DM thinp's use of bio_inc_remaining() is critical to ensure the original
parent discard bio isn't completed before sub-discards have.  DM thinp
needs this due to the extra quiescing that occurs, via multiple DM thinp
mappings, while processing large discards.  As such DM thinp must build
the async discard bio chain after some delay -- so bio_inc_remaining()
is used to enable DM thinp to take a reference on the original parent
discard bio for each mapping.  This allows the immediate use of
bio_endio() on that discard bio; but with the understanding that the
actual completion won't occur until each of the sub-discards'
per-mapping references are dropped.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Acked-by: Joe Thornber <ejt@redhat.com>

Given we don't yet support any feature flags in the dm-raid ondisk
metadata (see: 'features' member of 'struct dm_raid_superblock'),
add a check to ensure no flags are actually set, if any features are
set reject the activation of the RAID mapping.

This is to prevent possible data corruption in case of a kernel
downgrade when there'll potentially be feature flags set by a future
dm-raid target.
Signed-off-by: Heinz Mauelshagen <heinzm@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>

copy_params()'s use of __GFP_REPEAT for the __vmalloc() call doesn't make much
sense because vmalloc doesn't rely on costly high order allocations.
Signed-off-by: Michal Hocko <mhocko@suse.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>

Signed-off-by: Eric Engestrom <eric@engestrom.ch>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>

Also fix some typos and make all "smq" and "mq" references consistently
lowercase.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>

The primary motivation of this commit is to improve the scalability of
DM multipath on large NUMA systems where m->lock spinlock contention has
been proven to be a serious bottleneck on really fast storage.

The ability to atomically read a pointer, using lockless_dereference(),
is leveraged in this commit.  But all pointer writes are still protected
by the m->lock spinlock (which is fine since these all now occur in the
slow-path).

The following functions no longer require the m->lock spinlock in their
fast-path: multipath_busy(), __multipath_map(), and do_end_io()

And choose_pgpath() is modified to _not_ update m->current_pgpath unless
it also switches the path-group.  This is done to avoid needing to take
the m->lock everytime __multipath_map() calls choose_pgpath().
But m->current_pgpath will be reset if it is failed via fail_path().
Suggested-by: Jeff Moyer <jmoyer@redhat.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Tested-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>

Allows the 'work_mutex' member to no longer cross a cacheline.
Reviewed-by: Hannes Reinecke <hare@suse.com>
Tested-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>

The use of atomic_t for nr_valid_paths, pg_init_in_progress and
pg_init_count will allow relaxing the use of the m->lock spinlock.
Suggested-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Tested-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>

Mechanical change that doesn't make any real effort to reduce the use of
m->lock; that will come later (once atomics are used for counters, etc).
Suggested-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Tested-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>

Return statement at the end of a void function is useless.

The Coccinelle semantic patch used to make this change is as follows:
//<smpl>
@@
identifier f;
expression e;
@@
void f(...) {
<...
- return
  e;
...>
}
//</smpl>
Signed-off-by: Amitoj Kaur Chawla <amitoj1606@gmail.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>

Signed-off-by: Mike Snitzer <snitzer@redhat.com>

Needed in order to update the DM thinp code to use the new async
__blkdev_issue_discard() interface.

Commit 326e1dbb ("block: remove management of bi_remaining when
restoring original bi_end_io") made bio_inc_remaining() private to bio.c
because the only use-case that made sense was confined to the
bio_chain() interface.

Since that time DM thinp went on to use bio_chain() in its relatively
complex implementation of async discard support.  That implementation,
even when converted over to use the new async __blkdev_issue_discard()
interface, depends on deferred completion of the original discard bio --
which is most appropriately implemented using bio_inc_remaining().

DM thinp foolishly duplicated bio_inc_remaining(), local to dm-thin.c as
__bio_inc_remaining(), so re-exporting bio_inc_remaining() allows us to
put an end to that foolishness.

