OSDN Git Service

uclinux-h8/linux.git
8 years agodma-mapping: consolidate dma_{alloc,free}_{attrs,coherent}
Christoph Hellwig [Tue, 25 Aug 2015 04:32:01 +0000 (14:32 +1000)]
dma-mapping: consolidate dma_{alloc,free}_{attrs,coherent}

Since 2009 we have a nice asm-generic header implementing lots of DMA API
functions for architectures using struct dma_map_ops, but unfortunately
it's still missing a lot of APIs that all architectures still have to
duplicate.

This series consolidates the remaining functions, although we still need
arch opt outs for two of them as a few architectures have very
non-standard implementations.

This patch (of 5):

The coherent DMA allocator works the same over all architectures supporting
dma_map operations.

This patch consolidates them and converges the minor differences:

 - the debug_dma helpers are now called from all architectures, including
   those that were previously missing them
 - dma_alloc_from_coherent and dma_release_from_coherent are now always
   called from the generic alloc/free routines instead of the ops
   dma-mapping-common.h always includes dma-coherent.h to get the defintions
   for them, or the stubs if the architecture doesn't support this feature
 - checks for ->alloc / ->free presence are removed.  There is only one
   magic instead of dma_map_ops without them (mic_dma_ops) and that one
   is x86 only anyway.

Besides that only x86 needs special treatment to replace a default devices
if none is passed and tweak the gfp_flags.  An optional arch hook is provided
for that.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: Michal Simek <monstr@monstr.eu>
Cc: Jonas Bonn <jonas@southpole.se>
Cc: Chris Metcalf <cmetcalf@ezchip.com>
Cc: Guan Xuetao <gxt@mprc.pku.edu.cn>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
8 years agoselftests: enhance membarrier syscall test
Mathieu Desnoyers [Tue, 25 Aug 2015 04:32:01 +0000 (14:32 +1000)]
selftests: enhance membarrier syscall test

Update the membarrier syscall self-test to match the membarrier interface.
 Extend coverage of the interface.  Consider ENOSYS as a "SKIP" test,
since it is a valid configuration, but does not allow testing the system
call.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Pranith Kumar <bobby.prani@gmail.com>
Cc: Shuah Khan <shuahkh@osg.samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
8 years agoselftests: add membarrier syscall test
Pranith Kumar [Tue, 25 Aug 2015 04:32:01 +0000 (14:32 +1000)]
selftests: add membarrier syscall test

Add a self test for the membarrier system call.

Signed-off-by: Pranith Kumar <bobby.prani@gmail.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Shuah Khan <shuahkh@osg.samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
8 years agosys_membarrier(): system-wide memory barrier (generic, x86)
Mathieu Desnoyers [Tue, 25 Aug 2015 04:32:01 +0000 (14:32 +1000)]
sys_membarrier(): system-wide memory barrier (generic, x86)

Here is an implementation of a new system call, sys_membarrier(), which
executes a memory barrier on all threads running on the system.  It is
implemented by calling synchronize_sched().  It can be used to distribute
the cost of user-space memory barriers asymmetrically by transforming
pairs of memory barriers into pairs consisting of sys_membarrier() and a
compiler barrier.  For synchronization primitives that distinguish between
read-side and write-side (e.g.  userspace RCU [1], rwlocks), the read-side
can be accelerated significantly by moving the bulk of the memory barrier
overhead to the write-side.

The existing applications of which I am aware that would be improved by this
system call are as follows:

* Through Userspace RCU library (http://urcu.so)
  - DNS server (Knot DNS) https://www.knot-dns.cz/
  - Network sniffer (http://netsniff-ng.org/)
  - Distributed object storage (https://sheepdog.github.io/sheepdog/)
  - User-space tracing (http://lttng.org)
  - Network storage system (https://www.gluster.org/)
  - Virtual routers (https://events.linuxfoundation.org/sites/events/files/slides/DPDK_RCU_0MQ.pdf)
  - Financial software (https://lkml.org/lkml/2015/3/23/189)

Those projects use RCU in userspace to increase read-side speed and
scalability compared to locking.  Especially in the case of RCU used by
libraries, sys_membarrier can speed up the read-side by moving the bulk of
the memory barrier cost to synchronize_rcu().

* Direct users of sys_membarrier
  - core dotnet garbage collector (https://github.com/dotnet/coreclr/issues/198)

Microsoft core dotnet GC developers are planning to use the mprotect()
side-effect of issuing memory barriers through IPIs as a way to implement
Windows FlushProcessWriteBuffers() on Linux.  They are referring to
sys_membarrier in their github thread, specifically stating that
sys_membarrier() is what they are looking for.

