2 platform
  MAINTAINERS: add maintainer entry for ARM/OXNAS platform
  ARM: Add new mach-oxnas
  irqchip: versatile-fpga: add new compatible for OX810SE SoC
  ARM: uniphier: correct the call order of of_node_put()
  MAINTAINERS: fix stale TI DaVinci entries
  ...

Pull ARM SoC cleanups and fixes from Arnd Bergmann:
 "Traditionally we've had two separate branches for cleanups and
  non-critical bug fixes, but both of these got smaller with each
  release and the differences are rather unclear now, so it seems more
  appropriate to have a combined branch.

  The most notable change is for OMAP, which gets a small rework to
  simplify handling of the AUXDATA mechanism used on machines that are
  not completely DT based yet, along with other work that is used as
  preparation for dropping the legacy board files"

* tag 'armsoc-cleanups-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc:
  ARM: dts: exynos: Add interrupt line to MAX8997 PMIC on exynos4210-trats
  ARM: dts: exynos: Fix regulator name to avoid forbidden character on exynos4210-trats
  ARM: dts: exynos: Add MFC memory banks for Peach boards
  ARM: OMAP2+: n900 needs MMC slot names for legacy user space
  ARM: OMAP2+: Add more functions to pwm pdata for ir-rx51
  ARM: debug: remove extraneous DEBUG_HI3716_UART option
  ARM: OMAP2+: Simplify auxdata by using the generic match
  of/platform: Allow secondary compatible match in of_dev_lookup
  ARM: davinci: use IRQCHIP_DECLARE for cp_intc
  ARM: davinci: remove unused DA8XX_NUM_UARTS
  ARM: davinci: simplify call to of populate
  ARM: DaVinci USB: removed deprecated properties from MUSB config
  ARM: rockchip: Fix use of plain integer as NULL pointer
  ARM: realview: hide unused 'pmu_device' object
  soc: versatile: dynamically detect RealView HBI numbers

Pull s390 updates from Martin Schwidefsky:
 "The s390 patches for the 4.7 merge window have the usual bug fixes and
  cleanups, and the following new features:

   - An interface for dasd driver to query if a volume is online to
     another operating system

   - A new ioctl for the dasd driver to verify the format for a range of
     tracks

   - Following the example of x86 the struct fpu is now allocated with
     the task_struct

   - The 'report_error' interface for the PCI bus to send an
     adapter-error notification from user space to the service element
     of the machine"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux: (29 commits)
  s390/vmem: remove unused function parameter
  s390/vmem: fix identity mapping
  s390: add missing include statements
  s390: add missing declarations
  s390: make couple of variables and functions static
  s390/cache: remove superfluous locking
  s390/cpuinfo: simplify locking and skip offline cpus early
  s390/3270: hangup the 3270 tty after a disconnect
  s390/3270: handle reconnect of a tty with a different size
  s390/3270: avoid endless I/O loop with disconnected 3270 terminals
  s390/3270: fix garbled output on 3270 tty view
  s390/3270: fix view reference counting
  s390/3270: add missing tty_kref_put
  s390/dumpstack: implement and use return_address()
  s390/cpum_sf: Remove superfluous SMP function call
  s390/cpum_cf: Remove superfluous SMP function call
  s390/Kconfig: make z196 the default processor type
  s390/sclp: avoid compile warning in sclp_pci_report
  s390/fpu: allocate 'struct fpu' with the task_struct
  s390/crypto: cleanup and move the header with the cpacf definitions
  ...

My laptop that uses the intel 7680 iwlwifi module would no longer
connects to the network.  It would fail with a "Microcode SW error
detected." and spew out register state over and over again without ever
connecting to the network.

The cause is mis-merge in commit 909b27f7, where David seems to have
lost some of the changes to iwl_mvm_set_tx_cmd() from commit
5c08b0f5 ("iwlwifi: mvm: don't override the rate with the AMSDU
len").

The reason seems to be a conflict with commit d8fe4844 ("iwlwifi:
mvm: add support for new TX CMD API"), which touched a line adjacent to
the changes in 909b27f7.

