GitHub/mt8127/android_kernel_alcatel_ttab.git
15 years agox86, pat/mtrr: Rendezvous all the cpus for MTRR/PAT init
Suresh Siddha [Thu, 20 Aug 2009 01:05:36 +0000 (18:05 -0700)]
x86, pat/mtrr: Rendezvous all the cpus for MTRR/PAT init

SDM Vol 3a section titled "MTRR considerations in MP systems" specifies
the need for synchronizing the logical cpu's while initializing/updating
MTRR.

Currently Linux kernel does the synchronization of all cpu's only when
a single MTRR register is programmed/updated. During an AP online
(during boot/cpu-online/resume)  where we initialize all the MTRR/PAT registers,
we don't follow this synchronization algorithm.

This can lead to scenarios where during a dynamic cpu online, that logical cpu
is initializing MTRR/PAT with cache disabled (cr0.cd=1) etc while other logical
HT sibling continue to run (also with cache disabled because of cr0.cd=1
on its sibling).

Starting from Westmere, VMX transitions with cr0.cd=1 don't work properly
(because of some VMX performance optimizations) and the above scenario
(with one logical cpu doing VMX activity and another logical cpu coming online)
can result in system crash.

Fix the MTRR initialization by doing rendezvous of all the cpus. During
boot and resume, we delay the MTRR/PAT init for APs till all the
logical cpu's come online and the rendezvous process at the end of AP's bringup,
will initialize the MTRR/PAT for all AP's.

For dynamic single cpu online, we synchronize all the logical cpus and
do the MTRR/PAT init on the AP that is coming online.

Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
15 years agogeneric-ipi: Allow cpus not yet online to call smp_call_function with irqs disabled
Suresh Siddha [Thu, 20 Aug 2009 01:05:35 +0000 (18:05 -0700)]
generic-ipi: Allow cpus not yet online to call smp_call_function with irqs disabled

Because of deadlock possiblities smp_call_function() is not allowed to
be called with interrupts disabled. Add an exception for the cpu not
yet online, as no one else can send smp call function interrupt to this
cpu that is not yet online and as such deadlock condition is not possible.

Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Acked-by: Nick Piggin <npiggin@suse.de>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
15 years agox86: Use printk_once()
Marcin Slusarz [Sun, 9 Aug 2009 19:54:00 +0000 (21:54 +0200)]
x86: Use printk_once()

Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
LKML-Reference: <1249847649-11631-6-git-send-email-marcin.slusarz@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agox86/cpu: Clean up various files a bit
Alan Cox [Fri, 3 Jul 2009 23:35:45 +0000 (00:35 +0100)]
x86/cpu: Clean up various files a bit

No code changes except printk levels (although some of the K6
mtrr code might be clearer if there were a few as would
splitting out some of the intel cache code).

Signed-off-by: Alan Cox <alan@linux.intel.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agox86: Remove duplicated #include
Huang Weiyi [Sat, 11 Jul 2009 01:32:46 +0000 (09:32 +0800)]
x86: Remove duplicated #include

Remove duplicated #include in:

  arch/x86/kernel/dumpstack.c

Signed-off-by: Huang Weiyi <weiyi.huang@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agox86, ipi: Clean up safe_smp_processor_id() by using the cpu_has_apic() macro helper
Cyrill Gorcunov [Sun, 5 Jul 2009 16:01:54 +0000 (20:01 +0400)]
x86, ipi: Clean up safe_smp_processor_id() by using the cpu_has_apic() macro helper

We already use a lot of cpu_has_ helpers.
Lets do here the same for consistency.

Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Cc: Yinghai Lu <yinghai@kernel.org>
LKML-Reference: <20090705160154.GB4791@lenovo>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agox86: Clean up idt_descr and idt_tableby using NR_VECTORS instead of hardcoded number
Cyrill Gorcunov [Wed, 8 Jul 2009 18:03:53 +0000 (22:03 +0400)]
x86: Clean up idt_descr and idt_tableby using NR_VECTORS instead of hardcoded number

Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Cc: Yinghai Lu <yinghai@kernel.org>
LKML-Reference: <20090708180353.GH5301@lenovo>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agox86: Further clean up of mtrr/generic.c
Ingo Molnar [Sun, 5 Jul 2009 07:44:11 +0000 (09:44 +0200)]
x86: Further clean up of mtrr/generic.c

Yinghai noticed that i defined BIOS_BUG_MSG but added no
usage for it. The usage is to clean up this turd in generic.c:

printk(KERN_WARNING "WARNING: BIOS bug: VAR MTRR %d "
"contains strange UC entry under 1M, check "
"with your system vendor!\n", i);

Breaking printk lines in the middle looks ugly, is hard to read
and breaks 'git grep'. Use the BIOS_BUG_MSG instead.

Also complete the moving of structure definitions and variables
to the top of the file.

Reported-by: Yinghai Lu <yinghai@kernel.org>
LKML-Reference: <20090703164225.GA21447@elte.hu>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agox86: Clean up mtrr/main.c
Jaswinder Singh Rajput [Sat, 4 Jul 2009 02:26:28 +0000 (07:56 +0530)]
x86: Clean up mtrr/main.c

Fix following trivial style problems:

  ERROR: trailing whitespace X 25
  WARNING: Use #include <linux/uaccess.h> instead of <asm/uaccess.h>
  WARNING: Use #include <linux/kvm_para.h> instead of <asm/kvm_para.h>
  ERROR: do not initialise externals to 0 or NULL X 2
  ERROR: "foo * bar" should be "foo *bar" X 5
  ERROR: do not use assignment in if condition X 2
  WARNING: line over 80 characters X 8
  ERROR: return is not a function, parentheses are not required
  WARNING: braces {} are not necessary for any arm of this statement
  ERROR: space required before the open parenthesis '(' X 2
  ERROR: open brace '{' following function declarations go on the next line
  ERROR: space required after that ',' (ctx:VxV) X 8
  ERROR: space required before the open parenthesis '(' X 3
  ERROR: else should follow close brace '}'
  WARNING: space prohibited between function name and open parenthesis '('
  WARNING: EXPORT_SYMBOL(foo); should immediately follow its function/variable X 2

Also use pr_debug and pr_warning where possible.

total: 50 errors, 14 warnings

arch/x86/kernel/cpu/mtrr/main.o:

   text    data     bss     dec     hex filename
   3668     116    4156    7940    1f04 main.o.before
   3668     116    4156    7940    1f04 main.o.after

md5:
   e01af2fd28deef77c8d01e71acfbd365  main.o.before.asm
   e01af2fd28deef77c8d01e71acfbd365  main.o.after.asm

Suggested-by: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Jaswinder Singh Rajput <jaswinderrajput@gmail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Yinghai Lu <yinghai@kernel.org>
LKML-Reference: <20090703164225.GA21447@elte.hu>
Cc: Avi Kivity <avi@redhat.com> # Avi, please have a look at the kvm_para.h bit
[ More cleanups ]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agox86: Clean up mtrr/state.c
Jaswinder Singh Rajput [Sat, 4 Jul 2009 02:24:53 +0000 (07:54 +0530)]
x86: Clean up mtrr/state.c

Fix:

  WARNING: Use #include <linux/io.h> instead of <asm/io.h>
  WARNING: line over 80 characters X 4

arch/x86/kernel/cpu/mtrr/state.o:

   text    data     bss     dec     hex filename
    864       0       0     864     360 state.o.before
    864       0       0     864     360 state.o.after

md5:
   c5c4364b9aeac74d70111e1e49667a2c  state.o.before.asm
   c5c4364b9aeac74d70111e1e49667a2c  state.o.after.asm

Suggested-by: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Jaswinder Singh Rajput <jaswinderrajput@gmail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Yinghai Lu <yinghai@kernel.org>
LKML-Reference: <20090703164225.GA21447@elte.hu>
[ More cleanups ]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agox86: Clean up mtrr/mtrr.h
Jaswinder Singh Rajput [Sat, 4 Jul 2009 02:24:16 +0000 (07:54 +0530)]
x86: Clean up mtrr/mtrr.h

Fix:

  ERROR: do not use C99 // comments
  ERROR: "foo * bar" should be "foo *bar" X 2

Suggested-by: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Jaswinder Singh Rajput <jaswinderrajput@gmail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Yinghai Lu <yinghai@kernel.org>
LKML-Reference: <20090703164225.GA21447@elte.hu>
[ More tidyups ]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agox86: Clean up mtrr/if.c
Jaswinder Singh Rajput [Sat, 4 Jul 2009 02:23:40 +0000 (07:53 +0530)]
x86: Clean up mtrr/if.c

