Use bitmap_set() instead of using set_bit() for each bit.  This conversion
is valid because the bitmap is private in the function call and atomic
bitops were unnecessary.

This also includes minor change.
- Use bitmap_copy() for shorter typing
Signed-off-by: NAkinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

The IPMI watchdog timer clears or extends the timer on reboot/shutdown.
It was using the non-locking routine for setting the watchdog timer, but
this was causing race conditions.  Instead, use the locking version to
avoid the races.  It seems to work fine.
Signed-off-by: NCorey Minyard <cminyard@mvista.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Now that the the IPMI driver is using a tasklet, we can simplify the
locking in the driver and get rid of the message lock.
Signed-off-by: NCorey Minyard <cminyard@mvista.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

The part of the IPMI driver that delivered panic information to the event
log and extended the watchdog timeout during a panic was not properly
handling the messages.  It used static messages to avoid allocation, but
wasn't properly waiting for these, or wasn't properly handling the
refcounts.
Signed-off-by: NCorey Minyard <cminyard@mvista.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

The IPMI driver would release a lock, deliver a message, then relock.
This is obviously ugly, and this patch converts the message handler
interface to use a tasklet to schedule work.  This lets the receive
handler be called from an interrupt handler with interrupts enabled.
Signed-off-by: NCorey Minyard <cminyard@mvista.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

We currently time out and retry KCS transactions after 1 second of waiting
for IBF or OBF.  This appears to be too short for some hardware.  The IPMI
spec says "All system software wait loops should include error timeouts.
For simplicity, such timeouts are not shown explicitly in the flow
diagrams.  A five-second timeout or greater is recommended".  Change the
timeout to five seconds to satisfy the slow hardware.
Signed-off-by: NMatthew Garrett <mjg@redhat.com>
Signed-off-by: NCorey Minyard <cminyard@mvista.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Call the event handler immediately after starting the next message.

This change considerably decreases the IPMI transaction time (cuts off
~9ms for a single ipmitool transaction).
Signed-off-by: NSrinivas_Gowda <srinivas_g_gowda@dell.com>
Signed-off-by: NCorey Minyard <cminyard@mvista.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

crashkernel reservation need know the total memory size.  Current
get_total_mem simply use max_pfn - min_low_pfn.  It is wrong because it
will including memory holes in the middle.

Especially for kvm guest with memory > 0xe0000000, there's below in qemu
code: qemu split memory as below:

    if (ram_size >= 0xe0000000 ) {
        above_4g_mem_size = ram_size - 0xe0000000;
        below_4g_mem_size = 0xe0000000;
    } else {
        below_4g_mem_size = ram_size;
    }

So for 4G mem guest, seabios will insert a 512M usable region beyond of
4G.  Thus in above case max_pfn - min_low_pfn will be more than original
memsize.

Fixing this issue by using memblock_phys_mem_size() to get the total
memsize.
Signed-off-by: NDave Young <dyoung@redhat.com>
Reviewed-by: NWANG Cong <xiyou.wangcong@gmail.com>
Reviewed-by: NSimon Horman <horms@verge.net.au>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

When using crashkernel=2M-256M, the kernel doesn't give any warning.  This
is misleading sometimes.
Signed-off-by: NZhenzhong Duan <zhenzhong.duan@oracle.com>
Acked-by: N"Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

nommu platforms don't have very interesting swapper_pg_dir pointers and
usually just #define them to NULL, meaning that we can't include them in
the vmcoreinfo on the kexec crash path.

This patch only saves the swapper_pg_dir if we have an MMU.
Signed-off-by: NWill Deacon <will.deacon@arm.com>
Reviewed-by: NSimon Horman <horms@verge.net.au>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

This was marked as obsolete for quite a while now..  Now it is time to
remove it altogether.  And while doing this, get rid of first_cpu() as
well.  Also, remove the redundant setting of cpu_online_mask in
smp_prepare_cpus() because the generic code would have already set cpu 0
in cpu_online_mask.
Reported-by: NTony Luck <tony.luck@intel.com>
Signed-off-by: NSrivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

__any_online_cpu() is not optimal and also unnecessary.  So, replace its
use by faster cpumask_* operations.
Signed-off-by: NSrivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Venkatesh Pallipadi <venki@google.com>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Calculate a cpumask of CPUs with per-cpu pages in any zone and only send
an IPI requesting CPUs to drain these pages to the buddy allocator if they
actually have pages when asked to flush.

This patch saves 85%+ of IPIs asking to drain per-cpu pages in case of
severe memory pressure that leads to OOM since in these cases multiple,
possibly concurrent, allocation requests end up in the direct reclaim code
path so when the per-cpu pages end up reclaimed on first allocation
failure for most of the proceeding allocation attempts until the memory
pressure is off (possibly via the OOM killer) there are no per-cpu pages
on most CPUs (and there can easily be hundreds of them).

This also has the side effect of shortening the average latency of direct
reclaim by 1 or more order of magnitude since waiting for all the CPUs to
ACK the IPI takes a long time.

Tested by running "hackbench 400" on a 8 CPU x86 VM and observing the
difference between the number of direct reclaim attempts that end up in
drain_all_pages() and those were more then 1/2 of the online CPU had any
per-cpu page in them, using the vmstat counters introduced in the next
patch in the series and using proc/interrupts.

