This patch fixes "Missing a blank line after declarations" warnings.
Signed-off-by: NSeunghun Lee <waydi1@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The white space was all wrong here.  The case statements were indented
too far.  The if else blocks weren't indented at all.  There was a break
statement aligned with the else block and it confused my static checker
because "were curly braces intended" so that the break statement was
only on the else side?  Also I removed some commented out code.
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

There were a couple lines which were not indented far enough and it was
confusing.
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The for loop body wasn't indented so it upset my static checker.  Also
I removed an obsolete comment on the same line.
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

"*(p + 1)" and "len" are the same thing.  For reviewers who don't know
that, then this code is worrying because we cap "len", but pass
"*(p + 1)" to memcpy().

I have changed the code to use "len" throughout.
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Fold a line to make it less than 80 characters
Signed-off-by: NSarath Lakshman <sarathlakshman@slynux.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Change old way of ops->setsockopt or ops->getsockopt in kernel
to kernel_setsockopt or kernel_getsockopt.
Signed-off-by: NFredrick John Berchmans <fredrickprashanth@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

If CONFIG_FS_POSIX_ACL=n:

drivers/staging/lustre/lustre/llite/xattr.c: In function 'll_setxattr_common':
drivers/staging/lustre/lustre/llite/xattr.c:115:27: warning: unused variable 'rce' [-Wunused-variable]
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

drivers/staging/lustre/lustre/llite/rw26.c: In function 'll_direct_IO_26':
drivers/staging/lustre/lustre/llite/rw26.c:383:2: warning: format '%lu' expects argument of type 'long unsigned int', but argument 6 has type 'ssize_t' [-Wformat]
drivers/staging/lustre/lustre/llite/rw26.c:383:2: warning: format '%lu' expects argument of type 'long unsigned int', but argument 10 has type 'ssize_t' [-Wformat]

Join the quoted string split across lines to fix a checkpatch warning while
we're at it.
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

When CONFIG_SMP=n:

drivers/staging/lustre/include/linux/libcfs/linux/linux-mem.h:58:31: fatal error: libcfs/libcfs_cpu.h: No such file or directory

drivers/staging/lustre/lustre/libcfs/libcfs_cpu.c:78:1: error: redefinition of 'cfs_cpt_table_print'
drivers/staging/lustre/include/linux/libcfs/libcfs_cpu.h:109:1: note: previous definition of 'cfs_cpt_table_print' was here
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The obd_ioctl_getdata() function caps "data->ioc_len" at
OBD_MAX_IOCTL_BUFFER and then calls this obd_ioctl_is_invalid() to check
that the other values inside data are valid.

There are several lengths inside data but when they are added together
they must not be larger than "data->ioc_len".  The checks against
"(data->ioc_inllen1 > (1<<30))" are supposed to ensure that the addition
does not have an integer overflow.  But "(1<<30) * 4" actually can
overflow 32 bits, so the checks are insufficient.

I have changed it to "> OBD_MAX_IOCTL_BUFFER" instead.
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Pointer 'ni' checked for NULL at line 1569 may be passed to
function and may be dereferenced there by passing argument 1 to
function 'lnet_ni_notify_locked' at line 1621.
found by Klocwork Insight tool
Signed-off-by: NOleg Drokin <oleg.drokin@intel.com>
CC: Dmitry Eremin <dmitry.eremin@intel.com>
CC: Liang Zhen <liang.zhen@intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Null pointer 'cp' that comes from line 2544 may be dereferenced
at line 2618.
found by Klocwork Insight tool
Signed-off-by: NDmitry Eremin <dmitry.eremin@intel.com>
Reviewed-on: http://review.whamcloud.com/9386
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4629Reviewed-by: NJohn L. Hammond <john.hammond@intel.com>
Reviewed-by: NIsaac Huang <he.huang@intel.com>
Signed-off-by: NOleg Drokin <oleg.drokin@intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Local variable 'hash' is never used
found by Klocwork Insight tool
Signed-off-by: NDmitry Eremin <dmitry.eremin@intel.com>
Reviewed-on: http://review.whamcloud.com/9386
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4629Reviewed-by: NJohn L. Hammond <john.hammond@intel.com>
Reviewed-by: NIsaac Huang <he.huang@intel.com>
Signed-off-by: NOleg Drokin <oleg.drokin@intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

It should never be NULL because our interface list is up to date,
and even if it does, we'll just crash anyway so we are no better off.
Signed-off-by: NOleg Drokin <oleg.drokin@intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Insted of meddling directly in process environment variables
(which is also not possible on certain platforms due to not exported
symbols), create jobid_name proc file to represent this info
(to be filled by job scheduler epilogue).
Signed-off-by: NOleg Drokin <oleg.drokin@intel.com>
CC: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

In ll_ioctl_fiemap(), a user-supplied value is used to calculate a
length of a buffer which is later allocated with user data.
Signed-off-by: NVitaly Osipov <vitaly.osipov@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The asm-generic, big-endian version of zero_bytemask creates a mask of
bytes preceding the first zero-byte by left shifting ~0ul based on the
position of the first zero byte.

