GitHub/LineageOS/G12/android_kernel_amlogic_linux-4.9.git
11 years agomm: introduce compaction and migration for ballooned pages
Rafael Aquini [Wed, 12 Dec 2012 00:02:42 +0000 (16:02 -0800)]
mm: introduce compaction and migration for ballooned pages

Memory fragmentation introduced by ballooning might reduce significantly
the number of 2MB contiguous memory blocks that can be used within a guest,
thus imposing performance penalties associated with the reduced number of
transparent huge pages that could be used by the guest workload.

This patch introduces the helper functions as well as the necessary changes
to teach compaction and migration bits how to cope with pages which are
part of a guest memory balloon, in order to make them movable by memory
compaction procedures.

Signed-off-by: Rafael Aquini <aquini@redhat.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: introduce a common interface for balloon pages mobility
Rafael Aquini [Wed, 12 Dec 2012 00:02:38 +0000 (16:02 -0800)]
mm: introduce a common interface for balloon pages mobility

Memory fragmentation introduced by ballooning might reduce significantly
the number of 2MB contiguous memory blocks that can be used within a guest,
thus imposing performance penalties associated with the reduced number of
transparent huge pages that could be used by the guest workload.

This patch introduces a common interface to help a balloon driver on
making its page set movable to compaction, and thus allowing the system
to better leverage the compation efforts on memory defragmentation.

[akpm@linux-foundation.org: use PAGE_FLAGS_CHECK_AT_PREP, s/__balloon_page_flags/page_flags_cleared/, small cleanups]
[rientjes@google.com: allow balloon compaction for any system with memory compaction enabled, which is the defconfig]
Signed-off-by: Rafael Aquini <aquini@redhat.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Minchan Kim <minchan@kernel.org>
Signed-off-by: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: redefine address_space.assoc_mapping
Rafael Aquini [Wed, 12 Dec 2012 00:02:35 +0000 (16:02 -0800)]
mm: redefine address_space.assoc_mapping

Overhaul struct address_space.assoc_mapping renaming it to
address_space.private_data and its type is redefined to void*.  By this
approach we consistently name the .private_* elements from struct
address_space as well as allow extended usage for address_space
association with other data structures through ->private_data.

Also, all users of old ->assoc_mapping element are converted to reflect
its new name and type change (->private_data).

Signed-off-by: Rafael Aquini <aquini@redhat.com>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: adjust address_space_operations.migratepage() return code
Rafael Aquini [Wed, 12 Dec 2012 00:02:31 +0000 (16:02 -0800)]
mm: adjust address_space_operations.migratepage() return code

Memory fragmentation introduced by ballooning might reduce significantly
the number of 2MB contiguous memory blocks that can be used within a
guest, thus imposing performance penalties associated with the reduced
number of transparent huge pages that could be used by the guest workload.

This patch-set follows the main idea discussed at 2012 LSFMMS session:
"Ballooning for transparent huge pages" -- http://lwn.net/Articles/490114/
to introduce the required changes to the virtio_balloon driver, as well as
the changes to the core compaction & migration bits, in order to make
those subsystems aware of ballooned pages and allow memory balloon pages
become movable within a guest, thus avoiding the aforementioned
fragmentation issue

Following are numbers that prove this patch benefits on allowing
compaction to be more effective at memory ballooned guests.

Results for STRESS-HIGHALLOC benchmark, from Mel Gorman's mmtests suite,
running on a 4gB RAM KVM guest which was ballooning 512mB RAM in 64mB
chunks, at every minute (inflating/deflating), while test was running:

===BEGIN stress-highalloc

STRESS-HIGHALLOC
                 highalloc-3.7     highalloc-3.7
                     rc4-clean         rc4-patch
Pass 1          55.00 ( 0.00%)    62.00 ( 7.00%)
Pass 2          54.00 ( 0.00%)    62.00 ( 8.00%)
while Rested    75.00 ( 0.00%)    80.00 ( 5.00%)

MMTests Statistics: duration
                 3.7         3.7
           rc4-clean   rc4-patch
User         1207.59     1207.46
System       1300.55     1299.61
Elapsed      2273.72     2157.06

MMTests Statistics: vmstat
                                3.7         3.7
                          rc4-clean   rc4-patch
Page Ins                    3581516     2374368
Page Outs                  11148692    10410332
Swap Ins                         80          47
Swap Outs                      3641         476
Direct pages scanned          37978       33826
Kswapd pages scanned        1828245     1342869
Kswapd pages reclaimed      1710236     1304099
Direct pages reclaimed        32207       31005
Kswapd efficiency               93%         97%
Kswapd velocity             804.077     622.546
Direct efficiency               84%         91%
Direct velocity              16.703      15.682
Percentage direct scans          2%          2%
Page writes by reclaim        79252        9704
Page writes file              75611        9228
Page writes anon               3641         476
Page reclaim immediate        16764       11014
Page rescued immediate            0           0
Slabs scanned               2171904     2152448
Direct inode steals             385        2261
Kswapd inode steals          659137      609670
Kswapd skipped wait               1          69
THP fault alloc                 546         631
THP collapse alloc              361         339
THP splits                      259         263
THP fault fallback               98          50
THP collapse fail                20          17
Compaction stalls               747         499
Compaction success              244         145
Compaction failures             503         354
Compaction pages moved       370888      474837
Compaction move failure       77378       65259

===END stress-highalloc

This patch:

Introduce MIGRATEPAGE_SUCCESS as the default return code for
address_space_operations.migratepage() method and documents the expected
return code for the same method in failure cases.

Signed-off-by: Rafael Aquini <aquini@redhat.com>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agoarch/sparc/kernel/sys_sparc_64.c: s/COLOUR/COLOR/
Andrew Morton [Wed, 12 Dec 2012 00:02:29 +0000 (16:02 -0800)]
arch/sparc/kernel/sys_sparc_64.c: s/COLOUR/COLOR/

Consistently spell this word across arch/sparc/mm and arch/sparc/kernel.

Acked-by: David Miller <davem@davemloft.net>
Cc: Michel Lespinasse <walken@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: use vm_unmapped_area() in hugetlbfs on sparc64 architecture
Michel Lespinasse [Wed, 12 Dec 2012 00:02:25 +0000 (16:02 -0800)]
mm: use vm_unmapped_area() in hugetlbfs on sparc64 architecture

Update the sparc64 hugetlb_get_unmapped_area function to make use of
vm_unmapped_area() instead of implementing a brute force search.

Signed-off-by: Michel Lespinasse <walken@google.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Chris Metcalf <cmetcalf@tilera.com>
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>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: use vm_unmapped_area() on sparc64 architecture
Michel Lespinasse [Wed, 12 Dec 2012 00:02:21 +0000 (16:02 -0800)]
mm: use vm_unmapped_area() on sparc64 architecture

Update the sparc64 arch_get_unmapped_area[_topdown] functions to make use
of vm_unmapped_area() instead of implementing a brute force search.

[akpm@linux-foundation.org: remove now-unused COLOUR_ALIGN_DOWN()]
Signed-off-by: Michel Lespinasse <walken@google.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Chris Metcalf <cmetcalf@tilera.com>
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>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: use vm_unmapped_area() in hugetlbfs on tile architecture
Michel Lespinasse [Wed, 12 Dec 2012 00:02:17 +0000 (16:02 -0800)]
mm: use vm_unmapped_area() in hugetlbfs on tile architecture

Update the tile hugetlb_get_unmapped_area function to make use of
vm_unmapped_area() instead of implementing a brute force search.

[akpm@linux-foundation.org: fix build]
Signed-off-by: Michel Lespinasse <walken@google.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Chris Metcalf <cmetcalf@tilera.com>
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>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: use vm_unmapped_area() on sparc32 architecture
Michel Lespinasse [Wed, 12 Dec 2012 00:02:15 +0000 (16:02 -0800)]
mm: use vm_unmapped_area() on sparc32 architecture

Update the sparc32 arch_get_unmapped_area function to make use of
vm_unmapped_area() instead of implementing a brute force search.

[akpm@linux-foundation.org: fix build]
[akpm@linux-foundation.org: remove now-unused COLOUR_ALIGN()]
Signed-off-by: Michel Lespinasse <walken@google.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Paul Mundt <lethal@linux-sh.org>
Acked-by: "David S. Miller" <davem@davemloft.net>
Cc: Chris Metcalf <cmetcalf@tilera.com>
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>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: use vm_unmapped_area() on sh architecture
Michel Lespinasse [Wed, 12 Dec 2012 00:02:12 +0000 (16:02 -0800)]
mm: use vm_unmapped_area() on sh architecture

Update the sh arch_get_unmapped_area[_topdown] functions to make use of
vm_unmapped_area() instead of implementing a brute force search.

[akpm@linux-foundation.org: remove now-unused COLOUR_ALIGN_DOWN()]
Signed-off-by: Michel Lespinasse <walken@google.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Chris Metcalf <cmetcalf@tilera.com>
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>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: use vm_unmapped_area() on arm architecture
Michel Lespinasse [Wed, 12 Dec 2012 00:02:10 +0000 (16:02 -0800)]
mm: use vm_unmapped_area() on arm architecture

Update the arm arch_get_unmapped_area[_topdown] functions to make use of
vm_unmapped_area() instead of implementing a brute force search.

[akpm@linux-foundation.org: remove now-unused COLOUR_ALIGN_DOWN()]
Signed-off-by: Michel Lespinasse <walken@google.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Chris Metcalf <cmetcalf@tilera.com>
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>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: use vm_unmapped_area() on mips architecture
Michel Lespinasse [Wed, 12 Dec 2012 00:02:06 +0000 (16:02 -0800)]
mm: use vm_unmapped_area() on mips architecture

Update the mips arch_get_unmapped_area[_topdown] functions to make use of
vm_unmapped_area() instead of implementing a brute force search.

[akpm@linux-foundation.org: remove now-unused COLOUR_ALIGN_DOWN()]
Signed-off-by: Michel Lespinasse <walken@google.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Chris Metcalf <cmetcalf@tilera.com>
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>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: use vm_unmapped_area() in hugetlbfs on i386 architecture
Michel Lespinasse [Wed, 12 Dec 2012 00:02:02 +0000 (16:02 -0800)]
mm: use vm_unmapped_area() in hugetlbfs on i386 architecture

Update the i386 hugetlb_get_unmapped_area function to make use of
vm_unmapped_area() instead of implementing a brute force search.

[akpm@linux-foundation.org: fix build]
Signed-off-by: Michel Lespinasse <walken@google.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Chris Metcalf <cmetcalf@tilera.com>
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>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: use vm_unmapped_area() in hugetlbfs
Michel Lespinasse [Wed, 12 Dec 2012 00:02:00 +0000 (16:02 -0800)]
mm: use vm_unmapped_area() in hugetlbfs

Update the hugetlb_get_unmapped_area function to make use of
vm_unmapped_area() instead of implementing a brute force search.

Signed-off-by: Michel Lespinasse <walken@google.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Chris Metcalf <cmetcalf@tilera.com>
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>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: fix cache coloring on x86_64 architecture
Michel Lespinasse [Wed, 12 Dec 2012 00:01:56 +0000 (16:01 -0800)]
mm: fix cache coloring on x86_64 architecture

Fix the x86-64 cache alignment code to take pgoff into account.  Use the
x86 and MIPS cache alignment code as the basis for a generic cache
alignment function.

