Applying restrictive seccomp filter programs to large or diverse
codebases often requires handling threads which may be started early in
the process lifetime (e.g., by code that is linked in). While it is
possible to apply permissive programs prior to process start up, it is
difficult to further restrict the kernel ABI to those threads after that
point.

This change adds a new seccomp syscall flag to SECCOMP_SET_MODE_FILTER for
synchronizing thread group seccomp filters at filter installation time.

When calling seccomp(SECCOMP_SET_MODE_FILTER, SECCOMP_FILTER_FLAG_TSYNC,
filter) an attempt will be made to synchronize all threads in current's
threadgroup to its new seccomp filter program. This is possible iff all
threads are using a filter that is an ancestor to the filter current is
attempting to synchronize to. NULL filters (where the task is running as
SECCOMP_MODE_NONE) are also treated as ancestors allowing threads to be
transitioned into SECCOMP_MODE_FILTER. If prctrl(PR_SET_NO_NEW_PRIVS,
...) has been set on the calling thread, no_new_privs will be set for
all synchronized threads too. On success, 0 is returned. On failure,
the pid of one of the failing threads will be returned and no filters
will have been applied.

The race conditions against another thread are:
- requesting TSYNC (already handled by sighand lock)
- performing a clone (already handled by sighand lock)
- changing its filter (already handled by sighand lock)
- calling exec (handled by cred_guard_mutex)
The clone case is assisted by the fact that new threads will have their
seccomp state duplicated from their parent before appearing on the tasklist.

Holding cred_guard_mutex means that seccomp filters cannot be assigned
while in the middle of another thread's exec (potentially bypassing
no_new_privs or similar). The call to de_thread() may kill threads waiting
for the mutex.

Changes across threads to the filter pointer includes a barrier.

Based on patches by Will Drewry.
Suggested-by: NJulien Tinnes <jln@chromium.org>
Signed-off-by: NKees Cook <keescook@chromium.org>
Reviewed-by: NOleg Nesterov <oleg@redhat.com>
Reviewed-by: NAndy Lutomirski <luto@amacapital.net>

This changes the mode setting helper to allow threads to change the
seccomp mode from another thread. We must maintain barriers to keep
TIF_SECCOMP synchronized with the rest of the seccomp state.
Signed-off-by: NKees Cook <keescook@chromium.org>
Reviewed-by: NOleg Nesterov <oleg@redhat.com>
Reviewed-by: NAndy Lutomirski <luto@amacapital.net>

Normally, task_struct.seccomp.filter is only ever read or modified by
the task that owns it (current). This property aids in fast access
during system call filtering as read access is lockless.

Updating the pointer from another task, however, opens up race
conditions. To allow cross-thread filter pointer updates, writes to the
seccomp fields are now protected by the sighand spinlock (which is shared
by all threads in the thread group). Read access remains lockless because
pointer updates themselves are atomic.  However, writes (or cloning)
often entail additional checking (like maximum instruction counts)
which require locking to perform safely.

In the case of cloning threads, the child is invisible to the system
until it enters the task list. To make sure a child can't be cloned from
a thread and left in a prior state, seccomp duplication is additionally
moved under the sighand lock. Then parent and child are certain have
the same seccomp state when they exit the lock.

Based on patches by Will Drewry and David Drysdale.
Signed-off-by: NKees Cook <keescook@chromium.org>
Reviewed-by: NOleg Nesterov <oleg@redhat.com>
Reviewed-by: NAndy Lutomirski <luto@amacapital.net>

In preparation for adding seccomp locking, move filter creation away
from where it is checked and applied. This will allow for locking where
no memory allocation is happening. The validation, filter attachment,
and seccomp mode setting can all happen under the future locks.

