Update arch.apic_base before triggering recalculate_apic_map. Otherwise
the recalculation will work against the previous state of the APIC and
will fail to build the correct map when an APIC is hardware-enabled
again.

This fixes a regression of 1e08ec4a.

Cc: stable@vger.kernel.org
Signed-off-by: NJan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: NMarcelo Tosatti <mtosatti@redhat.com>

Pull ext4 fixes from Ted Ts'o:
 "A collection of bug fixes destined for stable and some printk cleanups
  and a patch so that instead of BUG'ing we use the ext4_error()
  framework to mark the file system is corrupted"

* tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4:
  ext4: add explicit casts when masking cluster sizes
  ext4: fix deadlock when writing in ENOSPC conditions
  jbd2: rename obsoleted msg JBD->JBD2
  jbd2: revise KERN_EMERG error messages
  jbd2: don't BUG but return ENOSPC if a handle runs out of space
  ext4: Do not reserve clusters when fs doesn't support extents
  ext4: fix del_timer() misuse for ->s_err_report
  ext4: check for overlapping extents in ext4_valid_extent_entries()
  ext4: fix use-after-free in ext4_mb_new_blocks
  ext4: call ext4_error_inode() if jbd2_journal_dirty_metadata() fails

Pull block fixes from Jens Axboe:
 - fix for a memory leak on certain unplug events
 - a collection of bcache fixes from Kent and Nicolas
 - a few null_blk fixes and updates form Matias
 - a marking of static of functions in the stec pci-e driver

* 'for-linus' of git://git.kernel.dk/linux-block:
  null_blk: support submit_queues on use_per_node_hctx
  null_blk: set use_per_node_hctx param to false
  null_blk: corrections to documentation
  null_blk: warning on ignored submit_queues param
  null_blk: refactor init and init errors code paths
  null_blk: documentation
  null_blk: mem garbage on NUMA systems during init
  drivers: block: Mark the functions as static in skd_main.c
  bcache: New writeback PD controller
  bcache: bugfix for race between moving_gc and bucket_invalidate
  bcache: fix for gc and writeback race
  bcache: bugfix - moving_gc now moves only correct buckets
  bcache: fix for gc crashing when no sectors are used
  bcache: Fix heap_peek() macro
  bcache: Fix for can_attach_cache()
  bcache: Fix dirty_data accounting
  bcache: Use uninterruptible sleep in writeback
  bcache: kthread don't set writeback task to INTERUPTIBLE
  block: fix memory leaks on unplugging block device
  bcache: fix sparse non static symbol warning

Pull cgroup fixes from Tejun Heo:
 "Two fixes.  One fixes a bug in the error path of cgroup_create().  The
  other changes cgrp->id lifetime rule so that the id doesn't get
  recycled before all controller states are destroyed.  This premature
  id recycling made memcg malfunction"

* 'for-3.13-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup:
  cgroup: don't recycle cgroup id until all csses' have been destroyed
  cgroup: fix cgroup_create() error handling path

Pull percpu fix from Tejun Heo:
 "A single commit to fix a spurious sparse warning coming from
  DEFINE_PER_CPU()'s hack to support the use of weak symbols.  Shouldn't
  cause observable behavior change"

* 'for-3.13-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/percpu:
  percpu: fix spurious sparse warnings from DEFINE_PER_CPU()

Pull libata fixes from Tejun Heo:
 "There's one interseting commit - "libata, freezer: avoid block device
  removal while system is frozen".  It's an ugly hack working around a
  deadlock condition between driver core resume and block layer device
  removal paths through freezer which was made more reproducible by
  writeback being converted to workqueue some releases ago.  The bug has
  nothing to do with libata but it's just an workaround which is easy to
  backport.  After discussion, Rafael and I seem to agree that we don't
  really need kernel freezables - both kthread and workqueue.  There are
  few specific workqueues which constitute PM operations and require
  freezing, which will be converted to use workqueue_set_max_active()
  instead.  All other kernel freezer uses are planned to be removed,
  followed by the removal of kthread and workqueue freezer support,
  hopefully.

