iio: hid-sensors: Add API to power on/off

Added an API to allow client drivers to turn ON and OFF sensors for
quick read. Added data_read as counting varaible instead of boolean,
so that sensor is powered off only when last user released it.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: hid-sensors: Pressure: adjust scale and offset

Using units and unit exponent to calculate scale which is compliant
to IIO ABI.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: hid-sensors: Inclinometer 3D: adjust scale and offset

Using units and unit exponent to calculate scale which is compliant
to IIO ABI.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: hid-sensors: Compass 3D: adjust scale and offset

Using units and unit exponent to calculate scale which is compliant
to IIO ABI.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: hid-sensors: ALS: adjust scale and offset

Using units and unit exponent to calculate scale which is compliant
to IIO ABI.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: hid-sensors: Gyro 3D : adjust scale and offset

Using units and unit exponent to calculate scale which is compliant
to IIO ABI.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: hid-sensors: Accelerometer 3D: adjust scale and offset

Using units and unit exponent to calculate scale which is compliant
to IIO ABI.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: hid-sensors: Add api to get poll value

Added interface to get poll value in milli-seconds. This value is
changed by changing sampling frequency. This API allows clients
to wait for at least some poll milli seconds before reading a new sample.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: hid-sensors: Convert units and exponent

HID sensor hub specify a default unit and alternative units. This
along with unit exponent can be used adjust scale. This change
change HID sensor data units to IIO defined units for each
sensor type. So in this way user space can use a simply use:
"(data + offset) * scale" to get final result.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: Add Freescale MMA8452Q 3-axis accelerometer driver

3-axis accelerometer sensor (2/4/8 g) with 12-bit resolution
and I2C interface

many extra features are unsupported (freefall detection, orientation
change, autosleep)

datasheet is here:
http://cache.freescale.com/files/sensors/doc/data_sheet/MMA8452Q.pdf

v2: (thanks to Jonathan Cameron)
* use ARRAY_SIZE()

Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: ak8975: Fix issue with the name

Fixing issue caused by clash of two patches, one adding a name with
the acpi enumeration patch.

54ab3e24 Beomho Seo          2014-04-02 @572   indio_dev->name = id->name;
d913971e Srinivas Pandruvada 2014-03-19  574   indio_dev->name = name;

The name added by commit 54ab3e24 is not required as this is already
added by taking care of case where id is NULL.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: Fix spelling of illuminance in gp2ap020a00f driver

Cc: Jacek Anaszewski <j.anaszewski@samsung.com>
Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: Add ABI documentation for illuminance

has been in use for a long time, but never documented;
state that this measurement should be in lux (drivers may feel
different about this)

Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: Add intensity including ir and both modifiers to ABI documentation ex-post

intensity has been in use for a long time but never documented
it is beneficial to document what this is supposed to be

Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: Add Melexis mlx90614 contact-less infrared temperature sensor driver

I2C-controlled sensor measures ambient and object temperatuer

see
http://www.melexis.com/Infrared-Thermometer-Sensors/Infrared-Thermometer-Sensors/MLX90614-615.aspx

Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: Add ABI documentation for object and ambient modifiers

Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: Add TEMP_AMBIENT and TEMP_OBJECT channel modifiers

useful for contactless temperature sensors to distinguish
between the ambient temperature and the temperature of the object

Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: Add Freescale MPL115A2 pressure / temperature sensor driver

I2C-controlled sensor with 10-bit pressure and temperature measurement

datasheet: http://cache.freescale.com/files/sensors/doc/data_sheet/MPL3115A2.pdf

v2:
* use devm_iio_device_register()

Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio:magnetometer:mag3110: Enable user offset calibration

add IIO_CHAN_INFO_CALIBBIAS

Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio:magnetometer:mag3110: Fix initialization of ctrl_reg1

DR bits need to be shifted; since MAG3110_CTRL_DR_DEFAULT is
zero, the change has no effect

Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

staging: iio: ad799x: remove some unneeded IS_ERR() checks

My static checker is upset that we check IS_ERR(t->reg) when we know it
is not an ERR_PTR.