In the test sceanrio, this was seen to save around 3600 global
IPIs after trigerring an OOM on a concurrent workload:

$ cat /proc/vmstat | tail -n 2
pcp_global_drain 0
pcp_global_ipi_saved 0

$ cat /proc/interrupts | grep CAL
CAL:          1          2          1          2
          2          2          2          2   Function call interrupts

$ hackbench 400
[OOM messages snipped]

$ cat /proc/vmstat | tail -n 2
pcp_global_drain 3647
pcp_global_ipi_saved 3642

$ cat /proc/interrupts | grep CAL
CAL:          6         13          6          3
          3          3         1 2          7   Function call interrupts

Please note that if the global drain is removed from the direct reclaim
path as a patch from Mel Gorman currently suggests this should be replaced
with an on_each_cpu_cond invocation.
Signed-off-by: NGilad Ben-Yossef <gilad@benyossef.com>
Acked-by: NMel Gorman <mel@csn.ul.ie>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Acked-by: NChristoph Lameter <cl@linux.com>
Acked-by: NPeter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Andi Kleen <andi@firstfloor.org>
Acked-by: NMichal Nazarewicz <mina86@mina86.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

In several code paths, such as when unmounting a file system (but not
only) we send an IPI to ask each cpu to invalidate its local LRU BHs.

For multi-cores systems that have many cpus that may not have any LRU BH
because they are idle or because they have not performed any file system
accesses since last invalidation (e.g.  CPU crunching on high perfomance
computing nodes that write results to shared memory or only using
filesystems that do not use the bh layer.) This can lead to loss of
performance each time someone switches the KVM (the virtual keyboard and
screen type, not the hypervisor) if it has a USB storage stuck in.

This patch attempts to only send an IPI to cpus that have LRU BH.
Signed-off-by: NGilad Ben-Yossef <gilad@benyossef.com>
Acked-by: NPeter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

flush_all() is called for each kmem_cache_destroy().  So every cache being
destroyed dynamically ends up sending an IPI to each CPU in the system,
regardless if the cache has ever been used there.

For example, if you close the Infinband ipath driver char device file, the
close file ops calls kmem_cache_destroy().  So running some infiniband
config tool on one a single CPU dedicated to system tasks might interrupt
the rest of the 127 CPUs dedicated to some CPU intensive or latency
sensitive task.

I suspect there is a good chance that every line in the output of "git
grep kmem_cache_destroy linux/ | grep '\->'" has a similar scenario.

This patch attempts to rectify this issue by sending an IPI to flush the
per cpu objects back to the free lists only to CPUs that seem to have such
objects.

The check which CPU to IPI is racy but we don't care since asking a CPU
without per cpu objects to flush does no damage and as far as I can tell
the flush_all by itself is racy against allocs on remote CPUs anyway, so
if you required the flush_all to be determinstic, you had to arrange for
locking regardless.

Without this patch the following artificial test case:

$ cd /sys/kernel/slab
$ for DIR in *; do cat $DIR/alloc_calls > /dev/null; done

produces 166 IPIs on an cpuset isolated CPU. With it it produces none.

The code path of memory allocation failure for CPUMASK_OFFSTACK=y
config was tested using fault injection framework.
Signed-off-by: NGilad Ben-Yossef <gilad@benyossef.com>
Acked-by: NChristoph Lameter <cl@linux.com>
Cc: Chris Metcalf <cmetcalf@tilera.com>
Acked-by: NPeter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: Matt Mackall <mpm@selenic.com>
Cc: Sasha Levin <levinsasha928@gmail.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Mel Gorman <mel@csn.ul.ie>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Avi Kivity <avi@redhat.com>
Cc: Michal Nazarewicz <mina86@mina86.org>
Cc: Kosaki Motohiro <kosaki.motohiro@gmail.com>
Cc: Milton Miller <miltonm@bga.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Add the on_each_cpu_cond() function that wraps on_each_cpu_mask() and
calculates the cpumask of cpus to IPI by calling a function supplied as a
parameter in order to determine whether to IPI each specific cpu.

The function works around allocation failure of cpumask variable in
CONFIG_CPUMASK_OFFSTACK=y by itereating over cpus sending an IPI a time
via smp_call_function_single().

The function is useful since it allows to seperate the specific code that
decided in each case whether to IPI a specific cpu for a specific request
from the common boilerplate code of handling creating the mask, handling
failures etc.

[akpm@linux-foundation.org: s/gfpflags/gfp_flags/]
[akpm@linux-foundation.org: avoid double-evaluation of `info' (per Michal), parenthesise evaluation of `cond_func']
[akpm@linux-foundation.org: s/CPU/CPUs, use all 80 cols in comment]
Signed-off-by: NGilad Ben-Yossef <gilad@benyossef.com>
Cc: Chris Metcalf <cmetcalf@tilera.com>
Cc: Christoph Lameter <cl@linux-foundation.org>
Acked-by: NPeter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: Matt Mackall <mpm@selenic.com>
Cc: Sasha Levin <levinsasha928@gmail.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Avi Kivity <avi@redhat.com>
Acked-by: NMichal Nazarewicz <mina86@mina86.org>
Cc: Kosaki Motohiro <kosaki.motohiro@gmail.com>
Cc: Milton Miller <miltonm@bga.com>
Reviewed-by: N"Srivatsa S. Bhat" <srivatsa.bhat@linux.vnet.ibm.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

We have lots of infrastructure in place to partition multi-core systems
such that we have a group of CPUs that are dedicated to specific task:
cgroups, scheduler and interrupt affinity, and cpuisol= boot parameter.
Still, kernel code will at times interrupt all CPUs in the system via IPIs
for various needs.  These IPIs are useful and cannot be avoided
altogether, but in certain cases it is possible to interrupt only specific
CPUs that have useful work to do and not the entire system.

