platform/kernel/linux-stable.git
13 years agoACPI, APEI, Add APEI generic error status printing support
Huang Ying [Tue, 7 Dec 2010 02:22:30 +0000 (10:22 +0800)]
ACPI, APEI, Add APEI generic error status printing support

In APEI, Hardware error information reported by firmware to Linux
kernel is in the data structure of APEI generic error status (struct
acpi_hes_generic_status).  While now printk is used by Linux kernel to
report hardware error information to user space.

So, this patch adds printing support for the data structure, so that
the corresponding hardware error information can be reported to user
space via printk.

PCIe AER information printing is not implemented yet.  Will refactor the
original PCIe AER information printing code to avoid code duplicating.

The output format is as follow:

<error record> :=
APEI generic hardware error status
severity: <integer>, <severity string>
section: <integer>, severity: <integer>, <severity string>
flags: <integer>
<section flags strings>
fru_id: <uuid string>
fru_text: <string>
section_type: <section type string>
<section data>

<severity string>* := recoverable | fatal | corrected | info

<section flags strings># :=
[primary][, containment warning][, reset][, threshold exceeded]\
[, resource not accessible][, latent error]

<section type string> := generic processor error | memory error | \
PCIe error | unknown, <uuid string>

<section data> :=
<generic processor section data> | <memory section data> | \
<pcie section data> | <null>

<generic processor section data> :=
[processor_type: <integer>, <proc type string>]
[processor_isa: <integer>, <proc isa string>]
[error_type: <integer>
<proc error type strings>]
[operation: <integer>, <proc operation string>]
[flags: <integer>
<proc flags strings>]
[level: <integer>]
[version_info: <integer>]
[processor_id: <integer>]
[target_address: <integer>]
[requestor_id: <integer>]
[responder_id: <integer>]
[IP: <integer>]

<proc type string>* := IA32/X64 | IA64

<proc isa string>* := IA32 | IA64 | X64

<processor error type strings># :=
[cache error][, TLB error][, bus error][, micro-architectural error]

<proc operation string>* := unknown or generic | data read | data write | \
instruction execution

<proc flags strings># :=
[restartable][, precise IP][, overflow][, corrected]

<memory section data> :=
[error_status: <integer>]
[physical_address: <integer>]
[physical_address_mask: <integer>]
[node: <integer>]
[card: <integer>]
[module: <integer>]
[bank: <integer>]
[device: <integer>]
[row: <integer>]
[column: <integer>]
[bit_position: <integer>]
[requestor_id: <integer>]
[responder_id: <integer>]
[target_id: <integer>]
[error_type: <integer>, <mem error type string>]

<mem error type string>* :=
unknown | no error | single-bit ECC | multi-bit ECC | \
single-symbol chipkill ECC | multi-symbol chipkill ECC | master abort | \
target abort | parity error | watchdog timeout | invalid address | \
mirror Broken | memory sparing | scrub corrected error | \
scrub uncorrected error

<pcie section data> :=
[port_type: <integer>, <pcie port type string>]
[version: <integer>.<integer>]
[command: <integer>, status: <integer>]
[device_id: <integer>:<integer>:<integer>.<integer>
slot: <integer>
secondary_bus: <integer>
vendor_id: <integer>, device_id: <integer>
class_code: <integer>]
[serial number: <integer>, <integer>]
[bridge: secondary_status: <integer>, control: <integer>]

<pcie port type string>* := PCIe end point | legacy PCI end point | \
unknown | unknown | root port | upstream switch port | \
downstream switch port | PCIe to PCI/PCI-X bridge | \
PCI/PCI-X to PCIe bridge | root complex integrated endpoint device | \
root complex event collector

Where, [] designate corresponding content is optional

All <field string> description with * has the following format:

field: <integer>, <field string>

Where value of <integer> should be the position of "string" in <field
string> description. Otherwise, <field string> will be "unknown".

All <field strings> description with # has the following format:

field: <integer>
<field strings>

Where each string in <fields strings> corresponding to one set bit of
<integer>. The bit position is the position of "string" in <field
strings> description.

For more detailed explanation of every field, please refer to UEFI
specification version 2.3 or later, section Appendix N: Common
Platform Error Record.