Checking for IS_ERR() twice is often a sign of confusion and buggy code.
In this case, if the call to "ret = regulator_enable(st->vref);" fails,
then we call "regulator_disable(st->vref);" and that's a mistake because
"st->vref" is not enabled.

I fixed these problems and Hartmut Knaack pointed out a couple unneeded
IS_ERR() checks in ad799x_remove() so I have removed those as well.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Hartmut Knaack <knaack.h@gmx.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

staging: iio: Fix format string mismatch in ad7280a.c

Fix two format string mismatch in ad7280a.c

Signed-off-by: Masanari Iida <standby24x7@gmail.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: exynos_adc: do a reinit_completion before the conversion

Add reinit_completion() before the wait_for_completion_timeout in
raw_read() call.

Signed-off-by: Naveen Krishna Chatradhi <ch.naveen@samsung.com>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: exynos_adc: do a soft reset in case of timeout

Do a soft reset software if a timeout happens.
This is applicable only for ADC_V2.

Signed-off-by: Naveen Krishna Chatradhi <ch.naveen@samsung.com>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: exynos_adc: reduce timeout and use wait_for_completion_timeout

ADC module on Exynos5 SoCs runs at 600KSPS. At this conversion rate,
waiting for 1000 msecs is wasteful (incase of h/w failure).

Hence, reduce the time out to 100msecs and use
wait_for_completion_timeout() instead of
wait_for_completion_interruptible_timeout()

Signed-off-by: Naveen Krishna Chatradhi <ch.naveen@samsung.com>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: exynos_adc: rearrange clk and regulator enable/disable calls

This patch maintains the following order in
probe(), remove(), resume() and suspend() calls

regulator enable, clk prepare enable
...
clk disable unprepare, regulator disable

While at it,
1. enable the regulator before the iio_device_register()
2. handle the return values for enable/disable calls

Signed-off-by: Naveen Krishna Ch <ch.naveen@samsung.com>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: Added ABI description for quaternion

Added documentation for reading quaternion components for 3D rotations.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: hid-sensors: Added device rotation support

Added usage id processing for device rotation. This uses IIO
interfaces for triggered buffer to present data to user
mode.This uses HID sensor framework for registering callback
events from the sensor hub.
Data is exported to user space in the form of quaternion rotation
format.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

IIO: core: Add quaternion modifier

Added quaternion in the list of supported modifiers.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

IIO: core: Modify scan element type

The current scan element type uses the following format:
  [be|le]:[s|u]bits/storagebits[>>shift].
To specify multiple elements in this type, added a repeat value.
So new format is:
  [be|le]:[s|u]bits/storagebitsXr[>>shift].
Here r is specifying how may times, real/storage bits are repeating.

When X is value is 0 or 1, then repeat value is not used in the format,
and it will be same as existing format.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

IIO: core: Introduce read_raw_multi

This callback is introduced to overcome some limitations of existing
read_raw callback. The functionality of both existing read_raw and
read_raw_multi is similar, both are used to request values from the
device. The current read_raw callback allows only two return values.
The new read_raw_multi allows returning multiple values. Instead of
passing just address of val and val2, it passes length and pointer
to values. Depending on the type and length of passed buffer, iio
client drivers can return multiple values.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

devres: introduce API "devm_kmemdup

Introduce devm_kmemdup, which uses resource managed kmalloc.
There are several request from maintainers to add this instead
of using kmemdup.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

Merge 3.15-rc3 into staging-next

Linux 3.15-rc3

word-at-a-time: avoid undefined behaviour in zero_bytemask macro

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: Will Deacon <will.deacon@arm.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Merge branch 'safe-dirty-tlb-flush'

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

Merge branch 'for-linus' of git://git./linux/kernel/git/mason/linux-btrfs

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()

Merge branch 'fixes' of git://ftp.arm.linux.org.uk/~rmk/linux-arm

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

Merge tag 'arm64-fixes' of git://git./linux/kernel/git/arm64/linux

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

staging/lustre/llite: Fix a compile warning.