For extreme defensiveness, I've added a BUG_ON check for the calculated
size of the buffer allocation in case BPF_MAXINSN ever changes, which
shouldn't ever happen. The compiler should actually optimize out this
check since the test above it makes it impossible.
Signed-off-by: NKees Cook <keescook@chromium.org>
Reviewed-by: NOleg Nesterov <oleg@redhat.com>
Reviewed-by: NAndy Lutomirski <luto@amacapital.net>

Since seccomp transitions between threads requires updates to the
no_new_privs flag to be atomic, the flag must be part of an atomic flag
set. This moves the nnp flag into a separate task field, and introduces
accessors.
Signed-off-by: NKees Cook <keescook@chromium.org>
Reviewed-by: NOleg Nesterov <oleg@redhat.com>
Reviewed-by: NAndy Lutomirski <luto@amacapital.net>

Wires up the new seccomp syscall.
Signed-off-by: NKees Cook <keescook@chromium.org>
Reviewed-by: NOleg Nesterov <oleg@redhat.com>

Wires up the new seccomp syscall.
Signed-off-by: NKees Cook <keescook@chromium.org>
Reviewed-by: NOleg Nesterov <oleg@redhat.com>

This adds the new "seccomp" syscall with both an "operation" and "flags"
parameter for future expansion. The third argument is a pointer value,
used with the SECCOMP_SET_MODE_FILTER operation. Currently, flags must
be 0. This is functionally equivalent to prctl(PR_SET_SECCOMP, ...).

In addition to the TSYNC flag later in this patch series, there is a
non-zero chance that this syscall could be used for configuring a fixed
argument area for seccomp-tracer-aware processes to pass syscall arguments
in the future. Hence, the use of "seccomp" not simply "seccomp_add_filter"
for this syscall. Additionally, this syscall uses operation, flags,
and user pointer for arguments because strictly passing arguments via
a user pointer would mean seccomp itself would be unable to trivially
filter the seccomp syscall itself.
Signed-off-by: NKees Cook <keescook@chromium.org>
Reviewed-by: NOleg Nesterov <oleg@redhat.com>
Reviewed-by: NAndy Lutomirski <luto@amacapital.net>

Separates the two mode setting paths to make things more readable with
fewer #ifdefs within function bodies.
Signed-off-by: NKees Cook <keescook@chromium.org>
Reviewed-by: NOleg Nesterov <oleg@redhat.com>
Reviewed-by: NAndy Lutomirski <luto@amacapital.net>

To support splitting mode 1 from mode 2, extract the mode checking and
assignment logic into common functions.
Signed-off-by: NKees Cook <keescook@chromium.org>
Reviewed-by: NOleg Nesterov <oleg@redhat.com>
Reviewed-by: NAndy Lutomirski <luto@amacapital.net>

In preparation for having other callers of the seccomp mode setting
logic, split the prctl entry point away from the core logic that performs
seccomp mode setting.
Signed-off-by: NKees Cook <keescook@chromium.org>
Reviewed-by: NOleg Nesterov <oleg@redhat.com>
Reviewed-by: NAndy Lutomirski <luto@amacapital.net>

Add myself as seccomp maintainer.
Suggested-by: NJames Morris <jmorris@namei.org>
Signed-off-by: NKees Cook <keescook@chromium.org>

The asynchronous hash API allows initiating a hash calculation and
then performing other tasks, while waiting for the hash calculation
to complete.

This patch introduces usage of double buffering for simultaneous
hashing and reading of the next chunk of data from storage.

Changes in v3:
- better comments
Signed-off-by: NDmitry Kasatkin <d.kasatkin@samsung.com>
Signed-off-by: NMimi Zohar <zohar@linux.vnet.ibm.com>

Use of multiple-page collect buffers reduces:
1) the number of block IO requests
2) the number of asynchronous hash update requests

Second is important for HW accelerated hashing, because significant
amount of time is spent for preparation of hash update operation,
which includes configuring acceleration HW, DMA engine, etc...
Thus, HW accelerators are more efficient when working on large
chunks of data.

This patch introduces usage of multi-page collect buffers. Buffer size
can be specified using 'ahash_bufsize' module parameter. Default buffer
size is 4096 bytes.