The old x86 code will always align the mmap to aliasing boundaries,
even if the program mmaps the file with a non-zero pgoff.

If program A mmaps the file with pgoff 0, and program B mmaps the file
with pgoff 1.  The old code would align the mmaps, resulting in misaligned
pages:

  A:  0123
  B:  123

After this patch, they are aligned so the pages line up:

  A: 0123
  B:  123

Proposed by Rik van Riel.

Signed-off-by: Michel Lespinasse <walken@google.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Chris Metcalf <cmetcalf@tilera.com>
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>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: use vm_unmapped_area() on x86_64 architecture
Michel Lespinasse [Wed, 12 Dec 2012 00:01:52 +0000 (16:01 -0800)]
mm: use vm_unmapped_area() on x86_64 architecture

Update the x86_64 arch_get_unmapped_area[_topdown] functions to make use
of vm_unmapped_area() instead of implementing a brute force search.

Signed-off-by: Michel Lespinasse <walken@google.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Chris Metcalf <cmetcalf@tilera.com>
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>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: vm_unmapped_area() lookup function
Michel Lespinasse [Wed, 12 Dec 2012 00:01:49 +0000 (16:01 -0800)]
mm: vm_unmapped_area() lookup function

Implement vm_unmapped_area() using the rb_subtree_gap and highest_vm_end
information to look up for suitable virtual address space gaps.

struct vm_unmapped_area_info is used to define the desired allocation
request:
 - lowest or highest possible address matching the remaining constraints
 - desired gap length
 - low/high address limits that the gap must fit into
 - alignment mask and offset

Also update the generic arch_get_unmapped_area[_topdown] functions to make
use of vm_unmapped_area() instead of implementing a brute force search.

[akpm@linux-foundation.org: checkpatch fixes]
Signed-off-by: Michel Lespinasse <walken@google.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Chris Metcalf <cmetcalf@tilera.com>
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>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: rearrange vm_area_struct for fewer cache misses
Rik van Riel [Wed, 12 Dec 2012 00:01:44 +0000 (16:01 -0800)]
mm: rearrange vm_area_struct for fewer cache misses

The kernel walks the VMA rbtree in various places, including the page
fault path.  However, the vm_rb node spanned two cache lines, on 64 bit
systems with 64 byte cache lines (most x86 systems).

Rearrange vm_area_struct a little, so all the information we need to do a
VMA tree walk is in the first cache line.

[akpm@linux-foundation.org: checkpatch fixes]
Signed-off-by: Michel Lespinasse <walken@google.com>
Signed-off-by: Rik van Riel <riel@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Chris Metcalf <cmetcalf@tilera.com>
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>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: check rb_subtree_gap correctness
Michel Lespinasse [Wed, 12 Dec 2012 00:01:42 +0000 (16:01 -0800)]
mm: check rb_subtree_gap correctness

When CONFIG_DEBUG_VM_RB is enabled, check that rb_subtree_gap is correctly
set for every vma and that mm->highest_vm_end is also correct.

Also add an explicit 'bug' variable to track if browse_rb() detected any
invalid condition.

[akpm@linux-foundation.org: repair innovative coding-style inventions]
Signed-off-by: Michel Lespinasse <walken@google.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Chris Metcalf <cmetcalf@tilera.com>
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>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: augment vma rbtree with rb_subtree_gap
Michel Lespinasse [Wed, 12 Dec 2012 00:01:38 +0000 (16:01 -0800)]
mm: augment vma rbtree with rb_subtree_gap

Define vma->rb_subtree_gap as the largest gap between any vma in the
subtree rooted at that vma, and their predecessor.  Or, for a recursive
definition, vma->rb_subtree_gap is the max of:

 - vma->vm_start - vma->vm_prev->vm_end
 - rb_subtree_gap fields of the vmas pointed by vma->rb.rb_left and
   vma->rb.rb_right

This will allow get_unmapped_area_* to find a free area of the right
size in O(log(N)) time, instead of potentially having to do a linear
walk across all the VMAs.

Also define mm->highest_vm_end as the vm_end field of the highest vma,
so that we can easily check if the following gap is suitable.

This does have the potential to make unmapping VMAs more expensive,
especially for processes with very large numbers of VMAs, where the VMA
rbtree can grow quite deep.

Signed-off-by: Michel Lespinasse <walken@google.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Chris Metcalf <cmetcalf@tilera.com>
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>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agoselftests: add a test program for variable huge page sizes in mmap/shmget
Andi Kleen [Wed, 12 Dec 2012 00:01:36 +0000 (16:01 -0800)]
selftests: add a test program for variable huge page sizes in mmap/shmget

Also remove -Wextra because gcc-4.6 emits lots of irritating
signed/unsigned comparison warnings.

Signed-off-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: support more pagesizes for MAP_HUGETLB/SHM_HUGETLB
Andi Kleen [Wed, 12 Dec 2012 00:01:34 +0000 (16:01 -0800)]
mm: support more pagesizes for MAP_HUGETLB/SHM_HUGETLB

There was some desire in large applications using MAP_HUGETLB or
SHM_HUGETLB to use 1GB huge pages on some mappings, and stay with 2MB on
others.  This is useful together with NUMA policy: use 2MB interleaving
on some mappings, but 1GB on local mappings.

This patch extends the IPC/SHM syscall interfaces slightly to allow
specifying the page size.

It borrows some upper bits in the existing flag arguments and allows
encoding the log of the desired page size in addition to the *_HUGETLB
flag.  When 0 is specified the default size is used, this makes the
change fully compatible.

Extending the internal hugetlb code to handle this is straight forward.
Instead of a single mount it just keeps an array of them and selects the
right mount based on the specified page size.  When no page size is
specified it uses the mount of the default page size.

The change is not visible in /proc/mounts because internal mounts don't
appear there.  It also has very little overhead: the additional mounts
just consume a super block, but not more memory when not used.

I also exported the new flags to the user headers (they were previously
under __KERNEL__).  Right now only symbols for x86 and some other
architecture for 1GB and 2MB are defined.  The interface should already
work for all other architectures though.  Only architectures that define
multiple hugetlb sizes actually need it (that is currently x86, tile,
powerpc).  However tile and powerpc have user configurable hugetlb
sizes, so it's not easy to add defines.  A program on those
architectures would need to query sysfs and use the appropiate log2.

[akpm@linux-foundation.org: cleanups]
[rientjes@google.com: fix build]
[akpm@linux-foundation.org: checkpatch fixes]
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
Acked-by: Rik van Riel <riel@redhat.com>
Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Hillf Danton <dhillf@gmail.com>
Signed-off-by: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: hwpoison: fix action_result() to print out dirty/clean
Naoya Horiguchi [Wed, 12 Dec 2012 00:01:32 +0000 (16:01 -0800)]
mm: hwpoison: fix action_result() to print out dirty/clean

action_result() fails to print out "dirty" even if an error occurred on
a dirty pagecache, because when we check PageDirty in action_result() it
was cleared after page isolation even if it's dirty before error
handling.  This can break some applications that monitor this message,
so should be fixed.

There are several callers of action_result() except page_action(), but
either of them are not for LRU pages but for free pages or kernel pages,
so we don't have to consider dirty or not for them.

Note that PG_dirty can be set outside page locks as described in commit
6746aff74da2 ("HWPOISON: shmem: call set_page_dirty() with locked
page"), so this patch does not completely closes the race window, but
just narrows it.

Signed-off-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Reviewed-by: Andi Kleen <ak@linux.intel.com>
Cc: Tony Luck <tony.luck@intel.com>
Cc: "Jun'ichi Nomura" <j-nomura@ce.jp.nec.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agodmapool: make DMAPOOL_DEBUG detect corruption of free marker
Matthieu CASTET [Wed, 12 Dec 2012 00:01:31 +0000 (16:01 -0800)]
dmapool: make DMAPOOL_DEBUG detect corruption of free marker

This can help to catch the case where hardware is writing after dma free.

[akpm@linux-foundation.org: tidy code, fix comment, use sizeof(page->offset), use pr_err()]
Signed-off-by: Matthieu Castet <matthieu.castet@parrot.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm, oom: allow exiting threads to have access to memory reserves
David Rientjes [Wed, 12 Dec 2012 00:01:30 +0000 (16:01 -0800)]
mm, oom: allow exiting threads to have access to memory reserves

Exiting threads, those with PF_EXITING set, can pagefault and require
memory before they can make forward progress.  This happens, for instance,
when a process must fault task->robust_list, a userspace structure, before
detaching its memory.

These threads also aren't guaranteed to get access to memory reserves
unless oom killed or killed from userspace.  The oom killer won't grant
memory reserves if other threads are also exiting other than current and
stalling at the same point.  This prevents needlessly killing processes
when others are already exiting.

Instead of special casing all the possible situations between PF_EXITING
getting set and a thread detaching its mm where it may allocate memory,
which probably wouldn't get updated when a change is made to the exit
path, the solution is to give all exiting threads access to memory
reserves if they call the oom killer.  This allows them to quickly
allocate, detach its mm, and free the memory it represents.

Summary of Luigi's bug report:

: He had an oom condition where threads were faulting on task->robust_list
: and repeatedly called the oom killer but it would defer killing a thread
: because it saw other PF_EXITING threads.  This can happen anytime we need
: to allocate memory after setting PF_EXITING and before detaching our mm;
: if there are other threads in the same state then the oom killer won't do
: anything unless one of them happens to be killed from userspace.
:
: So instead of only deferring for PF_EXITING and !task->robust_list, it's
: better to just give them access to memory reserves to prevent a potential
: livelock so that any other faults that may be introduced in the future in
: the exit path don't cause the same problem (and hopefully we don't allow
: too many of those!).

Signed-off-by: David Rientjes <rientjes@google.com>
Acked-by: Minchan Kim <minchan@kernel.org>
Tested-by: Luigi Semenzato <semenzato@google.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agoDocumentation/cgroups/memory.txt: s/mem_cgroup_charge/mem_cgroup_change_common/
Jeff Liu [Wed, 12 Dec 2012 00:01:28 +0000 (16:01 -0800)]
Documentation/cgroups/memory.txt: s/mem_cgroup_charge/mem_cgroup_change_common/

mem_cgroup_charge_common() is invoked as the entry point for cgroup limits
charge rather than mem_cgroup_charge(), as the later has been removed for
years.  Update the cgroup/memory.txt to reflect this change.

Signed-off-by: Jie Liu <jeff.liu@oracle.com>
Cc: Ying Han <yinghan@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: thp: set the accessed flag for old pages on access fault
Will Deacon [Wed, 12 Dec 2012 00:01:27 +0000 (16:01 -0800)]
mm: thp: set the accessed flag for old pages on access fault

On x86 memory accesses to pages without the ACCESSED flag set result in
the ACCESSED flag being set automatically.  With the ARM architecture a
page access fault is raised instead (and it will continue to be raised
until the ACCESSED flag is set for the appropriate PTE/PMD).

For normal memory pages, handle_pte_fault will call pte_mkyoung
(effectively setting the ACCESSED flag).  For transparent huge pages,
pmd_mkyoung will only be called for a write fault.

This patch ensures that faults on transparent hugepages which do not
result in a CoW update the access flags for the faulting pmd.