Unfortunately, if the first (top) byte is zero, the output of
prep_zero_mask has only the top bit set, resulting in undefined C
behaviour as we shift left by an amount equal to the width of the type.
As it happens, GCC doesn't manage to spot this through the call to fls(),
but the issue remains if architectures choose to implement their shift
instructions differently.

An example would be arch/arm/ (AArch32), where LSL Rd, Rn, #32 results
in Rd == 0x0, whilst on arch/arm64 (AArch64) LSL Xd, Xn, #64 results in
Xd == Xn.

Rather than check explicitly for the problematic shift, this patch adds
an extra shift by 1, replacing fls with __fls. Since zero_bytemask is
never called with a zero argument (has_zero() is used to check the data
first), we don't need to worry about calling __fls(0), which is
undefined.

Cc: <stable@vger.kernel.org>
Cc: Victor Kamensky <victor.kamensky@linaro.org>
Signed-off-by: NWill Deacon <will.deacon@arm.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

This merges the patch to fix possible loss of dirty bit on munmap() or
madvice(DONTNEED).  If there are concurrent writers on other CPU's that
have the unmapped/unneeded page in their TLBs, their writes to the page
could possibly get lost if a third CPU raced with the TLB flush and did
a page_mkclean() before the page was fully written.

Admittedly, if you unmap() or madvice(DONTNEED) an area _while_ another
thread is still busy writing to it, you deserve all the lost writes you
could get.  But we kernel people hold ourselves to higher quality
standards than "crazy people deserve to lose", because, well, we've seen
people do all kinds of crazy things.

So let's get it right, just because we can, and we don't have to worry
about it.

* safe-dirty-tlb-flush:
  mm: split 'tlb_flush_mmu()' into tlb flushing and memory freeing parts

Pull btrfs fixes from Chris Mason.

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs:
  Btrfs: limit the path size in send to PATH_MAX
  Btrfs: correctly set profile flags on seqlock retry
  Btrfs: use correct key when repeating search for extent item
  Btrfs: fix inode caching vs tree log
  Btrfs: fix possible memory leaks in open_ctree()
  Btrfs: avoid triggering bug_on() when we fail to start inode caching task
  Btrfs: move btrfs_{set,clear}_and_info() to ctree.h
  btrfs: replace error code from btrfs_drop_extents
  btrfs: Change the hole range to a more accurate value.
  btrfs: fix use-after-free in mount_subvol()

Pull arm fixes from Russell King:
 "A number of fixes for the PJ4/iwmmxt changes which arm-soc forced me
  to take during the merge window.  This stuff should have been better
  tested and sorted out *before* the merge window"

* 'fixes' of git://ftp.arm.linux.org.uk/~rmk/linux-arm:
  ARM: 8042/1: iwmmxt: allow to build iWMMXt on Marvell PJ4B
  ARM: 8041/1: pj4: fix cpu_is_pj4 check
  ARM: 8040/1: pj4: properly detect existence of iWMMXt coprocessor
  ARM: 8039/1: pj4: enable iWMMXt only if CONFIG_IWMMXT is set
  ARM: 8038/1: iwmmxt: explicitly check for supported architectures

Pull arm64 fixes from Catalin Marinas:
 - compat renameat2 syscall wiring and __NR_compat_syscalls fix
 - TLB fix for transparent huge pages following switch to generic
   mmu_gather
 - spinlock initialisation for init_mm's context
 - move of_clk_init() earlier
 - Kconfig duplicate entry fix

* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
  arm64: init: Move of_clk_init to time_init
  arm64: initialize spinlock for init_mm's context
  arm64: debug: remove noisy, pointless warning
  arm64: mm: Add THP TLB entries to general mmu_gather
  arm64: add renameat2 compat syscall
  ARM64: Remove duplicated Kconfig entry for "kernel/power/Kconfig"
  arm64: __NR_compat_syscalls fix

Quiet the warning below in Lustre code.
Actually the warning is invalid since we either always assign
the symname in ll_readlink_internal or return an error there and
then the following rc check would assign symlink variable explicitly.