This patch set, inspired by discussions with Peter Zijlstra and Frederic
Weisbecker when testing the nohz task patch set, is a first stab at trying
to explore doing this by locating the places where such global IPI calls
are being made and turning the global IPI into an IPI for a specific group
of CPUs.  The purpose of the patch set is to get feedback if this is the
right way to go for dealing with this issue and indeed, if the issue is
even worth dealing with at all.  Based on the feedback from this patch set
I plan to offer further patches that address similar issue in other code
paths.

This patch creates an on_each_cpu_mask() and on_each_cpu_cond()
infrastructure API (the former derived from existing arch specific
versions in Tile and Arm) and uses them to turn several global IPI
invocation to per CPU group invocations.

Core kernel:

on_each_cpu_mask() calls a function on processors specified by cpumask,
which may or may not include the local processor.

You must not call this function with disabled interrupts or from a
hardware interrupt handler or from a bottom half handler.

arch/arm:

Note that the generic version is a little different then the Arm one:

1. It has the mask as first parameter
2. It calls the function on the calling CPU with interrupts disabled,
   but this should be OK since the function is called on the other CPUs
   with interrupts disabled anyway.

arch/tile:

The API is the same as the tile private one, but the generic version
also calls the function on the with interrupts disabled in UP case

This is OK since the function is called on the other CPUs
with interrupts disabled.
Signed-off-by: NGilad Ben-Yossef <gilad@benyossef.com>
Reviewed-by: NChristoph Lameter <cl@linux.com>
Acked-by: NChris Metcalf <cmetcalf@tilera.com>
Acked-by: NPeter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: Matt Mackall <mpm@selenic.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Sasha Levin <levinsasha928@gmail.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Avi Kivity <avi@redhat.com>
Acked-by: NMichal Nazarewicz <mina86@mina86.org>
Cc: Kosaki Motohiro <kosaki.motohiro@gmail.com>
Cc: Milton Miller <miltonm@bga.com>
Cc: Russell King <linux@arm.linux.org.uk>
Acked-by: NPeter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Most system calls taking flags first check that the flags passed in are
valid, and that helps userspace to detect when new flags are supported.

But swapon never did so: start checking now, to help if we ever want to
support more swap_flags in future.

It's difficult to get stray bits set in an int, and swapon is not widely
used, so this is most unlikely to break any userspace; but we can just
revert if it turns out to do so.
Signed-off-by: NHugh Dickins <hughd@google.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

The size of coredump files is limited by RLIMIT_CORE, however, allocating
large amounts of memory results in three negative consequences:

 - the coredumping process may be chosen for oom kill and quickly deplete
   all memory reserves in oom conditions preventing further progress from
   being made or tasks from exiting,

 - the coredumping process may cause other processes to be oom killed
   without fault of their own as the result of a SIGSEGV, for example, in
   the coredumping process, or

 - the coredumping process may result in a livelock while writing to the
   dump file if it needs memory to allocate while other threads are in
   the exit path waiting on the coredumper to complete.

This is fixed by implying __GFP_NORETRY in the page allocator for
coredumping processes when reclaim has failed so the allocations fail and
the process continues to exit.
Signed-off-by: NDavid Rientjes <rientjes@google.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

pmd_trans_unstable() should be called before pmd_offset_map() in the
locations where the mmap_sem is held for reading.
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Hugh Dickins <hughd@google.com>
Cc: Larry Woodman <lwoodman@redhat.com>
Cc: Ulrich Obergfell <uobergfe@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Mark Salter <msalter@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Holepunching filesystems ext4 and xfs are using truncate_inode_pages_range
but forgetting to unmap pages first (ocfs2 remembers).  This is not really
a bug, since races already require truncate_inode_page() to handle that
case once the page is locked; but it can be very inefficient if the file
being punched happens to be mapped into many vmas.

Provide a drop-in replacement truncate_pagecache_range() which does the
unmapping pass first, handling the awkward mismatch between arguments to
truncate_inode_pages_range() and arguments to unmap_mapping_range().

Note that holepunching does not unmap privately COWed pages in the range:
POSIX requires that we do so when truncating, but it's hard to justify,
difficult to implement without an i_size cutoff, and no filesystem is
attempting to implement it.
Signed-off-by: NHugh Dickins <hughd@google.com>
Cc: "Theodore Ts'o" <tytso@mit.edu>
Cc: Andreas Dilger <adilger.kernel@dilger.ca>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Ben Myers <bpm@sgi.com>
Cc: Alex Elder <elder@kernel.org>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Dave Chinner <david@fromorbit.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

bda7bad6 ("procfs: speed up /proc/pid/stat, statm") broke /proc/statm
- 'text' is printed twice by mistake.
Signed-off-by: NKAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Reported-by: NUlrich Drepper <drepper@gmail.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Pull trivial writeback fixes from Wu Fengguang:
 "They've been tested in linux-next for 20 days actually."