Changes in v3:
- kernel parameter replaced with module parameter
Signed-off-by: NDmitry Kasatkin <d.kasatkin@samsung.com>
Signed-off-by: NMimi Zohar <zohar@linux.vnet.ibm.com>

Async hash API allows the use of HW acceleration for hash calculation.
It may give significant performance gain and/or reduce power consumption,
which might be very beneficial for battery powered devices.

This patch introduces hash calculation using ahash API. ahash performance
depends on the data size and the particular HW. Depending on the specific
system, shash performance may be better.

This patch defines 'ahash_minsize' module parameter, which is used to
define the minimal file size to use with ahash.  If this minimum file size
is not set or the file is smaller than defined by the parameter, shash will
be used.

Changes in v3:
- kernel parameter replaced with module parameter
- pr_crit replaced with pr_crit_ratelimited
- more comment changes - Mimi

Changes in v2:
- ima_ahash_size became as ima_ahash
- ahash pre-allocation moved out from __init code to be able to use
  ahash crypto modules. Ahash allocated once on the first use.
- hash calculation falls back to shash if ahash allocation/calculation fails
- complex initialization separated from variable declaration
- improved comments
Signed-off-by: NDmitry Kasatkin <d.kasatkin@samsung.com>
Signed-off-by: NMimi Zohar <zohar@linux.vnet.ibm.com>

Replace spaces in op keyword labels in log output since userspace audit tools
can't parse orphaned keywords.
Reported-by: NSteve Grubb <sgrubb@redhat.com>
Signed-off-by: NRichard Guy Briggs <rgb@redhat.com>
Signed-off-by: NMimi Zohar <zohar@linux.vnet.ibm.com>

process_measurement() always calls ima_template_desc_current(),
including when an IMA policy has not been defined.

This patch delays template descriptor lookup until action is
determined.
Signed-off-by: NDmitry Kasatkin <d.kasatkin@samsung.com>
Signed-off-by: NMimi Zohar <zohar@linux.vnet.ibm.com>

Before 2.6.39 inode->i_readcount was maintained by IMA. It was not atomic
and protected using spinlock. For 2.6.39, i_readcount was converted to
atomic and maintaining was moved VFS layer. Spinlock for some unclear
reason was replaced by i_mutex.

After analyzing the code, we came to conclusion that i_mutex locking is
unnecessary, especially when an IMA policy has not been defined.

This patch removes i_mutex locking from ima_rdwr_violation_check().
Signed-off-by: NDmitry Kasatkin <d.kasatkin@samsung.com>
Signed-off-by: NMimi Zohar <zohar@linux.vnet.ibm.com>

Pull ext4 bugfixes from Ted Ts'o:
 "More bug fixes for ext4 -- most importantly, a fix for a bug
  introduced in 3.15 that can end up triggering a file system corruption
  error after a journal replay.

  It shouldn't lead to any actual data corruption, but it is scary and
  can force file systems to be remounted read-only, etc"

* tag 'ext4_for_linus_stable' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4:
  ext4: fix potential null pointer dereference in ext4_free_inode
  ext4: fix a potential deadlock in __ext4_es_shrink()
  ext4: revert commit which was causing fs corruption after journal replays
  ext4: disable synchronous transaction batching if max_batch_time==0
  ext4: clarify ext4_error message in ext4_mb_generate_buddy_error()
  ext4: clarify error count warning messages
  ext4: fix unjournalled bg descriptor while initializing inode bitmap

Pull clock driver fixes from Mike Turquette:
 "This batch of fixes is for a handful of clock drivers from Allwinner,
  Samsung, ST & TI.  Most of them are of the "this hardware won't work
  without this fix" variety, including patches that fix platforms that
  did not boot under certain configurations.  Other fixes are the result
  of changes to the clock core introduced in 3.15 that had subtle
  impacts on the clock drivers.