David missed the fact that "info->driver_data[0]" had become
"skb_info->driver_data[0]".  Then he removed the skb_info because it was
unused.

This just re-updates iwl_mvm_set_tx_cmd() with the lost two lines.
Reported-and-tested-by: NLinus Torvalds <torvalds@linux-foundation.org>
Reported-by: NReinoud Koornstra <reinoudkoornstra@gmail.com>
Cc: Luciano Coelho <luciano.coelho@intel.com>
Cc: Kalle Valo <kvalo@codeaurora.org>
Cc: David Miller <davem@davemloft.net>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Pull misc vfs cleanups from Al Viro:
 "Assorted cleanups and fixes all over the place"

* 'work.misc' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  coredump: only charge written data against RLIMIT_CORE
  coredump: get rid of coredump_params->written
  ecryptfs_lookup(): try either only encrypted or plaintext name
  ecryptfs: avoid multiple aliases for directories
  bpf: reject invalid names right in ->lookup()
  __d_alloc(): treat NULL name as QSTR("/", 1)
  mtd: switch ubi_open_volume_path() to vfs_stat()
  mtd: switch open_mtd_by_chdev() to use of vfs_stat()

Pull iov_iter cleanups from Al Viro.

* 'work.iov_iter' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  fold checks into iterate_and_advance()
  rw_verify_area(): saner calling conventions
  aio: remove a pointless assignment

Pull parallel lookup fixups from Al Viro:
 "Fix for xfs parallel readdir (turns out the cxfs exposure was not
  enough to catch all problems), and a reversion of btrfs back to
  ->iterate() until the fs/btrfs/delayed-inode.c gets fixed"

* 'work.lookups' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  xfs: concurrent readdir hangs on data buffer locks
  Revert "btrfs: switch to ->iterate_shared()"

There's a three-process deadlock involving shared/exclusive barriers
and inverted lock orders in the directory readdir implementation.
It's a pre-existing problem with lock ordering, exposed by the
VFS parallelisation code.

process 1               process 2               process 3
---------               ---------               ---------
readdir
iolock(shared)
  get_leaf_dents
    iterate entries
       ilock(shared)
       map, lock and read buffer
       iunlock(shared)
       process entries in buffer
       .....
                                                readdir
                                                iolock(shared)
                                                  get_leaf_dents
                                                    iterate entries
                                                      ilock(shared)
                                                      map, lock buffer
                                                      <blocks>
                        finish ->iterate_shared
                        file_accessed()
                          ->update_time
                            start transaction
                            ilock(excl)
                            <blocks>
        .....
        finishes processing buffer
        get next buffer
          ilock(shared)
          <blocks>

And that's the deadlock.

Fix this by dropping the current buffer lock in process 1 before
trying to map the next buffer. This means we keep the lock order of
ilock -> buffer lock intact and hence will allow process 3 to make
progress and drop it's ilock(shared) once it is done.
Reported-by: NXiong Zhou <xzhou@redhat.com>
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

This reverts commit 972b241f.
Quoth Chris:
	didn't take the delayed inode stuff into account
	it got an rbtree of items and it pulls things out
	so in shared mode, its hugely racey
	sorry, lets revert and fix it for real inside of btrfs
Signed-off-by: NChris Mason <clm@fb.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Pull cifs iovec cleanups from Al Viro.

* 'sendmsg.cifs' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  cifs: don't bother with kmap on read_pages side
  cifs_readv_receive: use cifs_read_from_socket()
  cifs: no need to wank with copying and advancing iovec on recvmsg side either
  cifs: quit playing games with draining iovecs
  cifs: merge the hash calculation helpers

Pull remaining vfs xattr work from Al Viro:
 "The rest of work.xattr (non-cifs conversions)"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  btrfs: Switch to generic xattr handlers
  ubifs: Switch to generic xattr handlers
  jfs: Switch to generic xattr handlers
  jfs: Clean up xattr name mapping
  gfs2: Switch to generic xattr handlers
  ceph: kill __ceph_removexattr()
  ceph: Switch to generic xattr handlers
  ceph: Get rid of d_find_alias in ceph_set_acl