* tag 'writeback-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/wfg/linux:
  writeback: Remove outdated comment
  fs: Remove bogus wait in write_inode_now()

Pull ext4 updates for 3.4 from Ted Ts'o:
 "Ext4 commits for 3.3 merge window; mostly cleanups and bug fixes

  The changes to export dirty_writeback_interval are from Artem's s_dirt
  cleanup patch series.  The same is true of the change to remove the
  s_dirt helper functions which never got used by anyone in-tree.  I've
  run these changes by Al Viro, and am carrying them so that Artem can
  more easily fix up the rest of the file systems during the next merge
  window.  (Originally we had hopped to remove the use of s_dirt from
  ext4 during this merge window, but his patches had some bugs, so I
  ultimately ended dropping them from the ext4 tree.)"

* tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4: (66 commits)
  vfs: remove unused superblock helpers
  mm: export dirty_writeback_interval
  ext4: remove useless s_dirt assignment
  ext4: write superblock only once on unmount
  ext4: do not mark superblock as dirty unnecessarily
  ext4: correct ext4_punch_hole return codes
  ext4: remove restrictive checks for EOFBLOCKS_FL
  ext4: always set then trimmed blocks count into len
  ext4: fix trimmed block count accunting
  ext4: fix start and len arguments handling in ext4_trim_fs()
  ext4: update s_free_{inodes,blocks}_count during online resize
  ext4: change some printk() calls to use ext4_msg() instead
  ext4: avoid output message interleaving in ext4_error_<foo>()
  ext4: remove trailing newlines from ext4_msg() and ext4_error() messages
  ext4: add no_printk argument validation, fix fallout
  ext4: remove redundant "EXT4-fs: " from uses of ext4_msg
  ext4: give more helpful error message in ext4_ext_rm_leaf()
  ext4: remove unused code from ext4_ext_map_blocks()
  ext4: rewrite punch hole to use ext4_ext_remove_space()
  jbd2: cleanup journal tail after transaction commit
  ...

Pull Ceph updates for 3.4-rc1 from Sage Weil:
 "Alex has been busy.  There are a range of rbd and libceph cleanups,
  especially surrounding device setup and teardown, and a few critical
  fixes in that code.  There are more cleanups in the messenger code,
  virtual xattrs, a fix for CRC calculation/checks, and lots of other
  miscellaneous stuff.

  There's a patch from Amon Ott to make inos behave a bit better on
  32-bit boxes, some decode check fixes from Xi Wang, and network
  throttling fix from Jim Schutt, and a couple RBD fixes from Josh
  Durgin.

  No new functionality, just a lot of cleanup and bug fixing."

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client: (65 commits)
  rbd: move snap_rwsem to the device, rename to header_rwsem
  ceph: fix three bugs, two in ceph_vxattrcb_file_layout()
  libceph: isolate kmap() call in write_partial_msg_pages()
  libceph: rename "page_shift" variable to something sensible
  libceph: get rid of zero_page_address
  libceph: only call kernel_sendpage() via helper
  libceph: use kernel_sendpage() for sending zeroes
  libceph: fix inverted crc option logic
  libceph: some simple changes
  libceph: small refactor in write_partial_kvec()
  libceph: do crc calculations outside loop
  libceph: separate CRC calculation from byte swapping
  libceph: use "do" in CRC-related Boolean variables
  ceph: ensure Boolean options support both senses
  libceph: a few small changes
  libceph: make ceph_tcp_connect() return int
  libceph: encapsulate some messenger cleanup code
  libceph: make ceph_msgr_wq private
  libceph: encapsulate connection kvec operations
  libceph: move prepare_write_banner()
  ...

Pull ext3, UDF, and quota fixes from Jan Kara:
 "A couple of ext3 & UDF fixes and also one improvement in quota
  locking."

* 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs:
  ext3: fix start and len arguments handling in ext3_trim_fs()
  udf: Fix deadlock in udf_release_file()
  udf: Fix file entry logicalBlocksRecorded
  udf: Fix handling of i_blocks
  quota: Make quota code not call tty layer with dqptr_sem held
  udf: Init/maintain file entry checkpoint field
  ext3: Update ctime in ext3_splice_branch() only when needed
  ext3: Don't call dquot_free_block() if we don't update anything
  udf: Remove unnecessary OOM messages

Pull 9p changes for the 3.4 merge window from Eric Van Hensbergen.

* tag 'for-linus-3.4-merge-window' of git://git.kernel.org/pub/scm/linux/kernel/git/ericvh/v9fs:
  9p: statfs should not override server f_type
  net/9p: handle flushed Tclunk/Tremove
  net/9p: don't allow Tflush to be interrupted

In d_materialise_unique() there are 3 subcases to the 'aliased dentry'
case; in two subcases the inode i_lock is properly released but this
does not occur in the -ELOOP subcase.

This seems to have been introduced by commit 18367501 ("fix loop
checks in d_materialise_unique()").
Signed-off-by: NMichel Lespinasse <walken@google.com>
Cc: stable@vger.kernel.org # v3.0+
[ Added a comment, and moved the unlock to where we generate the -ELOOP,
  which seems to be more natural.

  You probably can't actually trigger this without a buggy network file
  server - d_materialize_unique() is for finding aliases on non-local
  filesystems, and the d_ancestor() case is for a hardlinked directory
  loop.