Fix:

  WARNING: Use #include <linux/uaccess.h> instead of <asm/uaccess.h>
  ERROR: trailing whitespace X 7
  ERROR: trailing statements should be on next line X 3
  WARNING: line over 80 characters X 5
  ERROR: space required before the open parenthesis '('

arch/x86/kernel/cpu/mtrr/if.o:

   text    data     bss     dec     hex filename
   2239       4       0    2243     8c3 if.o.before
   2239       4       0    2243     8c3 if.o.after

md5:
   78d1f2aa4843ec6509c18e2dee54bc7f  if.o.before.asm
   78d1f2aa4843ec6509c18e2dee54bc7f  if.o.after.asm

Suggested-by: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Jaswinder Singh Rajput <jaswinderrajput@gmail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Yinghai Lu <yinghai@kernel.org>
LKML-Reference: <20090703164225.GA21447@elte.hu>
[ More cleanups to make the code more consistent. ]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agox86: Clean up mtrr/generic.c
Jaswinder Singh Rajput [Sat, 4 Jul 2009 02:23:00 +0000 (07:53 +0530)]
x86: Clean up mtrr/generic.c

Fix following trivial style problems:

  ERROR: trailing whitespace X 4
  WARNING: Use #include <linux/io.h> instead of <asm/io.h>
  WARNING: braces {} are not necessary for single statement blocks X 3
  ERROR: "foo * bar" should be "foo *bar"
  WARNING: line over 80 characters X 6
  ERROR: "foo * bar" should be "foo *bar"
  ERROR: spaces required around that '=' (ctx:VxO)
  ERROR: space required before that '-' (ctx:OxV)
  WARNING: suspect code indent for conditional statements (8, 12)
  ERROR: spaces required around that '=' (ctx:VxV)
  ERROR: do not initialise statics to 0 or NULL
  ERROR: space prohibited after that open parenthesis '(' X 2
  ERROR: space prohibited before that close parenthesis ')' X 2
  ERROR: trailing statements should be on next line
  ERROR: return is not a function, parentheses are not required

Also use pr_debug and pr_warning where possible.

arch/x86/kernel/cpu/mtrr/generic.o:

   text    data     bss     dec     hex filename
   5652      77    4224    9953    26e1 generic.o.before
   5652      77    4220    9949    26dd generic.o.after

The md5 changed:
   b34d6c045f06daa4ed092b90cc760e8f  generic.o.before.asm
   a490c6251cfd8442fbffecc0e09a573d  generic.o.after.asm

Because mtrr_state moved from data to bss, changing its
offsets - and also because __LINE__ numbers changed.

Suggested-by: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Jaswinder Singh Rajput <jaswinderrajput@gmail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Yinghai Lu <yinghai@kernel.org>
LKML-Reference: <20090703164225.GA21447@elte.hu>
[ Further cleanups to make the code more consistent ]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agox86: Clean up mtrr/cyrix.c
Jaswinder Singh Rajput [Sat, 4 Jul 2009 02:22:08 +0000 (07:52 +0530)]
x86: Clean up mtrr/cyrix.c

Fix trivial style problems:

  WARNING: Use #include <linux/io.h> instead of <asm/io.h>
  WARNING: line over 80 characters
  ERROR: do not initialise statics to 0 or NULL
  ERROR: space prohibited after that open parenthesis '(' X 2
  ERROR: space prohibited before that close parenthesis ')' X 2
  ERROR: trailing whitespace X 2
  ERROR: trailing statements should be on next line
  ERROR: do not use C99 // comments X 2

arch/x86/kernel/cpu/mtrr/cyrix.o:

   text    data     bss     dec     hex filename
   1637      32       8    1677     68d cyrix.o.before
   1637      32       8    1677     68d cyrix.o.after

md5:
   6f52abd06905be3f4cabb5239f9b0ff0  cyrix.o.before.asm
   6f52abd06905be3f4cabb5239f9b0ff0  cyrix.o.after.asm

Suggested-by: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Jaswinder Singh Rajput <jaswinderrajput@gmail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Yinghai Lu <yinghai@kernel.org>
LKML-Reference: <20090703164225.GA21447@elte.hu>
[ Made the code more consistent ]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agox86: Clean up mtrr/cleanup.c
Jaswinder Singh Rajput [Sat, 4 Jul 2009 02:21:32 +0000 (07:51 +0530)]
x86: Clean up mtrr/cleanup.c

Fix trivial style problems:

  WARNING: Use #include <linux/uaccess.h> instead of <asm/uaccess.h>
  WARNING: Use #include <linux/kvm_para.h> instead of <asm/kvm_para.h>

Also, nr_mtrr_spare_reg should be unsigned long.

arch/x86/kernel/cpu/mtrr/cleanup.o:

   text    data     bss     dec     hex filename
   6241    8992    2056   17289    4389 cleanup.o.before
   6241    8992    2056   17289    4389 cleanup.o.after

The md5 has changed:
   1a7a27513aef1825236daf29110fe657  cleanup.o.before.asm
   bcea358efa2532b6020e338e158447af  cleanup.o.after.asm

Because a WARN_ON()'s __LINE__ value changed by 3 lines.

Suggested-by: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Jaswinder Singh Rajput <jaswinderrajput@gmail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Yinghai Lu <yinghai@kernel.org>
LKML-Reference: <20090703164225.GA21447@elte.hu>
[ Did lots of other cleanups to make the code look more consistent. ]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agox86: Clean up mtrr/centaur.c
Jaswinder Singh Rajput [Sat, 4 Jul 2009 02:20:44 +0000 (07:50 +0530)]
x86: Clean up mtrr/centaur.c

Remove dead code and fix trivial style problems:

  ERROR: trailing whitespace X 2
  WARNING: line over 80 characters X 3
  ROR: trailing whitespace
  ERROR: do not use C99 // comments X 2

arch/x86/kernel/cpu/mtrr/centaur.o:

   text    data     bss     dec     hex filename
    605      32      68     705     2c1 centaur.o.before
    605      32      68     705     2c1 centaur.o.after

md5:
   a4865ea98ce3c163bb1d376a3949b3e3  centaur.o.before.asm
   a4865ea98ce3c163bb1d376a3949b3e3  centaur.o.after.asm

Suggested-by: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Jaswinder Singh Rajput <jaswinderrajput@gmail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Yinghai Lu <yinghai@kernel.org>
LKML-Reference: <20090703164225.GA21447@elte.hu>
[ Standardized comments, DocBook, curly braces, newlines. ]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agox86: Clean up mtrr/amd.c:
Jaswinder Singh Rajput [Sat, 4 Jul 2009 02:20:00 +0000 (07:50 +0530)]
x86: Clean up mtrr/amd.c:

Fix trivial style problems :

  ERROR: trailing whitespace
  WARNING: line over 80 characters
  ERROR: do not use C99 // comments

arch/x86/kernel/cpu/mtrr/amd.o:

   text    data     bss     dec     hex filename
    501      32       0     533     215 amd.o.before
    501      32       0     533     215 amd.o.after

md5:
   62f795eb840ee2d17b03df89e789e76c  amd.o.before.asm
   62f795eb840ee2d17b03df89e789e76c  amd.o.after.asm

Suggested-by: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Jaswinder Singh Rajput <jaswinderrajput@gmail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Yinghai Lu <yinghai@kernel.org>
LKML-Reference: <20090703164225.GA21447@elte.hu>
[ Also restructured comments to be standard, removed stray return,
  converted function description to DocBook style, etc. ]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agoMerge branch 'linus' into x86/cleanups
Ingo Molnar [Sat, 4 Jul 2009 09:00:38 +0000 (11:00 +0200)]
Merge branch 'linus' into x86/cleanups

Merge reason: We were on an older pre-rc1 base, move to almost-rc2.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agoMerge branch 'for-linus' of git://git.infradead.org/users/eparis/notify
Linus Torvalds [Thu, 2 Jul 2009 23:54:07 +0000 (16:54 -0700)]
Merge branch 'for-linus' of git://git.infradead.org/users/eparis/notify

* 'for-linus' of git://git.infradead.org/users/eparis/notify:
  fs/notify/inotify: decrement user inotify count on close

15 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/mason/btrfs-unstable
Linus Torvalds [Thu, 2 Jul 2009 23:52:38 +0000 (16:52 -0700)]
Merge git://git./linux/kernel/git/mason/btrfs-unstable