  There are no fixes to the clock framework core in this pull request"

* tag 'clk-fixes-for-linus' of git://git.linaro.org/people/mike.turquette/linux:
  clk: spear3xx: Set proper clock parent of uart1/2
  clk: spear3xx: Use proper control register offset
  clk: qcom: HDMI source sel is 3 not 2
  clk: sunxi: fix devm_ioremap_resource error detection code
  clk: s2mps11: Fix double free corruption during driver unbind
  clk: ti: am43x: Fix boot with CONFIG_SOC_AM33XX disabled
  clk: exynos5420: Remove aclk66_peric from the clock tree description
  clk/exynos5250: fix bit number for tv sysmmu clock
  clk: s3c64xx: Hookup SPI clocks correctly
  clk: samsung: exynos4: Remove SRC_MASK_ISP gates
  clk: samsung: add more aliases for s3c24xx
  clk: samsung: fix several typos to fix boot on s3c2410
  clk: ti: set CLK_SET_RATE_NO_REPARENT for ti,mux-clock
  clk: ti: am43x: Fix boot with CONFIG_SOC_AM33XX disabled
  clk: ti: dra7: return error code in failure case
  clk: ti: apll: not allocating enough data

Pull ARM SoC fixes from Olof Johansson:
 "This week's arm-soc fixes:

   - Another set of OMAP fixes
     * Clock fixes
     * Restart handling
     * PHY regulators
     * SATA hwmod data for DRA7
     + Some trivial fixes and removal of a bit of dead code
   - Exynos fixes
     * A bunch of clock fixes
     * Some SMP fixes
     * Exynos multi-core timer: register as clocksource and fix ftrace.
     + a few other minor fixes

  There's also a couple more patches, and at91 fix for USB caused by
  common clock conversion, and more MAINTAINERS entries for shmobile.

  We're definitely switching to only regression fixes from here on out,
  we've been a little less strict than usual up until now"

* tag 'fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (26 commits)
  ARM: at91: at91sam9x5: add clocks for usb device
  ARM: EXYNOS: Register cpuidle device only on exynos4210 and 5250
  ARM: dts: Add clock property for mfc_pd in exynos5420
  clk: exynos5420: Add IDs for clocks used in PD mfc
  ARM: EXYNOS: Add support for clock handling in power domain
  ARM: OMAP2+: Remove non working OMAP HDMI audio initialization
  ARM: imx: fix shared gate clock
  ARM: dts: Update the parent for Audss clocks in Exynos5420
  ARM: EXYNOS: Update secondary boot addr for secure mode
  ARM: dts: Fix TI CPSW Phy mode selection on IGEP COM AQUILA.
  ARM: dts: am335x-evmsk: Enable the McASP FIFO for audio
  ARM: dts: am335x-evm: Enable the McASP FIFO for audio
  ARM: OMAP2+: Make GPMC skip disabled devices
  ARM: OMAP2+: create dsp device only on OMAP3 SoCs
  ARM: dts: dra7-evm: Make VDDA_1V8_PHY supply always on
  ARM: DRA7/AM43XX: fix header definition for omap44xx_restart
  ARM: OMAP2+: clock/dpll: fix _dpll_test_fint arithmetics overflow
  ARM: DRA7: hwmod: Add SYSCONFIG for usb_otg_ss
  ARM: DRA7: hwmod: Fixup SATA hwmod
  ARM: OMAP3: PRM/CM: Add back macros used by TI DSP/Bridge driver
  ...