In file included from /home/green/bk/linux/drivers/staging/lustre/lustre/llite/../include/linux/lustre_compat25.h:41:0,
                 from /home/green/bk/linux/drivers/staging/lustre/lustre/llite/../include/linux/lvfs.h:48,
                 from /home/green/bk/linux/drivers/staging/lustre/lustre/llite/../include/lvfs.h:45,
                 from /home/green/bk/linux/drivers/staging/lustre/lustre/llite/../include/obd_support.h:41,
                 from /home/green/bk/linux/drivers/staging/lustre/lustre/llite/../include/obd_class.h:40,
                 from /home/green/bk/linux/drivers/staging/lustre/lustre/llite/../include/linux/lustre_lite.h:49,
                 from /home/green/bk/linux/drivers/staging/lustre/lustre/llite/../include/lustre_lite.h:45,
                 from /home/green/bk/linux/drivers/staging/lustre/lustre/llite/symlink.c:42:
/home/green/bk/linux/drivers/staging/lustre/lustre/llite/symlink.c: In function ‘ll_follow_link’:
/home/green/bk/linux/include/linux/namei.h:88:29: warning: ‘symname’ may be used uninitialized in this function [-Wmaybe-uninitialized]
  nd->saved_names[nd->depth] = path;
                             ^
/home/green/bk/linux/drivers/staging/lustre/lustre/llite/symlink.c:123:8: note: ‘symname’ was declared here
  char *symname;
        ^
Signed-off-by: NOleg Drokin <oleg.drokin@intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Pull irq fixes from Thomas Gleixner:
 "A slighlty large fix for a subtle issue in the CPU hotplug code of
  certain ARM SoCs, where the not yet online cpu needs to setup the cpu
  local timer and needs to set the interrupt affinity to itself.
  Setting interrupt affinity to a not online cpu is prohibited and
  therefor the timer interrupt ends up on the wrong cpu, which leads to
  nasty complications.

  The SoC folks tried to hack around that in the SoC code in some more
  than nasty ways.  The proper solution is to have a way to enforce the
  affinity setting to a not online cpu.  The core patch to the genirq
  code provides that facility and the follow up patches make use of it
  in the GIC interrupt controller and the exynos timer driver.

  The change to the core code has no implications to existing users,
  except for the rename of the locked function and therefor the
  necessary fixup in mips/cavium.  Aside of that, no runtime impact is
  possible, as none of the existing interrupt chips implements anything
  which depends on the force argument of the irq_set_affinity()
  callback"

* 'irq-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  clocksource: Exynos_mct: Register clock event after request_irq()
  clocksource: Exynos_mct: Use irq_force_affinity() in cpu bringup
  irqchip: Gic: Support forced affinity setting
  genirq: Allow forcing cpu affinity of interrupts

Pull tty/serial fixes from Greg KH:
 "Here are a few tty/serial fixes for 3.15-rc3 that resolve a number of
  reported issues in the 8250 and samsung serial drivers, as well as a
  character loss fix for the tty core that was caused by the lock
  removal patches a release ago"

* tag 'tty-3.15-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty:
  serial_core: fix uart PORT_UNKNOWN handling
  serial: samsung: Change barrier() to cpu_relax() in console output
  serial: samsung: don't check config for every character
  serial: samsung: Use the passed in "port", fixing kgdb w/ no console
  serial: 8250: Fix thread unsafe __dma_tx_complete function
  8250_core: Fix unwanted TX chars write
  tty: Fix race condition between __tty_buffer_request_room and flush_to_ldisc

Pull staging / IIO driver fixes from Greg KH:
 "Here are some small staging and IIO driver fixes for 3.15-rc3.

  Nothing major at all, just some assorted issues that people have
  reported"

* tag 'staging-3.15-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging:
  staging: comedi: usbdux: bug fix for accessing 'ao_chanlist' in private data
  iio: adc: mxs-lradc: fix warning when buidling on avr32
  iio: cm36651: Fix i2c client leak and possible NULL pointer dereference
  iio: querying buffer scan_mask should return 0/1
  staging:iio:ad2s1200 fix a missing break
  iio: adc: at91_adc: correct default shtim value
  ARM: at91: at91sam9260: change at91_adc name
  ARM: at91: at91sam9g45: change at91_adc name
  iio: cm32181: Fix read integration time function
  iio: adc: at91_adc: Repair broken platform_data support

lov_fiemap() does not take consider its @vallen parameter, which is
the max buffer size the caller can hold for the fiemap extents.