* git://git.kernel.org/pub/scm/linux/kernel/git/mason/btrfs-unstable:
  Btrfs: fix error message formatting
  Btrfs: fix use after free in btrfs_start_workers fail path
  Btrfs: honor nodatacow/sum mount options for new files
  Btrfs: update backrefs while dropping snapshot
  Btrfs: account for space we may use in fallocate
  Btrfs: fix the file clone ioctl for preallocated extents
  Btrfs: don't log the inode in file_write while growing the file

15 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-rc-fixes-2.6
Linus Torvalds [Thu, 2 Jul 2009 23:52:25 +0000 (16:52 -0700)]
Merge git://git./linux/kernel/git/jejb/scsi-rc-fixes-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-rc-fixes-2.6:
  [SCSI] cxgb3i: fix connection error when vlan is enabled
  [SCSI] FC transport: Locking fix for common-code FC pass-through patch
  [SCSI] zalon: fix oops on attach failure
  [SCSI] fnic: use DMA_BIT_MASK(nn) instead of deprecated DMA_nnBIT_MASK
  [SCSI] fnic: remove redundant BUG_ONs and fix checks on unsigned
  [SCSI] ibmvscsi: Fix module load hang

15 years agoMerge git://git.infradead.org/iommu-2.6
Linus Torvalds [Thu, 2 Jul 2009 23:51:09 +0000 (16:51 -0700)]
Merge git://git.infradead.org/iommu-2.6

* git://git.infradead.org/iommu-2.6: (38 commits)
  intel-iommu: Don't keep freeing page zero in dma_pte_free_pagetable()
  intel-iommu: Introduce first_pte_in_page() to simplify PTE-setting loops
  intel-iommu: Use cmpxchg64_local() for setting PTEs
  intel-iommu: Warn about unmatched unmap requests
  intel-iommu: Kill superfluous mapping_lock
  intel-iommu: Ensure that PTE writes are 64-bit atomic, even on i386
  intel-iommu: Make iommu=pt work on i386 too
  intel-iommu: Performance improvement for dma_pte_free_pagetable()
  intel-iommu: Don't free too much in dma_pte_free_pagetable()
  intel-iommu: dump mappings but don't die on pte already set
  intel-iommu: Combine domain_pfn_mapping() and domain_sg_mapping()
  intel-iommu: Introduce domain_sg_mapping() to speed up intel_map_sg()
  intel-iommu: Simplify __intel_alloc_iova()
  intel-iommu: Performance improvement for domain_pfn_mapping()
  intel-iommu: Performance improvement for dma_pte_clear_range()
  intel-iommu: Clean up iommu_domain_identity_map()
  intel-iommu: Remove last use of PHYSICAL_PAGE_MASK, for reserving PCI BARs
  intel-iommu: Make iommu_flush_iotlb_psi() take pfn as argument
  intel-iommu: Change aligned_size() to aligned_nrpages()
  intel-iommu: Clean up intel_map_sg(), remove domain_page_mapping()
  ...

15 years agox86: add boundary check for 32bit res before expand e820 resource to alignment
Yinghai Lu [Wed, 1 Jul 2009 19:32:18 +0000 (12:32 -0700)]
x86: add boundary check for 32bit res before expand e820 resource to alignment

fix hang with HIGHMEM_64G and 32bit resource.  According to hpa and
Linus, use (resource_size_t)-1 to fend off big ranges.

Analyzed by hpa

Reported-and-tested-by: Mikael Pettersson <mikpe@it.uu.se>
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agox86: fix power-of-2 round_up/round_down macros
Linus Torvalds [Thu, 2 Jul 2009 19:05:10 +0000 (12:05 -0700)]
x86: fix power-of-2 round_up/round_down macros

These macros had two bugs:
 - the type of the mask was not correctly expanded to the full size of
   the argument being expanded, resulting in possible loss of high bits
   when mixing types.
 - the alignment argument was evaluated twice, despite the macro looking
   like a fancy function (but it really does need to be a macro, since
   it works on arbitrary integer types)

Noticed by Peter Anvin, and with a fix that is a modification of his
suggestion (bug noticed by Yinghai Lu).

Cc: Peter Anvin <hpa@zytor.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoBtrfs: fix error message formatting
Hu Tao [Thu, 2 Jul 2009 17:55:45 +0000 (13:55 -0400)]
Btrfs: fix error message formatting

Make an error msg look nicer by inserting a space between number and word.

Signed-off-by: Hu Tao <hu.taoo@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
15 years agoBtrfs: fix use after free in btrfs_start_workers fail path
Jiri Slaby [Thu, 2 Jul 2009 17:50:58 +0000 (13:50 -0400)]
Btrfs: fix use after free in btrfs_start_workers fail path

worker memory is already freed on one fail path in btrfs_start_workers,
but is still dereferenced. Switch the dereference and kfree.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
15 years agoBtrfs: honor nodatacow/sum mount options for new files
Chris Mason [Thu, 2 Jul 2009 16:26:06 +0000 (12:26 -0400)]
Btrfs: honor nodatacow/sum mount options for new files

The btrfs attr patches unconditionally inherited the inode flags field
without honoring nodatacow and nodatasum.  This fix makes sure
we properly record the nodatacow/sum mount options in new inodes.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
15 years agoBtrfs: update backrefs while dropping snapshot
Yan Zheng [Sun, 28 Jun 2009 01:07:35 +0000 (21:07 -0400)]
Btrfs: update backrefs while dropping snapshot

The new backref format has restriction on type of backref item.  If a tree
block isn't referenced by its owner tree, full backrefs must be used for the
pointers in it. When a tree block loses its owner tree's reference, backrefs
for the pointers in it should be updated to full backrefs. Current
btrfs_drop_snapshot misses the code that updates backrefs, so it's unsafe for
general use.

This patch adds backrefs update code to btrfs_drop_snapshot.  It isn't a
problem in the restricted form btrfs_drop_snapshot is used today, but for
general snapshot deletion this update is required.

Signed-off-by: Yan Zheng <zheng.yan@oracle.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
15 years agoBtrfs: account for space we may use in fallocate
Josef Bacik [Sun, 28 Jun 2009 01:07:34 +0000 (21:07 -0400)]
Btrfs: account for space we may use in fallocate

Using Eric Sandeen's xfstest for fallocate, you can easily trigger a ENOSPC
panic on btrfs.  This is because we do not account for data we may use when
doing the fallocate.  This patch fixes the problem by properly reserving space,
and then just freeing it when we are done.  The reservation stuff was made with
delalloc in mind, so its a little crude for this case, but it keeps the box
from panicing.

Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
15 years agoBtrfs: fix the file clone ioctl for preallocated extents
Chris Mason [Sun, 28 Jun 2009 01:07:03 +0000 (21:07 -0400)]
Btrfs: fix the file clone ioctl for preallocated extents

15 years agoBtrfs: don't log the inode in file_write while growing the file
Chris Mason [Sun, 28 Jun 2009 01:06:22 +0000 (21:06 -0400)]
Btrfs: don't log the inode in file_write while growing the file

15 years agofs/notify/inotify: decrement user inotify count on close
Keith Packard [Thu, 2 Jul 2009 04:56:38 +0000 (21:56 -0700)]
fs/notify/inotify: decrement user inotify count on close

The per-user inotify_devs value is incremented each time a new file is
allocated, but never decremented. This led to inotify_init failing after a
limited number of calls.

Signed-off-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Eric Paris <eparis@redhat.com>
15 years agointel-iommu: Don't keep freeing page zero in dma_pte_free_pagetable()
David Woodhouse [Thu, 2 Jul 2009 11:02:34 +0000 (12:02 +0100)]
intel-iommu: Don't keep freeing page zero in dma_pte_free_pagetable()

Check dma_pte_present() and only free the page if there _is_ one.
Kind of surprising that there was no warning about this.

Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
15 years agointel-iommu: Introduce first_pte_in_page() to simplify PTE-setting loops
David Woodhouse [Thu, 2 Jul 2009 10:21:16 +0000 (11:21 +0100)]
intel-iommu: Introduce first_pte_in_page() to simplify PTE-setting loops

On Wed, 2009-07-01 at 16:59 -0700, Linus Torvalds wrote:
> I also _really_ hate how you do
>
>         (unsigned long)pte >> VTD_PAGE_SHIFT ==
>         (unsigned long)first_pte >> VTD_PAGE_SHIFT

Kill this, in favour of just looking to see if the incremented pte
pointer has 'wrapped' onto the next page. Which means we have to check
it _after_ incrementing it, not before.

Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
15 years agoFRV: Add basic performance counter support
David Howells [Wed, 1 Jul 2009 23:46:21 +0000 (00:46 +0100)]
FRV: Add basic performance counter support

Add basic performance counter support to the FRV arch.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoFRV: Implement atomic64_t
David Howells [Wed, 1 Jul 2009 23:46:16 +0000 (00:46 +0100)]
FRV: Implement atomic64_t

Implement atomic64_t and its ops for FRV.  Tested with the following patch:

diff --git a/arch/frv/kernel/setup.c b/arch/frv/kernel/setup.c
index 55e4fab..086d50d 100644
--- a/arch/frv/kernel/setup.c
+++ b/arch/frv/kernel/setup.c
@@ -746,6 +746,52 @@ static void __init parse_cmdline_early(char *cmdline)

 } /* end parse_cmdline_early() */

+static atomic64_t xxx;
+
+static void test_atomic64(void)
+{
+ atomic64_set(&xxx, 0x12300000023LL);
+
+ mb();
+ BUG_ON(atomic64_read(&xxx) != 0x12300000023LL);
+ mb();
+ if (atomic64_inc_return(&xxx) != 0x12300000024LL)
+ BUG();
+ mb();
+ BUG_ON(atomic64_read(&xxx) != 0x12300000024LL);
+ mb();
+ if (atomic64_sub_return(0x36900000050LL, &xxx) != -0x2460000002cLL)
+ BUG();
+ mb();
+ BUG_ON(atomic64_read(&xxx) != -0x2460000002cLL);
+ mb();
+ if (atomic64_dec_return(&xxx) != -0x2460000002dLL)
+ BUG();
+ mb();
+ BUG_ON(atomic64_read(&xxx) != -0x2460000002dLL);
+ mb();
+ if (atomic64_add_return(0x36800000001LL, &xxx) != 0x121ffffffd4LL)
+ BUG();
+ mb();
+ BUG_ON(atomic64_read(&xxx) != 0x121ffffffd4LL);
+ mb();
+ if (atomic64_cmpxchg(&xxx, 0x123456789abcdefLL, 0x121ffffffd4LL) != 0x121ffffffd4LL)
+ BUG();
+ mb();
+ BUG_ON(atomic64_read(&xxx) != 0x121ffffffd4LL);
+ mb();
+ if (atomic64_cmpxchg(&xxx, 0x121ffffffd4LL, 0x123456789abcdefLL) != 0x121ffffffd4LL)
+ BUG();
+ mb();
+ BUG_ON(atomic64_read(&xxx) != 0x123456789abcdefLL);
+ mb();
+ if (atomic64_xchg(&xxx, 0xabcdef123456789LL) != 0x123456789abcdefLL)
+ BUG();
+ mb();
+ BUG_ON(atomic64_read(&xxx) != 0xabcdef123456789LL);
+ mb();
+}
+
 /*****************************************************************************/
 /*
  *
@@ -845,6 +891,8 @@ void __init setup_arch(char **cmdline_p)
 // asm volatile("movgs %0,timerd" :: "r"(10000000));
 // __set_HSR(0, __get_HSR(0) | HSR0_ETMD);

+ test_atomic64();
+
 } /* end setup_arch() */

 #if 0

Note that this doesn't cover all the trivial wrappers, but does cover all the
substantial implementations.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agointel-iommu: Use cmpxchg64_local() for setting PTEs
David Woodhouse [Wed, 1 Jul 2009 19:27:03 +0000 (20:27 +0100)]
intel-iommu: Use cmpxchg64_local() for setting PTEs

Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
15 years agointel-iommu: Warn about unmatched unmap requests
David Woodhouse [Wed, 1 Jul 2009 18:27:53 +0000 (19:27 +0100)]
intel-iommu: Warn about unmatched unmap requests

This would have found the bug in i386 pci_unmap_addr() a long time ago.
We shouldn't just silently return without doing anything.

Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
15 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/lethal/sh-2.6
Linus Torvalds [Wed, 1 Jul 2009 18:46:30 +0000 (11:46 -0700)]
Merge git://git./linux/kernel/git/lethal/sh-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/lethal/sh-2.6:
  sh: LCDC dcache flush for deferred io
  sh: Fix compiler error and include the definition of IS_ERR_VALUE
  sh: re-add LCDC fbdev support to the Migo-R defconfig
  sh: fix se7724 ceu names
  sh: ms7724se: Enable sh_eth in defconfig.
  arch/sh/boards/mach-se/7206/io.c: Remove unnecessary semicolons
  sh: ms7724se: Add sh_eth support
  nommu: provide follow_pfn().
  sh: Kill off unused DEBUG_BOOTMEM symbol.
  perf_counter tools: add cpu_relax()/rmb() definitions for sh.
  sh64: Hook up page fault events for software perf counters.
  sh: Hook up page fault events for software perf counters.
  sh: make set_perf_counter_pending() static inline.
  clocksource: sh_tmu: Make undefined TCOR behaviour less undefined.

15 years agointel-iommu: Kill superfluous mapping_lock
David Woodhouse [Wed, 1 Jul 2009 18:30:28 +0000 (19:30 +0100)]
intel-iommu: Kill superfluous mapping_lock

Since we're using cmpxchg64() anyway (because that's the only way to do
an atomic 64-bit store on i386), we might as well ditch the extra
locking and just use cmpxchg64() to ensure that we don't add the page
twice.

Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
15 years agosh: LCDC dcache flush for deferred io
Paul Mundt [Wed, 1 Jul 2009 06:50:31 +0000 (06:50 +0000)]
sh: LCDC dcache flush for deferred io

Since writenotify on uncached vmas is unsupported in 2.6.31,
live with cached framebuffer memory in the deferred io
case for now and flush the dcache before forcing refresh.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
Acked-by: Magnus damm <damm@igel.co.jp>
15 years agosh: Fix compiler error and include the definition of IS_ERR_VALUE
Matt Fleming [Wed, 1 Jul 2009 18:14:18 +0000 (18:14 +0000)]
sh: Fix compiler error and include the definition of IS_ERR_VALUE

When arch/sh/include/asm/syscall_32.h is included from a file that
doesn't also include linux/err.h the following error is produced,

In file included from /home/matt/src/kernels/sh-2.6/arch/sh/include/asm/syscall.h:5,
                 from kernel/trace/trace_syscalls.c:3:
/home/matt/src/kernels/sh-2.6/arch/sh/include/asm/syscall_32.h: In function 'syscall_get_error':
/home/matt/src/kernels/sh-2.6/arch/sh/include/asm/syscall_32.h:28: error: implicit declaration of function 'IS_ERR_VALUE'
make[2]: *** [kernel/trace/trace_syscalls.o] Error 1
make[1]: *** [kernel/trace] Error 2
make: *** [kernel] Error 2

Signed-off-by: Matt Fleming <matt@console-pimps.org>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
15 years agokernel-doc: move ignoring kmemcheck
Randy Dunlap [Mon, 29 Jun 2009 21:54:11 +0000 (14:54 -0700)]
kernel-doc: move ignoring kmemcheck

Somehow I managed to generate a diff that put these 2 lines
into the wrong function:  should have been in dump_struct()
instead of in dump_enum().

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoMerge git://git.infradead.org/mtd-2.6
Linus Torvalds [Wed, 1 Jul 2009 18:25:46 +0000 (11:25 -0700)]
Merge git://git.infradead.org/mtd-2.6

* git://git.infradead.org/mtd-2.6:
  mtd: nand: fix build failure and incorrect return from omap_wait()
  mtd: Use BLOCK_NIL consistently in NFTL/INFTL
  mtd: m25p80 timeout too short for worst-case m25p16 devices
  mtd: atmel_nand: Fix typo s/parititions/partitions/
  mtd: cmdlineparts: Use 64-bit format when printing a debug message.
  mtd: maps: Remove BUS_ID_SIZE from integrator_flash
  jffs2: fix another potential leak on error path in scan.c

15 years agointel-iommu: Ensure that PTE writes are 64-bit atomic, even on i386
David Woodhouse [Wed, 1 Jul 2009 18:21:24 +0000 (19:21 +0100)]
intel-iommu: Ensure that PTE writes are 64-bit atomic, even on i386

Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
15 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi...
Linus Torvalds [Wed, 1 Jul 2009 18:20:46 +0000 (11:20 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/mszeredi/fuse

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/fuse:
  fuse: invalidation reverse calls
  fuse: allow umask processing in userspace
  fuse: fix bad return value in fuse_file_poll()
  fuse: fix return value of fuse_dev_write()

15 years agoFix iommu address space allocation
David Woodhouse [Wed, 1 Jul 2009 17:49:06 +0000 (18:49 +0100)]
Fix iommu address space allocation

This fixes kernel.org bug #13584. The IOVA code attempted to optimise
the insertion of new ranges into the rbtree, with the unfortunate result
that some ranges just didn't get inserted into the tree at all. Then
those ranges would be handed out more than once, and things kind of go
downhill from there.