Signed-off-by: Will Deacon <will.deacon@arm.com>
Cc: Chris Metcalf <cmetcalf@tilera.com>
Acked-by: Kirill A. Shutemov <kirill@shutemov.name>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Cc: Ni zhan Chen <nizhan.chen@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm, highmem: get virtual address of the page using PKMAP_ADDR()
Joonsoo Kim [Wed, 12 Dec 2012 00:01:24 +0000 (16:01 -0800)]
mm, highmem: get virtual address of the page using PKMAP_ADDR()

In flush_all_zero_pkmaps(), we have an index of the pkmap associated with
the page.  Using this index, we can simply get virtual address of the
page.  So change it.

Signed-off-by: Joonsoo Kim <js1304@gmail.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Reviewed-by: Minchan Kim <minchan@kernel.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm, highmem: remove page_address_pool list
Joonsoo Kim [Wed, 12 Dec 2012 00:01:23 +0000 (16:01 -0800)]
mm, highmem: remove page_address_pool list

We can find free page_address_map instance without the page_address_pool.
So remove it.

Signed-off-by: Joonsoo Kim <js1304@gmail.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Reviewed-by: Minchan Kim <minchan@kernel.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm, highmem: remove useless pool_lock
Joonsoo Kim [Wed, 12 Dec 2012 00:01:20 +0000 (16:01 -0800)]
mm, highmem: remove useless pool_lock

The pool_lock protects the page_address_pool from concurrent access.  But,
access to the page_address_pool is already protected by kmap_lock.  So
remove it.

Signed-off-by: Joonsoo Kim <js1304@gmail.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Reviewed-by: Minchan Kin <minchan@kernel.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm, highmem: use PKMAP_NR() to calculate an index of pkmap
Joonsoo Kim [Wed, 12 Dec 2012 00:01:17 +0000 (16:01 -0800)]
mm, highmem: use PKMAP_NR() to calculate an index of pkmap

To calculate an index of pkmap, using PKMAP_NR() is more understandable
and maintainable, so change it.

Signed-off-by: Joonsoo Kim <js1304@gmail.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Reviewed-by: Minchan Kim <minchan@kernel.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: do not call frontswap_init() during swapoff
Cesar Eduardo Barros [Wed, 12 Dec 2012 00:01:14 +0000 (16:01 -0800)]
mm: do not call frontswap_init() during swapoff

The call to frontswap_init() was added within enable_swap_info(), which
was called not only during sys_swapon, but also to reinsert the swap_info
into the swap_list in case of failure of try_to_unuse() within
sys_swapoff.  This means that frontswap_init() might be called more than
once for the same swap area.

While as far as I could see no frontswap implementation has any problem
with it (and in fact, all the ones I found ignore the parameter passed to
frontswap_init), this could change in the future.

To prevent future problems, move the call to frontswap_init() to outside
the code shared between sys_swapon and sys_swapoff.

Signed-off-by: Cesar Eduardo Barros <cesarb@cesarb.net>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Acked-by: Dan Magenheimer <dan.magenheimer@oracle.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Rik van Riel <riel@redhat.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Hugh Dickins <hughd@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: refactor reinsert of swap_info in sys_swapoff()
Cesar Eduardo Barros [Wed, 12 Dec 2012 00:01:13 +0000 (16:01 -0800)]
mm: refactor reinsert of swap_info in sys_swapoff()

The block within sys_swapoff() which re-inserts the swap_info into the
swap_list in case of failure of try_to_unuse() reads a few values outside
the swap_lock.  While this is safe at that point, it is subtle code.

Simplify the code by moving the reading of these values to a separate
function, refactoring it a bit so they are read from within the swap_lock.
 This is easier to understand, and matches better the way it worked before
I unified the insertion of the swap_info from both sys_swapon and
sys_swapoff.

This change should make no functional difference.  The only real change is
moving the read of two or three structure fields to within the lock
(frontswap_map_get() is nothing more than a read of p->frontswap_map).

Signed-off-by: Cesar Eduardo Barros <cesarb@cesarb.net>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Dan Magenheimer <dan.magenheimer@oracle.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Rik van Riel <riel@redhat.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Hugh Dickins <hughd@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm,vmscan: only evict file pages when we have plenty
Rik van Riel [Wed, 12 Dec 2012 00:01:10 +0000 (16:01 -0800)]
mm,vmscan: only evict file pages when we have plenty

If we have more inactive file pages than active file pages, we skip
scanning the active file pages altogether, with the idea that we do not
want to evict the working set when there is plenty of streaming IO in the
cache.

However, the code forgot to also skip scanning anonymous pages in that
situation.  That leads to the curious situation of keeping the active file
pages protected from being paged out when there are lots of inactive file
pages, while still scanning and evicting anonymous pages.

This patch fixes that situation, by only evicting file pages when we have
plenty of them and most are inactive.

[akpm@linux-foundation.org: adjust comment layout]
Signed-off-by: Rik van Riel <riel@redhat.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: add comment on storage key dirty bit semantics
Jan Kara [Wed, 12 Dec 2012 00:01:09 +0000 (16:01 -0800)]
mm: add comment on storage key dirty bit semantics

Add comments that dirty bit in storage key gets set whenever page content
is changed.  Hopefully if someone will use this function, he'll have a
look at one of the two places where we comment on this.

Signed-off-by: Jan Kara <jack@suse.cz>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Hugh Dickins <hughd@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm/memory_hotplug.c: update start_pfn in zone and pg_data when spanned_pages == 0.
Tang Chen [Wed, 12 Dec 2012 00:01:07 +0000 (16:01 -0800)]
mm/memory_hotplug.c: update start_pfn in zone and pg_data when spanned_pages == 0.

If we hot-remove memory only and leave the cpus alive, the corresponding
node will not be removed.  But the node_start_pfn and node_spanned_pages
in pg_data will be reset to 0.  In this case, when we hot-add the memory
back next time, the node_start_pfn will always be 0 because no pfn is less
than 0.  After that, if we hot-remove the memory again, it will cause
kernel panic in function find_biggest_section_pfn() when it tries to scan
all the pfns.

The zone will also have the same problem.

This patch sets start_pfn to the start_pfn of the section being added when
spanned_pages of the zone or pg_data is 0.

  ---How to reproduce---

1. hot-add a container with some memory and cpus;
2. hot-remove the container's memory, and leave cpus there;
3. hot-add these memory again;
4. hot-remove them again;

then, the kernel will panic.

  ---Call trace---

  BUG: unable to handle kernel paging request at 00000fff82a8cc38
  IP: [<ffffffff811c0d55>] find_biggest_section_pfn+0xe5/0x180
  ......
  Call Trace:
   [<ffffffff811c1124>] __remove_zone+0x184/0x1b0
   [<ffffffff811c11dc>] __remove_section+0x8c/0xb0
   [<ffffffff811c12e7>] __remove_pages+0xe7/0x120
   [<ffffffff81654f7c>] arch_remove_memory+0x2c/0x80
   [<ffffffff81655bb6>] remove_memory+0x56/0x90
   [<ffffffff813da0c8>] acpi_memory_device_remove_memory+0x48/0x73
   [<ffffffff813da55a>] acpi_memory_device_notify+0x153/0x274
   [<ffffffff813b6786>] acpi_ev_notify_dispatch+0x41/0x5f
   [<ffffffff813a3867>] acpi_os_execute_deferred+0x27/0x34
   [<ffffffff81090589>] process_one_work+0x219/0x680
   [<ffffffff810923be>] worker_thread+0x12e/0x320
   [<ffffffff81098396>] kthread+0xc6/0xd0
   [<ffffffff8167c7c4>] kernel_thread_helper+0x4/0x10
  ......
  ---[ end trace 96d845dbf33fee11 ]---

Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Wen Congyang <wency@cn.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agoslub, hotplug: ignore unrelated node's hot-adding and hot-removing
Lai Jiangshan [Wed, 12 Dec 2012 00:01:05 +0000 (16:01 -0800)]
slub, hotplug: ignore unrelated node's hot-adding and hot-removing

SLUB only focuses on the nodes which have normal memory and it ignores the
other node's hot-adding and hot-removing.

Aka: if some memory of a node which has no onlined memory is online, but
this new memory onlined is not normal memory (for example, highmem), we
should not allocate kmem_cache_node for SLUB.

And if the last normal memory is offlined, but the node still has memory,
we should remove kmem_cache_node for that node.  (The current code delays
it when all of the memory is offlined)

So we only do something when marg->status_change_nid_normal > 0.
marg->status_change_nid is not suitable here.

The same problem doesn't exist in SLAB, because SLAB allocates kmem_list3
for every node even the node don't have normal memory, SLAB tolerates
kmem_list3 on alien nodes.  SLUB only focuses on the nodes which have
normal memory, it don't tolerate alien kmem_cache_node.  The patch makes
SLUB become self-compatible and avoids WARNs and BUGs in rare conditions.

Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Rob Landley <rob@landley.net>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Jiang Liu <jiang.liu@huawei.com>
Cc: Kay Sievers <kay.sievers@vrfy.org>
Cc: Greg Kroah-Hartman <gregkh@suse.de>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Wen Congyang <wency@cn.fujitsu.com>
Acked-by: Christoph Lameter <cl@linux.com>
Cc: Pekka Enberg <penberg@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomemory_hotplug: fix possible incorrect node_states[N_NORMAL_MEMORY]
Lai Jiangshan [Wed, 12 Dec 2012 00:01:03 +0000 (16:01 -0800)]
memory_hotplug: fix possible incorrect node_states[N_NORMAL_MEMORY]

Currently memory_hotplug only manages the node_states[N_HIGH_MEMORY], it
forgets to manage node_states[N_NORMAL_MEMORY].  This may cause
node_states[N_NORMAL_MEMORY] to become incorrect.

Example, if a node is empty before online, and we online a memory which is
in ZONE_NORMAL.  And after online, node_states[N_HIGH_MEMORY] is correct,
but node_states[N_NORMAL_MEMORY] is incorrect, the online code doesn't set
the new online node to node_states[N_NORMAL_MEMORY].

The same thing will happen when offlining (the offline code doesn't clear
the node from node_states[N_NORMAL_MEMORY] when needed).  Some memory
managment code depends node_states[N_NORMAL_MEMORY], so we have to fix up
the node_states[N_NORMAL_MEMORY].

We add node_states_check_changes_online() and
node_states_check_changes_offline() to detect whether
node_states[N_HIGH_MEMORY] and node_states[N_NORMAL_MEMORY] are changed
while hotpluging.

Also add @status_change_nid_normal to struct memory_notify, thus the
memory hotplug callbacks know whether the node_states[N_NORMAL_MEMORY] are
changed.  (We can add a @flags and reuse @status_change_nid instead of
introducing @status_change_nid_normal, but it will add much more
complexity in memory hotplug callback in every subsystem.  So introducing
@status_change_nid_normal is better and it doesn't change the sematics of
@status_change_nid)

Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Rob Landley <rob@landley.net>
Cc: Jiang Liu <jiang.liu@huawei.com>
Cc: Kay Sievers <kay.sievers@vrfy.org>
Cc: Greg Kroah-Hartman <gregkh@suse.de>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Wen Congyang <wency@cn.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomemory-hotplug: allocate zone's pcp before onlining pages
Wen Congyang [Wed, 12 Dec 2012 00:01:01 +0000 (16:01 -0800)]
memory-hotplug: allocate zone's pcp before onlining pages

We use __free_page() to put a page to buddy system when onlining pages.
__free_page() will store NR_FREE_PAGES in zone's pcp.vm_stat_diff, so we
should allocate zone's pcp before onlining pages, otherwise we will lose
some free pages.