Signed-off-by: Huang Ying <ying.huang@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
13 years agoAdd CPER PCIe error section structure and constants definition
Huang Ying [Tue, 7 Dec 2010 02:22:29 +0000 (10:22 +0800)]
Add CPER PCIe error section structure and constants definition

On some machine, PCIe error is reported via APEI (ACPI Platform Error
Interface).  The error data is passed from firmware to Linux via CPER
PCIe error section structure.

This patch adds CPER PCIe error section structure and constants
definition.

Signed-off-by: Huang Ying <ying.huang@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
13 years agoLinux 2.6.37-rc5 v2.6.37-rc5
Linus Torvalds [Tue, 7 Dec 2010 04:09:04 +0000 (20:09 -0800)]
Linux 2.6.37-rc5

13 years agoMerge branch 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jlbec...
Linus Torvalds [Tue, 7 Dec 2010 04:08:25 +0000 (20:08 -0800)]
Merge branch 'upstream-linus' of git://git./linux/kernel/git/jlbec/ocfs2

* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jlbec/ocfs2:
  ocfs2_connection_find() returns pointer to bad structure
  ocfs2: char is not always signed
  Ocfs2: Stop tracking a negative dentry after dentry_iput().
  ocfs2: fix memory leak
  fs/ocfs2/dlm: Use GFP_ATOMIC under spin_lock

13 years agoARM: tegra: fix regression from addruart rewrite
Olof Johansson [Tue, 7 Dec 2010 01:36:32 +0000 (19:36 -0600)]
ARM: tegra: fix regression from addruart rewrite

Commit 0ea129300982 ("arm: return both physical and virtual addresses
from addruart") took out the test for MMU on/off but didn't switch the
ldr instructions to no longer be conditionals based on said test.

Fix that.

Signed-off-by: Olof Johansson <olof@lixom.net>
Acked-by: Colin Cross <ccross@android.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agoMerge branch 'pm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/suspe...
Linus Torvalds [Mon, 6 Dec 2010 23:51:14 +0000 (15:51 -0800)]
Merge branch 'pm-fixes' of git://git./linux/kernel/git/rafael/suspend-2.6

* 'pm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/suspend-2.6:
  PM / Hibernate: Fix memory corruption related to swap
  PM / Hibernate: Use async I/O when reading compressed hibernation image

13 years agoPM / Hibernate: Fix memory corruption related to swap
Rafael J. Wysocki [Fri, 3 Dec 2010 21:57:45 +0000 (22:57 +0100)]
PM / Hibernate: Fix memory corruption related to swap

There is a problem that swap pages allocated before the creation of
a hibernation image can be released and used for storing the contents
of different memory pages while the image is being saved.  Since the
kernel stored in the image doesn't know of that, it causes memory
corruption to occur after resume from hibernation, especially on
systems with relatively small RAM that need to swap often.

This issue can be addressed by keeping the GFP_IOFS bits clear
in gfp_allowed_mask during the entire hibernation, including the
saving of the image, until the system is finally turned off or
the hibernation is aborted.  Unfortunately, for this purpose
it's necessary to rework the way in which the hibernate and
suspend code manipulates gfp_allowed_mask.

This change is based on an earlier patch from Hugh Dickins.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Reported-by: Ondrej Zary <linux@rainbow-software.org>
Acked-by: Hugh Dickins <hughd@google.com>
Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: stable@kernel.org
13 years agoMerge master.kernel.org:/home/rmk/linux-2.6-arm
Linus Torvalds [Mon, 6 Dec 2010 22:49:51 +0000 (14:49 -0800)]
Merge master.kernel.org:/home/rmk/linux-2.6-arm