Pull ARM fixes from Russell King:
 "Another round of fixes for ARM:
   - a set of kprobes fixes from Jon Medhurst
   - fix the revision checking for the L2 cache which wasn't noticed to
     have been broken"

* 'fixes' of git://ftp.arm.linux.org.uk/~rmk/linux-arm:
  ARM: l2c: fix revision checking
  ARM: kprobes: Fix test code compilation errors for ARMv4 targets
  ARM: kprobes: Disallow instructions with PC and register specified shift
  ARM: kprobes: Prevent known test failures stopping other tests running

Pull m68k fixes from Geert Uytterhoeven:
 "Summary:
  - Fix for a boot regression introduced in v3.16-rc1,
  - Fix for a build issue in -next"

Christoph Hellwig questioned why mach_random_get_entropy should be
exported to modules, and Geert explains that random_get_entropy() is
called by at least the crypto layer and ends up using it on m68k.  On
most other architectures it just uses get_cycles() (which is typically
inlined and doesn't need exporting),

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/geert/linux-m68k:
  m68k: Export mach_random_get_entropy to modules
  m68k: Fix boot regression on machines with RAM at non-zero

Pull parisc fixes from Helge Deller:
 "The major patch in here is one which fixes the fanotify_mark() syscall
  in the compat layer of the 64bit parisc kernel.  It went unnoticed so
  long, because the calling syntax when using a 64bit parameter in a
  32bit syscall is quite complex and even worse, it may be even
  different if you call syscall() or the glibc wrapper.  This patch
  makes the kernel accept the calling convention when called by the
  glibc wrapper.

  The other two patches are trivial and remove unused headers, #includes
  and adds the serial ports of the fastest C8000 workstation to the
  parisc-kernel internal hardware database"

* 'parisc-3.16-5' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux:
  parisc: drop unused defines and header includes
  parisc: fix fanotify_mark() syscall on 32bit compat kernel
  parisc: add serial ports of C8000/1GHz machine to hardware database

The uarts only work when the parent is ras_ahb_clk. The stale 3.5
based ST tree does this in the board file.

Add it to the clk init function. Not pretty, but the mess there is
amazing anyway.
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Acked-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NMike Turquette <mturquette@linaro.org>

The control register is at offset 0x10, not 0x0. This is wreckaged
since commit 5df33a62 (SPEAr: Switch to common clock framework).
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: stable@vger.kernel.org
Acked-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NMike Turquette <mturquette@linaro.org>

Signed-off-by: NHelge Deller <deller@gmx.de>
Cc: stable@vger.kernel.org # 3.13+

On parisc we can not use the existing compat implementation for fanotify_mark()
because for the 64bit mask parameter the higher and lower 32bits are ordered
differently than what the compat function expects from big endian
architectures.

Specifically:
It finally turned out, that on hppa we end up with different assignments
of parameters to kernel arguments depending on if we call the glibc
wrapper function
 int fanotify_mark (int __fanotify_fd, unsigned int __flags,
                    uint64_t __mask, int __dfd, const char *__pathname);
or directly calling the syscall manually
 syscall(__NR_fanotify_mark, ...)

Reason is, that the syscall() function is implemented as C-function and
because we now have the sysno as first parameter in front of the other
parameters the compiler will unexpectedly add an empty paramenter in
front of the u64 value to ensure the correct calling alignment for 64bit
values.
This means, on hppa you can't simply use syscall() to call the kernel
fanotify_mark() function directly, but you have to use the glibc
function instead.

This patch fixes the kernel in the hppa-arch specifc coding to adjust
the parameters in a way as if userspace calls the glibc wrapper function
fanotify_mark().
Signed-off-by: NHelge Deller <deller@gmx.de>
Cc: stable@vger.kernel.org # 3.13+

Signed-off-by: NHelge Deller <deller@gmx.de>
Cc: stable@vger.kernel.org # 3.13+

Pull slave-dmaengine fixes from Vinod Koul:
 "We have two small fixes.  First one from Daniel to handle 0-length
  packets for usb cppi dma.  Second by Russell for imx-sdam cyclic
  residue reporting"

* 'fixes' of git://git.infradead.org/users/vkoul/slave-dma:
  Update imx-sdma cyclic handling to report residue
  dma: cppi41: handle 0-length packets

Merge tag 'samsung-fixes-3' of git://git.kernel.org/pub/scm/linux/kernel/git/kgene/linux-samsung into fixes

Merge "Samsung fixes-3 for 3.16" from Kukjin Kim:

Samsung fixes-3 for v3.16
- update the parent for Auudss clock because kernel will be hang
  during late boot if the parent clock is disabled in bootloader.
- enable clk handing in power domain because while power domain
  on/off, its regarding clock source will be reset and it causes
  a problem so need to handle it.
- add mux clocks to be used by power domain for exynos5420-mfc
  during power domain on/off and property in device tree also.
- register cpuidle only for exynos4210 and exynos5250 because a
  system failure will be happened on other exynos SoCs.