All said, bio_inc_remaining() should really only be used in conjunction
with bio_chain().  It isn't intended for generic bio reference counting.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Acked-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Jens Axboe <axboe@fb.com>

Commit 38f25255 ("block: add __blkdev_issue_discard") incorrectly
disallowed the early return of -EOPNOTSUPP if the device doesn't support
discard (or secure discard).  This early return of -EOPNOTSUPP has
always been part of blkdev_issue_discard() interface so there isn't a
good reason to break that behaviour -- especially when it can be easily
reinstated.

The nuance of allowing early return of -EOPNOTSUPP vs disallowing late
return of -EOPNOTSUPP is: if the overall device never advertised support
for discards and one is issued to the device it is beneficial to inform
the caller that discards are not supported via -EOPNOTSUPP.  But if a
device advertises discard support it means that at least a subset of the
device does have discard support -- but it could be that discards issued
to some regions of a stacked device will not be supported.  In that case
the late return of -EOPNOTSUPP must be disallowed.

Fixes: 38f25255 ("block: add __blkdev_issue_discard")
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Jens Axboe <axboe@fb.com>

blk_account_io_start does not need to be wrapped with blk_do_io_stat
ais it already checks for that condition.
Signed-off-by: Michael Callahan <michaelcallahan@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>

This is a version of blkdev_issue_discard which doesn't wait for
the I/O to complete, but instead allows the caller to submit
the final bio and/or chain it to others.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Ming Lin <ming.l@ssi.samsung.com>
Signed-off-by: Sagi Grimberg <sagig@grimberg.me>
Reviewed-by: Ming Lei <tom.leiming@gmail.com>
Signed-off-by: Jens Axboe <axboe@fb.com>

It can be replaced with a combination of bio_chain and submit_bio_wait.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Ming Lin <ming.l@ssi.samsung.com>
Signed-off-by: Sagi Grimberg <sagig@grimberg.me>
Reviewed-by: Ming Lei <tom.leiming@gmail.com>
Signed-off-by: Jens Axboe <axboe@fb.com>

Pull thermal fixes from Eduardo Valentin:
 "A couple of minor fixes for the thermal subsystem.

  Specifics in this pull request:

   - Fixes in hisilicon thermal driver
   - More fixes of unsigned to int type change in thermal_core.c"

* 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/evalenti/linux-soc-thermal:
  thermal: use %d to print S32 parameters
  thermal: hisilicon: increase temperature resolution

Pull powerpc fixes from Michael Ellerman:
 "A few more powerpc fixes for 4.6:

   - cxl: Keep IRQ mappings on context teardown from Michael Neuling

   - cxl: Poll for outstanding IRQs when detaching a context from
     Michael Neuling

   - Wire up preadv2 and pwritev2 syscalls from Rui Salvaterra"

* tag 'powerpc-4.6-4' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:
  powerpc: wire up preadv2 and pwritev2 syscalls
  cxl: Poll for outstanding IRQs when detaching a context
  cxl: Keep IRQ mappings on context teardown

Pull EDAC fix from Borislav Petkov:
 "Make sure sb_edac and i7core_edac do not terminate MCE processing on
  the decoding callchain prematurely"

* tag 'edac_fix_for_4.6' of git://git.kernel.org/pub/scm/linux/kernel/git/bp/bp:
  EDAC: i7core, sb_edac: Don't return NOTIFY_BAD from mce_decoder callback

Pull power management fixes from Rafael Wysocki:
 "One revert of a recent cpufreq commit that introduced a regression and
  a fix for intel_pstate's Turbo Activation Ratio handling code.

  Specifics:

   - Revert cpufreq commit that attempted to fix a problem in the
     ondemand/conservative governor code, but did that incorrectly and
     introduced another problem instead (Rafael Wysocki).