Pull cifs updates from Steve French:
 "Various small CIFS and SMB3 fixes (including some for stable)"

* 'for-next' of git://git.samba.org/sfrench/cifs-2.6:
  remove directory incorrectly tries to set delete on close on non-empty directories
  Update cifs.ko version to 2.09
  fs/cifs: correctly to anonymous authentication for the NTLM(v2) authentication
  fs/cifs: correctly to anonymous authentication for the NTLM(v1) authentication
  fs/cifs: correctly to anonymous authentication for the LANMAN authentication
  fs/cifs: correctly to anonymous authentication via NTLMSSP
  cifs: remove any preceding delimiter from prefix_path
  cifs: Use file_dentry()

Pull libata updates from Tejun Heo:
 "Trivial changes except for special case timeout bumping.

  I have two more libata branches which depend on SCSI and dmaengine
  tree respectively.  I'll send pull requests for them once the
  prerequisite trees are pulled in"

* 'for-4.7' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/libata:
  libata-scsi: use %*ph to dump small buffers
  treewide: Fix typos in libata.xml
  libata-core: Allow longer timeout for drive spinup from PUIS
  libata: Fixup awkward whitespace in warning by removing line continuation.

Merge tag 'regulator-fix-can-change-voltage' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator

Pull regulator fix from Mark Brown:
 "Fix build warnings from regulator_can_change_voltage()

  Cut down on noise for mainstream users of the API and people
  doing build testing by dropping the deprecated flag from
  regulator_can_change_voltage() as it triggers even on the
  EXPORT_SYMBOL_GPL() which affects all builds rather than just
  the remaining drivers with calls to it (for which fixes are
  currently pending).

  The function remains deprecated and is expected to be removed
  entirely in v4.8"

* tag 'regulator-fix-can-change-voltage' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator:
  regulator: Silence build warnings from regulator_can_change_voltage()

Pull GPIO updates from Linus Walleij:
 "This is the bulk of GPIO changes for kernel cycle v4.7:

  Core infrastructural changes:

   - Support for natively single-ended GPIO driver stages.

     This means that if the hardware has registers to configure open
     drain or open source configuration, we use that rather than (as we
     did before) try to emulate it by switching the line to an input to
     get high impedance.

     This is also documented throughly in Documentation/gpio/driver.txt
     for those of you who did not understand one word of what I just
     wrote.

   - Start to do away with the unnecessarily complex and unitelligible
     ARCH_REQUIRE_GPIOLIB and ARCH_WANT_OPTIONAL_GPIOLIB, another
     evolutional artifact from the time when the GPIO subsystem was
     unmaintained.

     Archs can now just select GPIOLIB and be done with it, cleanups to
     arches will trickle in for the next kernel.  Some minor archs ACKed
     the changes immediately so these are included in this pull request.

   - Advancing the use of the data pointer inside the GPIO device for
     storing driver data by switching the PowerPC, Super-H Unicore and
     a few other subarches or subsystem drivers in ALSA SoC, Input,
     serial, SSB, staging etc to use it.

   - The initialization now reads the input/output state of the GPIO
     lines, so that each GPIO descriptor knows - if this callback is
     implemented - whether the line is input or output.  This also
     reflects nicely in userspace "lsgpio".

   - It is now possible to name GPIO producer names, line names, from
     the device tree.  (Platform data has been supported for a while).
     I bet we will get a similar mechanism for ACPI one of those days.
     This makes is possible to get sensible producer names for e.g.
     GPIO rails in "lsgpio" in userspace.

  New drivers:

   - New driver for the Loongson1.

   - The XLP driver now supports Broadcom Vulcan ARM64.

   - The IT87 driver now supports IT8620 and IT8628.

   - The PCA953X driver now supports Galileo Gen2.

  Driver improvements:

   - MCP23S08 was switched to use the gpiolib irqchip helpers and now
     also suppors level-triggered interrupts.