* tag 'samsung-fixes-3' of git://git.kernel.org/pub/scm/linux/kernel/git/kgene/linux-samsung:
  ARM: EXYNOS: Register cpuidle device only on exynos4210 and 5250
  ARM: dts: Add clock property for mfc_pd in exynos5420
  clk: exynos5420: Add IDs for clocks used in PD mfc
  ARM: EXYNOS: Add support for clock handling in power domain
  ARM: dts: Update the parent for Audss clocks in Exynos5420
Signed-off-by: NOlof Johansson <olof@lixom.net>

Pull USB fixes from Greg KH:
 "Here are some small USB fixes, PHY driver fixes (they ended up in this
  tree for lack of somewhere else to put them), and some new USB device
  ids"

* tag 'usb-3.16-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb:
  phy: omap-usb2: Balance pm_runtime_enable() on probe failure and remove
  phy: core: Fix error path in phy_create()
  drivers: phy: phy-samsung-usb2.c: Add missing MODULE_DEVICE_TABLE
  phy: omap-usb2: fix devm_ioremap_resource error detection code
  phy: sun4i: depend on RESET_CONTROLLER
  USB: serial: ftdi_sio: Add Infineon Triboard
  USB: ftdi_sio: Add extra PID.
  usb: option: Add ID for Telewell TW-LTE 4G v2
  USB: cp210x: add support for Corsair usb dongle

Pull tty/serial fixes from Greg KH:
 "Here are some small serial fixes that resolve some reported problems
  that started in 3.15 with some serial drivers.

  And there's a new dt binding for a serial driver, which was all that
  was needed for the renesas serial driver"

* tag 'tty-3.16-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty:
  serial: sh-sci: Add device tree support for r8a7{778,740,3a4} and sh73a0
  serial: imx: Fix build breakage
  serial: arc_uart: Use uart_circ_empty() for open-coded comparison
  serial: Test for no tx data on tx restart

Pull char/misc driver fixes from Greg KH:
 "Here are two hyperv driver fixes, and one i8k driver fix for 3.16"

* tag 'char-misc-3.16-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc:
  i8k: Fix non-SMP operation
  Drivers: hv: util: Fix a bug in the KVP code
  Drivers: hv: vmbus: Fix a bug in the channel callback dispatch code

Pull IIO fixes from Greg KH:
 "Here are some IIO driver fixes for 3.16-rc5.  Nothing major, just
  resolves some minor issues that have been reported"

* tag 'staging-3.16-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging:
  iio: ti_am335x_adc: Fix: Use same step id at FIFOs both ends
  iio:tcs3472: Check for buffer enabled and locking
  iio: hid-sensor-prox: Fix return values
  iio: hid-sensor-gyro-3d: Fix return values
  iio: hid-sensor-als: Fix return values
  iio: hid-sensor-magn-3d: Fix return values
  iio: hid-sensor-accel-3d: Fix return values
  iio: hid-sensor-press: Fix return values

Fix potential null pointer dereferencing problem caused by e43bb4e6
("ext4: decrement free clusters/inodes counters when block group declared bad")
Reported-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NNamjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: NAshish Sangwan <a.sangwan@samsung.com>
Signed-off-by: NTheodore Ts'o <tytso@mit.edu>
Reviewed-by: NLukas Czerner <lczerner@redhat.com>

This fixes the following lockdep complaint:

[ INFO: possible circular locking dependency detected ]
3.16.0-rc2-mm1+ #7 Tainted: G           O  
-------------------------------------------------------
kworker/u24:0/4356 is trying to acquire lock:
 (&(&sbi->s_es_lru_lock)->rlock){+.+.-.}, at: [<ffffffff81285fff>] __ext4_es_shrink+0x4f/0x2e0

but task is already holding lock:
 (&ei->i_es_lock){++++-.}, at: [<ffffffff81286961>] ext4_es_insert_extent+0x71/0x180

which lock already depends on the new lock.