To explain the benefit of this scheme, let's introduce two example threads:

Thread A (non-frequent, e.g. executing liburcu synchronize_rcu())
Thread B (frequent, e.g. executing liburcu
rcu_read_lock()/rcu_read_unlock())

In a scheme where all smp_mb() in thread A are ordering memory accesses
with respect to smp_mb() present in Thread B, we can change each
smp_mb() within Thread A into calls to sys_membarrier() and each
smp_mb() within Thread B into compiler barriers "barrier()".

Before the change, we had, for each smp_mb() pairs:

Thread A                    Thread B
previous mem accesses       previous mem accesses
smp_mb()                    smp_mb()
following mem accesses      following mem accesses

After the change, these pairs become:

Thread A                    Thread B
prev mem accesses           prev mem accesses
sys_membarrier()            barrier()
follow mem accesses         follow mem accesses

As we can see, there are two possible scenarios: either Thread B memory
accesses do not happen concurrently with Thread A accesses (1), or they
do (2).

1) Non-concurrent Thread A vs Thread B accesses:

Thread A                    Thread B
prev mem accesses
sys_membarrier()
follow mem accesses
                            prev mem accesses
                            barrier()
                            follow mem accesses

In this case, thread B accesses will be weakly ordered. This is OK,
because at that point, thread A is not particularly interested in
ordering them with respect to its own accesses.

2) Concurrent Thread A vs Thread B accesses

Thread A                    Thread B
prev mem accesses           prev mem accesses
sys_membarrier()            barrier()
follow mem accesses         follow mem accesses

In this case, thread B accesses, which are ensured to be in program
order thanks to the compiler barrier, will be "upgraded" to full
smp_mb() by synchronize_sched().

* Benchmarks

On Intel Xeon E5405 (8 cores)
(one thread is calling sys_membarrier, the other 7 threads are busy
looping)

1000 non-expedited sys_membarrier calls in 33s =3D 33 milliseconds/call.

* User-space user of this system call: Userspace RCU library

Both the signal-based and the sys_membarrier userspace RCU schemes
permit us to remove the memory barrier from the userspace RCU
rcu_read_lock() and rcu_read_unlock() primitives, thus significantly
accelerating them. These memory barriers are replaced by compiler
barriers on the read-side, and all matching memory barriers on the
write-side are turned into an invocation of a memory barrier on all
active threads in the process. By letting the kernel perform this
synchronization rather than dumbly sending a signal to every process
threads (as we currently do), we diminish the number of unnecessary wake
ups and only issue the memory barriers on active threads. Non-running
threads do not need to execute such barrier anyway, because these are
implied by the scheduler context switches.

Results in liburcu:

Operations in 10s, 6 readers, 2 writers:

memory barriers in reader:    1701557485 reads, 2202847 writes
signal-based scheme:          9830061167 reads,    6700 writes
sys_membarrier:               9952759104 reads,     425 writes
sys_membarrier (dyn. check):  7970328887 reads,     425 writes

The dynamic sys_membarrier availability check adds some overhead to
the read-side compared to the signal-based scheme, but besides that,
sys_membarrier slightly outperforms the signal-based scheme. However,
this non-expedited sys_membarrier implementation has a much slower grace
period than signal and memory barrier schemes.

Besides diminishing the number of wake-ups, one major advantage of the
membarrier system call over the signal-based scheme is that it does not
need to reserve a signal. This plays much more nicely with libraries,
and with processes injected into for tracing purposes, for which we
cannot expect that signals will be unused by the application.

An expedited version of this system call can be added later on to speed
up the grace period. Its implementation will likely depend on reading
the cpu_curr()->mm without holding each CPU's rq lock.

This patch adds the system call to x86 and to asm-generic.

[1] http://urcu.so

membarrier(2) man page:

MEMBARRIER(2)              Linux Programmer's Manual             MEMBARRIER(2)

NAME
       membarrier - issue memory barriers on a set of threads

SYNOPSIS
       #include <linux/membarrier.h>

       int membarrier(int cmd, int flags);

DESCRIPTION
       The cmd argument is one of the following:

       MEMBARRIER_CMD_QUERY
              Query  the  set  of  supported commands. It returns a bitmask of
              supported commands.

       MEMBARRIER_CMD_SHARED
              Execute a memory barrier on all threads running on  the  system.
              Upon  return from system call, the caller thread is ensured that
              all running threads have passed through a state where all memory
              accesses  to  user-space  addresses  match program order between
              entry to and return from the system  call  (non-running  threads
              are de facto in such a state). This covers threads from all pro=E2=80=90
              cesses running on the system.  This command returns 0.

       The flags argument needs to be 0. For future extensions.