* master.kernel.org:/home/rmk/linux-2.6-arm:
  ARM: 6524/1: GIC irq desciptor bug fix
  ARM: 6523/1: iop: ensure sched_clock() is notrace
  ARM: 6456/1: Fix for building DEBUG with sa11xx_base.c as a module.
  ARM: 6519/1: kuser: Fix incorrect cmpxchg syscall in kuser helpers
  ARM: 6505/1: kprobes: Don't HAVE_KPROBES when CONFIG_THUMB2_KERNEL is selected
  ARM: 6508/1: vexpress: Correct data alignment in headsmp.S for CONFIG_THUMB2_KERNEL
  ARM: 6507/1: RealView: Correct data alignment in headsmp.S for CONFIG_THUMB2_KERNEL
  ARM: 6504/1: Thumb-2: Fix long-distance conditional branches in head.S for Thumb-2.
  ARM: 6503/1: Thumb-2: Restore sensible zImage header layout for CONFIG_THUMB2_KERNEL
  ARM: 6502/1: Thumb-2: Fix CONFIG_THUMB2_KERNEL breakage in compressed/head.S
  ARM: 6501/1: Thumb-2: Correct data alignment for CONFIG_THUMB2_KERNEL in mm/proc-v7.S
  ARM: 6500/1: Thumb-2: Correct data alignment for CONFIG_THUMB2_KERNEL in kernel/head.S
  ARM: 6499/1: Thumb-2: Correct data alignment for CONFIG_THUMB2_KERNEL in bootp/init.S
  ARM: 6498/1: vfp: Correct data alignment for CONFIG_THUMB2_KERNEL
  ARM: 6497/1: kexec: Correct data alignment for CONFIG_THUMB2_KERNEL
  ARM: 6496/1: GIC: Do not try to register more then NR_IRQS interrupts
  ARM: cns3xxx: Fix build with CONFIG_PCI=y

13 years agoPM / Hibernate: Use async I/O when reading compressed hibernation image
Bojan Smojver [Thu, 25 Nov 2010 22:41:39 +0000 (23:41 +0100)]
PM / Hibernate: Use async I/O when reading compressed hibernation image

This is a fix for reading LZO compressed image using async I/O.
Essentially, instead of having just one page into which we keep
reading blocks from swap, we allocate enough of them to cover the
largest compressed size and then let block I/O pick them all up. Once
we have them all (and here we wait), we decompress them, as usual.
Obviously, the very first block we still pick up synchronously,
because we need to know the size of the lot before we pick up the
rest.

Also fixed the copyright line, which I've forgotten before.

Signed-off-by: Bojan Smojver <bojan@rexursive.com>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
13 years agoMerge branch 'for-rmk-fixes' of git://git.infradead.org/users/cbou/linux-cns3xxx
Russell King [Mon, 6 Dec 2010 15:39:23 +0000 (15:39 +0000)]
Merge branch 'for-rmk-fixes' of git://git.infradead.org/users/cbou/linux-cns3xxx

13 years agoARM: 6524/1: GIC irq desciptor bug fix
Chao Xie [Mon, 6 Dec 2010 06:01:10 +0000 (07:01 +0100)]
ARM: 6524/1: GIC irq desciptor bug fix

gic_set_cpu will directly use irq_desc[]. If CONFIG_SPARSE_IRQ is
enabled, there is no irq_desc[]. So we need use irq_to_desc(irq) to
get the descriptor for irq.

Signed-off-by: Chao Xie <chao.xie@marvell.com>
Acked-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
13 years agoMerge branch 'slab/urgent' of git://git.kernel.org/pub/scm/linux/kernel/git/penberg...
Linus Torvalds [Mon, 6 Dec 2010 00:45:02 +0000 (16:45 -0800)]
Merge branch 'slab/urgent' of git://git./linux/kernel/git/penberg/slab-2.6

* 'slab/urgent' of git://git.kernel.org/pub/scm/linux/kernel/git/penberg/slab-2.6:
  slub: Fix a crash during slabinfo -v

13 years agoMerge branch 'rc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuil...
Linus Torvalds [Mon, 6 Dec 2010 00:41:13 +0000 (16:41 -0800)]
Merge branch 'rc-fixes' of git://git./linux/kernel/git/mmarek/kbuild-2.6

* 'rc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild-2.6:
  initramfs: Really fix build break on symbol-prefixed archs
  [media] Fix Kconfig errors due to two visible menus
  i2c/algos: convert Kconfig to use the menu's `visible' keyword
  media/video: convert Kconfig to use the menu's `visible' keyword
  Revert "i2c: Fix Kconfig dependencies"
  kconfig: regen parser
  kconfig: add an option to determine a menu's visibility

13 years agoMerge branch 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/kyle/parisc-2.6
Linus Torvalds [Mon, 6 Dec 2010 00:40:31 +0000 (16:40 -0800)]
Merge branch 'fixes' of git://git./linux/kernel/git/kyle/parisc-2.6

* 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/kyle/parisc-2.6:
  parisc: Fix GSC PS/2 driver name for keyboard and mouse
  parisc: KittyHawk LCD fix
  parisc: convert the rest of the irq handlers to simple/percpu
  parisc: fix dino/gsc interrupts
  parisc: remove redundant initialization in sigsegv path of sys_rt_sigreturn

13 years agoRevert "vfs: show unreachable paths in getcwd and proc"
Eric W. Biederman [Sun, 5 Dec 2010 23:51:21 +0000 (15:51 -0800)]
Revert "vfs: show unreachable paths in getcwd and proc"

Because it caused a chroot ttyname regression in 2.6.36.