   - Fix incorrect decoding of MSR contents related to the Turbo
     Activation Ratio (TAR) handling in the intel_pstate driver
     (Srinivas Pandruvada)"

* tag 'pm+acpi-4.6-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  cpufreq: intel_pstate: Fix processing for turbo activation ratio
  Revert "cpufreq: governor: Fix negative idle_time when configured with CONFIG_HZ_PERIODIC"

Pull MMC fixes from Ulf Hansson:
 "Here are a two MMC host fixes:

  - sdhci-acpi: Reduce Baytrail eMMC/SD/SDIO hangs

  - sunxi: Disable eMMC HS-DDR for Allwinner A80"

* tag 'mmc-v4.6-rc4' of git://git.linaro.org/people/ulf.hansson/mmc:
  mmc: sunxi: Disable eMMC HS-DDR (MMC_CAP_1_8V_DDR) for Allwinner A80
  mmc: sdhci-acpi: Reduce Baytrail eMMC/SD/SDIO hangs

Pull drm fixes from Dave Airlie:
 "A few fixes all over the place:

  radeon is probably the biggest standout, it's a fix for screen
  corruption or hung black outputs so I thought it was worth pulling in.

  Otherwise some amdgpu power control fixes, some misc vmwgfx fixes, one
  etnaviv fix, one virtio-gpu fix, two DP MST fixes, and a single TTM
  fix"

* 'drm-fixes' of git://people.freedesktop.org/~airlied/linux:
  drm/vmwgfx: Fix order of operation
  drm/vmwgfx: use vmw_cmd_dx_cid_check for query commands.
  drm/vmwgfx: Enable SVGA_3D_CMD_DX_SET_PREDICATION
  drm/amdgpu: disable vm interrupts with vm_fault_stop=2
  drm/amdgpu: print a message if ATPX dGPU power control is missing
  Revert "drm/amdgpu: disable runtime pm on PX laptops without dGPU power control"
  drm/radeon: fix vertical bars appear on monitor (v2)
  drm/ttm: fix kref count mess in ttm_bo_move_to_lru_tail
  drm/virtio: send vblank event after crtc updates
  drm/dp/mst: Restore primary hub guid on resume
  drm/dp/mst: Get validated port ref in drm_dp_update_payload_part1()
  drm/etnaviv: don't move linear memory window on 3D cores without MC2.0

Pull rdma fixes from Doug Ledford:
 "Final set of -rc fixes for 4.6.

  I've collected up a number of patches that are all pretty small with
  the exception of only a couple.  The hfi1 driver has a number of
  important patches, and it is what really drives the line count of this
  pull request up.  These are all small and I've got this kernel built
  and running in the test lab (I have most of the hardware, I think nes
  is the only thing in this patch set that I can't say I've personally
  tested and have up and running).

  Summary:

   - A number of collected fixes for oopses, memory corruptions,
     deadlocks, etc.  All of these fixes are small (many only 5-10
     lines), obvious, and tested.

   - Fix for the security issue related to the use of write for
     bi-directional communications"

* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dledford/rdma:
  RDMA/nes: don't leak skb if carrier down
  IB/security: Restrict use of the write() interface
  IB/hfi1: Use kernel default llseek for ui device
  IB/hfi1: Don't attempt to free resources if initialization failed
  IB/hfi1: Fix missing lock/unlock in verbs drain callback
  IB/rdmavt: Fix send scheduling
  IB/hfi1: Prevent unpinning of wrong pages
  IB/hfi1: Fix deadlock caused by locking with wrong scope
  IB/hfi1: Prevent NULL pointer deferences in caching code
  MAINTAINERS: Update iser/isert maintainer contact info
  IB/mlx5: Expose correct max_sge_rd limit
  RDMA/iw_cxgb4: Fix bar2 virt addr calculation for T4 chips
  iw_cxgb4: handle draining an idle qp
  iw_cxgb3: initialize ibdev.iwcm->ifname for port mapping
  iw_cxgb4: initialize ibdev.iwcm->ifname for port mapping
  IB/core: Don't drain non-existent rq queue-pair
  IB/core: Fix oops in ib_cache_gid_set_default_gid

Merge fixes from Andrew Morton:
 "20 fixes"