       All memory accesses performed  in  program  order  from  each  targeted
       thread is guaranteed to be ordered with respect to sys_membarrier(). If
       we use the semantic "barrier()" to represent a compiler barrier forcing
       memory  accesses  to  be performed in program order across the barrier,
       and smp_mb() to represent explicit memory barriers forcing full  memory
       ordering  across  the barrier, we have the following ordering table for
       each pair of barrier(), sys_membarrier() and smp_mb():

       The pair ordering is detailed as (O: ordered, X: not ordered):

                              barrier()   smp_mb() sys_membarrier()
              barrier()          X           X            O
              smp_mb()           X           O            O
              sys_membarrier()   O           O            O

RETURN VALUE
       On success, these system calls return zero.  On error, -1 is  returned,
       and errno is set appropriately. For a given command, with flags
       argument set to 0, this system call is guaranteed to always return the
       same value until reboot.

ERRORS
       ENOSYS System call is not implemented.

       EINVAL Invalid arguments.

Linux                             2015-04-15                     MEMBARRIER(2)

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Reviewed-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Nicholas Miell <nmiell@comcast.net>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Alan Cox <gnomes@lxorguk.ukuu.org.uk>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Stephen Hemminger <stephen@networkplumber.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: David Howells <dhowells@redhat.com>
Cc: Pranith Kumar <bobby.prani@gmail.com>
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
Cc: Shuah Khan <shuahkh@osg.samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
8 years agomm, madvise: use vma_is_anonymous() to check for anon VMA
Kirill A. Shutemov [Tue, 25 Aug 2015 04:32:00 +0000 (14:32 +1000)]
mm, madvise: use vma_is_anonymous() to check for anon VMA

!vma->vm_file is not reliable to detect anon VMA, because not all
drivers bother set it. Let's use vma_is_anonymous() instead.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Acked-by: Minchan Kim <minchan@kernel.org>
Reviewed-by: Oleg Nesterov <oleg@redhat.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
8 years agomm: use vma_is_anonymous() in create_huge_pmd() and wp_huge_pmd()
Kirill A. Shutemov [Tue, 25 Aug 2015 04:32:00 +0000 (14:32 +1000)]
mm: use vma_is_anonymous() in create_huge_pmd() and wp_huge_pmd()

Let's use helper rather than direct check of vma->vm_ops to distinguish
anonymous VMA.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Reviewed-by: Oleg Nesterov <oleg@redhat.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
8 years agomm: make sure all file VMAs have ->vm_ops set
Kirill A. Shutemov [Tue, 25 Aug 2015 04:32:00 +0000 (14:32 +1000)]
mm: make sure all file VMAs have ->vm_ops set

We rely on vma->vm_ops == NULL to detect anonymous VMA: see
vma_is_anonymous(), but some drivers doesn't set ->vm_ops.

As a result we can end up with anonymous page in private file mapping.
That should not lead to serious misbehaviour, but nevertheless is wrong.

Let's fix by setting up dummy ->vm_ops for file mmapping if f_op->mmap()
didn't set its own.

The patch also adds sanity check into __vma_link_rb(). It will help
catch broken VMAs which inserted directly into mm_struct via
insert_vm_struct().

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Reviewed-by: Oleg Nesterov <oleg@redhat.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
8 years agomm-mpx-add-vm_flags_t-vm_flags-arg-to-do_mmap_pgoff-fix-checkpatch-fixes
Andrew Morton [Tue, 25 Aug 2015 04:32:00 +0000 (14:32 +1000)]
mm-mpx-add-vm_flags_t-vm_flags-arg-to-do_mmap_pgoff-fix-checkpatch-fixes

WARNING: Possible unwrapped commit description (prefer a maximum 75 chars per line)
#5:
>> mm/nommu.c:1248:30: error: 'vm_flags' redeclared as different kind of symbol

WARNING: please, no spaces at the start of a line
#28: FILE: mm/nommu.c:1248:
+       unsigned long capabilities, result;$

WARNING: please, no spaces at the start of a line
#37: FILE: mm/nommu.c:1266:
+       vm_flags |= determine_vm_flags(file, prot, flags, capabilities);$

total: 0 errors, 3 warnings, 16 lines checked

./patches/mm-mpx-add-vm_flags_t-vm_flags-arg-to-do_mmap_pgoff-fix.patch has style problems, please review.

NOTE: If any of the errors are false positives, please report
      them to the maintainer, see CHECKPATCH in MAINTAINERS.