As of 2.6.36 ttyname does not work in a chroot.  It has already been
reported that screen breaks, and for me this breaks an automated
distribution testsuite, that I need to preserve the ability to run the
existing binaries on for several more years.  glibc 2.11.3 which has a
fix for this is not an option.

The root cause of this breakage is:

    commit 8df9d1a4142311c084ffeeacb67cd34d190eff74
    Author: Miklos Szeredi <mszeredi@suse.cz>
    Date:   Tue Aug 10 11:41:41 2010 +0200

    vfs: show unreachable paths in getcwd and proc

    Prepend "(unreachable)" to path strings if the path is not reachable
    from the current root.

    Two places updated are
     - the return string from getcwd()
     - and symlinks under /proc/$PID.

    Other uses of d_path() are left unchanged (we know that some old
    software crashes if /proc/mounts is changed).

Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
So remove the nice sounding, but ultimately ill advised change to how
/proc/fd symlinks work.

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agoparisc: Fix GSC PS/2 driver name for keyboard and mouse
Guy Martin [Tue, 2 Nov 2010 15:23:08 +0000 (15:23 +0000)]
parisc: Fix GSC PS/2 driver name for keyboard and mouse

Fix kernel warnings caused by the driver name of GSC PS/2 containing '/'.

The following warnings are observed on a K410 system :

[   10.700000] name 'GSC PS/2 keyboard'
[   10.732000] ------------[ cut here ]------------
[   10.772000] WARNING: at fs/proc/generic.c:323
[   10.828000] Modules linked in:
[   10.916000]
[   10.916000]      YZrvWESTHLNXBCVMcbcbcbcbOGFRQPDI
[   10.936000] PSW: 00000000000001000000000000001111 Not tainted
[   10.992000] r00-03  0004000f 104fe3e0 10201ea0 00000000
[   11.060000] r04-07  4fc405c8 00000006 4fc405c8 4fc40694
[   11.124000] r08-11  4fc40708 10438aa0 00000001 1043bfc8
[   11.184000] r12-15  104ff2a0 104ff2a0 4fc38634 104ff2a0
[   11.248000] r16-19  f0001570 10479af0 f000006c 1044fe50
[   11.308000] r20-23  00000000 00000028 104cd858 00000000
[   11.372000] r24-27  ffffffff 0000000e 1044fe10 1043bbe0
[   11.436000] r28-31  0000002b 00000078 4fc40800 0000000d
[   11.496000] sr00-03  00000000 00000000 00000000 00000000
[   11.560000] sr04-07  00000000 00000000 00000000 00000000
[   11.624000]
[   11.688000] IASQ: 00000000 00000000 IAOQ: 10201ea0 10201ea4
[   11.704000]  IIR: 03ffe01f    ISR: 00000000  IOR: 0000000d
[   11.772000]  CPU:        0   CR30: 4fc40000 CR31: f01043b0
[   11.836000]  ORIG_R28: 4fc40940
[   11.904000]  IAOQ[0]: __xlate_proc_name+0x90/0xd0
[   11.940000]  IAOQ[1]: __xlate_proc_name+0x94/0xd0
[   11.996000]  RP(r2): __xlate_proc_name+0x90/0xd0
[   12.052000] Backtrace:
[   12.108000]  [<10257790>] vsnprintf+0x290/0x4f4
[   12.136000]
[   12.188000] ---[ end trace 91bf6ece17e322dd ]---
[   12.208000] serio: GSC PS/2 keyboard port at 0x0001c000 irq 19 @ 10:12:7
[   12.264000] name 'GSC PS/2 mouse'
[   12.344000] ------------[ cut here ]------------
[   12.384000] WARNING: at fs/proc/generic.c:323
[   12.436000] Modules linked in:
[   12.524000]
[   12.528000]      YZrvWESTHLNXBCVMcbcbcbcbOGFRQPDI
[   12.544000] PSW: 00000000000001000000000000001111 Tainted: G        W
[   12.600000] r00-03  0004000f 104fe3e0 10201ea0 00000000
[   12.680000] r04-07  4fc405c8 00000006 4fc405c8 4fc40694
[   12.740000] r08-11  4fc40708 10438aa0 00000001 1043bfc8
[   12.804000] r12-15  104ff2a0 104ff2a0 4fc38634 104ff2a0
[   12.868000] r16-19  f0001570 10479af0 f000006c 1044fe50
[   12.928000] r20-23  00000000 00000025 104cd858 00000000
[   12.992000] r24-27  ffffffff 0000000e 1044fe10 1043bbe0
[   13.056000] r28-31  00000028 00000078 4fc40800 0000000d
[   13.116000] sr00-03  00000000 00000000 00000000 00000000
[   13.180000] sr04-07  00000000 00000000 00000000 00000000
[   13.244000]
[   13.308000] IASQ: 00000000 00000000 IAOQ: 10201ea0 10201ea4
[   13.324000]  IIR: 03ffe01f    ISR: 00000000  IOR: 0000000d
[   13.392000]  CPU:        0   CR30: 4fc40000 CR31: f01043b0
[   13.456000]  ORIG_R28: 4fc40940
[   13.524000]  IAOQ[0]: __xlate_proc_name+0x90/0xd0
[   13.560000]  IAOQ[1]: __xlate_proc_name+0x94/0xd0
[   13.616000]  RP(r2): __xlate_proc_name+0x90/0xd0
[   13.672000] Backtrace:
[   13.728000]  [<10257790>] vsnprintf+0x290/0x4f4
[   13.756000]
[   13.808000] ---[ end trace 91bf6ece17e322de ]---
[   13.828000] serio: GSC PS/2 mouse port at 0x00020100 irq 19 @ 10:12:8