  Others are device-specific fixes.  The most notable is the addition of
  NO_NCQ_TRIM which is used to disable queued TRIM commands to Micro
  M500 SSDs which otherwise suffers data corruption"

* 'for-3.13-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/libata:
  libata, freezer: avoid block device removal while system is frozen
  libata: implement ATA_HORKAGE_NO_NCQ_TRIM and apply it to Micro M500 SSDs
  libata: disable a disk via libata.force params
  ahci: bail out on ICH6 before using AHCI BAR
  ahci: imx: Explicitly clear IMX6Q_GPR13_SATA_MPLL_CLK_EN
  libata: add ATA_HORKAGE_BROKEN_FPDMA_AA quirk for Seagate Momentus SpinPoint M8

Commit 2171364d ("powerpc: Add HWCAP2 aux entry") introduced a new
AT_ auxv entry type AT_HWCAP2 but failed to update AT_VECTOR_SIZE_BASE
accordingly.
Signed-off-by: NArd Biesheuvel <ard.biesheuvel@linaro.org>
Fixes: 2171364d (powerpc: Add HWCAP2 aux entry)
Cc: stable@vger.kernel.org
Acked-by: NMichael Neuling <michael@neuling.org>
Cc: Nishanth Aravamudan <nacc@linux.vnet.ibm.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Pull SELinux fixes from James Morris.

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/linux-security:
  selinux: selinux_setprocattr()->ptrace_parent() needs rcu_read_lock()
  selinux: fix broken peer recv check

Pull ext2 fix from Jan Kara:
 "One simple fix of oops in ext2 which was recently hit by Christoph"

* 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs:
  ext2: Fix oops in ext2_get_block() called from ext2_quota_write()

Pull infiniband fixes from Roland Dreier:
 "Last batch of InfiniBand/RDMA changes for 3.13 / 2014:
   - Additional checks for uverbs to ensure forward compatibility,
     handle malformed input better.
   - Fix potential use-after-free in iWARP connection manager.
   - Make a function static"

* tag 'rdma-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband:
  IB/uverbs: Check access to userspace response buffer in extended command
  IB/uverbs: Check input length in flow steering uverbs
  IB/uverbs: Set error code when fail to consume all flow_spec items
  IB/uverbs: Check reserved fields in create_flow
  IB/uverbs: Check comp_mask in destroy_flow
  IB/uverbs: Check reserved field in extended command header
  IB/uverbs: New macro to set pointers to NULL if length is 0 in INIT_UDATA()
  IB/core: const'ify inbuf in struct ib_udata
  RDMA/iwcm: Don't touch cm_id after deref in rem_ref
  RDMA/cxgb4: Make _c4iw_write_mem_dma() static

selinux_setprocattr() does ptrace_parent(p) under task_lock(p),
but task_struct->alloc_lock doesn't pin ->parent or ->ptrace,
this looks confusing and triggers the "suspicious RCU usage"
warning because ptrace_parent() does rcu_dereference_check().

And in theory this is wrong, spin_lock()->preempt_disable()
doesn't necessarily imply rcu_read_lock() we need to access
the ->parent.
Reported-by: NEvan McNabb <emcnabb@redhat.com>
Signed-off-by: NOleg Nesterov <oleg@redhat.com>
Cc: stable@vger.kernel.org
Signed-off-by: NPaul Moore <pmoore@redhat.com>

Fix a broken networking check. Return an error if peer recv fails.  If
secmark is active and the packet recv succeeds the peer recv error is
ignored.
Signed-off-by: NChad Hanson <chanson@trustedcs.com>
Cc: stable@vger.kernel.org
Signed-off-by: NPaul Moore <pmoore@redhat.com>

Pull drm fixes from Dave Airlie:
 "Xmas fixes pull, all small nothing major, intel, radeon, one ttm
  regression, and one build fix"