Please run checkpatch prior to sending patches

Cc: "Kirill A. Shutemov" <kirill@shutemov.name>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
8 years agomm-mpx-add-vm_flags_t-vm_flags-arg-to-do_mmap_pgoff-fix
Kirill A. Shutemov [Tue, 25 Aug 2015 04:31:59 +0000 (14:31 +1000)]
mm-mpx-add-vm_flags_t-vm_flags-arg-to-do_mmap_pgoff-fix

   mm/nommu.c: In function 'do_mmap':
>> mm/nommu.c:1248:30: error: 'vm_flags' redeclared as different kind of symbol
     unsigned long capabilities, vm_flags, result;
                                 ^
   mm/nommu.c:1241:15: note: previous definition of 'vm_flags' was here
       vm_flags_t vm_flags,

Reported-by: Wu Fengguang <fengguang.wu@intel.com>
Reported-by: Oleg Nesterov <oleg@redhat.com>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Tested-by: Mark Salter <msalter@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
8 years agomm, mpx: add "vm_flags_t vm_flags" arg to do_mmap_pgoff()
Oleg Nesterov [Tue, 25 Aug 2015 04:31:59 +0000 (14:31 +1000)]
mm, mpx: add "vm_flags_t vm_flags" arg to do_mmap_pgoff()

Add the additional "vm_flags_t vm_flags" argument to do_mmap_pgoff(),
rename it to do_mmap(), and re-introduce do_mmap_pgoff() as a simple
wrapper on top of do_mmap().  Perhaps we should update the callers of
do_mmap_pgoff() and kill it later.

This way mpx_mmap() can simply call do_mmap(vm_flags => VM_MPX) and do not
play with vm internals.

After this change mmap_region() has a single user outside of mmap.c,
arch/tile/mm/elf.c:arch_setup_additional_pages().  It would be nice to
change arch/tile/ and unexport mmap_region().

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
Tested-by: Dave Hansen <dave.hansen@linux.intel.com>
Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
8 years agomm: mark most vm_operations_struct const
Kirill A. Shutemov [Tue, 25 Aug 2015 04:31:59 +0000 (14:31 +1000)]
mm: mark most vm_operations_struct const

With two exceptions (drm/qxl and drm/radeon) all vm_operations_struct
structs should be constant.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Reviewed-by: Oleg Nesterov <oleg@redhat.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
8 years agofs-seq_file-convert-int-seq_vprint-seq_printf-etc-returns-to-void-fix-fix
Andrew Morton [Tue, 25 Aug 2015 04:31:59 +0000 (14:31 +1000)]
fs-seq_file-convert-int-seq_vprint-seq_printf-etc-returns-to-void-fix-fix

fix drivers/iommu/omap-iommu-debug.c for fs-seq_file-convert-int-seq_vprint-seq_printf-etc-returns-to-void.patch

Cc: Mark Brown <broonie@kernel.org>
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Cc: Joe Perches <joe@perches.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Joerg Roedel <jroedel@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
8 years agofs-seq_file-convert-int-seq_vprint-seq_printf-etc-returns-to-void-fix
Andrew Morton [Tue, 25 Aug 2015 04:31:59 +0000 (14:31 +1000)]
fs-seq_file-convert-int-seq_vprint-seq_printf-etc-returns-to-void-fix

Cc: Al Viro <viro@ZenIV.linux.org.uk>
Cc: Joe Perches <joe@perches.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
8 years agofs/seq_file: convert int seq_vprint/seq_printf/etc... returns to void
Joe Perches [Tue, 25 Aug 2015 04:31:58 +0000 (14:31 +1000)]
fs/seq_file: convert int seq_vprint/seq_printf/etc... returns to void

The seq_<foo> function return values were frequently misused.

See: commit 1f33c41c03da ("seq_file: Rename seq_overflow() to
     seq_has_overflowed() and make public")

All uses of these return values have been removed, so convert the
return types to void.

Miscellanea:

o Move seq_put_decimal_<type> and seq_escape prototypes closer the
  other seq_vprintf prototypes
o Reorder seq_putc and seq_puts to return early on overflow
o Add argument names to seq_vprintf and seq_printf
o Update the seq_escape kernel-doc
o Convert a couple of leading spaces to tabs in seq_escape

Signed-off-by: Joe Perches <joe@perches.com>
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Cc: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
8 years agonamei: fix warning while make xmldocs caused by namei.c
Masanari Iida [Tue, 25 Aug 2015 04:31:58 +0000 (14:31 +1000)]
namei: fix warning while make xmldocs caused by namei.c

Fix the following warnings:

Warning(.//fs/namei.c:2422): No description found for parameter 'nd'
Warning(.//fs/namei.c:2422): Excess function parameter 'nameidata'
description in 'path_mountpoint'

Signed-off-by: Masanari Iida <standby24x7@gmail.com>
Acked-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
8 years agoMerge branch 'akpm-current/current'
Stephen Rothwell [Fri, 28 Aug 2015 07:00:28 +0000 (17:00 +1000)]
Merge branch 'akpm-current/current'