Signed-off-by: Guy Martin <gmsoft@tuxicoman.be>
Acked-by: Helge Deller <deller@gmx.de>
Signed-off-by: Kyle McMartin <kyle@mcmartin.ca>
13 years agoparisc: KittyHawk LCD fix
Guy Martin [Sat, 6 Nov 2010 16:24:29 +0000 (16:24 +0000)]
parisc: KittyHawk LCD fix

K class aka KittyHawk don't have LED support on their LCD. Installing
HP-UX confirmed this. The current led_wq fills the LCD with black
characters each time it runs.

The patch prevents the led_wq workqueue and its proc entry to be
created for KittyHawk machines.

It also increase min_cmd_delay as currently, one character out of two
is lost when a string is sent to the LCD.

Signed-off-by: Guy Martin <gmsoft@tuxicoman.be>
Signed-off-by: Kyle McMartin <kyle@mcmartin.c>
13 years agoparisc: convert the rest of the irq handlers to simple/percpu
James Bottomley [Fri, 3 Dec 2010 02:01:05 +0000 (02:01 +0000)]
parisc: convert the rest of the irq handlers to simple/percpu

The generic conversion eliminates the spurious no_ack and no_end
routines, converts all the cascaded handlers to handle_simple_irq() and
makes iosapic use a modified handle_percpu_irq() to become the same as
the CPU irq's.  This isn't an essential change, but it eliminates the
mask/unmask overhead of handle_level_irq().

Signed-off-by: James Bottomley <James.Bottomley@suse.de>
Tested-by: Helge Deller <deller@gmx.de>
Signed-off-by: Kyle McMartin <kyle@mcmartin.ca>
13 years agoparisc: fix dino/gsc interrupts
James Bottomley [Thu, 2 Dec 2010 23:36:47 +0000 (23:36 +0000)]
parisc: fix dino/gsc interrupts

The essential problem we're currently having is that dino (and gsc) is a
cascaded CPU interrupt.  Under the old __do_IRQ() handler, our CPU
interrupts basically did an ack followed by an end.  In the new scheme,
we replaced them with level handlers which do a mask, an ack and then an
unmask (but no end).  Instead, with the renaming of end to eoi, we
actually want to call the percpu flow handlers, because they actually
have all the characteristics we want.

This patch does the conversion and gets my C360 booting again.