 Possible unsafe locking scenario:

       CPU0                    CPU1
       ----                    ----
  lock(&ei->i_es_lock);
                               lock(&(&sbi->s_es_lru_lock)->rlock);
                               lock(&ei->i_es_lock);
  lock(&(&sbi->s_es_lru_lock)->rlock);

 *** DEADLOCK ***

6 locks held by kworker/u24:0/4356:
 #0:  ("writeback"){.+.+.+}, at: [<ffffffff81071d00>] process_one_work+0x180/0x560
 #1:  ((&(&wb->dwork)->work)){+.+.+.}, at: [<ffffffff81071d00>] process_one_work+0x180/0x560
 #2:  (&type->s_umount_key#22){++++++}, at: [<ffffffff811a9c74>] grab_super_passive+0x44/0x90
 #3:  (jbd2_handle){+.+...}, at: [<ffffffff812979f9>] start_this_handle+0x189/0x5f0
 #4:  (&ei->i_data_sem){++++..}, at: [<ffffffff81247062>] ext4_map_blocks+0x132/0x550
 #5:  (&ei->i_es_lock){++++-.}, at: [<ffffffff81286961>] ext4_es_insert_extent+0x71/0x180

stack backtrace:
CPU: 0 PID: 4356 Comm: kworker/u24:0 Tainted: G           O   3.16.0-rc2-mm1+ #7
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
Workqueue: writeback bdi_writeback_workfn (flush-253:0)
 ffffffff8213dce0 ffff880014b07538 ffffffff815df0bb 0000000000000007
 ffffffff8213e040 ffff880014b07588 ffffffff815db3dd ffff880014b07568
 ffff880014b07610 ffff88003b868930 ffff88003b868908 ffff88003b868930
Call Trace:
 [<ffffffff815df0bb>] dump_stack+0x4e/0x68
 [<ffffffff815db3dd>] print_circular_bug+0x1fb/0x20c
 [<ffffffff810a7a3e>] __lock_acquire+0x163e/0x1d00
 [<ffffffff815e89dc>] ? retint_restore_args+0xe/0xe
 [<ffffffff815ddc7b>] ? __slab_alloc+0x4a8/0x4ce
 [<ffffffff81285fff>] ? __ext4_es_shrink+0x4f/0x2e0
 [<ffffffff810a8707>] lock_acquire+0x87/0x120
 [<ffffffff81285fff>] ? __ext4_es_shrink+0x4f/0x2e0
 [<ffffffff8128592d>] ? ext4_es_free_extent+0x5d/0x70
 [<ffffffff815e6f09>] _raw_spin_lock+0x39/0x50
 [<ffffffff81285fff>] ? __ext4_es_shrink+0x4f/0x2e0
 [<ffffffff8119760b>] ? kmem_cache_alloc+0x18b/0x1a0
 [<ffffffff81285fff>] __ext4_es_shrink+0x4f/0x2e0
 [<ffffffff812869b8>] ext4_es_insert_extent+0xc8/0x180
 [<ffffffff812470f4>] ext4_map_blocks+0x1c4/0x550
 [<ffffffff8124c4c4>] ext4_writepages+0x6d4/0xd00
	...
Reported-by: NMinchan Kim <minchan@kernel.org>
Signed-off-by: NTheodore Ts'o <tytso@mit.edu>
Reported-by: NMinchan Kim <minchan@kernel.org>
Cc: stable@vger.kernel.org
Cc: Zheng Liu <gnehzuil.liu@gmail.com>