Introduced after 2.6.25 by ddf02886cbe665d67ca750750196ea5bf524b10b
("PCI: iova RB tree setup tweak").

Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Cc: mark gross <mgross@linux.intel.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: stable@kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoFix pci_unmap_addr() et al on i386.
David Woodhouse [Wed, 1 Jul 2009 17:34:52 +0000 (18:34 +0100)]
Fix pci_unmap_addr() et al on i386.

We can run a 32-bit kernel on boxes with an IOMMU, so we need
pci_unmap_addr() etc. to work -- without it, drivers will leak mappings.

To be honest, this whole thing looks like it's more pain than it's
worth; I'm half inclined to remove the no-op #else case altogether.

But this is the minimal fix, which just does the right thing if
CONFIG_DMAR is set.

Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Cc: stable@kernel.org [ for 2.6.30 ]
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoelf: fix one check-after-use
Amerigo Wang [Wed, 1 Jul 2009 05:06:26 +0000 (01:06 -0400)]
elf: fix one check-after-use

Check before use it.

Signed-off-by: WANG Cong <amwang@redhat.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: David Howells <dhowells@redhat.com>
Acked-by: Roland McGrath <roland@redhat.com>
Acked-by: James Morris <jmorris@namei.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agointel-iommu: Make iommu=pt work on i386 too
David Woodhouse [Wed, 1 Jul 2009 17:56:16 +0000 (18:56 +0100)]
intel-iommu: Make iommu=pt work on i386 too

Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
15 years agoMerge branch 'for-linus' of git://git.kernel.dk/linux-2.6-block
Linus Torvalds [Wed, 1 Jul 2009 17:41:09 +0000 (10:41 -0700)]
Merge branch 'for-linus' of git://git.kernel.dk/linux-2.6-block

* 'for-linus' of git://git.kernel.dk/linux-2.6-block:
  cfq-iosched: remove redundant check for NULL cfqq in cfq_set_request()
  blocK: Restore barrier support for md and probably other virtual devices.
  block: get rid of queue-private command filter
  block: Create bip slabs with embedded integrity vectors
  cfq-iosched: get rid of the need for __GFP_NOFAIL in cfq_find_alloc_queue()
  cfq-iosched: move cfqq initialization out of cfq_find_alloc_queue()
  Trivial typo fixes in Documentation/block/data-integrity.txt.

15 years agoMerge branch 'for-linus' of git://neil.brown.name/md
Linus Torvalds [Wed, 1 Jul 2009 17:31:26 +0000 (10:31 -0700)]
Merge branch 'for-linus' of git://neil.brown.name/md

* 'for-linus' of git://neil.brown.name/md:
  md: use interruptible wait when duration is controlled by userspace.
  md/raid5: suspend shouldn't affect read requests.
  md: tidy up error paths in md_alloc
  md: fix error path when duplicate name is found on md device creation.
  md: avoid dereferencing NULL pointer when accessing suspend_* sysfs attributes.
  md: Use new topology calls to indicate alignment and I/O sizes

15 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Wed, 1 Jul 2009 17:29:26 +0000 (10:29 -0700)]
Merge git://git./linux/kernel/git/davem/net-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6: (31 commits)
  Revert "ipv4: arp announce, arp_proxy and windows ip conflict verification"
  igb: return PCI_ERS_RESULT_DISCONNECT on permanent error
  e1000e: io_error_detected callback should return PCI_ERS_RESULT_DISCONNECT
  e1000: return PCI_ERS_RESULT_DISCONNECT on permanent error
  e1000: fix unmap bug
  igb: fix unmap length bug
  ixgbe: fix unmap length bug
  ixgbe: Fix link capabilities during adapter resets
  ixgbe: Fix device capabilities of 82599 single speed fiber NICs.
  ixgbe: Fix SFP log messages
  usbnet: Remove private stats structure
  usbnet: Use netdev stats structure
  smsc95xx: Use netdev stats structure
  rndis_host: Use netdev stats structure
  net1080: Use netdev stats structure
  dm9601: Use netdev stats structure
  cdc_eem: Use netdev stats structure
  ipv4: Fix fib_trie rebalancing, part 3
  bnx2x: Fix the behavior of ethtool when ONBOOT=no
  sctp: xmit sctp packet always return no route error
  ...

15 years agokmemleak: Fix scheduling-while-atomic bug
Ingo Molnar [Wed, 1 Jul 2009 07:43:53 +0000 (09:43 +0200)]
kmemleak: Fix scheduling-while-atomic bug

One of the kmemleak changes caused the following
scheduling-while-holding-the-tasklist-lock regression on x86:

BUG: sleeping function called from invalid context at mm/kmemleak.c:795
in_atomic(): 1, irqs_disabled(): 0, pid: 1737, name: kmemleak
2 locks held by kmemleak/1737:
 #0:  (scan_mutex){......}, at: [<c10c4376>] kmemleak_scan_thread+0x45/0x86
 #1:  (tasklist_lock){......}, at: [<c10c3bb4>] kmemleak_scan+0x1a9/0x39c
Pid: 1737, comm: kmemleak Not tainted 2.6.31-rc1-tip #59266
Call Trace:
 [<c105ac0f>] ? __debug_show_held_locks+0x1e/0x20
 [<c102e490>] __might_sleep+0x10a/0x111
 [<c10c38d5>] scan_yield+0x17/0x3b
 [<c10c3970>] scan_block+0x39/0xd4
 [<c10c3bc6>] kmemleak_scan+0x1bb/0x39c
 [<c10c4331>] ? kmemleak_scan_thread+0x0/0x86
 [<c10c437b>] kmemleak_scan_thread+0x4a/0x86
 [<c104d73e>] kthread+0x6e/0x73
 [<c104d6d0>] ? kthread+0x0/0x73
 [<c100959f>] kernel_thread_helper+0x7/0x10
kmemleak: 834 new suspected memory leaks (see /sys/kernel/debug/kmemleak)

The bit causing it is highly dubious:

static void scan_yield(void)
{
        might_sleep();

        if (time_is_before_eq_jiffies(next_scan_yield)) {
                schedule();
                next_scan_yield = jiffies + jiffies_scan_yield;
        }
}

It called deep inside the codepath and in a conditional way,
and that is what crapped up when one of the new scan_block()
uses grew a tasklist_lock dependency.

This minimal patch removes that yielding stuff and adds the
proper cond_resched().

The background scanning thread could probably also be reniced
to +10.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Acked-by: Pekka Enberg <penberg@cs.helsinki.fi>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agocfq-iosched: remove redundant check for NULL cfqq in cfq_set_request()
Shan Wei [Wed, 1 Jul 2009 10:41:14 +0000 (12:41 +0200)]
cfq-iosched: remove redundant check for NULL cfqq in cfq_set_request()

With the changes for falling back to an oom_cfqq, we never fail
to find/allocate a queue in cfq_get_queue(). So remove the check.

Signed-off-by: Shan Wei <shanwei@cn.fujitsu.com>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
15 years agoblocK: Restore barrier support for md and probably other virtual devices.
NeilBrown [Tue, 30 Jun 2009 07:35:44 +0000 (09:35 +0200)]
blocK: Restore barrier support for md and probably other virtual devices.

The next_ordered flag is only meaningful for devices that use __make_request.
So move the test against next_ordered out of generic code and in to
__make_request

Since this test was added, barriers have not worked on md or any
devices that don't use __make_request and so don't bother to set
next_ordered.  (dm explicitly sets something other than
QUEUE_ORDERED_NONE since
  commit 99360b4c18f7675b50d283301d46d755affe75fd
but notes in the comments that it is otherwise meaningless).

Cc: Ken Milmore <ken.milmore@googlemail.com>
Cc: stable@kernel.org
Signed-off-by: NeilBrown <neilb@suse.de>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
15 years agoblock: get rid of queue-private command filter
Jens Axboe [Fri, 26 Jun 2009 14:27:10 +0000 (16:27 +0200)]
block: get rid of queue-private command filter

The initial patches to support this through sysfs export were broken
and have been if 0'ed out in any release. So lets just kill the code
and reclaim some space in struct request_queue, if anyone would later
like to fixup the sysfs bits, the git history can easily restore
the removed bits.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
15 years agoblock: Create bip slabs with embedded integrity vectors
Martin K. Petersen [Fri, 26 Jun 2009 13:37:49 +0000 (15:37 +0200)]
block: Create bip slabs with embedded integrity vectors

This patch restores stacking ability to the block layer integrity
infrastructure by creating a set of dedicated bip slabs.  Each bip slab
has an embedded bio_vec array at the end.  This cuts down on memory
allocations and also simplifies the code compared to the original bvec
version.  Only the largest bip slab is backed by a mempool.  The pool is
contained in the bio_set so stacking drivers can ensure forward
progress.

Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Jens Axboe <axboe@carl.(none)>
15 years agocfq-iosched: get rid of the need for __GFP_NOFAIL in cfq_find_alloc_queue()
Jens Axboe [Tue, 30 Jun 2009 07:34:12 +0000 (09:34 +0200)]
cfq-iosched: get rid of the need for __GFP_NOFAIL in cfq_find_alloc_queue()

Setup an emergency fallback cfqq that we allocate at IO scheduler init
time. If the slab allocation fails in cfq_find_alloc_queue(), we'll just
punt IO to that cfqq instead. This ensures that cfq_find_alloc_queue()
never fails without having to ensure free memory.

On cfqq lookup, always try to allocate a new cfqq if the given cfq io
context has the oom_cfqq assigned. This ensures that we only temporarily
punt to this shared queue.

Reviewed-by: Jeff Moyer <jmoyer@redhat.com>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
15 years agocfq-iosched: move cfqq initialization out of cfq_find_alloc_queue()
Jens Axboe [Fri, 26 Jun 2009 08:44:34 +0000 (10:44 +0200)]
cfq-iosched: move cfqq initialization out of cfq_find_alloc_queue()

We're going to be needing that init code outside of that function
to get rid of the __GFP_NOFAIL in cfqq allocation.

Reviewed-by: Jeff Moyer <jmoyer@redhat.com>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
15 years agoTrivial typo fixes in Documentation/block/data-integrity.txt.
Andre Noll [Thu, 25 Jun 2009 11:03:04 +0000 (13:03 +0200)]
Trivial typo fixes in Documentation/block/data-integrity.txt.

Signed-off-by: Andre Noll <maan@systemlinux.org>
Acked-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
15 years agosh: re-add LCDC fbdev support to the Migo-R defconfig
Magnus Damm [Wed, 1 Jul 2009 05:16:31 +0000 (05:16 +0000)]
sh: re-add LCDC fbdev support to the Migo-R defconfig

Re-add LCDC fbdev support to the Migo-R defconfig.

Signed-off-by: Magnus Damm <damm@igel.co.jp>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
15 years agosh: fix se7724 ceu names
Magnus Damm [Wed, 1 Jul 2009 04:55:35 +0000 (04:55 +0000)]
sh: fix se7724 ceu names

Use "ceu0" and "ceu1" as CEU names instead of "ceu".
This fixes "memchunk" kernel command line selection
on the solution engine 7724 board.

With this patch applied use "memchunk.ceu0=1m" or
"memchunk.ceu1=1m" on kernel command line to override
physically memory size to one meg for CEU0 or CEU1.

Signed-off-by: Magnus Damm <damm@igel.co.jp>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
15 years agomd: use interruptible wait when duration is controlled by userspace.
NeilBrown [Wed, 1 Jul 2009 03:15:35 +0000 (13:15 +1000)]
md: use interruptible wait when duration is controlled by userspace.

User space can set various limits on an md array so that resync waits
when it gets to a certain point, or so that I/O is blocked for a short
while.
When md is waiting against one of these limit, it should use an
interruptible wait so as not to add to the load average, and so are
not to trigger a warning if the wait goes on for too long.

Signed-off-by: NeilBrown <neilb@suse.de>
15 years agomd/raid5: suspend shouldn't affect read requests.
NeilBrown [Wed, 1 Jul 2009 03:15:35 +0000 (13:15 +1000)]
md/raid5: suspend shouldn't affect read requests.

md allows write to regions on an array to be suspended temporarily.
This allows user-space to participate is aspects of reshape.
In particular, data can be copied with not risk of a race.
We should not be blocking read requests though, so don't.

Cc: stable@kernel.org
Signed-off-by: NeilBrown <neilb@suse.de>
15 years agoRevert "ipv4: arp announce, arp_proxy and windows ip conflict verification"
Eric W. Biederman [Tue, 30 Jun 2009 16:27:17 +0000 (16:27 +0000)]
Revert "ipv4: arp announce, arp_proxy and windows ip conflict verification"

This reverts commit 73ce7b01b4496a5fbf9caf63033c874be692333f.

After discovering that we don't listen to gratuitious arps in 2.6.30
I tracked the failure down to this commit.

The patch makes absolutely no sense.  RFC2131 RFC3927 and RFC5227.
are all in agreement that an arp request with sip == 0 should be used
for the probe (to prevent learning) and an arp request with sip == tip
should be used for the gratitous announcement that people can learn
from.

It appears the author of the broken patch got those two cases confused
and modified the code to drop all gratuitous arp traffic.  Ouch!

Cc: stable@kernel.org
Signed-off-by: Eric W. Biederman <ebiederm@aristanetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoigb: return PCI_ERS_RESULT_DISCONNECT on permanent error
Alexander Duyck [Tue, 30 Jun 2009 12:46:34 +0000 (12:46 +0000)]
igb: return PCI_ERS_RESULT_DISCONNECT on permanent error

PCI drivers that implement the io_error_detected callback should return
PCI_ERS_RESULT_DISCONNECT if the state passed in is
pci_channel_io_perm_failure.  This patch fixes the issue for igb.

Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoe1000e: io_error_detected callback should return PCI_ERS_RESULT_DISCONNECT
Mike Mason [Tue, 30 Jun 2009 12:45:53 +0000 (12:45 +0000)]
e1000e: io_error_detected callback should return PCI_ERS_RESULT_DISCONNECT

on permanent failure

PCI drivers that implement the io_error_detected callback
should return PCI_ERS_RESULT_DISCONNECT if the state
passed in is pci_channel_io_perm_failure.  This state is not
checked in many of the network drivers.

This patch fixes the omission in the e1000e driver.

Signed-off-by: Mike Mason <mmlnx@us.ibm.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoe1000: return PCI_ERS_RESULT_DISCONNECT on permanent error
Andre Detsch [Tue, 30 Jun 2009 12:46:13 +0000 (12:46 +0000)]
e1000: return PCI_ERS_RESULT_DISCONNECT on permanent error

PCI drivers that implement the io_error_detected callback
should return PCI_ERS_RESULT_DISCONNECT if the state
passed in is pci_channel_io_perm_failure.  This state is
not checked in many of the network drivers.

The patch fixes the omission in the e1000 driver.

Based on Mike Mason's similar patch for e1000e.

Signed-off-by: Andre Detsch <adetsch@br.ibm.com>
CC: Mike Mason <mmlnx@us.ibm.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoe1000: fix unmap bug
Jesse Brandeburg [Tue, 30 Jun 2009 12:45:34 +0000 (12:45 +0000)]
e1000: fix unmap bug