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: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Merge branch 'irq-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

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

Merge tag 'tty-3.15-rc3' of git://git./linux/kernel/git/gregkh/tty

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

Merge tag 'staging-3.15-rc3' of git://git./linux/kernel/git/gregkh/staging

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

staging/lustre/llite: prevent buffer overflow in fiemap

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: Bobi Jam <bobijam.xu@intel.com>
Reviewed-on: http://review.whamcloud.com/9834
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4619
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Patrick Farrell <paf@cray.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre: Fix unsafe userspace access in many proc files

Apparently we are pretty bad about verifying our buffers passed
from userspace.

Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Reviewed-on: http://review.whamcloud.com/9059
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4563
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre: pass fsync() range through RPC/IO stack

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: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-on: http://review.whamcloud.com/8626
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4388
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre: Always clamp cdls_delay between min and max

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: Ryan Haasken <haasken@cray.com>
Reviewed-on: http://review.whamcloud.com/9503
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4711
Reviewed-by: Chris Horn <hornc@cray.com>
Reviewed-by: Ann Koehler <amk@cray.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/osc: Update inode timestamp for lockless IO as well

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: Swapnil Pimpale <spimpale@ddn.com>
Reviewed-on: http://review.whamcloud.com/8797
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3868
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre: remove assertion of spin_is_locked()

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: Li Xi <lixi@ddn.com>
Signed-off-by: James Simmons <uja.ornl@gmail.com>
Reviewed-on: http://review.whamcloud.com/8144
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4199
Reviewed-by: Alexey Lyashkov <alexey_lyashkov@xyratex.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/llite: remove dead code

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: John L. Hammond <john.hammond@intel.com>
Reviewed-on: http://review.whamcloud.com/9767
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/mdc: use cl_max_mds_md to pack getattr RPC

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: wang di <di.wang@intel.com>
Reviewed-on: http://review.whamcloud.com/9862
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4847
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/clio: Solve a race in cl_lock_put

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: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-on: http://review.whamcloud.com/9881
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4558
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/lloop: avoid panic during blockdev_info

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: Bob Glossman <bob.glossman@intel.com>
Reviewed-on: http://review.whamcloud.com/9888
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4863
Reviewed-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/llite: Do not rate limit dirty page discard warning

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: Ryan Haasken <haasken@cray.com>
Reviewed-on: http://review.whamcloud.com/9752
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4799
Reviewed-by: Cory Spitz <spitzcor@cray.com>
Reviewed-by: Ann Koehler <amk@cray.com>
Reviewed-by: Chris Horn <hornc@cray.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/clio: replace semaphore with mutex

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: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-on: http://review.whamcloud.com/9095
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4257
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre: replace semaphores with mutexes

It's just optimization. The mutex subsystem is slightly faster
and has better scalability for contended workloads.

Remove the lustre_lock and it's accessor functions l_lock(),
l_unlock(), l_lock_init(), and l_has_lock() since they have
not been used by the code since Lustre 1.6.

Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-on: http://review.whamcloud.com/9294
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4588
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/libcfs: add CPU table functions for uniprocessor

Some CPU table functions for uniprocessor architecture is
missing.

Signed-off-by: Li Xi <lixi@ddn.com>
Signed-off-by: James Simmons <uja.ornl@gmail.com>
Reviewed-on: http://review.whamcloud.com/8873
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4199
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Liang Zhen <liang.zhen@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Merge tag 'driver-core-3.15-rc3' of git://git./linux/kernel/git/gregkh/driver-core

Pull driver core fixes from Greg KH:
 "Here are some kernfs fixes for 3.15-rc3 that resolve some reported
  problems.  Nothing huge, but all needed"

* tag 'driver-core-3.15-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core:
  s390/ccwgroup: Fix memory corruption
  kernfs: add back missing error check in kernfs_fop_mmap()
  kernfs: fix a subdir count leak

staging/lustre: remove lustre/include/ioctl.h

Remove the nowhere included header lustre/include/ioctl.h.

Signed-off-by: John L. Hammond <john.hammond@intel.com>
Reviewed-on: http://review.whamcloud.com/9757
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/lnet: NI shutdown may loop forever

lnet_shutdown_lndnis() may enter endless loop if there is a busy NI,
this is injected by LNet SMP improvements. It's fixed in this patch.