* 'drm-fixes' of git://people.freedesktop.org/~airlied/linux:
  drm/ttm: Fix swapin regression
  gpu: fix qxl missing crc32_le
  drm/radeon: fix asic gfx values for scrapper asics
  drm/i915: Use the correct GMCH_CTRL register for Sandybridge+
  drm/radeon: check for 0 count in speaker allocation and SAD code
  drm/radeon/dpm: disable ss on Cayman
  drm/radeon/dce6: set correct number of audio pins
  drm/i915: get a PC8 reference when enabling the power well
  drm/i915: change CRTC assertion on LCPLL disable
  drm/i915: Fix erroneous dereference of batch_obj inside reset_status
  drm/i915: Prevent double unref following alloc failure during execbuffer

Pull virtio balloon driver fixes from Rusty Russell:
 "Refactoring broke the balloon driver, and fixing kallsyms on ARM broke
  some (non-ARM) MMUless setups, so we're making that fix ARM-only for
  now.

  Unfortunately, the ARM refactoring which broke kallsyms/perf was
  CC:stable, so the fix (which broken non-ARM) was also CC:stable, so
  now the partial reversion is also CC:stable..."

* tag 'fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rusty/linux:
  scripts/link-vmlinux.sh: only filter kernel symbols for arm
  virtio_balloon: update_balloon_size(): update correct field

Merge tag 'drm-intel-fixes-2013-12-18' of git://people.freedesktop.org/~danvet/drm-intel into drm-fixes

Besides the 2 fixes for tricky corner cases in gem from Chris I've
promised already two patche from Paulo to fix pc8 warnings (both ported
from -next, bug report from Dave Jones) and one patch from to fix vga
enable/disable on snb+. That one is a really old bug, but apparently it
can cause machine hangs if you try hard enough with vgacon/efifb handover.

* tag 'drm-intel-fixes-2013-12-18' of git://people.freedesktop.org/~danvet/drm-intel:
  drm/i915: Use the correct GMCH_CTRL register for Sandybridge+
  drm/i915: get a PC8 reference when enabling the power well
  drm/i915: change CRTC assertion on LCPLL disable
  drm/i915: Fix erroneous dereference of batch_obj inside reset_status
  drm/i915: Prevent double unref following alloc failure during execbuffer

- fix for a long standing corruption bug on some Trinity/Richland parts.
- Stability fix for cayman dpm
- audio fixes for dce6+

* 'drm-fixes-3.13' of git://people.freedesktop.org/~agd5f/linux:
  drm/radeon: fix asic gfx values for scrapper asics
  drm/radeon: check for 0 count in speaker allocation and SAD code
  drm/radeon/dpm: disable ss on Cayman
  drm/radeon/dce6: set correct number of audio pins

Commit "drm/ttm: Don't move non-existing data" didn't take the
swapped-out corner case into account. This patch corrects that.
Fixes blank screen after attempted suspend / hibernate on vmwgfx.
Signed-off-by: NThomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Fix build error: qxl uses crc32 functions so it needs to select
CRC32.

Also use angle quotes around a kernel header file name.

drivers/built-in.o: In function `qxl_display_read_client_monitors_config':
(.text+0x19d754): undefined reference to `crc32_le'
Signed-off-by: NRandy Dunlap <rdunlap@infradead.org>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Pull ARM SoC fixes from Olof Johansson:
 "Much smaller batch of fixes this week.

  Biggest one is a revert of an OMAP display change that removed some
  non-DT pinmux code that was still needed for 3.13 to get DSI displays
  to work.

  There's also a fix that resolves some misdescribed GPIO controller
  resources on shmobile.  The rest are mostly smaller fixes, a couple of
  MAINTAINERS updates, etc"

* tag 'fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc:
  Revert "ARM: OMAP2+: Remove legacy mux code for display.c"
  MAINTAINERS: Add keystone clock drivers
  MAINTAINERS: Add keystone git tree information
  ARM: s3c64xx: dt: Fix boot failure due to double clock initialization
  ARM: shmobile: r8a7790: Fix GPIO resources in DTS
  irqchip: renesas-intc-irqpin: Fix register bitfield shift calculation
  ARM: shmobile: lager: phy fixup needs CONFIG_PHYLIB