8 years agoMerge remote-tracking branch 'nvdimm/libnvdimm-for-next'
Stephen Rothwell [Fri, 28 Aug 2015 06:21:01 +0000 (16:21 +1000)]
Merge remote-tracking branch 'nvdimm/libnvdimm-for-next'

8 years agoMerge remote-tracking branch 'rtc/rtc-next'
Stephen Rothwell [Fri, 28 Aug 2015 06:09:40 +0000 (16:09 +1000)]
Merge remote-tracking branch 'rtc/rtc-next'

8 years agoMerge remote-tracking branch 'coresight/next'
Stephen Rothwell [Fri, 28 Aug 2015 06:09:36 +0000 (16:09 +1000)]
Merge remote-tracking branch 'coresight/next'

8 years agoMerge remote-tracking branch 'livepatching/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 06:08:32 +0000 (16:08 +1000)]
Merge remote-tracking branch 'livepatching/for-next'

8 years agoMerge remote-tracking branch 'access_once/linux-next'
Stephen Rothwell [Fri, 28 Aug 2015 06:08:30 +0000 (16:08 +1000)]
Merge remote-tracking branch 'access_once/linux-next'

8 years agoMerge remote-tracking branch 'y2038/y2038'
Stephen Rothwell [Fri, 28 Aug 2015 06:07:24 +0000 (16:07 +1000)]
Merge remote-tracking branch 'y2038/y2038'

8 years agoMerge remote-tracking branch 'kselftest/next'
Stephen Rothwell [Fri, 28 Aug 2015 06:06:31 +0000 (16:06 +1000)]
Merge remote-tracking branch 'kselftest/next'

8 years agoMerge remote-tracking branch 'llvmlinux/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 06:06:23 +0000 (16:06 +1000)]
Merge remote-tracking branch 'llvmlinux/for-next'

8 years agoMerge remote-tracking branch 'clk/clk-next'
Stephen Rothwell [Fri, 28 Aug 2015 06:05:07 +0000 (16:05 +1000)]
Merge remote-tracking branch 'clk/clk-next'

8 years agoMerge remote-tracking branch 'userns/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 05:51:04 +0000 (15:51 +1000)]
Merge remote-tracking branch 'userns/for-next'

8 years agoMerge remote-tracking branch 'pwm/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 05:49:58 +0000 (15:49 +1000)]
Merge remote-tracking branch 'pwm/for-next'

8 years agoMerge remote-tracking branch 'gpio/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 05:48:11 +0000 (15:48 +1000)]
Merge remote-tracking branch 'gpio/for-next'

8 years agoMerge remote-tracking branch 'vhost/linux-next'
Stephen Rothwell [Fri, 28 Aug 2015 05:42:09 +0000 (15:42 +1000)]
Merge remote-tracking branch 'vhost/linux-next'

8 years agoMerge remote-tracking branch 'pinctrl/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 05:41:05 +0000 (15:41 +1000)]
Merge remote-tracking branch 'pinctrl/for-next'

8 years agoMerge remote-tracking branch 'target-updates/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 05:39:05 +0000 (15:39 +1000)]
Merge remote-tracking branch 'target-updates/for-next'

8 years agoMerge remote-tracking branch 'scsi/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 05:37:12 +0000 (15:37 +1000)]
Merge remote-tracking branch 'scsi/for-next'

8 years agoMerge remote-tracking branch 'cgroup/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 05:18:41 +0000 (15:18 +1000)]
Merge remote-tracking branch 'cgroup/for-next'

8 years agoMerge remote-tracking branch 'kdbus/kdbus'
Stephen Rothwell [Fri, 28 Aug 2015 05:02:45 +0000 (15:02 +1000)]
Merge remote-tracking branch 'kdbus/kdbus'

8 years agoMerge remote-tracking branch 'char-misc/char-misc-next'
Stephen Rothwell [Fri, 28 Aug 2015 04:59:11 +0000 (14:59 +1000)]
Merge remote-tracking branch 'char-misc/char-misc-next'

8 years agoMerge remote-tracking branch 'staging/staging-next'
Stephen Rothwell [Fri, 28 Aug 2015 04:57:19 +0000 (14:57 +1000)]
Merge remote-tracking branch 'staging/staging-next'

8 years agoMerge remote-tracking branch 'usb/usb-next'
Stephen Rothwell [Fri, 28 Aug 2015 04:54:50 +0000 (14:54 +1000)]
Merge remote-tracking branch 'usb/usb-next'

8 years agoMerge remote-tracking branch 'tty/tty-next'
Stephen Rothwell [Fri, 28 Aug 2015 04:53:02 +0000 (14:53 +1000)]
Merge remote-tracking branch 'tty/tty-next'