Signed-off-by: Liang Zhen <liang.zhen@intel.com>
Reviewed-on: http://review.whamcloud.com/9706
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4780
Reviewed-by: Isaac Huang <he.huang@intel.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/libcfs: fix issues found by Klocwork Insight tool

sscanf format specification '%u' expects type 'unsigned int *'
for 'u', but parameter 3 has a different type 'int*'.

Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-on: http://review.whamcloud.com/9400
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4629
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/mdc: fix issue found by Klocwork Insight tool

Pointer 'mod' checked for NULL at line 160 may be dereferenced at line 208.

Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-on: http://review.whamcloud.com/9387
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4629
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/ldlm: fix NULL pointer dereference

Pointer '*exp' returned from call to function 'class_conn2export'
at line 523 may be NULL and may be dereferenced at line 543.

Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-on: http://review.whamcloud.com/9323
ntel-bug-id: https://jira.hpdd.intel.com/browse/LU-4629
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre: shrink lu_object_header by 8 bytes on x86_64

Locate the loh_flags and loh_ref fields together in lu_object_header
to avoid holes and shrink the structure by 8 bytes.

Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-on: http://review.whamcloud.com/9185
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3059
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: Liang Zhen <liang.zhen@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/llite: access layout version under a lock

We used to access layout version under the protection of ldlm
lock, this introduces extra overhead for dlm lock matching.

In this patch, lli_layout_lock is introduced to access the layout
version. Also, when a layout lock is losing, we should tear down
mmap of the correspoding inode to avoid stale data accessing in the
future.

This is part of technical verification of replication.

Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-on: http://review.whamcloud.com/8689
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3254
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

stagaing/lustre: Improve statahead debug messages

The statahead debug messages include the pid of the current
process in their body.  This is both redudant (because all
lustre log messages contain the pid), and sometimes downright
misleading.  For instance the messages would say something like
"stopping statahead thread 3446".  One would probably think
that 3446 is the pid of the process that is being stopped,
but in fact it was the pid of the caller issuing the stop signal.

We remove all superfluous pids from the messages.

Next we have the ll_statahead_thread() and the ll_agl_thread() record
their respective pids in their respective ptlrpc_thread structures.
This allows to print the pid of the thread that we are trying to
stop (which is actually useful info) from other threads, such as those
calling ll_stop_statahead().

Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-on: http://review.whamcloud.com/9360
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4624
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/llite: Avoid statahead thread start/stop deadlocks

The statahead and statahead agl threads blindly set their
thread state to SVC_RUNNING without checking the state first.  If, for
instance, another thread sets the state to SVC_STOPPING that
stop signal will now have been lost.  Deadlock ensues.

We also partly improve the sai reference counting, because a race exists
where the ll_stop_statahead thread can drop the default reference, and
the statahead thread can exit and drop its reference as well.  With no
references on the sai, the final put will poison and free the buffer.  The
original do_statahead_enter() function may then continue to access
the buffer after it is freed because it did not take a reference of its
own.  We add a local reference to address that.

Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-on: http://review.whamcloud.com/9358
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4624
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre: Limit reply buffer size

When allocating a reply buffer for the striping information don't
assume the unlikely worst case.  Instead, assume the common case
and size the buffer based on the observed default ea/cookie size.

The default size is initialized to a single stripe and allowed to
grow up to an entire page if needed.  This means that for smallish
filesystems (less than ~21 OSTs) where the worst case striping
information can fit in a single page there is effectively no
change.  Only for larger filesystem will the default be less than
the maximum.  This has a number of advantages.

* By limiting the default reply buffer size we avoid always
  vmalloc()'ing the buffer because it exceeds four pages in size
  and instead kmalloc() it.  This prevents the client from
  thrashing on the global vmalloc() spin lock.

* A reply buffer of exactly the right size (no larger) is allocated
  in the overflow case.  These larger reply buffers are still
  unlikely to exceed the 16k limit where a vmalloc() will occur.

* Saves memory in the common case.  Wide striped files exceeded
  the default are expected to be the exception.

The reason this patch works is because the ptlrpc layer is smart
enough to reallocate the reply buffer when an overflow occurs.
Therefore the client doesn't have to drop the incoming reply and
send a new request with a larger reply buffer.