   - 74x164 and RCAR now supports the .set_multiple() callback

   - AMDPT was converted to use generic GPIO.

   - TC3589x, TPS65218, SX150X, F7188X, MENZ127, VX855, WM831X, WM8994
     support the new single ended callback for open drain and in some
     cases open source.

   - Implement the .get_direction() callback for a few more drivers like
     PL061, Xgene.

  Cleanups:

   - Paul Gortmaker combed through the drivers and de-modularized those
     who are not really modules.

   - Move the GPIO poweroff DT bindings to the power subdir where they
     belong.

   - Rename gpio-generic.c to gpio-mmio.c, which is much more to the
     point.  That's what it is handling, nothing more, nothing less"

* tag 'gpio-v4.7-1' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-gpio: (126 commits)
  MIPS: do away with ARCH_[WANT_OPTIONAL|REQUIRE]_GPIOLIB
  gpio: zevio: make it explicitly non-modular
  gpio: timberdale: make it explicitly non-modular
  gpio: stmpe: make it explicitly non-modular
  gpio: sodaville: make it explicitly non-modular
  pinctrl: sh-pfc: Let gpio_chip.to_irq() return zero on error
  gpio: dwapb: Add ACPI device ID for DWAPB GPIO controller on X-Gene platforms
  gpio: dt-bindings: add wd,mbl-gpio bindings
  gpio: of: make it possible to name GPIO lines
  gpio: make gpiod_to_irq() return negative for NO_IRQ
  gpio: xgene: implement .get_direction()
  gpio: xgene: Enable ACPI support for X-Gene GFC GPIO driver
  gpio: tegra: Implement gpio_get_direction callback
  gpio: set up initial state from .get_direction()
  gpio: rename gpio-generic.c into gpio-mmio.c
  gpio: generic: fix GPIO_GENERIC_PLATFORM is set to module case
  gpio: dwapb: add gpio-signaled acpi event support
  gpio: dwapb: convert device node to fwnode
  gpio: dwapb: remove name from dwapb_port_property
  gpio/qoriq: select IRQ_DOMAIN
  ...

Pull HID updates from Jiri Kosina:
 "No biggies this time:

   - micro-optimization of implement() in HID core parses, from Dmitry
     Torokhov

   - thingm driver cleanups from Heiner Kallweit

   - fine-graining detection of distance and tilt axes in wacom driver
     from Jason Gerecke

   - New hid-asus driver, currently supporting X205TA and VivoBook
     E200HA, from Yusuke Fujimaki"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid:
  HID: wacom: Add fuzz factor to distance and tilt axes
  HID: usbhid: quirks for Corsair RGB keyboard & mice (K70R, K95RGB, M65RGB, K70RGB, K65RGB)
  HID: thingm: remove not needed error message
  HID: thingm: set new flag LED_HW_PLUGGABLE
  HID: thingm: factor out duplicated code to thingm_init_led
  HID: simplify implement() a bit
  HID: asus: add support for VivoBook E200HA
  HID: hidraw: silence an uninitialized variable warning
  HID: roccat: silence an uninitialized variable warning
  HID: Asus X205TA keyboard driver
  HID: hidraw: switch to using memdup_user

Pull livepatching updates from Jiri Kosina:

 - remove of our own implementation of architecture-specific relocation
   code and leveraging existing code in the module loader to perform
   arch-dependent work, from Jessica Yu.

   The relevant patches have been acked by Rusty (for module.c) and
   Heiko (for s390).

 - live patching support for ppc64le, which is a joint work of Michael
   Ellerman and Torsten Duwe.  This is coming from topic branch that is
   share between livepatching.git and ppc tree.