[mhocko@suse.cz: make zone_pcp_reset independent of MEMORY_HOTREMOVE]
Signed-off-by: Wen Congyang <wency@cn.fujitsu.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Jiang Liu <liuj97@gmail.com>
Cc: Len Brown <len.brown@intel.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Christoph Lameter <cl@linux.com>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Dave Hansen <dave@linux.vnet.ibm.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Michal Hocko <mhocko@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomemory-hotplug, mm/sparse.c: clear the memory to store struct page
Wen Congyang [Wed, 12 Dec 2012 00:00:59 +0000 (16:00 -0800)]
memory-hotplug, mm/sparse.c: clear the memory to store struct page

If sparse memory vmemmap is enabled, we can't free the memory to store
struct page when a memory device is hotremoved, because we may store
struct page in the memory to manage the memory which doesn't belong to
this memory device.  When we hotadded this memory device again, we will
reuse this memory to store struct page, and struct page may contain some
obsolete information, and we will get bad-page state:

  init_memory_mapping: [mem 0x80000000-0x9fffffff]
  Built 2 zonelists in Node order, mobility grouping on.  Total pages: 547617
  Policy zone: Normal
  BUG: Bad page state in process bash  pfn:9b6dc
  page:ffffea0002200020 count:0 mapcount:0 mapping:          (null) index:0xfdfdfdfdfdfdfdfd
  page flags: 0x2fdfdfdfd5df9fd(locked|referenced|uptodate|dirty|lru|active|slab|owner_priv_1|private|private_2|writeback|head|tail|swapcache|reclaim|swapbacked|unevictable|uncached|compound_lock)
  Modules linked in: netconsole acpiphp pci_hotplug acpi_memhotplug loop kvm_amd kvm microcode tpm_tis tpm tpm_bios evdev psmouse serio_raw i2c_piix4 i2c_core parport_pc parport processor button thermal_sys ext3 jbd mbcache sg sr_mod cdrom ata_generic virtio_net ata_piix virtio_blk libata virtio_pci virtio_ring virtio scsi_mod
  Pid: 988, comm: bash Not tainted 3.6.0-rc7-guest #12
  Call Trace:
   [<ffffffff810e9b30>] ? bad_page+0xb0/0x100
   [<ffffffff810ea4c3>] ? free_pages_prepare+0xb3/0x100
   [<ffffffff810ea668>] ? free_hot_cold_page+0x48/0x1a0
   [<ffffffff8112cc08>] ? online_pages_range+0x68/0xa0
   [<ffffffff8112cba0>] ? __online_page_increment_counters+0x10/0x10
   [<ffffffff81045561>] ? walk_system_ram_range+0x101/0x110
   [<ffffffff814c4f95>] ? online_pages+0x1a5/0x2b0
   [<ffffffff8135663d>] ? __memory_block_change_state+0x20d/0x270
   [<ffffffff81356756>] ? store_mem_state+0xb6/0xf0
   [<ffffffff8119e482>] ? sysfs_write_file+0xd2/0x160
   [<ffffffff8113769a>] ? vfs_write+0xaa/0x160
   [<ffffffff81137977>] ? sys_write+0x47/0x90
   [<ffffffff814e2f25>] ? async_page_fault+0x25/0x30
   [<ffffffff814ea239>] ? system_call_fastpath+0x16/0x1b
  Disabling lock debugging due to kernel taint

This patch clears the memory to store struct page to avoid unexpected error.

Signed-off-by: Wen Congyang <wency@cn.fujitsu.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Jiang Liu <liuj97@gmail.com>
Cc: Minchan Kim <minchan.kim@gmail.com>
Acked-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Reported-by: Vasilis Liaskovitis <vasilis.liaskovitis@profitbricks.com>
Cc: Dave Hansen <dave@linux.vnet.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomemory-hotplug: suppress "Device nodeX does not have a release() function" warning
Yasuaki Ishimatsu [Wed, 12 Dec 2012 00:00:57 +0000 (16:00 -0800)]
memory-hotplug: suppress "Device nodeX does not have a release() function" warning

When calling unregister_node(), the function shows following message at
device_release().

"Device 'node2' does not have a release() function, it is broken and must
be fixed."

The reason is node's device struct does not have a release() function.

So the patch registers node_device_release() to the device's release()
function for suppressing the warning message.  Additionally, the patch
adds memset() to initialize a node struct into register_node().  Because
the node struct is part of node_devices[] array and it cannot be freed by
node_device_release().  So if system reuses the node struct, it has a
garbage.

Signed-off-by: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Signed-off-by: Wen Congyang <wency@cn.fujitsu.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Jiang Liu <liuj97@gmail.com>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agonuma: convert static memory to dynamically allocated memory for per node device
Wen Congyang [Wed, 12 Dec 2012 00:00:56 +0000 (16:00 -0800)]
numa: convert static memory to dynamically allocated memory for per node device

We use a static array to store struct node.  In many cases, we don't have
too many nodes, and some memory will be unused.  Convert it to per-device
dynamically allocated memory.

Signed-off-by: Wen Congyang <wency@cn.fujitsu.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Jiang Liu <liuj97@gmail.com>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomemory-hotplug: fix NR_FREE_PAGES mismatch
Wen Congyang [Wed, 12 Dec 2012 00:00:52 +0000 (16:00 -0800)]
memory-hotplug: fix NR_FREE_PAGES mismatch

NR_FREE_PAGES will be wrong after offlining pages.  We add/dec
NR_FREE_PAGES like this now:

1. move all pages in buddy system to MIGRATE_ISOLATE, and dec NR_FREE_PAGES

2. don't add NR_FREE_PAGES when it is freed and the migratetype is
   MIGRATE_ISOLATE

3. dec NR_FREE_PAGES when offlining isolated pages.

4. add NR_FREE_PAGES when undoing isolate pages.

When we come to step 3, all pages are in MIGRATE_ISOLATE list, and
NR_FREE_PAGES are right.  When we come to step4, all pages are not in
buddy system, so we don't change NR_FREE_PAGES in this step, but we change
NR_FREE_PAGES in step3.  So NR_FREE_PAGES is wrong after offlining pages.
So there is no need to change NR_FREE_PAGES in step3.

This patch also fixs a problem in step2: if the migratetype is
MIGRATE_ISOLATE, we should not add NR_FRR_PAGES when we remove pages from
pcppages.

Signed-off-by: Wen Congyang <wency@cn.fujitsu.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Jiang Liu <liuj97@gmail.com>
Cc: Len Brown <len.brown@intel.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Christoph Lameter <cl@linux.com>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Dave Hansen <dave@linux.vnet.ibm.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Cc: Jianguo Wu <wujianguo106@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomemory-hotplug: auto offline page_cgroup when onlining memory block failed
Wen Congyang [Wed, 12 Dec 2012 00:00:49 +0000 (16:00 -0800)]
memory-hotplug: auto offline page_cgroup when onlining memory block failed

When a memory block is onlined, we will try allocate memory on that node
to store page_cgroup.  If onlining the memory block failed, we don't
offline the page cgroup, and we have no chance to offline this page cgroup
unless the memory block is onlined successfully again.  It will cause that
we can't hot-remove the memory device on that node, because some memory is
used to store page cgroup.  If onlining the memory block is failed, there
is no need to stort page cgroup for this memory.  So auto offline
page_cgroup when onlining memory block failed.

Signed-off-by: Wen Congyang <wency@cn.fujitsu.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Jiang Liu <liuj97@gmail.com>
Cc: Len Brown <len.brown@intel.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Christoph Lameter <cl@linux.com>
Cc: Minchan Kim <minchan.kim@gmail.com>
Acked-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Dave Hansen <dave@linux.vnet.ibm.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomemory-hotplug: update mce_bad_pages when removing the memory
Wen Congyang [Wed, 12 Dec 2012 00:00:47 +0000 (16:00 -0800)]
memory-hotplug: update mce_bad_pages when removing the memory

When we hotremove a memory device, we will free the memory to store struct
page.  If the page is hwpoisoned page, we should decrease mce_bad_pages.

[akpm@linux-foundation.org: cleanup ifdefs]
Signed-off-by: Wen Congyang <wency@cn.fujitsu.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Jiang Liu <liuj97@gmail.com>
Cc: Len Brown <len.brown@intel.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Christoph Lameter <cl@linux.com>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Dave Hansen <dave@linux.vnet.ibm.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomemory-hotplug: skip HWPoisoned page when offlining pages
Wen Congyang [Wed, 12 Dec 2012 00:00:45 +0000 (16:00 -0800)]
memory-hotplug: skip HWPoisoned page when offlining pages

hwpoisoned may be set when we offline a page by the sysfs interface
/sys/devices/system/memory/soft_offline_page or
/sys/devices/system/memory/hard_offline_page. If we don't clear
this flag when onlining pages, this page can't be freed, and will
not in free list. So we can't offline these pages again. So we
should skip such page when offlining pages.

Signed-off-by: Wen Congyang <wency@cn.fujitsu.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Jiang Liu <liuj97@gmail.com>
Cc: Len Brown <len.brown@intel.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Christoph Lameter <cl@linux.com>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomemory hotplug: suppress "Device memoryX does not have a release() function" warning
Yasuaki Ishimatsu [Wed, 12 Dec 2012 00:00:44 +0000 (16:00 -0800)]
memory hotplug: suppress "Device memoryX does not have a release() function" warning

When calling remove_memory_block(), the function shows following message
at device_release().

"Device 'memory528' does not have a release() function, it is broken and
must be fixed."

The reason is memory_block's device struct does not have a release()
function.

So the patch registers memory_block_release() to the device's release()
function for suppressing the warning message.  Additionally, the patch
moves kfree(mem) into the release function since the release function is
prepared as a means to free a memory_block struct.

Signed-off-by: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Acked-by: David Rientjes <rientjes@google.com>
Cc: Jiang Liu <liuj97@gmail.com>
Cc: Minchan Kim <minchan.kim@gmail.com>
Acked-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Wen Congyang <wency@cn.fujitsu.com>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agothp: cleanup: introduce mk_huge_pmd()
Bob Liu [Wed, 12 Dec 2012 00:00:41 +0000 (16:00 -0800)]
thp: cleanup: introduce mk_huge_pmd()

Introduce mk_huge_pmd() to simplify the code

Signed-off-by: Bob Liu <lliubbo@gmail.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Ni zhan Chen <nizhan.chen@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agothp: introduce hugepage_vma_check()
Bob Liu [Wed, 12 Dec 2012 00:00:39 +0000 (16:00 -0800)]
thp: introduce hugepage_vma_check()

Multiple places do the same check.

Signed-off-by: Bob Liu <lliubbo@gmail.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Ni zhan Chen <nizhan.chen@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: introduce mm_find_pmd()
Bob Liu [Wed, 12 Dec 2012 00:00:37 +0000 (16:00 -0800)]
mm: introduce mm_find_pmd()

Several place need to find the pmd by(mm_struct, address), so introduce a
function to simplify it.

[akpm@linux-foundation.org: fix warning]
Signed-off-by: Bob Liu <lliubbo@gmail.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Ni zhan Chen <nizhan.chen@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agothp: clean up __collapse_huge_page_isolate
Bob Liu [Wed, 12 Dec 2012 00:00:34 +0000 (16:00 -0800)]
thp: clean up __collapse_huge_page_isolate

There are duplicated places using release_pte_pages().
And release_all_pte_pages() can be removed.