It's also worth mentioning that the reply buffer always contains
a significant amount of extra padding because they are rounded up
to the nearest power of two.  This means that even files striped
wider than the default have a good chance of fitting in the
allocated reply buffer.

Also remove client eadatasize check in mdt xattr packing because
as said above client can handle -EOVERFLOW.

Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-on: http://review.whamcloud.com/6339
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3338
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/lov: to not hold sub locks at initialization

Otherwise, it will cause deadlock because it essentially holds
some sub locks and then to request others in an arbitrary order.

Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-on: http://review.whamcloud.com/9152
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/llite: deadlock taking lli_trunc_sem during file write

File write before io loop will take lli_trun_sem read semaphore to
protect osc_extent, while after generic_file_aio_write() done, it
could possible need to kill suid or sgid, which will call
ll_setattr_raw() to change the inode's attribute, and it does not
involve size.

So the ll_truc_sem write semaphore should be constrained
around ll_setattr_ost() to not come across the lli_trunc_sem read
semaphore get from the normal file write path.

Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Reviewed-on: http://review.whamcloud.com/9267
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4627
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Merge tag 'usb-3.15-rc3' of git://git./linux/kernel/git/gregkh/usb

Pull USB fixes from Greg KH:
 "Here are a number of USB fixes for 3.15-rc3.  The majority are gadget
  fixes, as we didn't get any of those in for 3.15-rc2.  The others are
  all over the place, and there's a number of new device id addtions as
  well."

* tag 'usb-3.15-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb: (35 commits)
  usb: option: add and update a number of CMOTech devices
  usb: option: add Alcatel L800MA
  usb: option: add Olivetti Olicard 500
  usb: qcserial: add Sierra Wireless MC7305/MC7355
  usb: qcserial: add Sierra Wireless MC73xx
  usb: qcserial: add Sierra Wireless EM7355
  USB: io_ti: fix firmware download on big-endian machines
  usb/xhci: fix compilation warning when !CONFIG_PCI && !CONFIG_PM
  xhci: extend quirk for Renesas cards
  xhci: Switch Intel Lynx Point ports to EHCI on shutdown.
  usb: xhci: Prefer endpoint context dequeue pointer over stopped_trb
  phy: core: make NULL a valid phy reference if !CONFIG_GENERIC_PHY
  phy: fix kernel oops in phy_lookup()
  phy: restore OMAP_CONTROL_PHY dependencies
  phy: exynos: fix building as a module
  USB: serial: fix sysfs-attribute removal deadlock
  usb: wusbcore: fix panic in wusbhc_chid_set
  usb: wusbcore: convert nested lock to use spin_lock instead of spin_lock_irq
  uwb: don't call spin_unlock_irq in a USB completion handler
  usb: chipidea: coordinate usb phy initialization for different phy type
  ...

staging/lustre/llite: issue OST_SYNC for fsync()

The last parameter @datasync of fsync() has following indication:
* if datasync=0, we'd always flush data and metadata
* if datasync=1, we'd always flush data while does not flush modifed
  metadata unless that metadata is needed in order to allow a
  subsequent data retrieval to be correctly handled. For example, a
  change to the file size would require a metadata flush.

Lustre client can not tell the difference easily, and would issue
MDS_SYNC and OST_SYNC in all cases.

Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Reviewed-on: http://review.whamcloud.com/8684
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4388
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre: fix permission problem of setfacl

Setxattr does not check the permission when setting ACL xattrs. This
will cause security problem because any user can walk around
permission checking by changing ACL rules.

Signed-off-by: Li Xi <lixi@ddn.com>
Reviewed-on: http://review.whamcloud.com/9473
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4704
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/hsm: HSM requests not delivered

The total size of an HSM archive request may exceed the
desired (LNET) message. When this happens, it can hang
the client and not allow the archive request to succeed.

Before we know the total size of the hsm_action_items, we
need to limit the size of the reguest. Doing this limits
the number of items that can be sent in one archive request.
We'e reduced the size allowed for the user archive request
to MDS_MAXREQSIZE/3.