8 years agoMerge remote-tracking branch 'driver-core/driver-core-next'
Stephen Rothwell [Fri, 28 Aug 2015 04:35:42 +0000 (14:35 +1000)]
Merge remote-tracking branch 'driver-core/driver-core-next'

8 years agoMerge remote-tracking branch 'ipmi/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 04:34:38 +0000 (14:34 +1000)]
Merge remote-tracking branch 'ipmi/for-next'

8 years agoMerge remote-tracking branch 'leds/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 04:33:51 +0000 (14:33 +1000)]
Merge remote-tracking branch 'leds/for-next'

8 years agoMerge remote-tracking branch 'regmap/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 04:30:18 +0000 (14:30 +1000)]
Merge remote-tracking branch 'regmap/for-next'

8 years agoMerge remote-tracking branch 'chrome-platform/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 04:29:32 +0000 (14:29 +1000)]
Merge remote-tracking branch 'chrome-platform/for-next'

8 years agoMerge remote-tracking branch 'drivers-x86/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 04:28:43 +0000 (14:28 +1000)]
Merge remote-tracking branch 'drivers-x86/for-next'

8 years agoMerge remote-tracking branch 'workqueues/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 04:10:06 +0000 (14:10 +1000)]
Merge remote-tracking branch 'workqueues/for-next'

8 years agoMerge remote-tracking branch 'percpu/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 03:52:11 +0000 (13:52 +1000)]
Merge remote-tracking branch 'percpu/for-next'

8 years agoMerge remote-tracking branch 'xen-tip/linux-next'
Stephen Rothwell [Fri, 28 Aug 2015 03:37:51 +0000 (13:37 +1000)]
Merge remote-tracking branch 'xen-tip/linux-next'

8 years agoMerge remote-tracking branch 'kvm-ppc/kvm-ppc-next'
Stephen Rothwell [Fri, 28 Aug 2015 03:35:21 +0000 (13:35 +1000)]
Merge remote-tracking branch 'kvm-ppc/kvm-ppc-next'

8 years agoMerge remote-tracking branch 'kvm-arm/next'
Stephen Rothwell [Fri, 28 Aug 2015 03:34:26 +0000 (13:34 +1000)]
Merge remote-tracking branch 'kvm-arm/next'

8 years agoMerge remote-tracking branch 'kvm/linux-next'
Stephen Rothwell [Fri, 28 Aug 2015 03:33:17 +0000 (13:33 +1000)]
Merge remote-tracking branch 'kvm/linux-next'

8 years agoMerge remote-tracking branch 'rcu/rcu/next'
Stephen Rothwell [Fri, 28 Aug 2015 03:16:17 +0000 (13:16 +1000)]
Merge remote-tracking branch 'rcu/rcu/next'

8 years agoMerge remote-tracking branch 'irqchip/irqchip/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 03:16:13 +0000 (13:16 +1000)]
Merge remote-tracking branch 'irqchip/irqchip/for-next'

8 years agoMerge remote-tracking branch 'edac-amd/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 03:15:28 +0000 (13:15 +1000)]
Merge remote-tracking branch 'edac-amd/for-next'

8 years agoMerge remote-tracking branch 'tip/auto-latest'
Stephen Rothwell [Fri, 28 Aug 2015 02:58:04 +0000 (12:58 +1000)]
Merge remote-tracking branch 'tip/auto-latest'

8 years agoMerge remote-tracking branch 'spi/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 02:56:01 +0000 (12:56 +1000)]
Merge remote-tracking branch 'spi/for-next'

8 years agoMerge remote-tracking branch 'mailbox/mailbox-for-next'
Stephen Rothwell [Fri, 28 Aug 2015 02:54:58 +0000 (12:54 +1000)]
Merge remote-tracking branch 'mailbox/mailbox-for-next'

8 years agoMerge remote-tracking branch 'dt-rh/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 02:53:28 +0000 (12:53 +1000)]
Merge remote-tracking branch 'dt-rh/for-next'

8 years agoMerge remote-tracking branch 'audit/next'
Stephen Rothwell [Fri, 28 Aug 2015 02:51:54 +0000 (12:51 +1000)]
Merge remote-tracking branch 'audit/next'

8 years agoMerge remote-tracking branch 'trivial/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 02:36:56 +0000 (12:36 +1000)]
Merge remote-tracking branch 'trivial/for-next'

8 years agoMerge remote-tracking branch 'jc_docs/docs-next'
Stephen Rothwell [Fri, 28 Aug 2015 02:36:10 +0000 (12:36 +1000)]
Merge remote-tracking branch 'jc_docs/docs-next'

8 years agoMerge remote-tracking branch 'dwmw2-iommu/master'
Stephen Rothwell [Fri, 28 Aug 2015 02:34:58 +0000 (12:34 +1000)]
Merge remote-tracking branch 'dwmw2-iommu/master'