  But we should be robust in the case of such buggy servers anyway. ]
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Pull s390 patches part 2 from Martin Schwidefsky:
 "Some minor improvements and one additional feature for the 3.4 merge
  window: Hendrik added perf support for the s390 CPU counters."

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux:
  [S390] register cpu devices for SMP=n
  [S390] perf: add support for s390x CPU counters
  [S390] oprofile: Allow multiple users of the measurement alert interrupt
  [S390] qdio: log all adapter characteristics
  [S390] Remove unncessary export of arch_pick_mmap_layout

Pull UML changes from Richard Weinberger:
 "Mostly bug fixes and cleanups"

* 'for-linus-3.4-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rw/uml: (35 commits)
  um: Update defconfig
  um: Switch to large mcmodel on x86_64
  MTD: Relax dependencies
  um: Wire CONFIG_GENERIC_IO up
  um: Serve io_remap_pfn_range()
  Introduce CONFIG_GENERIC_IO
  um: allow SUBARCH=x86
  um: most of the SUBARCH uses can be killed
  um: deadlock in line_write_interrupt()
  um: don't bother trying to rebuild CHECKFLAGS for USER_OBJS
  um: use the right ifdef around exports in user_syms.c
  um: a bunch of headers can be killed by using generic-y
  um: ptrace-generic.h doesn't need user.h
  um: kill HOST_TASK_PID
  um: remove pointless include of asm/fixmap.h from asm/pgtable.h
  um: asm-offsets.h might as well come from underlying arch...
  um: merge processor_{32,64}.h a bit...
  um: switch close_chan() to struct line
  um: race fix: initialize delayed_work *before* registering IRQ
  um: line->have_irq is never checked...
  ...

Pull arch/microblaze fixes from Michal Simek

* 'next' of git://git.monstr.eu/linux-2.6-microblaze:
  microblaze: Handle TLB skip size dynamically
  microblaze: Introduce TLB skip size
  microblaze: Improve TLB calculation for small systems
  microblaze: Extend space for compiled-in FDT to 32kB
  microblaze: Clear all MSR flags on the first kernel instruction
  microblaze: Use node name instead of compatible string
  microblaze: Fix mapin_ram function
  microblaze: Highmem support
  microblaze: Use active regions
  microblaze: Show more detailed information about memory
  microblaze: Introduce fixmap
  microblaze: mm: Fix lowmem max memory size limits
  microblaze: mm: Use ZONE_DMA instead of ZONE_NORMAL
  microblaze: trivial: Fix typo fault in timer.c
  microblaze: Use vsprintf extention %pf with builtin_return_address
  microblaze: Add PVR version string for MB 8.20.b and 8.30.a
  microblaze: Fix makefile to work with latest toolchain
  microblaze: Fix typo in early_printk.c

Pull ARM platform updates from Russell King:
 "This covers platform stuff for platforms I have a direct interest in
  (iow, I have the hardware).  Essentially:
   - as we no longer support any other Acorn platforms other than RiscPC
     anymore, we can collect all that code into mach-rpc.
   - convert Acorn expansion card stuff to use IRQ allocation functions,
     and get rid of NO_IRQ from there.
   - cleanups to the ebsa110 platform to move some private stuff out of
     its header files.
   - large amount of SA11x0 updates:
   - conversion of private DMA implementation to DMA engine support
     (this actually gives us greater flexibility in drivers over the old
     API.)
   - re-worked ucb1x00 updates - convert to genirq, remove sa11x0
     dependencies, fix various minor issues
   - move platform specific sa11x0 framebuffer data into platform files
     in arch/arm instead of keeping this in the driver itself
   - update sa11x0 IrDA driver for DMA engine, and allow it to use DMA
     for SIR transmissions as well as FIR
   - rework sa1111 support for genirq, and irq allocation
   - fix sa1111 IRQ support so it works again
   - use sparse IRQ support

  After this, I have one more pull request remaining from my current
  set, which I think is going to be the most problematical as it
  generates 8 conflicts."

Fixed up the trivial conflict in arch/arm/mach-rpc/Makefile as per
Russell.

* 'platforms' of git://git.linaro.org/people/rmk/linux-arm: (125 commits)
  ARM: 7343/1: sa11x0: convert to sparse IRQ
  ARM: 7342/2: sa1100: prepare for sparse irq conversion
  ARM: 7341/1: input: prepare jornada720 keyboard and ts for sa11x0 sparse irq
  ARM: 7340/1: rtc: sa1100: include mach/irqs.h instead of asm/irq.h
  ARM: sa11x0: remove unused DMA controller definitions
  ARM: sa11x0: remove old SoC private DMA driver
  USB: sa1111: add hcd .reset method
  USB: sa1111: add OHCI shutdown methods
  USB: sa1111: reorganize ohci-sa1111.c
  USB: sa1111: get rid of nasty printk(KERN_DEBUG "%s: ...", __FILE__)
  USB: sa1111: sparse and checkpatch cleanups
  ARM: sa11x0: don't static map sa1111
  ARM: sa1111: use dev_err() rather than printk()
  ARM: sa1111: cleanup sub-device registration and unregistration
  ARM: sa1111: only setup DMA for DMA capable devices
  ARM: sa1111: register sa1111 devices with dmabounce in bus notifier
  ARM: sa1111: move USB interface register definitions to ohci-sa1111.c
  ARM: sa1111: move PCMCIA interface register definitions to sa1111_generic.c
  ARM: sa1111: move PS/2 interface register definitions to sa1111p2.c
  ARM: sa1111: delete unused physical GPIO register definitions
  ...