Signed-off-by: James Nunez <james.a.nunez@intel.com>
Reviewed-on: http://review.whamcloud.com/9393
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4639
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/hsm: count NULL terminator in hai_zero/hal_size

If fsname is 8-byte aligned, hai_zero fails to count the ending NULL
terminator causing hai to directly attached after fsname and future
hai_zero will return a different position for first hai.

Signed-off-by: Peng Tao <bergwolf@gmail.com>
Reviewed-on: http://review.whamcloud.com/9431
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4689
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/obdclass: remove uses of lov_stripe_md

Remove the unused function llog_obd_add(). Remove the unused count and
parameters from llog_cancel(). Move dump_lsm() from obdclass to
the only module that uses it (lov). Remove obd_lov.h.

Signed-off-by: John L. Hammond <john.hammond@intel.com>
Reviewed-on: http://review.whamcloud.com/8545
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/lov: remove unused lov llog code

Remove the unused functions lov_llog_init(), lov_llog_finish(),
their supporting functions, and the file lov_log.c.

Signed-off-by: John L. Hammond <john.hammond@intel.com>
Reviewed-on: http://review.whamcloud.com/8539
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre: quiet console permission error messages

Quiet some common console error messages for permission errors
that can be hit in common cases.

Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-on: http://review.whamcloud.com/8988
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4522
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Faccini Bruno <bruno.faccini@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre: restore __GFP_WAIT flag to memalloc calls

In Lustre 2.4, the flags passed to the memory allocation functions are
translated from CFS enumeration values types to the kernel GFP
values by calling cfs_alloc_flags_to_gfp(). This function adds
__GFP_WAIT to all flags except CFS_ALLOC_ATOMIC. In 2.5, when
the cfs wrappers were dropped, cfs_alloc_flags_to_gfp() was
removed and the CFS_ALLOC_xxxx was simply replaced with __GFP_xxxx.
This means that most memory allocation calls are missing the
__GFP_WAIT flag. The result is that Lustre experiences more ENOMEM
errors, many of which the higher levels of Lustre do not handle
robustly.
Notes GFP_NOFS = __GFP_WAIT | __GFP_IO. So the patch replaces
__GFP_IO with GFP_NOFS.
Patch does not add __GFP_WAIT to GFP_IOFS. GFP_IOFS was not used in
Lustre 2.4 so it has never been used with __GFP_WAIT.

Signed-off-by: Ann Koehler <amk@cray.com>
Signed-off-by: Emoly Liu <emoly.liu@intel.com>
Reviewed-on: http://review.whamcloud.com/9223
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4357
Reviewed-by: Liang Zhen <liang.zhen@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/ptlrpc: add rpc_cache

Add rpc_cache for allocating ptlrpc_requests.

Xyratex-bug-id: MRP-689
Signed-off-by: Andriy Skulysh <Andriy_Skulysh@xyratex.com>
Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-on: http://review.whamcloud.com/6874
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2424
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/gss: fix few issues found by Klocwork Insight tool

Array 'message_buf' of size 500 may use index value(s) -1

Object 'enc_key.data' was freed at line 164 after being freed
by calling 'free' at line 150. Also there are 3 similar errors
on line(s) 164.

Suspicious dereference of pointer 'vmsg' before NULL check at
line 187. Also there are 2 similar errors on line(s) 196, 205.

Suspicious dereference of pointer 'rmsg' before NULL check at
line 191. Also there are 2 similar errors on line(s) 200, 209.

Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-on: http://review.whamcloud.com/9274
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4629
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/ptlrpc: don't try to recover no_recov connection

If a connection has been stopped with ptlrpc_pinger_del_import() and
marked obd_no_recov, don't reconnect in ptlrpc_disconnect_import() if
the import is already disconnected.  Otherwise, without the pinger it
will just wait there indefinitely for the reconnection that will never
happen.

Put the obd_no_recov check inside ptlrpc_import_in_recovery() so that
any threads waiting on the connection to recover would also be broken
out of their sleep if obd_no_recov is set.

Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-on: http://review.whamcloud.com/8996
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4413
Reviewed-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Reviewed-by: wangdi <di.wang@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/clio: clear nowait flag agl lock re-enqueue

The LDLM_FL_BLOCK_NOWAIT flag should be cleared when re-enqueue
the agl lock as normal glimpse, otherwise, it won't get size back
if there is conflicting locks on other client.

Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-on: http://review.whamcloud.com/9249
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4597
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Ned Bass <bass6@llnl.gov>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/ldlm: Hold lock when clearing flag

This patch moves lock's skip flag clearing from lru-delete to
lru-add code to prevent clearing lock's flag without resource lock
protection.

Signed-off-by: Li Xi <lixi@ddn.com>
Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Reviewed-on: http://review.whamcloud.com/8772
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4269
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/lnet: Dropped messages are not accounted correctly

LNET messages that are dropped are not accounted for correctly in
/proc/sys/lnet/stats. What I assume to be a simple typo is causing
drop_length to be double-counted and drop_count to never be
incremented.

Signed-off-by: Matt Ezell <ezellma@ornl.gov>
Reviewed-on: http://review.whamcloud.com/9096
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4577
Reviewed-by: James Nunez <james.a.nunez@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Isaac Huang <he.huang@intel.com>
Reviewed-by: Liang Zhen <liang.zhen@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/osc: don't activate deactivated obd_import

In ptlrpc_activate_import(), obd_import->imp_deactive should
be checked if it is deactivated, otherwise it will trigger an
LBUG in ptlrpc_invalidate_import():

  ptlrpc_invalidate_import() ASSERTION(imp->imp_invalid) failed

Signed-off-by: Hongchao Zhang <hongchao.zhang@intel.com>
Reviewed-on: http://review.whamcloud.com/8747
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4386
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/ptlrpc: Remove log message about export timer update

Function ptlrpc_update_export_timer generates lots of D_HA level log
messages whenever the export timer gets updated. Those log messages
are found little use for issue investigations, and it will take space
in the Lustre log buffer. We are removing it now.

Xyratex-bug-id: MRP-733
Signed-off-by: Cheng Shao <cheng_shao@xyratex.com>
Reviewed-on: http://review.whamcloud.com/9147
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4590
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/lustre/ptlrpc: Fix assertion failure of null_alloc_rs()

lustre_get_emerg_rs() set the size of the reply buffer to zero
by mistake, which will cause LBUG in null_alloc_rs() when memory
pressure is high. This patch fix this problem and adds a size
check to avoid the problem of insufficient buffer size.

Signed-off-by: Li Xi <lixi@ddn.com>
Signed-off-by: Patrick Farrell <paf@cray.com>
Reviewed-on: http://review.whamcloud.com/8200
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3680
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Merge tag 'pm+acpi-3.15-rc3' of git://git./linux/kernel/git/rafael/linux-pm

Pull ACPI and power management fixes from Rafael Wysocki:
 "These include a fix for a recent ACPI regression related to device
  notifications, intel_idle fix related to IvyTown support, fix for a
  buffer size issue in ACPICA, PM core fix related to the "freeze" sleep
  state, four fixes for various types of breakage in cpufreq drivers, a
  PNP workaround for a wrong memory region size in ACPI tables, and a
  fix and cleanup for the ACPI tools Makefile.

  Specifics:

   - Fix for broken ACPI notifications on some systems caused by a
     recent ACPI hotplug commit that blocked the propagation of unknown
     type notifications to device drivers inadvertently.

   - intel_idle fix to make the IvyTown C-states handling (added
     recently) work as intended which now is broken due to missing
     braces.  From Christoph Jaeger.

   - ACPICA fix to make it allocate buffers of the right sizes for the
     Generic Serial Bus operation region access.  From Lv Zheng.

   - PM core fix unblocking cpuidle before entering the "freeze" sleep
     state which causes that state to be able to actually save more
     energy than runtime idle.

   - Configuration and build fixes for the highbank and powernv cpufreq
     drivers from Kefeng Wang and Srivatsa S Bhat.

   - Coccinelle warning fix related to error pointers for the unicore32
     cpufreq driver from Duan Jiong.

   - Integer overflow fix for the ppc-corenet cpufreq driver from Geert
     Uytterhoeven.

   - Workaround for BIOSes that don't report the entire Intel MCH area
     in their ACPI tables from Bjorn Helgaas.