8 years agoMerge remote-tracking branch 'iommu/next'
Stephen Rothwell [Fri, 28 Aug 2015 02:33:46 +0000 (12:33 +1000)]
Merge remote-tracking branch 'iommu/next'

8 years agoMerge remote-tracking branch 'security/next'
Stephen Rothwell [Fri, 28 Aug 2015 02:19:06 +0000 (12:19 +1000)]
Merge remote-tracking branch 'security/next'

8 years agoMerge remote-tracking branch 'regulator/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 02:17:08 +0000 (12:17 +1000)]
Merge remote-tracking branch 'regulator/for-next'

8 years agoMerge remote-tracking branch 'omap_dss2/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 02:15:13 +0000 (12:15 +1000)]
Merge remote-tracking branch 'omap_dss2/for-next'

8 years agoMerge remote-tracking branch 'battery/master'
Stephen Rothwell [Fri, 28 Aug 2015 02:14:27 +0000 (12:14 +1000)]
Merge remote-tracking branch 'battery/master'

8 years agoMerge remote-tracking branch 'backlight/for-backlight-next'
Stephen Rothwell [Fri, 28 Aug 2015 02:13:42 +0000 (12:13 +1000)]
Merge remote-tracking branch 'backlight/for-backlight-next'

8 years agoMerge remote-tracking branch 'mfd/for-mfd-next'
Stephen Rothwell [Fri, 28 Aug 2015 02:12:35 +0000 (12:12 +1000)]
Merge remote-tracking branch 'mfd/for-mfd-next'

8 years agoMerge remote-tracking branch 'md/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 02:11:45 +0000 (12:11 +1000)]
Merge remote-tracking branch 'md/for-next'

8 years agoMerge remote-tracking branch 'kgdb/kgdb-next'
Stephen Rothwell [Fri, 28 Aug 2015 02:10:39 +0000 (12:10 +1000)]
Merge remote-tracking branch 'kgdb/kgdb-next'

8 years agoMerge remote-tracking branch 'mmc-uh/next'
Stephen Rothwell [Fri, 28 Aug 2015 02:09:28 +0000 (12:09 +1000)]
Merge remote-tracking branch 'mmc-uh/next'

8 years agoMerge remote-tracking branch 'device-mapper/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 02:08:36 +0000 (12:08 +1000)]
Merge remote-tracking branch 'device-mapper/for-next'

8 years agoMerge remote-tracking branch 'block/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 01:51:41 +0000 (11:51 +1000)]
Merge remote-tracking branch 'block/for-next'

8 years agoMerge remote-tracking branch 'input/next'
Stephen Rothwell [Fri, 28 Aug 2015 01:50:29 +0000 (11:50 +1000)]
Merge remote-tracking branch 'input/next'

8 years agoMerge remote-tracking branch 'modules/modules-next'
Stephen Rothwell [Fri, 28 Aug 2015 01:49:44 +0000 (11:49 +1000)]
Merge remote-tracking branch 'modules/modules-next'

8 years agoMerge remote-tracking branch 'sound-asoc/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 01:35:10 +0000 (11:35 +1000)]
Merge remote-tracking branch 'sound-asoc/for-next'

8 years agoMerge remote-tracking branch 'sound/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 01:34:17 +0000 (11:34 +1000)]
Merge remote-tracking branch 'sound/for-next'

8 years agoMerge remote-tracking branch 'drm-misc/topic/drm-misc'
Stephen Rothwell [Fri, 28 Aug 2015 01:32:38 +0000 (11:32 +1000)]
Merge remote-tracking branch 'drm-misc/topic/drm-misc'

8 years agoMerge remote-tracking branch 'drm-intel/for-linux-next'
Stephen Rothwell [Fri, 28 Aug 2015 01:31:26 +0000 (11:31 +1000)]
Merge remote-tracking branch 'drm-intel/for-linux-next'

8 years agoMerge remote-tracking branch 'drm/drm-next'
Stephen Rothwell [Fri, 28 Aug 2015 01:29:43 +0000 (11:29 +1000)]
Merge remote-tracking branch 'drm/drm-next'

8 years agoMerge remote-tracking branch 'crypto/master'
Stephen Rothwell [Fri, 28 Aug 2015 01:27:07 +0000 (11:27 +1000)]
Merge remote-tracking branch 'crypto/master'

8 years agoMerge remote-tracking branch 'l2-mtd/master'
Stephen Rothwell [Fri, 28 Aug 2015 01:26:20 +0000 (11:26 +1000)]
Merge remote-tracking branch 'l2-mtd/master'