Pull networking fixes from David Miller:
 1) Name string overrun fix in gianfar driver from Joe Perches.

 2) VHOST bug fixes from Michael S. Tsirkin and Nadav Har'El

 3) Fix dependencies on xt_LOG netfilter module, from Pablo Neira Ayuso.

 4) Fix RCU locking in xt_CT, also from Pablo Neira Ayuso.

 5) Add a parameter to skb_add_rx_frag() so we can fix the truesize
    adjustments in the drivers that use it.  The individual drivers
    aren't fixed by this commit, but will be dealt with using follow-on
    commits.  From Eric Dumazet.

 6) Add some device IDs to qmi_wwan driver, from Andrew Bird.

 7) Fix a potential rcu_read_lock() imbalancein rt6_fill_node().  From
    Eric Dumazet.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net:
  net: fix a potential rcu_read_lock() imbalance in rt6_fill_node()
  net: add a truesize parameter to skb_add_rx_frag()
  gianfar: Fix possible overrun and simplify interrupt name field creation
  USB: qmi_wwan: Add ZTE (Vodafone) K3570-Z and K3571-Z net interfaces
  USB: option: Ignore ZTE (Vodafone) K3570/71 net interfaces
  USB: qmi_wwan: Add ZTE (Vodafone) K3565-Z and K4505-Z net interfaces
  qlcnic: Bug fix for LRO
  netfilter: nf_conntrack: permanently attach timeout policy to conntrack
  netfilter: xt_CT: fix assignation of the generic protocol tracker
  netfilter: xt_CT: missing rcu_read_lock section in timeout assignment
  netfilter: cttimeout: fix dependency with l4protocol conntrack module
  netfilter: xt_LOG: use CONFIG_IP6_NF_IPTABLES instead of CONFIG_IPV6
  vhost: fix release path lockdep checks
  vhost: don't forget to schedule()
  tools/virtio: stub out strong barriers
  tools/virtio: add linux/hrtimer.h stub
  tools/virtio: add linux/module.h stub

Pull "ARM: device tree work" from Arnd Bergmann:
 "Most of these patches convert code from using static platform data to
  describing the hardware in the device tree.  This is only the first
  half of the changes for v3.4 because a lot of patches for this topic
  came in the last week before the merge window.

  Signed-off-by: Arnd Bergmann <arnd@arndb.de>"

Fix up trivial conflicts in arch/arm/mach-vexpress/{Kconfig,core.h}

* tag 'dt' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (86 commits)
  Document: devicetree: add OF documents for arch-mmp
  ARM: dts: append DTS file of pxa168
  ARM: mmp: append OF support on pxa168
  ARM: mmp: enable rtc clk in pxa168
  i2c: pxa: add OF support
  serial: pxa: add OF support
  arm/dts: mt_ventoux: very basic support for TeeJet Mt.Ventoux board
  ARM: OMAP2+: Remove extra ifdefs for board-generic
  ARM: OMAP2+: Fix build error when only ARCH_OMAP2/3 or 4 is selected
  ASoC: DT: Add digital microphone binding to PAZ00 board.
  ARM: dt: Add ARM PMU to tegra*.dtsi
  ARM: at91: at91sam9x5cm/dt: add leds support
  ARM: at91: usb_a9g20/dt: add gpio-keys support
  ARM: at91: at91sam9m10g45ek/dt: add gpio-keys support
  ARM: at91: at91sam9m10g45ek/dt: add leds support
  ARM: at91: usb_a9g20/dt: add leds support
  ARM: at91/pio: add new PIO3 features
  ARM: at91: add sam9_smc.o to at91sam9x5 build
  ARM: at91/tc/clocksource: Add 32 bit variant to Timer Counter
  ARM: at91/tc: add device tree support to atmel_tclib
  ...

Pull "ARM: driver specific updates" from Arnd Bergmann:
 "These are all specific to some driver.  They are typically the
  platform side of a change in the drivers directory, such as adding a
  new driver or extending the interface to the platform.  In cases where
  there is no maintainer for the driver, or the maintainer prefers to
  have the platform changes in the same branch as the driver changes,
  the patches to the drivers are included as well.

  A much smaller set of driver updates that depend on other branches
  getting merged first will be sent later.

  The new export of tegra_chip_uid conflicts with other changes in
  fuse.c.  In rtc-sa1100.c, the global removal of IRQF_DISABLED
  conflicts with the cleanup of the interrupt handling of that driver.

  Signed-off-by: Arnd Bergmann <arnd@arndb.de>"

Fixed up aforementioned trivial conflicts.