This patch fixes this and limits the max mapped fiemap extent count
to fit in the preallocted buffer.

This patch also fixes a memory out of bound write issue when the
fiemap call is only for detecting the number of existing extent.
Signed-off-by: NBobi Jam <bobijam.xu@intel.com>
Reviewed-on: http://review.whamcloud.com/9834
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4619Reviewed-by: NFan Yong <fan.yong@intel.com>
Reviewed-by: NPatrick Farrell <paf@cray.com>
Signed-off-by: NOleg Drokin <oleg.drokin@intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Apparently we are pretty bad about verifying our buffers passed
from userspace.
Signed-off-by: NOleg Drokin <oleg.drokin@intel.com>
Reviewed-on: http://review.whamcloud.com/9059
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4563Reviewed-by: NDmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: NJames Simmons <uja.ornl@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The Linux VFS and Lustre OST_SYNC RPC are both capable of specifying
fsync() on a sub-extent of the file {start, end} instead of the full
file.  This allows less than the full amount of data to be flushed,
reducing or possibly eliminating the work needed before the syscall
can return.

However, the handling of sub-extent of the file for fsync was lost
with the move to CLIO on the client and OSD API on the server.  They
were ignoring the passed {start, end} and using {0, OBD_OBJECT_EOF}
instead.

Return the ability to pass a sub-extent for fsync() from the client,
to the specific stripes/OSTs that need the sync operation, and pass
it down to the OSD.  The ZFS OSD doesn't handle this yet, but there
is room for improvement in a separate patch.
Signed-off-by: NAndreas Dilger <andreas.dilger@intel.com>
Reviewed-on: http://review.whamcloud.com/8626
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4388Reviewed-by: NBobi Jam <bobijam@gmail.com>
Reviewed-by: NJinshan Xiong <jinshan.xiong@intel.com>
Signed-off-by: NOleg Drokin <oleg.drokin@intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

In libcfs_debug_vmsg2, cdls_delay is only clamped between the minimum
and the maximum when it is increased by multiplying by the backoff
factor.  It is not clamped when it is decreased by dividing by the
backoff factor.  This allows it to achieve values less than the
minimum, which allows a console message to be printed that should have
been skipped.  This patch moves the clamping outside of the else
statement, ensuring that cdls_delay is always between the min and the
max after the first time through libcfs_debug_vmsg2.
Signed-off-by: NRyan Haasken <haasken@cray.com>
Reviewed-on: http://review.whamcloud.com/9503
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4711Reviewed-by: NChris Horn <hornc@cray.com>
Reviewed-by: NAnn Koehler <amk@cray.com>
Reviewed-by: NAndreas Dilger <andreas.dilger@intel.com>
Signed-off-by: NOleg Drokin <oleg.drokin@intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Removed the checks for oi_lockless from osc_io_read_start() and
osc_io_write_start(). This patch also removes the unnecessary call to
cl_object_attr_get() in osc_io_write_start() before calling
cl_object_attr_set()
Signed-off-by: NSwapnil Pimpale <spimpale@ddn.com>
Reviewed-on: http://review.whamcloud.com/8797
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3868Reviewed-by: NJohn L. Hammond <john.hammond@intel.com>
Reviewed-by: NAndreas Dilger <andreas.dilger@intel.com>
Signed-off-by: NOleg Drokin <oleg.drokin@intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

spin_is_locked() is always false when the platform is
uniprocessor and CONFIG_DEBUG_SPINLOCK is not enabled.
This patch replaces its assertion by assert_spin_locked().
Signed-off-by: NLi Xi <lixi@ddn.com>
Signed-off-by: NJames Simmons <uja.ornl@gmail.com>
Reviewed-on: http://review.whamcloud.com/8144
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4199Reviewed-by: NAlexey Lyashkov <alexey_lyashkov@xyratex.com>
Reviewed-by: NAndreas Dilger <andreas.dilger@intel.com>
Signed-off-by: NOleg Drokin <oleg.drokin@intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

In llite remove unused declarations, parameters, types, and unused,
get-only, or set-only structure members. Add static and const
qualifiers to declarations where possible.
Signed-off-by: NJohn L. Hammond <john.hammond@intel.com>
Reviewed-on: http://review.whamcloud.com/9767
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675Reviewed-by: NLai Siyao <lai.siyao@intel.com>
Reviewed-by: NJinshan Xiong <jinshan.xiong@intel.com>
Signed-off-by: NOleg Drokin <oleg.drokin@intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

In some cases, cl_default_mds_easize might be zero, especially for
MDC connected to non-MDT0, then mdc might pack getattr RPC with
zero eadatasize.