Pull firewire fixlet from Stefan Richter:
 "A one-liner to reenable WRITE SAME over SBP-2 like in v3.8...v3.12.
  Buggy targets which could malfunction when being subjected to this
  command are already sufficiently protected by a scsi_level check in sd
  + SCSI core"

* tag 'firewire-fix' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394:
  firewire: sbp2: bring back WRITE SAME support

Pull SCSI target fixes from Nicholas Bellinger:
 "Mostly minor items this time around, the most notable being a FILEIO
  backend change to enforce hw_max_sectors based upon the current
  block_size to address a bug where large sized I/Os (> 1M) where being
  rejected"

* git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending:
  qla2xxx: Fix scsi_host leak on qlt_lport_register callback failure
  target: Remove extra percpu_ref_init
  target/file: Update hw_max_sectors based on current block_size
  iser-target: Move INIT_WORK setup into isert_create_device_ib_res
  iscsi-target: Fix incorrect np->np_thread NULL assignment
  qla2xxx: Fix schedule_delayed_work() for target timeout calculations
  iser-target: fix error return code in isert_create_device_ib_res()
  iscsi-target: Fix-up all zero data-length CDBs with R/W_BIT set
  target: Remove write-only stats fields and lock from struct se_node_acl
  iscsi-target: return -EINVAL on oversized configfs parameter

Pull AIO leak fixes from Ben LaHaise:
 "I've put these two patches plus Linus's change through a round of
  tests, and it passes millions of iterations of the aio numa
  migratepage test, as well as a number of repetitions of a few simple
  read and write tests.

  The first patch fixes the memory leak Kent introduced, while the
  second patch makes aio_migratepage() much more paranoid and robust"

* git://git.kvack.org/~bcrl/aio-next:
  aio/migratepages: make aio migrate pages sane
  aio: fix kioctx leak introduced by "aio: Fix a trinity splat"

Since commit 36bc08cc ("fs/aio: Add support to aio ring pages
migration") the aio ring setup code has used a special per-ring backing
inode for the page allocations, rather than just using random anonymous
pages.

However, rather than remembering the pages as it allocated them, it
would allocate the pages, insert them into the file mapping (dirty, so
that they couldn't be free'd), and then forget about them.  And then to
look them up again, it would mmap the mapping, and then use
"get_user_pages()" to get back an array of the pages we just created.

Now, not only is that incredibly inefficient, it also leaked all the
pages if the mmap failed (which could happen due to excessive number of
mappings, for example).

So clean it all up, making it much more straightforward.  Also remove
some left-overs of the previous (broken) mm_populate() usage that was
removed in commit d6c355c7 ("aio: fix race in ring buffer page
lookup introduced by page migration support") but left the pointless and
now misleading MAP_POPULATE flag around.
Tested-and-acked-by: NBenjamin LaHaise <bcrl@kvack.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

The arbitrary restriction on page counts offered by the core
migrate_page_move_mapping() code results in rather suspicious looking
fiddling with page reference counts in the aio_migratepage() operation.
To fix this, make migrate_page_move_mapping() take an extra_count parameter
that allows aio to tell the code about its own reference count on the page
being migrated.

While cleaning up aio_migratepage(), make it validate that the old page
being passed in is actually what aio_migratepage() expects to prevent
misbehaviour in the case of races.
Signed-off-by: NBenjamin LaHaise <bcrl@kvack.org>

e34ecee2 reworked the percpu reference
counting to correct a bug trinity found.  Unfortunately, the change lead
to kioctxes being leaked because there was no final reference count to
put.  Add that reference count back in to fix things.
Signed-off-by: NBenjamin LaHaise <bcrl@kvack.org>
Cc: stable@vger.kernel.org

In the case of both the submit_queues param and use_per_node_hctx param
are used. We limit the number af submit_queues to the number of online
nodes.