* emailed patches from Andrew Morton <akpm@linux-foundation.org>:
  Documentation/sysctl/vm.txt: update numa_zonelist_order description
  lib/stackdepot.c: allow the stack trace hash to be zero
  rapidio: fix potential NULL pointer dereference
  mm/memory-failure: fix race with compound page split/merge
  ocfs2/dlm: return zero if deref_done message is successfully handled
  Ananth has moved
  kcov: don't profile branches in kcov
  kcov: don't trace the code coverage code
  mm: wake kcompactd before kswapd's short sleep
  .mailmap: add Frank Rowand
  mm/hwpoison: fix wrong num_poisoned_pages accounting
  mm: call swap_slot_free_notify() with page lock held
  mm: vmscan: reclaim highmem zone if buffer_heads is over limit
  numa: fix /proc/<pid>/numa_maps for THP
  mm/huge_memory: replace VM_NO_THP VM_BUG_ON with actual VMA check
  mailmap: fix Krzysztof Kozlowski's misspelled name
  thp: keep huge zero page pinned until tlb flush
  mm: exclude HugeTLB pages from THP page_mapped() logic
  kexec: export OFFSET(page.compound_head) to find out compound tail page
  kexec: update VMCOREINFO for compound_order/dtor

Both of these drivers can return NOTIFY_BAD, but this terminates
processing other callbacks that were registered later on the chain.
Since the driver did nothing to log the error it seems wrong to prevent
other interested parties from seeing it. E.g. neither of them had even
bothered to check the type of the error to see if it was a memory error
before the return NOTIFY_BAD.
Signed-off-by: Tony Luck <tony.luck@intel.com>
Acked-by: Aristeu Rozanski <aris@redhat.com>
Acked-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Cc: linux-edac <linux-edac@vger.kernel.org>
Cc: <stable@vger.kernel.org>
Link: http://lkml.kernel.org/r/72937355dd92318d2630979666063f8a2853495b.1461864507.git.tony.luck@intel.comSigned-off-by: Borislav Petkov <bp@suse.de>

* pm-cpufreq-fixes:
  cpufreq: intel_pstate: Fix processing for turbo activation ratio
  Revert "cpufreq: governor: Fix negative idle_time when configured with CONFIG_HZ_PERIODIC"

A few fixes for 4.6.
- revert amdgpu PX commit that was previously reverted on the radeon side
- cleaned up version of the NI+ MC update display fix for radeon
- TTM kref fix

* 'drm-fixes-4.6' of git://people.freedesktop.org/~agd5f/linux:
  drm/amdgpu: disable vm interrupts with vm_fault_stop=2
  drm/amdgpu: print a message if ATPX dGPU power control is missing
  Revert "drm/amdgpu: disable runtime pm on PX laptops without dGPU power control"
  drm/radeon: fix vertical bars appear on monitor (v2)
  drm/ttm: fix kref count mess in ttm_bo_move_to_lru_tail

three misc vmwgfx fixes

* 'drm-vmwgfx-fixes' of git://people.freedesktop.org/~syeh/repos_linux:
  drm/vmwgfx: Fix order of operation
  drm/vmwgfx: use vmw_cmd_dx_cid_check for query commands.
  drm/vmwgfx: Enable SVGA_3D_CMD_DX_SET_PREDICATION

Pull x86 fixes from Ingo Molnar:
 "Two boot crash fixes and an IRQ handling crash fix"

* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/apic: Handle zero vector gracefully in clear_vector_irq()
  Revert "x86/mm/32: Set NX in __supported_pte_mask before enabling paging"
  xen/qspinlock: Don't kick CPU if IRQ is not initialized

Pull perf fixes from Ingo Molnar:
 "x86 PMU driver fixes plus a core code race fix"

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf/x86/intel: Fix incorrect lbr_sel_mask value
  perf/x86/intel/pt: Don't die on VMXON
  perf/core: Fix perf_event_open() vs. execve() race
  perf/x86/amd: Set the size of event map array to PERF_COUNT_HW_MAX
  perf/core: Make sysctl_perf_cpu_time_max_percent conform to documentation
  perf/x86/intel/rapl: Add missing Haswell model
  perf/x86/intel: Add model number for Skylake Server to perf

Pull locking fixes from Ingo Molnar:
 "Two lockdep fixes"

* 'locking-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  lockdep: Fix lock_chain::base size
  locking/lockdep: Fix ->irq_context calculation

Pull EFI fix from Ingo Molnar:
 "This fixes a bug in the efivars code"

* 'efi-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  efi: Fix out-of-bounds read in variable_matches()