If client is trying to access remote striped directory with
zero eadatasize, MDT will not return layout information of the
striped direcotry, which will be mis-regarded as non-striped
directory.

So we should use cl_max_mds_easize if cl_default_mds_easize is zero.
Signed-off-by: Nwang di <di.wang@intel.com>
Reviewed-on: http://review.whamcloud.com/9862
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4847Reviewed-by: NAndreas Dilger <andreas.dilger@intel.com>
Signed-off-by: NOleg Drokin <oleg.drokin@intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

It's not atomic to check the last reference and state of cl_lock
in cl_lock_put(). This can cause a problem that an using lock is
freed, if the process is preempted between atomic_dec_and_test()
and (lock->cll_state == CLS_FREEING).

This problem can be solved by holding a refcount by coh_locks. In
this case, it can be sure that if the lock refcount reaches zero,
nobody else can have any chance to use it again.
Signed-off-by: NJinshan Xiong <jinshan.xiong@intel.com>
Reviewed-on: http://review.whamcloud.com/9881
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4558Reviewed-by: NBobi Jam <bobijam@gmail.com>
Reviewed-by: NLai Siyao <lai.siyao@intel.com>
Signed-off-by: NOleg Drokin <oleg.drokin@intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Change the LL_IOC_LLOOP_INFO ioctl in the lustre lloop
device driver to return an error instead of causing
panics with LASSERT().
Signed-off-by: NBob Glossman <bob.glossman@intel.com>
Reviewed-on: http://review.whamcloud.com/9888
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4863Reviewed-by: NNathaniel Clark <nathaniel.l.clark@intel.com>
Reviewed-by: NAndreas Dilger <andreas.dilger@intel.com>
Signed-off-by: NOleg Drokin <oleg.drokin@intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Messages which are printed by ll_dirty_page_discard_warn() should not
be rate limited.  If they are rate limited, some files which may be
corrupted on client eviction will not be reported to the user.

This patch changes the CWARN to a CDEBUG to disable console message
rate limiting for this message.  The dirty page discard warnings are
already limited on a per-file basis by the function vvp_vmpage_error
which calls ll_dirty_page_discard_warn only if the ccc_object's
cob_discard_page_warned == 0.
Signed-off-by: NRyan Haasken <haasken@cray.com>
Reviewed-on: http://review.whamcloud.com/9752
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4799Reviewed-by: NCory Spitz <spitzcor@cray.com>
Reviewed-by: NAnn Koehler <amk@cray.com>
Reviewed-by: NChris Horn <hornc@cray.com>
Signed-off-by: NOleg Drokin <oleg.drokin@intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

According https://www.kernel.org/doc/Documentation/mutex-design.txt:
- the mutex subsystem is slightly faster and has better scalability
  for contended workloads. In terms of 'ops per CPU cycle', the
  semaphore kernel performed 551 ops/sec per 1% of CPU time used,
  while the mutex kernel performed 3825 ops/sec per 1% of CPU time
  used - it was 6.9 times more efficient.
- there are no fastpath tradeoffs, the mutex fastpath is just as
  tight as the semaphore fastpath. On x86, the locking fastpath is
  2 instructions.
- 'struct mutex' semantics are well-defined and are enforced if
  CONFIG_DEBUG_MUTEXES is turned on. Semaphores on the other hand
  have virtually no debugging code or instrumentation.

One more benefit of mutex is optimistic spinning. It try to spin for
acquisition when there are no pending waiters and the lock owner is
currently running on a (different) CPU. The rationale is that if the
lock owner is running, it is likely to release the lock soon.

This significantly reduce amount of context switches when locked
region is small and we have high contention.
Signed-off-by: NDmitry Eremin <dmitry.eremin@intel.com>
Reviewed-on: http://review.whamcloud.com/9095
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4257Reviewed-by: NAndreas Dilger <andreas.dilger@intel.com>
Reviewed-by: NJames Simmons <uja.ornl@gmail.com>
Signed-off-by: NOleg Drokin <oleg.drokin@intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>