If the submit_queues is a multiple of nr_online_nodes, its trivial. Simply map
them to the nodes. For example: 8 submit queues are mapped as node0[0,1],
node1[2,3], ...
If uneven, we are left with an uneven number of submit_queues that must be
mapped. These are mapped toward the first node and onward. E.g. 5
submit queues mapped onto 4 nodes are mapped as node0[0,1], node1[2], ...
Signed-off-by: NMatias Bjorling <m@bjorling.me>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

The defaults for the module is to instantiate itself with blk-mq and a
submit queue for each CPU node in the system.

To save resources, initialize instead with a single submit queue.
Signed-off-by: NMatias Bjorling <m@bjorling.me>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Randy Dunlap reported a couple of grammar errors and unfortunate usages of
socket/node/core.
Signed-off-by: NMatias Bjorling <m@bjorling.me>
Acked-by: NRandy Dunlap <rdunlap@infradead.org>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Commit 1bf49dd4 ("./Makefile: export initial ramdisk compression
config option") started setting the INITRD_COMPRESS environment variable
depending on which decompression models the kernel had available.

That is completely broken.

For example, we by default have CONFIG_RD_LZ4 enabled, and are able to
decompress such an initrd, but the user tools to *create* such an initrd
may not be availble.  So trying to tell dracut to generate an
lz4-compressed image just because we can decode such an image is
completely inappropriate.

Cc: J P <ppandit@redhat.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Jan Beulich <JBeulich@suse.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Pull xfs bugfixes from Ben Myers:
 "This contains fixes for some asserts
   related to project quotas, a memory leak, a hang when disabling group or
   project quotas before disabling user quotas, Dave's email address, several
   fixes for the alignment of file allocation to stripe unit/width geometry, a
   fix for an assertion with xfs_zero_remaining_bytes, and the behavior of
   metadata writeback in the face of IO errors.

   Details:
   - fix memory leak in xfs_dir2_node_removename
   - fix quota assertion in xfs_setattr_size
   - fix quota assertions in xfs_qm_vop_create_dqattach
   - fix for hang when disabling group and project quotas before
     disabling user quotas
   - fix Dave Chinner's email address in MAINTAINERS
   - fix for file allocation alignment
   - fix for assertion in xfs_buf_stale by removing xfsbdstrat
   - fix for alignment with swalloc mount option
   - fix for "retry forever" semantics on IO errors"

* tag 'xfs-for-linus-v3.13-rc5' of git://oss.sgi.com/xfs/xfs:
  xfs: abort metadata writeback on permanent errors
  xfs: swalloc doesn't align allocations properly
  xfs: remove xfsbdstrat error
  xfs: align initial file allocations correctly
  MAINTAINERS: fix incorrect mail address of XFS maintainer
  xfs: fix infinite loop by detaching the group/project hints from user dquot
  xfs: fix assertion failure at xfs_setattr_nonsize
  xfs: fix false assertion at xfs_qm_vop_create_dqattach
  xfs: fix memory leak in xfs_dir2_node_removename

Commit 597d795a ('mm: do not allocate page->ptl dynamically, if
spinlock_t fits to long') restructures some allocators that are compiled
even if USE_SPLIT_PTLOCKS arn't used.  It results in compilation
failure:

  mm/memory.c:4282:6: error: 'struct page' has no member named 'ptl'
  mm/memory.c:4288:12: error: 'struct page' has no member named 'ptl'

Add in the missing ifdef.

Fixes: 597d795a ('mm: do not allocate page->ptl dynamically, if spinlock_t fits to long')
Signed-off-by: NOlof Johansson <olof@lixom.net>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Hugh Dickins <hughd@google.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Pull ARC fix from Vineet Gupta:
 "Fix busted syscall table due to unistd header inclusion issue"

* tag 'arc-fixes-for-3.13-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/vgupta/arc:
  ARC: Allow conditional multiple inclusion of uapi/asm/unistd.h

Pull arm64 ptrace fix from Catalin Marinas.

* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
  arm64: ptrace: avoid using HW_BREAKPOINT_EMPTY for disabled events

Some pstore backing devices use on board flash as persistent
storage. These have limited numbers of write cycles so it
is a poor idea to use them from high frequency operations.
Signed-off-by: NTony Luck <tony.luck@intel.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Merge tag 'dmaengine-fixes-3.13-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/djbw/dmaengine

Pull dmaengine fixes from Dan Williams:

 - deprecation of net_dma to be removed in 3.14

 - crash regression fix in pl330 from the dmaengine_unmap rework

 - crash regression fix for any channel running raid ops without
   CONFIG_ASYNC_TX_DMA from dmaengine_unmap

 - memory leak regression in mv_xor from dmaengine_unmap

 - build warning regressions in mv_xor, fsldma, ppc4xx, txx9, and
   at_hdmac from dmaengine_unmap

 - sleep in atomic regression in dma_async_memcpy_pg_to_pg

 - new fix in mv_xor for handling channel initialization failures

* tag 'dmaengine-fixes-3.13-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/djbw/dmaengine:
  net_dma: mark broken
  dma: pl330: ensure DMA descriptors are zero-initialised
  dmaengine: fix sleep in atomic
  dmaengine: mv_xor: fix oops when channels fail to initialise
  dma: mv_xor: Use dmaengine_unmap_data for the self-tests
  dmaengine: fix enable for high order unmap pools
  dma: fix build warnings in txx9
  dmatest: fix build warning on mips
  dma: fix fsldma build warnings
  dma: fix build warnings in ppc4xx
  dmaengine: at_hdmac: remove unused function
  dma: mv_xor: remove mv_desc_get_dest_addr()

Pull KVM fixes from Paolo Bonzini:
 "The PPC folks had a large amount of changes queued for 3.13, and now
  they are fixing the bugs"

* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
  KVM: PPC: Book3S HV: Don't drop low-order page address bits
  powerpc: book3s: kvm: Don't abuse host r2 in exit path
  powerpc/kvm/booke: Fix build break due to stack frame size warning
  KVM: PPC: Book3S: PR: Enable interrupts earlier
  KVM: PPC: Book3S: PR: Make svcpu -> vcpu store preempt savvy
  KVM: PPC: Book3S: PR: Export kvmppc_copy_to|from_svcpu
  KVM: PPC: Book3S: PR: Don't clobber our exit handler id
  powerpc: kvm: fix rare but potential deadlock scene
  KVM: PPC: Book3S HV: Take SRCU read lock around kvm_read_guest() call
  KVM: PPC: Book3S HV: Make tbacct_lock irq-safe
  KVM: PPC: Book3S HV: Refine barriers in guest entry/exit
  KVM: PPC: Book3S HV: Fix physical address calculations

In struct page we have enough space to fit long-size page->ptl there,
but we use dynamically-allocated page->ptl if size(spinlock_t) is larger
than sizeof(int).

It hurts 64-bit architectures with CONFIG_GENERIC_LOCKBREAK, where
sizeof(spinlock_t) == 8, but it easily fits into struct page.
Signed-off-by: NKirill A. Shutemov <kirill.shutemov@linux.intel.com>
Acked-by: NHugh Dickins <hughd@google.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Commit 81c0a2bb ("mm: page_alloc: fair zone allocator policy") meant
to bring aging fairness among zones in system, but it was overzealous
and badly regressed basic workloads on NUMA systems.

Due to the way kswapd and page allocator interacts, we still want to
make sure that all zones in any given node are used equally for all
allocations to maximize memory utilization and prevent thrashing on the
highest zone in the node.

While the same principle applies to NUMA nodes - memory utilization is
obviously improved by spreading allocations throughout all nodes -
remote references can be costly and so many workloads prefer locality
over memory utilization.  The original change assumed that
zone_reclaim_mode would be a good enough predictor for that, but it
turned out to be as indicative as a coin flip.

Revert the NUMA aspect of the fairness until we can find a proper way to
make it configurable and agree on a sane default.
Signed-off-by: NJohannes Weiner <hannes@cmpxchg.org>
Reviewed-by: NMichal Hocko <mhocko@suse.cz>
Signed-off-by: NMel Gorman <mgorman@suse.de>
Cc: <stable@kernel.org> # 3.12
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>