Pull media fixes from Mauro Carvalho Chehab:
 "Some regression fixes:

   - videobuf2 core: avoid the risk of going past buffer on multi-planes
     and fix rw mode

   - fix support for 4K formats at V4L2 core

   - fix a trouble at davinci_fpe, caused by a bad patch

   - usbvision: revert a patch with a partial fixup.  The fixup patch
     was merged already, and this one has some issues"

* tag 'media/v4.6-4' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media:
  [media] vb2-memops: Fix over allocation of frame vectors
  [media] media: vb2: Fix regression on poll() for RW mode
  [media] v4l2-dv-timings.h: fix polarity for 4k formats
  [media] davinci_vpfe: Revert "staging: media: davinci_vpfe: remove,unnecessary ret variable"
  [media] usbvision: revert commit 588afcc1
  [media] videobuf2-v4l2: Verify planes array in buffer dequeueing
  [media] videobuf2-core: Check user space planes array in dqbuf

Pull sound fixes from Takashi Iwai:
 "Usually we get a big collection of fixes for ASoC once during rc.  And
  this is it.

  At this time, most of fixes are about Intel Skylake ASoC driver, which
  is a new and still on-going development.  Along with it, a slight
  large LOC is seen in legacy HD-audio driver, but it's merely a code
  move to the upper layer.

  Other than that, the rest are small or trivial fixes to various
  drivers, in addition to an ASoC dapm debugfs code fix"

* tag 'sound-4.6-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound: (24 commits)
  ALSA: hda - Update BCLK also at hotplug for i915 HSW/BDW
  ALSA: hda - Add dock support for ThinkPad X260
  ASoC: wm5102: Free compressed IRQ in CODEC remove
  ASoC: arizona: Free speaker thermal IRQs in CODEC remove
  ASoC: Intel: Skylake: Fix ibs/obs calc for non-integral sampling rates
  ASoC: Intel: sst: fix a loop timeout in sst_hsw_stream_reset()
  ASoC: Intel: Skylake: Fix to turn OFF codec power when entering S3
  ASoC: hdac_hdmi: Fix codec power state in S3 during playback
  ASoC: hdac_hdmi: Fix to use dev_pm ops instead soc pm
  ASoC: wm8962: Correct typo when setting DSPCLK rate
  ASoC: nau8825: Fix jack detection across suspend
  ASoC: Intel: Skylake: Fix DSP resource de-allocation
  ASoC: Intel: Skylake: Fix for unloading module only when it is loaded
  ASoC: Intel: Skylake: Fix kbuild dependency
  ASoC: dapm: Make sure we have a card when displaying component widgets
  ASoC: rt5640: Correct the digital interface data select
  ASoC: Intel: Skylake: remove call to pci_dev_put
  ASoC: Intel: Skylake: Call i915 exit last
  ASoC: Intel: Skylake: Unmap the address last
  ASoC: Intel: Skylake: Freeup properly on skl_dsp_free
  ...

Commit 3193913c ("mm: page_alloc: default node-ordering on 64-bit
NUMA, zone-ordering on 32-bit") changes the default value of
numa_zonelist_order.  Update the document.
Signed-off-by: Xishi Qiu <qiuxishi@huawei.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Do not bail out from depot_save_stack() if the stack trace has zero hash.
Initially depot_save_stack() silently dropped stack traces with zero
hashes, however there's actually no point in reserving this zero value.
Reported-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Alexander Potapenko <glider@google.com>
Acked-by: Andrey Ryabinin <aryabinin@virtuozzo.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

The change fixes improper check for a returned error value by
class_create() function, which on error returns ERR_PTR() value, thus the
original check always results in a dead code on error path.
Signed-off-by: Vladimir Zapolskiy <vz@mleia.com>
Signed-off-by: Alexandre Bounine <alexandre.bounine@idt.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

get_hwpoison_page() must recheck relation between head and tail pages.

n-horiguchi said: without this recheck, the race causes kernel to pin an
irrelevant page, and finally makes kernel crash for refcount mismatch.
Signed-off-by: Konstantin Khlebnikov <khlebnikov@yandex-team.ru>
Acked-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>