 - addition of livepatching documentation from Petr Mladek

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/livepatching:
  livepatch: make object/func-walking helpers more robust
  livepatch: Add some basic livepatch documentation
  powerpc/livepatch: Add live patching support on ppc64le
  powerpc/livepatch: Add livepatch stack to struct thread_info
  powerpc/livepatch: Add livepatch header
  livepatch: Allow architectures to specify an alternate ftrace location
  ftrace: Make ftrace_location_range() global
  livepatch: robustify klp_register_patch() API error checking
  Documentation: livepatch: outline Elf format and requirements for patch modules
  livepatch: reuse module loader code to write relocations
  module: s390: keep mod_arch_specific for livepatch modules
  module: preserve Elf information for livepatch modules
  Elf: add livepatch-specific Elf constants

Pull trivial tree updates from Jiri Kosina.

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial: (21 commits)
  gitignore: fix wording
  mfd: ab8500-debugfs: fix "between" in printk
  memstick: trivial fix of spelling mistake on management
  cpupowerutils: bench: fix "average"
  treewide: Fix typos in printk
  IB/mlx4: printk fix
  pinctrl: sirf/atlas7: fix printk spelling
  serial: mctrl_gpio: Grammar s/lines GPIOs/line GPIOs/, /sets/set/
  w1: comment spelling s/minmum/minimum/
  Blackfin: comment spelling s/divsor/divisor/
  metag: Fix misspellings in comments.
  ia64: Fix misspellings in comments.
  hexagon: Fix misspellings in comments.
  tools/perf: Fix misspellings in comments.
  cris: Fix misspellings in comments.
  c6x: Fix misspellings in comments.
  blackfin: Fix misspelling of 'register' in comment.
  avr32: Fix misspelling of 'definitions' in comment.
  treewide: Fix typos in printk
  Doc: treewide : Fix typos in DocBook/filesystem.xml
  ...

Pull networking updates from David Miller:
 "Highlights:

   1) Support SPI based w5100 devices, from Akinobu Mita.

   2) Partial Segmentation Offload, from Alexander Duyck.

   3) Add GMAC4 support to stmmac driver, from Alexandre TORGUE.

   4) Allow cls_flower stats offload, from Amir Vadai.

   5) Implement bpf blinding, from Daniel Borkmann.

   6) Optimize _ASYNC_ bit twiddling on sockets, unless the socket is
      actually using FASYNC these atomics are superfluous.  From Eric
      Dumazet.

   7) Run TCP more preemptibly, also from Eric Dumazet.

   8) Support LED blinking, EEPROM dumps, and rxvlan offloading in mlx5e
      driver, from Gal Pressman.

   9) Allow creating ppp devices via rtnetlink, from Guillaume Nault.

  10) Improve BPF usage documentation, from Jesper Dangaard Brouer.

  11) Support tunneling offloads in qed, from Manish Chopra.

  12) aRFS offloading in mlx5e, from Maor Gottlieb.

  13) Add RFS and RPS support to SCTP protocol, from Marcelo Ricardo
      Leitner.

  14) Add MSG_EOR support to TCP, this allows controlling packet
      coalescing on application record boundaries for more accurate
      socket timestamp sampling.  From Martin KaFai Lau.

  15) Fix alignment of 64-bit netlink attributes across the board, from
      Nicolas Dichtel.

  16) Per-vlan stats in bridging, from Nikolay Aleksandrov.

  17) Several conversions of drivers to ethtool ksettings, from Philippe
      Reynes.

  18) Checksum neutral ILA in ipv6, from Tom Herbert.

  19) Factorize all of the various marvell dsa drivers into one, from
      Vivien Didelot

  20) Add VF support to qed driver, from Yuval Mintz"

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next: (1649 commits)
  Revert "phy dp83867: Fix compilation with CONFIG_OF_MDIO=m"
  Revert "phy dp83867: Make rgmii parameters optional"
  r8169: default to 64-bit DMA on recent PCIe chips
  phy dp83867: Make rgmii parameters optional
  phy dp83867: Fix compilation with CONFIG_OF_MDIO=m
  bpf: arm64: remove callee-save registers use for tmp registers
  asix: Fix offset calculation in asix_rx_fixup() causing slow transmissions
  switchdev: pass pointer to fib_info instead of copy
  net_sched: close another race condition in tcf_mirred_release()
  tipc: fix nametable publication field in nl compat
  drivers: net: Don't print unpopulated net_device name
  qed: add support for dcbx.
  ravb: Add missing free_irq() calls to ravb_close()
  qed: Remove a stray tab
  net: ethernet: fec-mpc52xx: use phy_ethtool_{get|set}_link_ksettings
  net: ethernet: fec-mpc52xx: use phydev from struct net_device
  bpf, doc: fix typo on bpf_asm descriptions
  stmmac: hardware TX COE doesn't work when force_thresh_dma_mode is set
  net: ethernet: fs-enet: use phy_ethtool_{get|set}_link_ksettings
  net: ethernet: fs-enet: use phydev from struct net_device
  ...