8 years agoMerge remote-tracking branch 'rdma/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 01:09:42 +0000 (11:09 +1000)]
Merge remote-tracking branch 'rdma/for-next'

Initial roundup of 4.3 merge window candidates

8 years agoMerge remote-tracking branch 'bluetooth/master'
Stephen Rothwell [Fri, 28 Aug 2015 00:44:09 +0000 (10:44 +1000)]
Merge remote-tracking branch 'bluetooth/master'

8 years agoMerge remote-tracking branch 'net-next/master'
Stephen Rothwell [Fri, 28 Aug 2015 00:32:56 +0000 (10:32 +1000)]
Merge remote-tracking branch 'net-next/master'

8 years agoMerge remote-tracking branch 'slave-dma/next'
Stephen Rothwell [Fri, 28 Aug 2015 00:27:04 +0000 (10:27 +1000)]
Merge remote-tracking branch 'slave-dma/next'

8 years agoMerge remote-tracking branch 'dlm/next'
Stephen Rothwell [Fri, 28 Aug 2015 00:26:16 +0000 (10:26 +1000)]
Merge remote-tracking branch 'dlm/next'

8 years agoMerge remote-tracking branch 'thermal/next'
Stephen Rothwell [Fri, 28 Aug 2015 00:25:10 +0000 (10:25 +1000)]
Merge remote-tracking branch 'thermal/next'

8 years agoMerge remote-tracking branch 'idle/next'
Stephen Rothwell [Fri, 28 Aug 2015 00:25:01 +0000 (10:25 +1000)]
Merge remote-tracking branch 'idle/next'

8 years agosock: fix kernel doc error
Jean Sacren [Fri, 28 Aug 2015 00:05:49 +0000 (18:05 -0600)]
sock: fix kernel doc error

The symbol '__sk_reclaim' is not present in the current tree. Apparently
'__sk_reclaim' was meant to be '__sk_mem_reclaim', so fix it with the
right symbol name for the kernel doc.

Signed-off-by: Jean Sacren <sakiwit@gmail.com>
Cc: Hideo Aoki <haoki@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge remote-tracking branch 'pm/linux-next'
Stephen Rothwell [Fri, 28 Aug 2015 00:10:03 +0000 (10:10 +1000)]
Merge remote-tracking branch 'pm/linux-next'

8 years agoMerge remote-tracking branch 'libata/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 00:09:18 +0000 (10:09 +1000)]
Merge remote-tracking branch 'libata/for-next'

8 years agoMerge remote-tracking branch 'kbuild/for-next'
Stephen Rothwell [Fri, 28 Aug 2015 00:08:11 +0000 (10:08 +1000)]
Merge remote-tracking branch 'kbuild/for-next'

8 years agoMerge remote-tracking branch 'v4l-dvb/master'
Stephen Rothwell [Thu, 27 Aug 2015 23:56:06 +0000 (09:56 +1000)]
Merge remote-tracking branch 'v4l-dvb/master'

8 years agoMerge remote-tracking branch 'hwmon-staging/hwmon-next'
Stephen Rothwell [Thu, 27 Aug 2015 23:55:23 +0000 (09:55 +1000)]
Merge remote-tracking branch 'hwmon-staging/hwmon-next'

8 years agoMerge branch 'jdelvare-hwmon/master'
Stephen Rothwell [Thu, 27 Aug 2015 23:54:40 +0000 (09:54 +1000)]
Merge branch 'jdelvare-hwmon/master'

8 years agoMerge remote-tracking branch 'i2c/i2c/for-next'
Stephen Rothwell [Thu, 27 Aug 2015 23:51:34 +0000 (09:51 +1000)]
Merge remote-tracking branch 'i2c/i2c/for-next'

8 years agoMerge remote-tracking branch 'hid/for-next'
Stephen Rothwell [Thu, 27 Aug 2015 23:50:45 +0000 (09:50 +1000)]
Merge remote-tracking branch 'hid/for-next'

8 years agolibnvdimm, pmem: direct map legacy pmem by default
Dan Williams [Mon, 24 Aug 2015 23:20:23 +0000 (19:20 -0400)]
libnvdimm, pmem: direct map legacy pmem by default

The expectation is that the legacy / non-standard pmem discovery method
(e820 type-12) will only ever be used to describe small quantities of
persistent memory.  Larger capacities will be described via the ACPI
NFIT.  When "allocate struct page from pmem" support is added this default
policy can be overridden by assigning a legacy pmem namespace to a pfn
device, however this would be only be necessary if a platform used the
legacy mechanism to define a very large range.

Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
8 years agoMerge remote-tracking branch 'pci/next'
Stephen Rothwell [Thu, 27 Aug 2015 23:45:56 +0000 (09:45 +1000)]
Merge remote-tracking branch 'pci/next'