as reported by kerneloops.org

[  121.781161] ------------[ cut here ]------------
[  121.781171] WARNING: at lib/dma-debug.c:793 check_unmap+0x14e/0x577()
[  121.781173] Hardware name: S5520HC
[  121.781177] e1000 0000:0a:00.0: DMA-API: device driver tries to free DMA
memory it has not allocated [device address=0x00000001d688b0fa] [size=1522
bytes]
[  121.781180] Modules linked in: e1000 mdio  dca [last unloaded: ixgbe]
[  121.781187] Pid: 4793, comm: bash Tainted: P 2.6.30-master-06161113 #3
[  121.781190] Call Trace:
[  121.781195]  [<ffffffff8123056f>] ? check_unmap+0x14e/0x577
[  121.781201]  [<ffffffff81057a19>] warn_slowpath_common+0x77/0x8f
[  121.781205]  [<ffffffff81057ae1>] warn_slowpath_fmt+0x9f/0xa1
[  121.781212]  [<ffffffff81477ce2>] ? _spin_lock_irqsave+0x3f/0x49
[  121.781216]  [<ffffffff8122fa97>] ? get_hash_bucket+0x28/0x33
[  121.781220]  [<ffffffff8123056f>] check_unmap+0x14e/0x577
[  121.781225]  [<ffffffff810e4f48>] ? check_bytes_and_report+0x38/0xcb
[  121.781230]  [<ffffffff81230bbf>] debug_dma_unmap_page+0x80/0x92
[  121.781234]  [<ffffffff8122e549>] ? unmap_single+0x1a/0x4e
[  121.781239]  [<ffffffff813901e1>] ? __kfree_skb+0x74/0x78
[  121.781250]  [<ffffffffa00662ef>] pci_unmap_single+0x64/0x6d [e1000]
[  121.781259]  [<ffffffffa0066344>] e1000_clean_rx_ring+0x4c/0xbf [e1000]
[  121.781268]  [<ffffffffa00663df>] e1000_clean_all_rx_rings+0x28/0x36 [e1000]
[  121.781277]  [<ffffffffa0067464>] e1000_down+0x138/0x141 [e1000]
[  121.781286]  [<ffffffffa00681c2>] __e1000_shutdown+0x6b/0x198 [e1000]
[  121.781296]  [<ffffffffa0068405>] e1000_suspend+0x17/0x50 [e1000]
[  121.781301]  [<ffffffff81237665>] pci_legacy_suspend+0x3b/0xbe
[  121.781305]  [<ffffffff81237bc6>] pci_pm_suspend+0x3e/0xf1
[  121.781310]  [<ffffffff812eaf1c>] pm_op+0x57/0xde
[  121.781314]  [<ffffffff812eb444>] dpm_suspend_start+0x31e/0x470
[  121.781319]  [<ffffffff810877da>] suspend_devices_and_enter+0x3e/0x1a2
[  121.781323]  [<ffffffff81087a0f>] enter_state+0xd1/0x127
[  121.781327]  [<ffffffff8108717a>] state_store+0xa7/0xc9
[  121.781332]  [<ffffffff81221843>] kobj_attr_store+0x17/0x19
[  121.781336]  [<ffffffff8113c01e>] sysfs_write_file+0xe5/0x121
[  121.781341]  [<ffffffff810ed165>] vfs_write+0xab/0x105
[  121.781344]  [<ffffffff810ed279>] sys_write+0x47/0x6d
[  121.781349]  [<ffffffff81027aab>] system_call_fastpath+0x16/0x1b
[  121.781352] ---[ end trace 97bacaaac2ed7786 ]---

Fix is to correctly zero out internal ->dma value when unmapping
and make sure never to unmap unless there specifically was a mapping done.

Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoigb: fix unmap length bug
Jesse Brandeburg [Tue, 30 Jun 2009 12:45:15 +0000 (12:45 +0000)]
igb: fix unmap length bug

driver was mixing NET_IP_ALIGN count bytes in map/unmap calls
unevenly. Only map the bytes that the hardware might dma into

also fix unmap related bug where ->dma was not being cleared
after unmap

Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoixgbe: fix unmap length bug
Jesse Brandeburg [Tue, 30 Jun 2009 11:44:56 +0000 (11:44 +0000)]
ixgbe: fix unmap length bug

This patch addresses three WARN_ON statements from DMA-API debug code

ixgbe is mapping more than it unmaps, reduce the length of the map call and
remove the "used once" local variable.

found by Joerg Roedel <joerg.roedel@amd.com> in 2.6.30, so is a candidate
for -stable.

in addition, fix missing ->dma = 0 after unmap to prevent double free with
pci_unmap_single

and lastly, don't unmap (half) pages that aren't mapped.

Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
CC: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoixgbe: Fix link capabilities during adapter resets
Mallikarjuna R Chilakala [Tue, 30 Jun 2009 11:44:36 +0000 (11:44 +0000)]
ixgbe: Fix link capabilities during adapter resets

Adapter link advertisement capabilities were not persistent during
adapter resets. While configuring multispeed fiber link check for
phy autoneg_advertised settings before overwriting with default
link capabilities

Signed-off-by: Mallikarjuna R Chilakala <mallikarjuna.chilakala@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoixgbe: Fix device capabilities of 82599 single speed fiber NICs.
Mallikarjuna R Chilakala [Tue, 30 Jun 2009 11:44:16 +0000 (11:44 +0000)]
ixgbe: Fix device capabilities of 82599 single speed fiber NICs.

82599 single speed fiber modules only support 10G/Full. Return
proper device capabilities while querrying the adapter and error
while changing device advertisement/speed/duplex capabilities.

Signed-off-by: Mallikarjuna R Chilakala <mallikarjuna.chilakala@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoixgbe: Fix SFP log messages
Don Skidmore [Tue, 30 Jun 2009 11:43:55 +0000 (11:43 +0000)]
ixgbe: Fix SFP log messages

We had a wide range of log messages for the same sort of SFP
failure.  This patch makes them all more similar and less
confusing along with converting them to dev_err.

Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agousbnet: Remove private stats structure
Herbert Xu [Mon, 29 Jun 2009 16:54:12 +0000 (16:54 +0000)]
usbnet: Remove private stats structure

Now that nothing uses the private stats structure we can remove it.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agousbnet: Use netdev stats structure
Herbert Xu [Mon, 29 Jun 2009 16:53:28 +0000 (16:53 +0000)]
usbnet: Use netdev stats structure

Now that netdev has its own stats structure we should use that
instead.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agosmsc95xx: Use netdev stats structure
Herbert Xu [Mon, 29 Jun 2009 16:53:00 +0000 (16:53 +0000)]
smsc95xx: Use netdev stats structure

Now that netdev has its own stats structure we should use that
instead.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agorndis_host: Use netdev stats structure
Herbert Xu [Mon, 29 Jun 2009 16:52:26 +0000 (16:52 +0000)]
rndis_host: Use netdev stats structure

Now that netdev has its own stats structure we should use that
instead.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agonet1080: Use netdev stats structure
Herbert Xu [Mon, 29 Jun 2009 16:51:40 +0000 (16:51 +0000)]
net1080: Use netdev stats structure

Now that netdev has its own stats structure we should use that
instead.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agodm9601: Use netdev stats structure
Herbert Xu [Mon, 29 Jun 2009 16:50:51 +0000 (16:50 +0000)]
dm9601: Use netdev stats structure

Now that netdev has its own stats structure we should use that
instead.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agocdc_eem: Use netdev stats structure
Herbert Xu [Mon, 29 Jun 2009 16:49:40 +0000 (16:49 +0000)]
cdc_eem: Use netdev stats structure

Now that netdev has its own stats structure we should use that
instead.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agomd: tidy up error paths in md_alloc
NeilBrown [Wed, 1 Jul 2009 02:27:21 +0000 (12:27 +1000)]
md: tidy up error paths in md_alloc

As the recent bug in md_alloc showed, having a single exit path for
unlocking and putting is a good idea.  So restructure md_alloc to have
a single mutex_unlock and mddev_put, and use gotos where necessary.

Found-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: NeilBrown <neilb@suse.de>
15 years agomd: fix error path when duplicate name is found on md device creation.
NeilBrown [Wed, 1 Jul 2009 02:27:21 +0000 (12:27 +1000)]
md: fix error path when duplicate name is found on md device creation.

When an md device is created by name (rather than number) we need to
check that the name is not already in use.  If this check finds a
duplicate, we return an error without dropping the lock or freeing
the newly create mddev.
This patch fixes that.

Cc: stable@kernel.org
Found-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: NeilBrown <neilb@suse.de>
15 years agoMerge branch 'kmemleak' of git://linux-arm.org/linux-2.6
Linus Torvalds [Wed, 1 Jul 2009 02:04:53 +0000 (19:04 -0700)]
Merge branch 'kmemleak' of git://linux-arm.org/linux-2.6

* 'kmemleak' of git://linux-arm.org/linux-2.6:
  kmemleak: Inform kmemleak about pid_hash
  kmemleak: Do not warn if an unknown object is freed
  kmemleak: Do not report new leaked objects if the scanning was stopped
  kmemleak: Slightly change the policy on newly allocated objects
  kmemleak: Do not trigger a scan when reading the debug/kmemleak file
  kmemleak: Simplify the reports logged by the scanning thread
  kmemleak: Enable task stacks scanning by default
  kmemleak: Allow the early log buffer to be configurable.