Signed-off-by: Bob Liu <lliubbo@gmail.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Ni zhan Chen <nizhan.chen@gmail.com>
Acked-by: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: use IS_ENABLED(CONFIG_COMPACTION) instead of COMPACTION_BUILD
Kirill A. Shutemov [Wed, 12 Dec 2012 00:00:31 +0000 (16:00 -0800)]
mm: use IS_ENABLED(CONFIG_COMPACTION) instead of COMPACTION_BUILD

We don't need custom COMPACTION_BUILD anymore, since we have handy
IS_ENABLED().

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Acked-by: Minchan Kim <minchan@kernel.org>
Acked-by: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: use IS_ENABLED(CONFIG_NUMA) instead of NUMA_BUILD
Kirill A. Shutemov [Wed, 12 Dec 2012 00:00:29 +0000 (16:00 -0800)]
mm: use IS_ENABLED(CONFIG_NUMA) instead of NUMA_BUILD

We don't need custom NUMA_BUILD anymore, since we have handy
IS_ENABLED().

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Acked-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Acked-by: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm, memcg: make mem_cgroup_out_of_memory() static
David Rientjes [Wed, 12 Dec 2012 00:00:26 +0000 (16:00 -0800)]
mm, memcg: make mem_cgroup_out_of_memory() static

mem_cgroup_out_of_memory() is only referenced from within file scope, so
it can be marked static.

Signed-off-by: David Rientjes <rientjes@google.com>
Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Acked-by: Michal Hocko <mhocko@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: show migration types in show_mem
Rabin Vincent [Wed, 12 Dec 2012 00:00:24 +0000 (16:00 -0800)]
mm: show migration types in show_mem

This is useful to diagnose the reason for page allocation failure for
cases where there appear to be several free pages.

Example, with this alloc_pages(GFP_ATOMIC) failure:

 swapper/0: page allocation failure: order:0, mode:0x0
 ...
 Mem-info:
 Normal per-cpu:
 CPU    0: hi:   90, btch:  15 usd:  48
 CPU    1: hi:   90, btch:  15 usd:  21
 active_anon:0 inactive_anon:0 isolated_anon:0
  active_file:0 inactive_file:84 isolated_file:0
  unevictable:0 dirty:0 writeback:0 unstable:0
  free:4026 slab_reclaimable:75 slab_unreclaimable:484
  mapped:0 shmem:0 pagetables:0 bounce:0
 Normal free:16104kB min:2296kB low:2868kB high:3444kB active_anon:0kB
 inactive_anon:0kB active_file:0kB inactive_file:336kB unevictable:0kB
 isolated(anon):0kB isolated(file):0kB present:331776kB mlocked:0kB
 dirty:0kB writeback:0kB mapped:0kB shmem:0kB slab_reclaimable:300kB
 slab_unreclaimable:1936kB kernel_stack:328kB pagetables:0kB unstable:0kB
 bounce:0kB writeback_tmp:0kB pages_scanned:0 all_unreclaimable? no
 lowmem_reserve[]: 0 0

Before the patch, it's hard (for me, at least) to say why all these free
chunks weren't considered for allocation:

 Normal: 0*4kB 0*8kB 0*16kB 0*32kB 0*64kB 0*128kB 1*256kB 1*512kB
 1*1024kB 1*2048kB 3*4096kB = 16128kB

After the patch, it's obvious that the reason is that all of these are
in the MIGRATE_CMA (C) freelist:

 Normal: 0*4kB 0*8kB 0*16kB 0*32kB 0*64kB 0*128kB 1*256kB (C) 1*512kB
 (C) 1*1024kB (C) 1*2048kB (C) 3*4096kB (C) = 16128kB

Signed-off-by: Rabin Vincent <rabin.vincent@stericsson.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agowriteback: remove nr_pages_dirtied arg from balance_dirty_pages_ratelimited_nr()
Namjae Jeon [Wed, 12 Dec 2012 00:00:21 +0000 (16:00 -0800)]
writeback: remove nr_pages_dirtied arg from balance_dirty_pages_ratelimited_nr()

There is no reason to pass the nr_pages_dirtied argument, because
nr_pages_dirtied value from the caller is unused in
balance_dirty_pages_ratelimited_nr().

Signed-off-by: Namjae Jeon <linkinjeon@gmail.com>
Signed-off-by: Vivek Trivedi <vtrivedi018@gmail.com>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agoMerge tag 'devicetree-for-linus' of git://git.secretlab.ca/git/linux-2.6
Linus Torvalds [Tue, 11 Dec 2012 19:30:41 +0000 (11:30 -0800)]
Merge tag 'devicetree-for-linus' of git://git.secretlab.ca/git/linux-2.6

Pull device tree changes from Grant Likely:
 "Here are the DT changes I've got queued up for v3.8.  As described
  below, there are a lot of bug fixes here and documentation updates but
  nothing major:

  Bug fixes, little cleanups, and documentation changes.  The most
  invasive thing here touches a bunch of the arch directories to use a
  common build rule for .dtb files.  There are no major changes to
  functionality here other than a few new helper functions."

* tag 'devicetree-for-linus' of git://git.secretlab.ca/git/linux-2.6: (34 commits)
  arm64: Fix the dtbs target building
  mtd: nand: davinci: fix the binding documentation
  rtc: rtc-mv: Add the device tree binding documentation
  devicetree/bindings: Move gpio-leds binding into leds directory
  of/vendor-prefixes: add Imagination Technologies
  microblaze: use new common dtc rule
  c6x: use new common dtc rule
  openrisc: use new common dtc rule
  arm64: Add dtbs target for building all the enabled dtb files
  arm64: use new common dtc rule
  ARM: dt: change .dtb build rules to build in dts directory
  kbuild: centralize .dts->.dtb rule
  Fix build when CONFIG_W1_MASTER_GPIO=m b exporting "allnodes"
  of/spi: Honour "status=disabled" property of device
  of_mdio: Honour "status=disabled" property of device
  of_i2c: Honour "status=disabled" property of device
  powerpc: Fix fallout from device_node->name constification
  of: add 'const' for of_parse_phandle parameter *np
  Documentation: correct of_platform_populate() argument list
  script: dtc: clean generated files
  ...

11 years agoMerge tag 'irqdomain-for-linus' of git://git.secretlab.ca/git/linux-2.6
Linus Torvalds [Tue, 11 Dec 2012 19:30:08 +0000 (11:30 -0800)]
Merge tag 'irqdomain-for-linus' of git://git.secretlab.ca/git/linux-2.6

Pull irqdomain changes from Grant Likely:
 "Trivial changes to irqdomain.  An update to the documentation and make
  one of the error paths not quite so obnoxious."

* tag 'irqdomain-for-linus' of git://git.secretlab.ca/git/linux-2.6:
  irqdomain: update documentation
  irqdomain: stop screaming about preallocated irqdescs

11 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/bp/bp
Linus Torvalds [Tue, 11 Dec 2012 19:28:43 +0000 (11:28 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/bp/bp

Pull EDAC fixes from Borislav Petkov:

 - EDAC core error path fix, from Denis Kirjanov.

 - Generalization of AMD MCE bank names and some minor error reporting
   improvements.

 - EDAC core cleanups and simplifications, from Wei Yongjun.

 - amd64_edac fixes for sysfs-reported values, from Josh Hunt.

 - some heavy amd64_edac error reporting path shaving, leading to
   removing a bunch of code.

 - amd64_edac error injection method improvements.

 - EDAC core cleanups and fixes

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/bp/bp: (24 commits)
  EDAC, pci_sysfs: Use for_each_pci_dev to simplify the code
  EDAC: Handle error path in edac_mc_sysfs_init() properly
  MCE, AMD: Dump error status
  MCE, AMD: Report decoded error type first
  MCE, AMD: Dump CPU f/m/s triple with the error
  MCE, AMD: Remove functional unit references
  EDAC: Convert to use simple_open()
  EDAC, Calxeda highbank: Convert to use simple_open()
  EDAC: Fix mc size reported in sysfs
  EDAC: Fix csrow size reported in sysfs
  EDAC: Pass mci parent
  EDAC: Add memory controller flags
  amd64_edac: Fix csrows size and pages computation
  amd64_edac: Use DBAM_DIMM macro
  amd64_edac: Fix K8 chip select reporting
  amd64_edac: Reorganize error reporting path
  amd64_edac: Do not check whether error address is valid
  amd64_edac: Improve error injection
  amd64_edac: Cleanup error injection code
  amd64_edac: Small fixlets and cleanups
  ...

11 years agoMerge branch 'for-v3.8' of git://git.linaro.org/people/mszyprowski/linux-dma-mapping
Linus Torvalds [Tue, 11 Dec 2012 19:27:10 +0000 (11:27 -0800)]
Merge branch 'for-v3.8' of git://git.linaro.org/people/mszyprowski/linux-dma-mapping

Pull CMA and DMA-mapping update from Marek Szyprowski:
 "Another set of Contiguous Memory Allocator and DMA-mapping framework
  updates for v3.8.

  This pull request consists only of two patches.  The first fixes a
  long standing issue with dmapools (the code predates current GIT
  history), which forced all allocations to use GFP_ATOMIC flag,
  ignoring the flags passed by the caller.  The second patch changes CMA
  code to correctly use phys_addr_t type what enables support for LPAE
  systems."

* 'for-v3.8' of git://git.linaro.org/people/mszyprowski/linux-dma-mapping:
  drivers: cma: represent physical addresses as phys_addr_t
  mm: dmapool: use provided gfp flags for all dma_alloc_coherent() calls

11 years agoMerge tag 'clk-for-linus' of git://git.linaro.org/people/mturquette/linux
Linus Torvalds [Tue, 11 Dec 2012 19:25:08 +0000 (11:25 -0800)]
Merge tag 'clk-for-linus' of git://git.linaro.org/people/mturquette/linux

Pull clock framework changes from Mike Turquette:
 "The common clock framework changes for 3.8 are comprised of lots of
  fixes for existing platforms as well as new ports for some ARM
  platforms.  In addition there are new clk drivers for audio devices
  and MFDs."

Fix up trivial conflict in <linux/clk-provider.h> (removal of 'inline'
clashing with return type fixes)

* tag 'clk-for-linus' of git://git.linaro.org/people/mturquette/linux: (51 commits)
  MAINTAINERS: bad email address for Mike Turquette
  clk: introduce optional disable_unused callback
  clk: ux500: fix bit error
  clk: clock multiplexers may register out of order
  clk: ux500: Initial support for abx500 clock driver
  CLK: SPEAr: Remove unused dummy apb_pclk
  CLK: SPEAr: Correct index scanning done for clock synths
  CLK: SPEAr: Update clock rate table
  CLK: SPEAr: Add missing clocks
  CLK: SPEAr: Set CLK_SET_RATE_PARENT for few clocks
  CLK: SPEAr13xx: fix parent names of multiple clocks
  CLK: SPEAr13xx: Fix mux clock names
  CLK: SPEAr: Fix dev_id & con_id for multiple clocks
  clk: move IM-PD1 clocks to drivers/clk
  clk: make ICST driver handle the VCO registers
  clk: add GPLv2 headers to the Versatile clock files
  clk: mxs: Use a better name for the USB PHY clock
  clk: spear: Add stub functions for spear3[0|1|2]0_clk_init()
  CLK: clk-twl6040: fix return value check in twl6040_clk_probe()
  clk: ux500: Register nomadik keypad clock lookups for u8500
  ...