   - ACPI tools Makefile fix and cleanup from Thomas Renninger"

* tag 'pm+acpi-3.15-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  ACPI / notify: Do not block unknown type notifications in root handler
  PNP: Work around BIOS defects in Intel MCH area reporting
  cpufreq: highbank: fix ARM_HIGHBANK_CPUFREQ dependency warning
  cpufreq: ppc: Fix integer overflow in expression
  cpufreq, powernv: Fix build failure on UP
  cpufreq: unicore32: replace IS_ERR and PTR_ERR with PTR_ERR_OR_ZERO
  PM / suspend: Make cpuidle work in the "freeze" state
  intel_idle: fix IVT idle state table setting
  ACPICA: Fix buffer allocation issue for generic_serial_bus region accesses.
  tools/power/acpi: Minor bugfixes

staging: rtl8723au: Get rid of _rtw_free_network23a_nolock23a()

_rtw_free_network23a() and _rtw_free_network23a_nolock23a() are now
identical - one copy should do.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_free_network_queue23a() eliminate unused argument 'isfreeall'

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: _rtw_free_network23a(): Remove unused argument 'isfreeall'

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: _rtw_free_network23a(): Remove write-only variable 'lifetime'

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Fold _rtw_find_network23a() into rtw_find_network23a()

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Fold _rtw_free_network23a_queue23a() into rtw_free_network_queue23a()

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Fold _rtw_init_mlme_priv23a() into rtw_init_mlme_priv23a()

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove redundant checks of WPA OUI

All callers get here by searching for the WPA OUI first, so no point
checking for it once we get here.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Eliminate the 217th copy of the microsoft OUI

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Make sparse happy - make rtw_get_beacon_interval23a_from_ie return __le16 *

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: r8723au: Fix endian errors for frame control

The following locations in the code treat the frame control as u16 rather than
the correct __le16:

drivers/staging/rtl8723au/core/rtw_mlme_ext.c:2471:15:    expected unsigned short *fctrl
drivers/staging/rtl8723au/core/rtw_mlme_ext.c:2644:15:    expected unsigned short *fctrl
drivers/staging/rtl8723au/core/rtw_mlme_ext.c:2849:15:    expected unsigned short *fctrl
drivers/staging/rtl8723au/core/rtw_mlme_ext.c:2994:15:    expected unsigned short *fctrl
drivers/staging/rtl8723au/core/rtw_mlme_ext.c:3308:15:    expected unsigned short *fctrl
drivers/staging/rtl8723au/core/rtw_mlme_ext.c:3578:15:    expected unsigned short *fctrl
drivers/staging/rtl8723au/core/rtw_mlme_ext.c:3699:15:    expected unsigned short *fctrl
drivers/staging/rtl8723au/core/rtw_mlme_ext.c:3821:15:    expected unsigned short *fctrl
drivers/staging/rtl8723au/core/rtw_mlme_ext.c:3932:15:    expected unsigned short *fctrl
drivers/staging/rtl8723au/core/rtw_mlme_ext.c:4004:15:    expected unsigned short [usertype] *fctrl
drivers/staging/rtl8723au/core/rtw_mlme_ext.c:4194:15:    expected unsigned short *fctrl
drivers/staging/rtl8723au/core/rtw_xmit.c:911:23:    expected unsigned short [usertype] *fctrl
drivers/staging/rtl8723au/hal/rtl8723a_cmd.c:216:15:    expected unsigned short [usertype] *fctrl
drivers/staging/rtl8723au/hal/rtl8723a_cmd.c:313:15:    expected unsigned short [usertype] *fctrl
drivers/staging/rtl8723au/hal/rtl8723a_cmd.c:350:15:    expected unsigned short [usertype] *fctrl
drivers/staging/rtl8723au/hal/rtl8723a_cmd.c:419:15:    expected unsigned short [usertype] *fctrl
drivers/staging/rtl8723au/os_dep/ioctl_cfg80211.c:314:15:    expected unsigned short *fctrl
drivers/staging/rtl8723au/os_dep/ioctl_cfg80211.c:2357:15:    expected unsigned short *fctrl

Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>