Signed-off-by: James Bottomley <James.Bottomley@suse.de>
Signed-off-by: Kyle McMartin <kyle@mcmartin.ca>
13 years agoARM: 6523/1: iop: ensure sched_clock() is notrace
Rabin Vincent [Sat, 4 Dec 2010 05:20:52 +0000 (06:20 +0100)]
ARM: 6523/1: iop: ensure sched_clock() is notrace

Include sched.h to ensure sched_clock() has the notrace
annotation, and mark any functions it calls as notrace
too.
Include sched.h to ensure sched_clock() has the notrace
annotation, and mark any functions it calls as notrace
too.

Acked-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Rabin Vincent <rabin@rab.in>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
13 years agoARM: 6456/1: Fix for building DEBUG with sa11xx_base.c as a module.
Marcelo Roberto Jimenez [Mon, 18 Oct 2010 21:38:08 +0000 (22:38 +0100)]
ARM: 6456/1: Fix for building DEBUG with sa11xx_base.c as a module.

This patch fixes a compilation issue when compiling PCMCIA SA1100
support as a module with PCMCIA_DEBUG enabled. The symbol
soc_pcmcia_debug was not beeing exported.
ARM: pcmcia: Fix for building DEBUG with sa11xx_base.c as a module.

This patch fixes a compilation issue when compiling PCMCIA SA1100
support as a module with PCMCIA_DEBUG enabled. The symbol
soc_pcmcia_debug was not beeing exported.

Cc: <stable@kernel.org>
Signed-off-by: Marcelo Roberto Jimenez <mroberto@cpti.cetuc.puc-rio.br>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
13 years agoARM: 6519/1: kuser: Fix incorrect cmpxchg syscall in kuser helpers
Dave Martin [Wed, 1 Dec 2010 17:12:43 +0000 (18:12 +0100)]
ARM: 6519/1: kuser: Fix incorrect cmpxchg syscall in kuser helpers

The existing code invokes the syscall with rubbish in r7,
due to what looks like an incorrect literal load idiom.

Reviewed-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Dave Martin <dave.martin@linaro.org>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
13 years agoslub: Fix a crash during slabinfo -v
Tero Roponen [Wed, 1 Dec 2010 18:04:20 +0000 (20:04 +0200)]
slub: Fix a crash during slabinfo -v

Commit f7cb1933621bce66a77f690776a16fe3ebbc4d58 ("SLUB: Pass active
and inactive redzone flags instead of boolean to debug functions")
missed two instances of check_object(). This caused a lot of warnings
during 'slabinfo -v' finally leading to a crash:

  BUG ext4_xattr: Freepointer corrupt
  ...
  BUG buffer_head: Freepointer corrupt
  ...
  BUG ext4_alloc_context: Freepointer corrupt
  ...
  ...
  BUG: unable to handle kernel NULL pointer dereference at 0000000000000008
  IP: [<ffffffff810a291f>] file_sb_list_del+0x1c/0x35
  PGD 79d78067 PUD 79e67067 PMD 0
  Oops: 0002 [#1] SMP
  last sysfs file: /sys/kernel/slab/:t-0000192/validate

This patch fixes the problem by converting the two missed instances.

Acked-by: Christoph Lameter <cl@linux.com>
Signed-off-by: Tero Roponen <tero.roponen@gmail.com>
Signed-off-by: Pekka Enberg <penberg@kernel.org>
13 years agoMerge branch '2.6.37-rc4-pvhvm-fixes' of git://xenbits.xen.org/people/sstabellini...
Linus Torvalds [Fri, 3 Dec 2010 19:30:57 +0000 (11:30 -0800)]
Merge branch '2.6.37-rc4-pvhvm-fixes' of git://xenbits.xen.org/people/sstabellini/linux-pvhvm

* '2.6.37-rc4-pvhvm-fixes' of git://xenbits.xen.org/people/sstabellini/linux-pvhvm:
  xen: unplug the emulated devices at resume time
  xen: fix save/restore for PV on HVM guests with pirq remapping
  xen: resume the pv console for hvm guests too
  xen: fix MSI setup and teardown for PV on HVM guests
  xen: use PHYSDEVOP_get_free_pirq to implement find_unbound_pirq

13 years agoMerge branches 'upstream/core' and 'upstream/bugfix' of git://git.kernel.org/pub...
Linus Torvalds [Fri, 3 Dec 2010 18:08:52 +0000 (10:08 -0800)]
Merge branches 'upstream/core' and 'upstream/bugfix' of git://git./linux/kernel/git/jeremy/xen