11 years agoMerge tag 'pinctrl-for-v3.8' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw...
Linus Torvalds [Tue, 11 Dec 2012 19:21:33 +0000 (11:21 -0800)]
Merge tag 'pinctrl-for-v3.8' of git://git./linux/kernel/git/linusw/linux-pinctrl

Pull pinctrl changes from Linus Walleij:
 "These are the first and major pinctrl changes for the v3.8 merge
  cycle.  Some of this is used as merge base for other trees so I better
  be early on the trigger.

  As can be seen from the diffstat the major changes are:

  - A big conversion of the AT91 pinctrl driver and the associated ACKed
    platform changes under arch/arm/max-at91 and its device trees.  This
    has been coordinated with the AT91 maintainers to go in through the
    pinctrl tree.

  - A larger chunk of changes to the SPEAr drivers and the addition of
    the "plgpio" driver for the SPEAr as well.

  - The removal of the remnants of the Nomadik driver from the arch/arm
    tree and fusion of that into the Nomadik driver and platform data
    header files.

  - Some local movement in the Marvell MVEBU drivers, these now have
    their own subdirectory.

  - The addition of a chunk of code to gpiolib under drivers/gpio to
    register gpio-to-pin range mappings from the GPIO side of things.
    This has been requested by Grant Likely and is now implemented, it
    is particularly useful for device tree work.

  Then we have incremental updates all over the place, many of these are
  cleanups and fixes from Axel Lin who has done a great job of removing
  minor mistakes and compilation annoyances."

* tag 'pinctrl-for-v3.8' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl: (114 commits)
  ARM: mmp: select PINCTRL for ARCH_MMP
  pinctrl: Drop selecting PINCONF for MMP2, PXA168 and PXA910
  pinctrl: pinctrl-single: Fix error check condition
  pinctrl: SPEAr: Update error check for unsigned variables
  gpiolib: Fix use after free in gpiochip_add_pin_range
  gpiolib: rename pin range arguments
  pinctrl: single: support gpio request and free
  pinctrl: generic: add input schmitt disable parameter
  pinctrl/u300/coh901: stop spawning pinctrl from GPIO
  pinctrl/u300/coh901: let the gpio_chip register the range
  pinctrl: add function to retrieve range from pin
  gpiolib: return any error code from range creation
  pinctrl: make range registration defer properly
  gpiolib: rename find_pinctrl_*
  gpiolib: let gpiochip_add_pin_range() specify offset
  ARM: at91: pm9g45: add mmc support
  ARM: at91: Animeo IP: add mmc support
  ARM: at91: dt: add mmc pinctrl for Atmel reference boards
  ARM: at91: dt: at91sam9: add mmc pinctrl support
  ARM: at91/dts: add nodes for atmel hsmci controllers for atmel boards
  ...

11 years agoMerge tag 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck...
Linus Torvalds [Tue, 11 Dec 2012 19:20:34 +0000 (11:20 -0800)]
Merge tag 'hwmon-for-linus' of git://git./linux/kernel/git/groeck/linux-staging

Pull hwmon updates from Guenter Roeck:
 "New driver: DA9055

  Added/improved support for new chips in existing drivers: Z650/670,
  N550/570, ADS7830, AMD 16h family"

* tag 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging:
  hwmon: (da9055) Fix chan_mux[DA9055_ADC_ADCIN3] setting
  hwmon: DA9055 HWMON driver
  hwmon: (coretemp) List TjMax for Z650/670 and N550/570
  hwmon: (coretemp) Drop N4xx, N5xx, D4xx, D5xx CPUs from tjmax table
  hwmon: (coretemp) Use model table instead of if/else to identify CPU models
  hwmon: da9052: Use da9052_reg_update for rmw operations
  hwmon: (coretemp) Drop dependency on PCI for TjMax detection on Atom CPUs
  hwmon: (ina2xx) use module_i2c_driver to simplify the code
  hwmon: (ads7828) add support for ADS7830
  hwmon: (ads7828) driver cleanup
  x86,AMD: Power driver support for AMD's family 16h processors

11 years agoMerge tag 'mmc-updates-for-3.8-rc1' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Tue, 11 Dec 2012 19:19:09 +0000 (11:19 -0800)]
Merge tag 'mmc-updates-for-3.8-rc1' of git://git./linux/kernel/git/cjb/mmc

Pull MMC updates from Chris Ball:
 "MMC highlights for 3.8:

  Core:
   - Expose access to the eMMC RPMB ("Replay Protected Memory Block")
     area by extending the existing mmc_block ioctl.
   - Add SDIO powered-suspend DT properties to the core MMC DT binding.
   - Add no-1-8-v DT flag for boards where the SD controller reports
     that it supports 1.8V but the board itself has no way to switch to
     1.8V.
   - More work on switching to 1.8V UHS support using a vqmmc regulator.
   - Fix up a case where the slot-gpio helper may fail to reset the host
     controller properly if a card was removed during a transfer.
   - Fix several cases where a broken device could cause an infinite
     loop while we wait for a register to update.

  Drivers:
   - at91-mci: Remove obsolete driver, atmel-mci handles these devices
     now.
   - sdhci-dove: Allow using GPIOs for card-detect notifications.
   - sdhci-esdhc: Fix for recovering from ADMA errors on broken silicon.
   - sdhci-s3c: Add pinctrl support.
   - wmt-sdmmc: New driver for WonderMedia SD/MMC controllers."

* tag 'mmc-updates-for-3.8-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/cjb/mmc: (65 commits)
  mmc: sdhci: implement the .card_event() method
  mmc: extend the slot-gpio card-detection to use host's .card_event() method
  mmc: add a card-event host operation
  mmc: sdhci-s3c: Fix compilation warning
  mmc: sdhci-pci: Enable SDHCI_CAN_DO_HISPD for Ricoh SDHCI controller
  mmc: sdhci-dove: allow GPIOs to be used for card detection on Dove
  mmc: sdhci-dove: use two-stage initialization for sdhci-pltfm
  mmc: sdhci-dove: use devm_clk_get()
  mmc: eSDHC: Recover from ADMA errors
  mmc: dw_mmc: remove duplicated buswidth code
  mmc: dw_mmc: relocate where dw_mci_setup_bus() is called from
  mmc: Limit MMC speed to 52MHz if not HS200
  mmc: dw_mmc: use devres functions in dw_mmc
  mmc: sh_mmcif: remove unneeded clock connection ID
  mmc: sh_mobile_sdhi: remove unneeded clock connection ID
  mmc: sh_mobile_sdhi: fix clock frequency printing
  mmc: Remove redundant null check before kfree in bus.c
  mmc: Remove redundant null check before kfree in sdio_bus.c
  mmc: sdhci-imx-esdhc: use more devm_* functions
  mmc: dt: add no-1-8-v device tree flag
  ...

11 years agodrivers: cma: represent physical addresses as phys_addr_t
Vitaly Andrianov [Wed, 5 Dec 2012 14:29:25 +0000 (09:29 -0500)]
drivers: cma: represent physical addresses as phys_addr_t

This commit changes the CMA early initialization code to use phys_addr_t
for representing physical addresses instead of unsigned long.

Without this change, among other things, dma_declare_contiguous() simply
discards any memory regions whose address is not representable as unsigned
long.

This is a problem on 32-bit PAE machines where unsigned long is 32-bit
but physical address space is larger.

Signed-off-by: Vitaly Andrianov <vitalya@ti.com>
Signed-off-by: Cyril Chemparathy <cyril@ti.com>
Acked-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
11 years agomm: dmapool: use provided gfp flags for all dma_alloc_coherent() calls
Marek Szyprowski [Wed, 7 Nov 2012 14:37:07 +0000 (15:37 +0100)]
mm: dmapool: use provided gfp flags for all dma_alloc_coherent() calls

dmapool always calls dma_alloc_coherent() with GFP_ATOMIC flag,
regardless the flags provided by the caller. This causes excessive
pruning of emergency memory pools without any good reason. Additionaly,
on ARM architecture any driver which is using dmapools will sooner or
later  trigger the following error:
"ERROR: 256 KiB atomic DMA coherent pool is too small!
Please increase it with coherent_pool= kernel parameter!".
Increasing the coherent pool size usually doesn't help much and only
delays such error, because all GFP_ATOMIC DMA allocations are always
served from the special, very limited memory pool.

This patch changes the dmapool code to correctly use gfp flags provided
by the dmapool caller.

Reported-by: Soeren Moch <smoch@web.de>
Reported-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Soeren Moch <smoch@web.de>
Cc: stable@vger.kernel.org
11 years agoMAINTAINERS: bad email address for Mike Turquette
Mike Turquette [Tue, 4 Dec 2012 19:20:46 +0000 (11:20 -0800)]
MAINTAINERS: bad email address for Mike Turquette

Signed-off-by: Mike Turquette <mturquette@linaro.org>
11 years agoclk: introduce optional disable_unused callback
Mike Turquette [Tue, 4 Dec 2012 19:00:35 +0000 (11:00 -0800)]
clk: introduce optional disable_unused callback

Some gate clocks have special needs which must be handled during the
disable-unused clocks sequence.  These needs might be driven by software
due to the fact that we're disabling a clock outside of the normal
clk_disable path and a clk's enable_count will not be accurate.  On the
other hand a specific hardware programming sequence might need to be
followed for this corner case.

This change is needed for the upcoming OMAP port to the common clock
framework.  Specifically, it is undesirable to treat the disable-unused
path identically to the normal clk_disable path since other software
layers are involved.  In this case OMAP's clockdomain code throws WARNs
and bails early due to the clock's enable_count being set to zero.  A
custom callback mitigates this problem nicely.

Cc: Paul Walmsley <paul@pwsan.com>
Acked-by: Ulf Hansson <ulf.hansson@linaro.org>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
11 years agoLinux 3.7
Linus Torvalds [Tue, 11 Dec 2012 03:30:57 +0000 (19:30 -0800)]
Linux 3.7

11 years agoarm64: Fix the dtbs target building
Catalin Marinas [Wed, 21 Nov 2012 11:44:59 +0000 (11:44 +0000)]
arm64: Fix the dtbs target building

The arch/arm64/Makefile was not passing the right target to the
boot/dts/Makefile.

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Rob Herring <rob.herring@calxeda.com>
11 years agoInput: matrix-keymap - provide proper module license
Florian Fainelli [Mon, 10 Dec 2012 20:25:32 +0000 (12:25 -0800)]
Input: matrix-keymap - provide proper module license

The matrix-keymap module is currently lacking a proper module license,
add one so we don't have this module tainting the entire kernel.  This
issue has been present since commit 1932811f426f ("Input: matrix-keymap
- uninline and prepare for device tree support")

Signed-off-by: Florian Fainelli <florian@openwrt.org>
CC: stable@vger.kernel.org # v3.5+
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
Linus Torvalds [Tue, 11 Dec 2012 00:07:11 +0000 (16:07 -0800)]
Merge git://git./linux/kernel/git/davem/net

Pull networking fixes from David Miller:

 1) Netlink socket dumping had several missing verifications and checks.

    In particular, address comparisons in the request byte code
    interpreter could access past the end of the address in the
    inet_request_sock.

    Also, address family and address prefix lengths were not validated
    properly at all.

    This means arbitrary applications can read past the end of certain
    kernel data structures.

    Fixes from Neal Cardwell.