The btrfs_{set,remove}xattr inode operations check for a read-only root
(btrfs_root_readonly) before calling into generic_{set,remove}xattr.  If
this check is moved into __btrfs_setxattr, we can get rid of
btrfs_{set,remove}xattr.

This patch applies to mainline, I would like to keep it together with
the other xattr cleanups if possible, though.  Could you please review?

Thanks,
Andreas
Signed-off-by: NAndreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Ubifs internally uses special inodes for storing xattrs. Those inodes
had NULL {get,set,remove}xattr inode operations before this change, so
xattr operations on them would fail. The super block's s_xattr field
would also apply to those special inodes. However, the inodes are not
visible outside of ubifs, and so no xattr operations will ever be
carried out on them anyway.
Signed-off-by: NAndreas Gruenbacher <agruenba@redhat.com>
Reviewed-by: NRichard Weinberger <richard@nod.at>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Pull device mapper updates from Mike Snitzer:

 - based on Jens' 'for-4.7/core' to have DM thinp's discard support use
   bio_inc_remaining() and the block core's new async __blkdev_issue_discard()
   interface

 - make DM multipath's fast code-paths lockless, using lockless_deference,
   to significantly improve large NUMA performance when using blk-mq.
   The m->lock spinlock contention was a serious bottleneck.

 - a few other small code cleanups and Documentation fixes

* tag 'dm-4.7-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm:
  dm thin: unroll issue_discard() to create longer discard bio chains
  dm thin: use __blkdev_issue_discard for async discard support
  dm thin: remove __bio_inc_remaining() and switch to using bio_inc_remaining()
  dm raid: make sure no feature flags are set in metadata
  dm ioctl: drop use of __GFP_REPEAT in copy_params()'s __vmalloc() call
  dm stats: fix spelling mistake in Documentation
  dm cache: update cache-policies.txt now that mq is an alias for smq
  dm mpath: eliminate use of spinlock in IO fast-paths
  dm mpath: move trigger_event member to the end of 'struct multipath'
  dm mpath: use atomic_t for counting members of 'struct multipath'
  dm mpath: switch to using bitops for state flags
  dm thin: Remove return statement from void function
  dm: remove unused mapped_device argument from free_tio()

Pull block driver updates from Jens Axboe:
 "On top of the core pull request, this is the drivers pull request for
  this merge window.  This contains:

   - Switch drivers to the new write back cache API, and kill off the
     flush flags.  From me.

   - Kill the discard support for the STEC pci-e flash driver.  It's
     trivially broken, and apparently unmaintained, so it's safer to
     just remove it.  From Jeff Moyer.

   - A set of lightnvm updates from the usual suspects (Matias/Javier,
     and Simon), and fixes from Arnd, Jeff Mahoney, Sagi, and Wenwei
     Tao.

   - A set of updates for NVMe:

        - Turn the controller state management into a proper state
          machine.  From Christoph.

        - Shuffling of code in preparation for NVMe-over-fabrics, also
          from Christoph.

        - Cleanup of the command prep part from Ming Lin.

        - Rewrite of the discard support from Ming Lin.

        - Deadlock fix for namespace removal from Ming Lin.

        - Use the now exported blk-mq tag helper for IO termination.
          From Sagi.

        - Various little fixes from Christoph, Guilherme, Keith, Ming
          Lin, Wang Sheng-Hui.