* tag 'drivers' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (94 commits)
  ARM: SAMSUNG: change the name from s3c-sdhci to exynos4-sdhci
  mmc: sdhci-s3c: add platform data for the second capability
  ARM: SAMSUNG: support the second capability for samsung-soc
  ARM: EXYNOS: add support DMA for EXYNOS4X12 SoC
  ARM: EXYNOS: Add apb_pclk clkdev entry for mdma1
  ARM: EXYNOS: Enable MDMA driver
  regulator: Remove bq24022 regulator driver
  rtc: sa1100: add OF support
  pxa: magician/hx4700: Convert to gpio-regulator from bq24022
  ARM: OMAP3+: SmartReflex: fix error handling
  ARM: OMAP3+: SmartReflex: fix the use of debugfs_create_* API
  ARM: OMAP3+: SmartReflex: micro-optimization for sanity check
  ARM: OMAP3+: SmartReflex: misc cleanups
  ARM: OMAP3+: SmartReflex: move late_initcall() closer to its argument
  ARM: OMAP3+: SmartReflex: add missing platform_set_drvdata()
  ARM: OMAP3+: hwmod: add SmartReflex IRQs
  ARM: OMAP3+: SmartReflex: clear ERRCONFIG_VPBOUNDINTST only on a need
  ARM: OMAP3+: SmartReflex: Fix status masking in ERRCONFIG register
  ARM: OMAP3+: SmartReflex: Add a shutdown hook
  ARM: OMAP3+: SmartReflex Class3: disable errorgen before disable VP
  ...

Conflicts:
	arch/arm/mach-tegra/Makefile
	arch/arm/mach-tegra/fuse.c
	drivers/rtc/rtc-sa1100.c

Pull "remoteproc/rpmsg: new subsystem" from Arnd Bergmann:
 "This new subsystem provides a common way to talk to secondary
  processors on an SoC, e.g.  a DSP, GPU or service processor, using
  virtio as the transport.  In the long run, it should replace a few
  dozen vendor specific ways to do the same thing, which all never made
  it into the upstream kernel.  There is a broad agreement that rpmsg is
  the way to go here and several vendors have started working on
  replacing their own subsystems.

  Two branches each add one virtio protocol number.  Fortunately the
  numbers were agreed upon in advance, so there are only context
  changes.

  Signed-off-by: Arnd Bergmann <arnd@arndb.de>"

Fixed up trivial protocol number conflict due to the mentioned additions
next to each other.

* tag 'rpmsg' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (32 commits)
  remoteproc: cleanup resource table parsing paths
  remoteproc: remove the hardcoded vring alignment
  remoteproc/omap: remove the mbox_callback limitation
  remoteproc: remove the single rpmsg vdev limitation
  remoteproc: safer boot/shutdown order
  remoteproc: remoteproc_rpmsg -> remoteproc_virtio
  remoteproc: resource table overhaul
  rpmsg: fix build warning when dma_addr_t is 64-bit
  rpmsg: fix published buffer length in rpmsg_recv_done
  rpmsg: validate incoming message length before propagating
  rpmsg: fix name service endpoint leak
  remoteproc/omap: two Kconfig fixes
  remoteproc: make sure we're parsing a 32bit firmware
  remoteproc: s/big switch/lookup table/
  remoteproc: bail out if firmware has different endianess
  remoteproc: don't use virtio's weak barriers
  rpmsg: rename virtqueue_add_buf_gfp to virtqueue_add_buf
  rpmsg: depend on EXPERIMENTAL
  remoteproc: depend on EXPERIMENTAL
  rpmsg: add Kconfig menu
  ...

Conflicts:
	include/linux/virtio_ids.h

Pull "ARM: board specific updates" from Arnd Bergmann/Olof Johansson:
 "These changes are all specific to one board only.  We're trying to
  keep the number of board files low, but generally board level updates
  are ok on platforms that are working on moving towards DT based
  probing, which will eventually lead to removing them.

  The board-ams-delta.c board file gets a conflict between the removal
  of ams_delta_config and the addition of a lot of other data.  The
  Kconfig file has two changes in the same line, and in exynos, the
  power domain cleanup conflicts with the addition of the image sensor
  device.
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
  [olof: Amended a fix for a mismerge to board-omap4panda.c]
  Signed-off-by: Olof Johansson <olof@lixom.net>"

Fixed up some fairly trivial conflicts manually.

* tag 'boards' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (82 commits)
  i.MX35-PDK: Add Camera support
  ARM : mx35: 3ds-board: add framebuffer device
  pxa/hx4700: Remove pcmcia platform_device structure
  ARM: pxa/hx4700: Reduce sleep mode battery discharge by 35%
  ARM: pxa/hx4700: Remove unwanted request for GPIO105
  ARM: EXYNOS: support Exynos4210-bus Devfreq driver on Nuri board
  ARM: EXYNOS: Register JPEG on nuri
  ARM: EXYNOS: Register JPEG on universal_c210
  ARM: S5PV210: Enable JPEG on SMDKV210
  ARM: S5PV210: Add JPEG board definition
  ARM: EXYNOS: Enable JPEG on Origen
  ARM: EXYNOS: Enable JPEG on SMDKV310
  ARM: EXYNOS: Add __init attribute to universal_camera_init()
  ARM: EXYNOS: Add __init attribute to nuri_camera_init()
  ARM: S5PV210: Enable FIMC on SMDKC110
  ARM: S5PV210: Enable FIMC on SMDKV210
  ARM: S5PV210: Enable MFC on SMDKC110
  ARM: S5PV210: Enable MFC on SMDKV210
  ARM: EXYNOS: Enable G2D on SMDKV310
  ARM: tegra: update defconfig
  ...

Pull "ARM: SoC specific updates" from Arnd Bergmann:
 "These changes are all specific to an soc family or the code for one
  soc.  Lots of work for Tegra3 this time, but also a lot of other
  platforms.  There will be another (smaller) set of soc patches later
  in the merge window for stuff that has dependencies on external trees
  or that was sent just before the merge window opened.