 2) ip_check_defrag() operates in contexts where we're in the process
    of, or about to, input the packet into the real protocols
    (specifically macvlan and AF_PACKET snooping).

    Unfortunately, it does a pskb_may_pull() which can modify the
    backing packet data which is not legal if the SKB is shared.  It
    very much can be shared in this context.

    Deal with the possibility that the SKB is segmented by using
    skb_copy_bits().

    Fix from Johannes Berg based upon a report by Eric Leblond.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net:
  ipv4: ip_check_defrag must not modify skb before unsharing
  inet_diag: validate port comparison byte code to prevent unsafe reads
  inet_diag: avoid unsafe and nonsensical prefix matches in inet_diag_bc_run()
  inet_diag: validate byte code to prevent oops in inet_diag_bc_run()
  inet_diag: fix oops for IPv4 AF_INET6 TCP SYN-RECV state

11 years agomtd: nand: davinci: fix the binding documentation
Kumar, Anil [Fri, 7 Dec 2012 09:09:34 +0000 (14:39 +0530)]
mtd: nand: davinci: fix the binding documentation

Since the aemif driver conversion to DT along with
its movement to drivers/ folder is not yet done,
fix NAND binding documentation to have NAND specific
DT details only.

Signed-off-by: Kumar, Anil <anilkumar.v@ti.com>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
11 years agortc: rtc-mv: Add the device tree binding documentation
Gregory CLEMENT [Mon, 10 Dec 2012 10:42:02 +0000 (11:42 +0100)]
rtc: rtc-mv: Add the device tree binding documentation

The support was already written, but the binding documentation was
lacking.

Signed-off-by: Gregory CLEMENT <gregory.clement@free-electrons.com>
Acked-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
11 years agoRevert "revert "Revert "mm: remove __GFP_NO_KSWAPD""" and associated damage
Linus Torvalds [Mon, 10 Dec 2012 18:51:16 +0000 (10:51 -0800)]
Revert "revert "Revert "mm: remove __GFP_NO_KSWAPD""" and associated damage

This reverts commits a50915394f1fc02c2861d3b7ce7014788aa5066e and
d7c3b937bdf45f0b844400b7bf6fd3ed50bac604.

This is a revert of a revert of a revert.  In addition, it reverts the
even older i915 change to stop using the __GFP_NO_KSWAPD flag due to the
original commits in linux-next.

It turns out that the original patch really was bogus, and that the
original revert was the correct thing to do after all.  We thought we
had fixed the problem, and then reverted the revert, but the problem
really is fundamental: waking up kswapd simply isn't the right thing to
do, and direct reclaim sometimes simply _is_ the right thing to do.

When certain allocations fail, we simply should try some direct reclaim,
and if that fails, fail the allocation.  That's the right thing to do
for THP allocations, which can easily fail, and the GPU allocations want
to do that too.

So starting kswapd is sometimes simply wrong, and removing the flag that
said "don't start kswapd" was a mistake.  Let's hope we never revisit
this mistake again - and certainly not this many times ;)

Acked-by: Mel Gorman <mgorman@suse.de>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agoipv4: ip_check_defrag must not modify skb before unsharing
Johannes Berg [Sun, 9 Dec 2012 23:41:06 +0000 (23:41 +0000)]
ipv4: ip_check_defrag must not modify skb before unsharing

ip_check_defrag() might be called from af_packet within the
RX path where shared SKBs are used, so it must not modify
the input SKB before it has unshared it for defragmentation.
Use skb_copy_bits() to get the IP header and only pull in
everything later.

The same is true for the other caller in macvlan as it is
called from dev->rx_handler which can also get a shared SKB.

Reported-by: Eric Leblond <eric@regit.org>
Cc: stable@vger.kernel.org
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoRevert "mm: avoid waking kswapd for THP allocations when compaction is deferred or...
Linus Torvalds [Mon, 10 Dec 2012 18:47:45 +0000 (10:47 -0800)]
Revert "mm: avoid waking kswapd for THP allocations when compaction is deferred or contended"

This reverts commit 782fd30406ecb9d9b082816abe0c6008fc72a7b0.

We are going to reinstate the __GFP_NO_KSWAPD flag that has been
removed, the removal reverted, and then removed again.  Making this
commit a pointless fixup for a problem that was caused by the removal of
__GFP_NO_KSWAPD flag.

The thing is, we really don't want to wake up kswapd for THP allocations
(because they fail quite commonly under any kind of memory pressure,
including when there is tons of memory free), and these patches were
just trying to fix up the underlying bug: the original removal of
__GFP_NO_KSWAPD in commit c654345924f7 ("mm: remove __GFP_NO_KSWAPD")
was simply bogus.

Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agoinet_diag: validate port comparison byte code to prevent unsafe reads
Neal Cardwell [Sun, 9 Dec 2012 11:09:54 +0000 (11:09 +0000)]
inet_diag: validate port comparison byte code to prevent unsafe reads

Add logic to verify that a port comparison byte code operation
actually has the second inet_diag_bc_op from which we read the port
for such operations.

Previously the code blindly referenced op[1] without first checking
whether a second inet_diag_bc_op struct could fit there. So a
malicious user could make the kernel read 4 bytes beyond the end of
the bytecode array by claiming to have a whole port comparison byte
code (2 inet_diag_bc_op structs) when in fact the bytecode was not
long enough to hold both.

Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoinet_diag: avoid unsafe and nonsensical prefix matches in inet_diag_bc_run()
Neal Cardwell [Sat, 8 Dec 2012 19:43:23 +0000 (19:43 +0000)]
inet_diag: avoid unsafe and nonsensical prefix matches in inet_diag_bc_run()

Add logic to check the address family of the user-supplied conditional
and the address family of the connection entry. We now do not do
prefix matching of addresses from different address families (AF_INET
vs AF_INET6), except for the previously existing support for having an
IPv4 prefix match an IPv4-mapped IPv6 address (which this commit
maintains as-is).

This change is needed for two reasons:

(1) The addresses are different lengths, so comparing a 128-bit IPv6
prefix match condition to a 32-bit IPv4 connection address can cause
us to unwittingly walk off the end of the IPv4 address and read
garbage or oops.

(2) The IPv4 and IPv6 address spaces are semantically distinct, so a
simple bit-wise comparison of the prefixes is not meaningful, and
would lead to bogus results (except for the IPv4-mapped IPv6 case,
which this commit maintains).

Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoinet_diag: validate byte code to prevent oops in inet_diag_bc_run()
Neal Cardwell [Sat, 8 Dec 2012 19:43:22 +0000 (19:43 +0000)]
inet_diag: validate byte code to prevent oops in inet_diag_bc_run()

Add logic to validate INET_DIAG_BC_S_COND and INET_DIAG_BC_D_COND
operations.

Previously we did not validate the inet_diag_hostcond, address family,
address length, and prefix length. So a malicious user could make the
kernel read beyond the end of the bytecode array by claiming to have a
whole inet_diag_hostcond when the bytecode was not long enough to
contain a whole inet_diag_hostcond of the given address family. Or
they could make the kernel read up to about 27 bytes beyond the end of
a connection address by passing a prefix length that exceeded the
length of addresses of the given family.

Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoinet_diag: fix oops for IPv4 AF_INET6 TCP SYN-RECV state
Neal Cardwell [Sat, 8 Dec 2012 19:43:21 +0000 (19:43 +0000)]
inet_diag: fix oops for IPv4 AF_INET6 TCP SYN-RECV state

Fix inet_diag to be aware of the fact that AF_INET6 TCP connections
instantiated for IPv4 traffic and in the SYN-RECV state were actually
created with inet_reqsk_alloc(), instead of inet6_reqsk_alloc(). This
means that for such connections inet6_rsk(req) returns a pointer to a
random spot in memory up to roughly 64KB beyond the end of the
request_sock.

With this bug, for a server using AF_INET6 TCP sockets and serving
IPv4 traffic, an inet_diag user like `ss state SYN-RECV` would lead to
inet_diag_fill_req() causing an oops or the export to user space of 16
bytes of kernel memory as a garbage IPv6 address, depending on where
the garbage inet6_rsk(req) pointed.

Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agomm: vmscan: fix inappropriate zone congestion clearing
Johannes Weiner [Thu, 6 Dec 2012 20:23:25 +0000 (15:23 -0500)]
mm: vmscan: fix inappropriate zone congestion clearing

commit c702418f8a2f ("mm: vmscan: do not keep kswapd looping forever due
to individual uncompactable zones") removed zone watermark checks from
the compaction code in kswapd but left in the zone congestion clearing,
which now happens unconditionally on higher order reclaim.

This messes up the reclaim throttling logic for zones with
dirty/writeback pages, where zones should only lose their congestion
status when their watermarks have been restored.

Remove the clearing from the zone compaction section entirely.  The
preliminary zone check and the reclaim loop in kswapd will clear it if
the zone is considered balanced.

Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Reviewed-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agovfs: fix O_DIRECT read past end of block device
Linus Torvalds [Sat, 8 Dec 2012 00:48:39 +0000 (16:48 -0800)]
vfs: fix O_DIRECT read past end of block device

The direct-IO write path already had the i_size checks in mm/filemap.c,
but it turns out the read path did not, and removing the block size
checks in fs/block_dev.c (commit bbec0270bdd8: "blkdev_max_block: make
private to fs/buffer.c") removed the magic "shrink IO to past the end of
the device" code there.

Fix it by truncating the IO to the size of the block device, like the
write path already does.

NOTE! I suspect the write path would be *much* better off doing it this
way in fs/block_dev.c, rather than hidden deep in mm/filemap.c.  The
mm/filemap.c code is extremely hard to follow, and has various
conditionals on the target being a block device (ie the flag passed in
to 'generic_write_checks()', along with a conditional update of the
inode timestamp etc).

It is also quite possible that we should treat this whole block device
size as a "s_maxbytes" issue, and try to make the logic even more
generic.  However, in the meantime this is the fairly minimal targeted
fix.

Noted by Milan Broz thanks to a regression test for the cryptsetup
reencrypt tool.

Reported-and-tested-by: Milan Broz <mbroz@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
Linus Torvalds [Sat, 8 Dec 2012 01:00:57 +0000 (17:00 -0800)]
Merge git://git./linux/kernel/git/davem/net

Pull networking fixes from David Miller:
 "Two stragglers:

   1) The new code that adds new flushing semantics to GRO can cause SKB
      pointer list corruption, manage the lists differently to avoid the
      OOPS.  Fix from Eric Dumazet.

   2) When TCP fast open does a retransmit of data in a SYN-ACK or
      similar, we update retransmit state that we shouldn't triggering a
      WARN_ON later.  Fix from Yuchung Cheng."

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net:
  net: gro: fix possible panic in skb_gro_receive()
  tcp: bug fix Fast Open client retransmission

11 years agonet: gro: fix possible panic in skb_gro_receive()
Eric Dumazet [Thu, 6 Dec 2012 13:54:59 +0000 (13:54 +0000)]
net: gro: fix possible panic in skb_gro_receive()

commit 2e71a6f8084e (net: gro: selective flush of packets) added
a bug for skbs using frag_list. This part of the GRO stack is rarely
used, as it needs skb not using a page fragment for their skb->head.

Most drivers do use a page fragment, but some of them use GFP_KERNEL
allocations for the initial fill of their RX ring buffer.

napi_gro_flush() overwrite skb->prev that was used for these skb to
point to the last skb in frag_list.