   - Convert mtip32xx to use the now exported blk-mq tag iter function,
     from Keith"

* 'for-4.7/drivers' of git://git.kernel.dk/linux-block: (74 commits)
  lightnvm: reserved space calculation incorrect
  lightnvm: rename nr_pages to nr_ppas on nvm_rq
  lightnvm: add is_cached entry to struct ppa_addr
  lightnvm: expose gennvm_mark_blk to targets
  lightnvm: remove mgt targets on mgt removal
  lightnvm: pass dma address to hardware rather than pointer
  lightnvm: do not assume sequential lun alloc.
  nvme/lightnvm: Log using the ctrl named device
  lightnvm: rename dma helper functions
  lightnvm: enable metadata to be sent to device
  lightnvm: do not free unused metadata on rrpc
  lightnvm: fix out of bound ppa lun id on bb tbl
  lightnvm: refactor set_bb_tbl for accepting ppa list
  lightnvm: move responsibility for bad blk mgmt to target
  lightnvm: make nvm_set_rqd_ppalist() aware of vblks
  lightnvm: remove struct factory_blks
  lightnvm: refactor device ops->get_bb_tbl()
  lightnvm: introduce nvm_for_each_lun_ppa() macro
  lightnvm: refactor dev->online_target to global nvm_targets
  lightnvm: rename nvm_targets to nvm_tgt_type
  ...

Pull core block layer updates from Jens Axboe:
 "This is the core block IO changes for this merge window.  Nothing
  earth shattering in here, it's mostly just fixes.  In detail:

   - Fix for a long standing issue where wrong ordering in blk-mq caused
     order_to_size() to spew a warning.  From Bart.

   - Async discard support from Christoph.  Basically just splitting our
     sync interface into a submit + wait part.

   - Add a cleaner interface for flagging whether a device has a write
     back cache or not.  We've previously overloaded blk_queue_flush()
     with this, but let's make it more explicit.  Drivers cleaned up and
     updated in the drivers pull request.  From me.

   - Fix for a double check for whether IO accounting is enabled or not.
     From Michael Callahan.

   - Fix for the async discard from Mike Snitzer, reinstating the early
     EOPNOTSUPP return if the device doesn't support discards.

   - Also from Mike, export bio_inc_remaining() so dm can drop it's
     private copy of it.

   - From Ming Lin, add support for passing in an offset for request
     payloads.

   - Tag function export from Sagi, which will be used in NVMe in the
     drivers pull.

   - Two blktrace related fixes from Shaohua.

   - Propagate NOMERGE flag when making a request from a bio, also from
     Shaohua.

   - An optimization to not parse cgroup paths in blk-throttle, if we
     don't need to.  From Shaohua"

* 'for-4.7/core' of git://git.kernel.dk/linux-block:
  blk-mq: fix undefined behaviour in order_to_size()
  blk-throttle: don't parse cgroup path if trace isn't enabled
  blktrace: add missed mask name
  blktrace: delete garbage for message trace
  block: make bio_inc_remaining() interface accessible again
  block: reinstate early return of -EOPNOTSUPP from blkdev_issue_discard
  block: Minor blk_account_io_start usage cleanup
  block: add __blkdev_issue_discard
  block: remove struct bio_batch
  block: copy NOMERGE flag from bio to request
  block: add ability to flag write back caching on a device
  blk-mq: Export tagset iter function
  block: add offset in blk_add_request_payload()
  writeback: Fix performance regression in wb_over_bg_thresh()

Pull vfs cleanups from Al Viro:
 "More cleanups from Christoph"

* 'work.preadv2' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  nfsd: use RWF_SYNC
  fs: add RWF_DSYNC aand RWF_SYNC
  ceph: use generic_write_sync
  fs: simplify the generic_write_sync prototype
  fs: add IOCB_SYNC and IOCB_DSYNC
  direct-io: remove the offset argument to dio_complete
  direct-io: eliminate the offset argument to ->direct_IO
  xfs: eliminate the pos variable in xfs_file_dio_aio_write
  filemap: remove the pos argument to generic_file_direct_write
  filemap: remove pos variables in generic_file_read_iter