  The asoc tree added a few devices to the i.mx platform, which conflict
  with other devices added in the same place here.

  The tegra Makefile conflicts between a number of branches, mostly
  because of changes regarding localtimer.c, which was removed in the
  end.

  Signed-off-by: Arnd Bergmann <arnd@arndb.de>"

Fix up some trivial conflicts, including the mentioned Tegra Makefile.

* tag 'soc' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (51 commits)
  ARM: EXYNOS: fix cycle count for periodic mode of clock event timers
  ARM: EXYNOS: add support JPEG
  ARM: EXYNOS: Add DMC1, allow PPMU access for DMC
  ARM: SAMSUNG: Correct MIPI-CSIS io memory resource definition
  ARM: SAMSUNG: fix __init attribute on regarding s3c_set_platdata()
  ARM: SAMSUNG: Add __init attribute to samsung_bl_set()
  ARM: S5PV210: Add usb otg phy control
  ARM: S3C64XX: Add usb otg phy control
  ARM: EXYNOS: Enable l2 configuration through device tree
  ARM: EXYNOS: remove useless code to save/restore L2
  ARM: EXYNOS: save L2 settings during bootup
  ARM: S5P: add L2 early resume code
  ARM: EXYNOS: Add support AFTR mode on EXYNOS4210
  ARM: mx35: Setup the AIPS registers
  ARM: mx5: Use common function for configuring AIPS
  ARM: mx3: Setup AIPS registers
  ARM: mx3: Let mx31 and mx35 enter in LPM mode in WFI
  ARM: defconfig: imx_v6_v7: build in REGULATOR_FIXED_VOLTAGE
  ARM: imx: update imx_v6_v7_defconfig
  ARM: tegra: Demote EMC clock inconsistency BUG to WARN
  ...

Pull "ARM: timer cleanup work" from Arnd Bergmann:
 "These are split out from the generic soc and driver updates because
  there was a lot of conflicting work by multiple people.  Marc Zyngier
  worked on simplifying the "localtimer" interfaces, and some of the
  platforms are touching the same code as they move to device tree based
  booting.

  Signed-off-by: Arnd Bergmann <arnd@arndb.de>"

* tag 'timer' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (61 commits)
  ARM: tegra: select USB_ULPI if USB is selected
  arm/tegra: pcie: fix return value of function
  ARM: ux500: fix compilation after local timer rework
  ARM: shmobile: remove additional __io() macro use
  ARM: local timers: make the runtime registration interface mandatory
  ARM: local timers: convert MSM to runtime registration interface
  ARM: local timers: convert exynos to runtime registration interface
  ARM: smp_twd: remove old local timer interface
  ARM: imx6q: convert to twd_local_timer_register() interface
  ARM: highbank: convert to twd_local_timer_register() interface
  ARM: ux500: convert to twd_local_timer_register() interface
  ARM: shmobile: convert to twd_local_timer_register() interface
  ARM: tegra: convert to twd_local_timer_register() interface
  ARM: plat-versatile: convert to twd_local_timer_register() interface
  ARM: OMAP4: convert to twd_local_timer_register() interface
  ARM: smp_twd: add device tree support
  ARM: smp_twd: add runtime registration support
  ARM: local timers: introduce a new registration interface
  ARM: smp_twd: make local_timer_stop a symbol instead of a #define
  ARM: mach-shmobile: default to no earlytimer
  ...

Pull "ARM: global cleanups" from Arnd Bergmann:
 "Quite a bit of code gets removed, and some stuff moved around, mostly
  the old samsung s3c24xx stuff.  There should be no functional changes
  in this series otherwise.  Some cleanups have dependencies on other
  arm-soc branches and will be sent in the second round.

  Signed-off-by: Arnd Bergmann <arnd@arndb.de>"

Fixed up trivial conflicts mainly due to #include's being changes on
both sides.

* tag 'cleanup' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (121 commits)
  ep93xx: Remove unnecessary includes of ep93xx-regs.h
  ep93xx: Move EP93XX_SYSCON defines to SoC private header
  ep93xx: Move crunch code to mach-ep93xx directory
  ep93xx: Make syscon access functions private to SoC
  ep93xx: Configure GPIO ports in core code
  ep93xx: Move peripheral defines to local SoC header
  ep93xx: Convert the watchdog driver into a platform device.
  ep93xx: Use ioremap for backlight driver
  ep93xx: Move GPIO defines to gpio-ep93xx.h
  ep93xx: Don't use system controller defines in audio drivers
  ep93xx: Move PHYS_BASE defines to local SoC header file
  ARM: EXYNOS: Add clock register addresses for EXYNOS4X12 bus devfreq driver
  ARM: EXYNOS: add clock registers for exynos4x12-cpufreq
  PM / devfreq: update the name of EXYNOS clock registers that were omitted
  PM / devfreq: update the name of EXYNOS clock register
  ARM: EXYNOS: change the prefix S5P_ to EXYNOS4_ for clock
  ARM: EXYNOS: use static declaration on regarding clock
  ARM: EXYNOS: replace clock.c for other new EXYNOS SoCs
  ARM: OMAP2+: Fix build error after merge
  ARM: S3C24XX: remove call to s3c24xx_setup_clocks
  ...