Fix this using a separate field in struct napi_gro_cb to point to the
last fragment.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agotcp: bug fix Fast Open client retransmission
Yuchung Cheng [Thu, 6 Dec 2012 08:45:32 +0000 (08:45 +0000)]
tcp: bug fix Fast Open client retransmission

If SYN-ACK partially acks SYN-data, the client retransmits the
remaining data by tcp_retransmit_skb(). This increments lost recovery
state variables like tp->retrans_out in Open state. If loss recovery
happens before the retransmission is acked, it triggers the WARN_ON
check in tcp_fastretrans_alert(). For example: the client sends
SYN-data, gets SYN-ACK acking only ISN, retransmits data, sends
another 4 data packets and get 3 dupacks.

Since the retransmission is not caused by network drop it should not
update the recovery state variables. Further the server may return a
smaller MSS than the cached MSS used for SYN-data, so the retranmission
needs a loop. Otherwise some data will not be retransmitted until timeout
or other loss recovery events.

Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agommc: sdhci: implement the .card_event() method
Guennadi Liakhovetski [Tue, 4 Dec 2012 15:51:40 +0000 (16:51 +0100)]
mmc: sdhci: implement the .card_event() method

Extracting a part of the SDHCI card tasklet into a .card_event()
implementation allows SDHCI hosts to use generic card-detection
services, e.g. the GPIO slot function.

Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Reviewed-by: Shawn Guo <shawn.guo@linaro.org>
Signed-off-by: Chris Ball <cjb@laptop.org>
11 years agommc: extend the slot-gpio card-detection to use host's .card_event() method
Guennadi Liakhovetski [Tue, 4 Dec 2012 15:51:36 +0000 (16:51 +0100)]
mmc: extend the slot-gpio card-detection to use host's .card_event() method

The slot-gpio API provides a generic card-detection handler. To support a
wider range of hosts it has to call the host's card-event callback, if
implemented. Also increase the debounce interval to 200ms to match the
SDHCI driver.

Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Reviewed-by: Shawn Guo <shawn.guo@linaro.org>
Signed-off-by: Chris Ball <cjb@laptop.org>
11 years agommc: add a card-event host operation
Guennadi Liakhovetski [Tue, 4 Dec 2012 15:51:32 +0000 (16:51 +0100)]
mmc: add a card-event host operation

Some hosts need to perform additional actions upon card insertion or
ejection. Add a host operation to be called from card detection handlers.

Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Reviewed-by: Shawn Guo <shawn.guo@linaro.org>
Signed-off-by: Chris Ball <cjb@laptop.org>
11 years agommc: sdhci-s3c: Fix compilation warning
Sachin Kamat [Tue, 4 Dec 2012 11:33:07 +0000 (17:03 +0530)]
mmc: sdhci-s3c: Fix compilation warning

'sc' is used only when CONFIG_PM_RUNTIME is defined. Hence define it
conditionally.

Silences the following warning:
drivers/mmc/host/sdhci-s3c.c: In function ‘sdhci_s3c_notify_change’:
drivers/mmc/host/sdhci-s3c.c:378:20: warning: unused variable ‘sc’ [-Wunused-variable]

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Acked-by: Jaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
11 years agoMerge tag 'mmc-fixes-for-3.7' of git://git.kernel.org/pub/scm/linux/kernel/git/cjb/mmc
Linus Torvalds [Fri, 7 Dec 2012 17:15:20 +0000 (09:15 -0800)]
Merge tag 'mmc-fixes-for-3.7' of git://git./linux/kernel/git/cjb/mmc

Pull MMC fixes from Chris Ball:
 "Two small regression fixes:

   - sdhci-s3c: Fix runtime PM regression against 3.7-rc1
   - sh-mmcif: Fix oops against 3.6"

* tag 'mmc-fixes-for-3.7' of git://git.kernel.org/pub/scm/linux/kernel/git/cjb/mmc:
  mmc: sh-mmcif: avoid oops on spurious interrupts (second try)
  Revert misapplied "mmc: sh-mmcif: avoid oops on spurious interrupts"
  mmc: sdhci-s3c: fix missing clock for gpio card-detect

12 years agotmpfs: fix shared mempolicy leak
Mel Gorman [Wed, 5 Dec 2012 22:01:41 +0000 (14:01 -0800)]
tmpfs: fix shared mempolicy leak

This fixes a regression in 3.7-rc, which has since gone into stable.

Commit 00442ad04a5e ("mempolicy: fix a memory corruption by refcount
imbalance in alloc_pages_vma()") changed get_vma_policy() to raise the
refcount on a shmem shared mempolicy; whereas shmem_alloc_page() went
on expecting alloc_page_vma() to drop the refcount it had acquired.
This deserves a rework: but for now fix the leak in shmem_alloc_page().

Hugh: shmem_swapin() did not need a fix, but surely it's clearer to use
the same refcounting there as in shmem_alloc_page(), delete its onstack
mempolicy, and the strange mpol_cond_copy() and __mpol_cond_copy() -
those were invented to let swapin_readahead() make an unknown number of
calls to alloc_pages_vma() with one mempolicy; but since 00442ad04a5e,
alloc_pages_vma() has kept refcount in balance, so now no problem.

Reported-and-tested-by: Tommi Rantala <tt.rantala@gmail.com>
Signed-off-by: Mel Gorman <mgorman@suse.de>
Signed-off-by: Hugh Dickins <hughd@google.com>
Cc: stable@vger.kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agomm: vmscan: do not keep kswapd looping forever due to individual uncompactable zones
Johannes Weiner [Tue, 4 Dec 2012 16:11:31 +0000 (11:11 -0500)]
mm: vmscan: do not keep kswapd looping forever due to individual uncompactable zones

When a zone meets its high watermark and is compactable in case of
higher order allocations, it contributes to the percentage of the node's
memory that is considered balanced.

This requirement, that a node be only partially balanced, came about
when kswapd was desparately trying to balance tiny zones when all bigger
zones in the node had plenty of free memory.  Arguably, the same should
apply to compaction: if a significant part of the node is balanced
enough to run compaction, do not get hung up on that tiny zone that
might never get in shape.

When the compaction logic in kswapd is reached, we know that at least
25% of the node's memory is balanced properly for compaction (see
zone_balanced and pgdat_balanced).  Remove the individual zone checks
that restart the kswapd cycle.

Otherwise, we may observe more endless looping in kswapd where the
compaction code loops back to reclaim because of a single zone and
reclaim does nothing because the node is considered balanced overall.

See for example

  https://bugzilla.redhat.com/show_bug.cgi?id=866988

Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Reported-and-tested-by: Thorsten Leemhuis <fedora@leemhuis.info>
Reported-by: Jiri Slaby <jslaby@suse.cz>
Tested-by: John Ellson <john.ellson@comcast.net>
Tested-by: Zdenek Kabelac <zkabelac@redhat.com>
Tested-by: Bruno Wolff III <bruno@wolff.to>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agomm: compaction: validate pfn range passed to isolate_freepages_block
Mel Gorman [Thu, 6 Dec 2012 19:01:14 +0000 (19:01 +0000)]
mm: compaction: validate pfn range passed to isolate_freepages_block

Commit 0bf380bc70ec ("mm: compaction: check pfn_valid when entering a
new MAX_ORDER_NR_PAGES block during isolation for migration") added a
check for pfn_valid() when isolating pages for migration as the scanner
does not necessarily start pageblock-aligned.

Since commit c89511ab2f8f ("mm: compaction: Restart compaction from near
where it left off"), the free scanner has the same problem.  This patch
makes sure that the pfn range passed to isolate_freepages_block() is
within the same block so that pfn_valid() checks are unnecessary.

In answer to Henrik's wondering why others have not reported this:
reproducing this requires a large enough hole with the right aligment to
have compaction walk into a PFN range with no memmap.  Size and
alignment depends in the memory model - 4M for FLATMEM and 128M for
SPARSEMEM on x86.  It needs a "lucky" machine.

Reported-by: Henrik Rydberg <rydberg@euromail.se>
Signed-off-by: Mel Gorman <mgorman@suse.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agommc: sdhci-pci: Enable SDHCI_CAN_DO_HISPD for Ricoh SDHCI controller
Madhvapathi Sriram [Mon, 15 Oct 2012 04:47:30 +0000 (04:47 +0000)]
mmc: sdhci-pci: Enable SDHCI_CAN_DO_HISPD for Ricoh SDHCI controller

The Ricoh SDHCI controllers support Highspeed clocks as evident from
the ricoh_mmc_probe_slot() settings. Hence, SDHCI_CAN_DO_HISPD needs
to be set to enable SDIO client drivers to set/enable high speed clock
settings

Signed-off-by: Madhvapathi Sriram <Madhvapathi.Sriram@csr.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
12 years agommc: sdhci-dove: allow GPIOs to be used for card detection on Dove
Russell King [Thu, 22 Nov 2012 23:55:51 +0000 (23:55 +0000)]
mmc: sdhci-dove: allow GPIOs to be used for card detection on Dove

This commit taken from Rabeeh's Cubox kernel and re-worked for DT;
Sebastian Hasselbrath is believed to be the original author.

Some Cuboxes require a GPIO for card detection; this implements the
optional GPIO support for card detection.  This GPIO is logic 0 for
card inserted.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Chris Ball <cjb@laptop.org>
12 years agommc: sdhci-dove: use two-stage initialization for sdhci-pltfm
Russell King [Thu, 22 Nov 2012 23:55:30 +0000 (23:55 +0000)]
mmc: sdhci-dove: use two-stage initialization for sdhci-pltfm

We need to use the two-stage initialization for sdhci-pltfm if we're
going to do anything extra at initialization time.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Chris Ball <cjb@laptop.org>
12 years agommc: sdhci-dove: use devm_clk_get()
Russell King [Thu, 22 Nov 2012 23:55:10 +0000 (23:55 +0000)]
mmc: sdhci-dove: use devm_clk_get()

Use devm_clk_get() rather than clk_get() to make cleanup paths more simple.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Chris Ball <cjb@laptop.org>
12 years agommc: eSDHC: Recover from ADMA errors
Haijun Zhang [Tue, 4 Dec 2012 02:41:28 +0000 (10:41 +0800)]
mmc: eSDHC: Recover from ADMA errors

A-003500: False ADMA Error might be reported when ADMA is used for
multiple block read command with Stop at Block Gap. If PROCTL[SABGREQ]
is set when the particular block's data is received by the System side
logic before entire block (with CRC) data is received by the SD side
logic, and also if ADMA descriptor line is fetched at the same time,
then DMA engine might report false ADMA error. eSDHC might not be able
to Continue (PROCTL[CREQ]=1) after Stop at Block Gap.
This issue will impact the eSDHC IP VVN2.3.

Signed-off-by: Haijun Zhang <Haijun.Zhang@freescale.com>
Signed-off-by: Jerry Huang <Chang-Ming.Huang@freescale.com>
Acked-by: Anton Vorontsov <cbouatmailru@gmail.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
12 years agommc: dw_mmc: remove duplicated buswidth code
Jaehoon Chung [Thu, 8 Nov 2012 08:35:31 +0000 (17:35 +0900)]
mmc: dw_mmc: remove duplicated buswidth code

ctype is using 1-bit buswidth mode by default.

Signed-off-by: Jaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Acked-by: James Hogan <james.hogan@imgtec.com>
Signed-off-by: Chris Ball <cjb@laptop.org>