15 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/agk/linux-2.6-dm
Linus Torvalds [Wed, 1 Jul 2009 02:04:14 +0000 (19:04 -0700)]
Merge git://git./linux/kernel/git/agk/linux-2.6-dm

* git://git.kernel.org/pub/scm/linux/kernel/git/agk/linux-2.6-dm:
  dm table: fix blk_stack_limits arg to use bytes not sectors
  dm exception store: really fix type lookup

15 years agoMerge branch 'perfcounters-fixes-for-linus' of git://git.kernel.org/pub/scm/linux...
Linus Torvalds [Wed, 1 Jul 2009 02:02:59 +0000 (19:02 -0700)]
Merge branch 'perfcounters-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'perfcounters-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (47 commits)
  perf report: Add --symbols parameter
  perf report: Add --comms parameter
  perf report: Add --dsos parameter
  perf_counter tools: Adjust only prelinked symbol's addresses
  perf_counter: Provide a way to enable counters on exec
  perf_counter tools: Reduce perf stat measurement overhead/skew
  perf stat: Use percentages for scaling output
  perf_counter, x86: Update x86_pmu after WARN()
  perf stat: Micro-optimize the code: memcpy is only required if no event is selected and !null_run
  perf stat: Improve output
  perf stat: Fix multi-run stats
  perf stat: Add -n/--null option to run without counters
  perf_counter tools: Remove dead code
  perf_counter: Complete counter swap
  perf report: Print sorted callchains per histogram entries
  perf_counter tools: Prepare a small callchain framework
  perf record: Fix unhandled io return value
  perf_counter tools: Add alias for 'l1d' and 'l1i'
  perf-report: Add bare minimum PERF_EVENT_READ parsing
  perf-report: Add modes for inherited stats and no-samples
  ...

15 years agoMerge branch 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6
Linus Torvalds [Wed, 1 Jul 2009 02:01:52 +0000 (19:01 -0700)]
Merge branch 'release' of git://git./linux/kernel/git/aegl/linux-2.6

* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6:
  Add Fenghua Yu as temporary co-maintainer for ia64
  [IA64] address compiler warnings perfmon.c/salinfo.c
  [IA64] Remove unnecessary semicolons
  [IA64] sprintf should not be used with same source & destination address

15 years agoMN10300: Wire up new syscalls
David Howells [Tue, 30 Jun 2009 21:33:15 +0000 (22:33 +0100)]
MN10300: Wire up new syscalls

Wire up new syscalls rt_tgsigqueueinfo and perf_counter_open.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoFRV: Wire up new syscalls
David Howells [Tue, 30 Jun 2009 21:24:54 +0000 (22:24 +0100)]
FRV: Wire up new syscalls

Wire up new syscalls rt_tgsigqueueinfo and perf_counter_open.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agohostfs: set maximum filesize in superblock for proper LFS support
Wolfgang Illmeyer [Tue, 30 Jun 2009 18:41:44 +0000 (11:41 -0700)]
hostfs: set maximum filesize in superblock for proper LFS support

Maximum file size for hostfs mounts defaults to 2GB, so bigger files cannot be
read/written through hostfs. This patch initializes the maximum file size to
MAX_LFS_SIZE.

Addresses http://bugzilla.kernel.org/show_bug.cgi?id=13531

Signed-off-by: Wolfgang Illmeyer <wolfgang@illmeyer.com>
Cc: Jeff Dike <jdike@addtoit.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agofloppy: fix lock imbalance
Jiri Slaby [Tue, 30 Jun 2009 18:41:44 +0000 (11:41 -0700)]
floppy: fix lock imbalance

A crappy macro prevents us unlocking on a fail path.

Expand the macro and unlock appropriatelly.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agobfin: delay IRQ registration until driver is ready
Mike Frysinger [Tue, 30 Jun 2009 18:41:43 +0000 (11:41 -0700)]
bfin: delay IRQ registration until driver is ready

Make sure we do not actually request the RTC IRQ until the device driver
is fully ready to handle and process any interrupt.  This way a spurious
interrupt won't crash the system (which may happen if the bootloader was
poking the RTC right before booting Linux).

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoatyfb: fix alignment for block writes
Ville Syrjala [Tue, 30 Jun 2009 18:41:42 +0000 (11:41 -0700)]
atyfb: fix alignment for block writes

Block writes require 64 byte alignment.  Since block writes could be used
with SGRAM or WRAM also refine the memory type detection to check for
either type before deciding to use the 64 byte alignment.

Signed-off-by: Ville Syrjala <syrjala@sci.fi>
Tested-by: Mikulas Patocka <mpatocka@redhat.com>
Cc: Krzysztof Helt <krzysztof.h1@poczta.fm>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoatyfb: fix HP OmniBook 500 reboot hang
Ville Syrjala [Tue, 30 Jun 2009 18:41:40 +0000 (11:41 -0700)]
atyfb: fix HP OmniBook 500 reboot hang

Apparently HP OmniBook 500's BIOS doesn't like the way atyfb reprograms
the hardware. The BIOS will simply hang after a reboot. Fix the problem
by restoring the hardware to it's original state on reboot.

Signed-off-by: Ville Syrjala <syrjala@sci.fi>
Cc: Mikulas Patocka <mpatocka@redhat.com>
Cc: Krzysztof Helt <krzysztof.h1@poczta.fm>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agogpio: pl061: fix IRQ handling for GPIOs >= PL061_GPIO_NR
Baruch Siach [Tue, 30 Jun 2009 18:41:39 +0000 (11:41 -0700)]
gpio: pl061: fix IRQ handling for GPIOs >= PL061_GPIO_NR

IRQ handling is wrong for any GPIO >= PL061_GPIO_NR.

Fix this by implementing and using a proper .to_irq method.

Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Cc: David Brownell <david-b@pacbell.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agogpio: pl061: fix probe error handling code
Baruch Siach [Tue, 30 Jun 2009 18:41:38 +0000 (11:41 -0700)]
gpio: pl061: fix probe error handling code

Note that IRQ has not been initialized when kmalloc() fails.

Also, use DECLARE_BITMAP() to make the code clearer.

Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Cc: David Brownell <david-b@pacbell.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agox86: only clear node_states for 64bit
Yinghai Lu [Tue, 30 Jun 2009 18:41:37 +0000 (11:41 -0700)]
x86: only clear node_states for 64bit

Nathan reported that

| commit 73d60b7f747176dbdff826c4127d22e1fd3f9f74
| Author: Yinghai Lu <yinghai@kernel.org>
| Date:   Tue Jun 16 15:33:00 2009 -0700
|
|    page-allocator: clear N_HIGH_MEMORY map before we set it again
|
|    SRAT tables may contains nodes of very small size.  The arch code may
|    decide to not activate such a node.  However, currently the early boot
|    code sets N_HIGH_MEMORY for such nodes.  These nodes therefore seem to be
|    active although these nodes have no present pages.
|
|    For 64bit N_HIGH_MEMORY == N_NORMAL_MEMORY, so that works for 64 bit too

unintentionally and incorrectly clears the cpuset.mems cgroup attribute on
an i386 kvm guest, meaning that cpuset.mems can not be used.

Fix this by only clearing node_states[N_NORMAL_MEMORY] for 64bit only.
and need to do save/restore for that in find_zone_movable_pfn

Reported-by: Nathan Lynch <ntl@pobox.com>
Tested-by: Nathan Lynch <ntl@pobox.com>
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Cc: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@elte.hu>,
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agocpusets: document adding/removing cpus to cpuset elaborately
Nikanth Karthikesan [Tue, 30 Jun 2009 18:41:36 +0000 (11:41 -0700)]
cpusets: document adding/removing cpus to cpuset elaborately

By writing a tasks's pid to the file, a process adds that task to that
cgroup/cpuset.  But to add a cpu/mem to a cpuset, the new list of cpus
should be written to the cpuset.mems file which would replace the old list
of cpus.  Make this clearer in the documentation.

Signed-off-by: Nikanth Karthikesan <knikanth@suse.de>
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Acked-by: Paul Menage <menage@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agomm: prevent balance_dirty_pages() from doing too much work
Richard Kennedy [Tue, 30 Jun 2009 18:41:35 +0000 (11:41 -0700)]
mm: prevent balance_dirty_pages() from doing too much work

balance_dirty_pages can overreact and move all of the dirty pages to
writeback unnecessarily.

balance_dirty_pages makes its decision to throttle based on the number of
dirty plus writeback pages that are over the calculated limit,so it will
continue to move pages even when there are plenty of pages in writeback
and less than the threshold still dirty.

This allows it to overshoot its limits and move all the dirty pages to
writeback while waiting for the drives to catch up and empty the writeback
list.

A simple fio test easily demonstrates this problem.

fio --name=f1 --directory=/disk1 --size=2G -rw=write --name=f2 --directory=/disk2 --size=1G --rw=write --startdelay=10

This is the simplest fix I could find, but I'm not entirely sure that it
alone will be enough for all cases.  But it certainly is an improvement on
my desktop machine writing to 2 disks.

Do we need something more for machines with large arrays where
bdi_threshold * number_of_drives is greater than the dirty_ratio ?

Signed-off-by: Richard Kennedy <richard@rsk.demon.co.uk>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>