* 'upstream/core' of git://git.kernel.org/pub/scm/linux/kernel/git/jeremy/xen:
  xen: allocate irq descs on any NUMA node
  xen: prevent crashes with non-HIGHMEM 32-bit kernels with largeish memory
  xen: use default_idle
  xen: clean up "extra" memory handling some more

* 'upstream/bugfix' of git://git.kernel.org/pub/scm/linux/kernel/git/jeremy/xen:
  xen: x86/32: perform initial startup on initial_page_table
  xen: don't bother to stop other cpus on shutdown/reboot

13 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6
Linus Torvalds [Fri, 3 Dec 2010 17:20:41 +0000 (09:20 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/tiwai/sound-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6:
  ASoC: omap: N810: Don't select CONFIG_OMAP_MUX but make it as dependency
  ALSA: hda: Use "alienware" model quirk for another SSID
  ASoC: WM8731: Fix incorrect mask for bypass path disable
  s6105-ipcam: fix compilation
  s6000-pcm: fix compilation
  s6000-i2s: fix compilation
  ASoC: Fix missing spin_unlock_irqrestore
  ALSA: Fix SNDCTL_DSP_RESET ioctl for OSS emulation
  ASoC: Add missing dev_set_drvdata in p1022_ds_probe
  ASoC: Add missing dev_set_drvdata in mpc8610_hpcd_probe
  ASoC: Remove unneeded !! operations while checking return value of nuc900_checkready
  ASoC: Fix compile error for nuc900-pcm.c
  ASoC: Fix prototype for nuc900_ac97_probe and nuc900_ac97_remove
  ASoC: Fix compile error for nuc900-ac97.c
  ALSA: hda: Use BIOS auto-parsing instead of existing model quirk for MEDION MD2

13 years agoMerge branch 'linux-next' of git://git.infradead.org/ubi-2.6
Linus Torvalds [Fri, 3 Dec 2010 17:20:21 +0000 (09:20 -0800)]
Merge branch 'linux-next' of git://git.infradead.org/ubi-2.6

* 'linux-next' of git://git.infradead.org/ubi-2.6:
  UBI: fix corrupted PEB detection for NOR flash

13 years agoMerge branch 'sh-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Fri, 3 Dec 2010 17:19:54 +0000 (09:19 -0800)]
Merge branch 'sh-fixes-for-linus' of git://git./linux/kernel/git/lethal/sh-2.6

* 'sh-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/lethal/sh-2.6:
  sh: se/7724: Remove FSI/B of GPIO init code
  sh: se/7724: Update clock framework of FSI clock to non-legacy
  sh: Assume new page cache pages have dirty dcache lines.
  sh: boards: mach-se: use IS_ERR() instead of NULL check
  sh: Add div6_reparent_clks to clock framework for FSI
  dma: shdma: add a MODULE_ALIAS() to allow module autoloading

13 years agoMerge branch 'fbdev-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Fri, 3 Dec 2010 17:19:31 +0000 (09:19 -0800)]
Merge branch 'fbdev-fixes-for-linus' of git://git./linux/kernel/git/lethal/fbdev-2.6

* 'fbdev-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/lethal/fbdev-2.6:
  lxfb: Maintain video processor palette through suspend/resume
  video: da8xx: Register IRQ as last thing in driver probing.
  framebuffer: fix fbcmap.c kernel-doc warning

13 years agoMN10300: Implement asm/syscall.h
David Howells [Fri, 3 Dec 2010 15:59:50 +0000 (15:59 +0000)]
MN10300: Implement asm/syscall.h

Implement asm/syscall.h for the MN10300 arch.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agoUBI: fix corrupted PEB detection for NOR flash
Artem Bityutskiy [Thu, 2 Dec 2010 04:34:01 +0000 (06:34 +0200)]
UBI: fix corrupted PEB detection for NOR flash

My new shiny code for corrupted PEB detection has NOR specific bug.
We tread PEB as corrupted and preserve it, if

1. EC header is OK.
2. VID header is corrupted.
3. data area is not "all 0xFFs"

In case of NOR we have 'nor_erase_prepare()' quirk, which invalidates
the headers before erasing the PEB. And we invalidate first the VID
header, and then the EC header. So if a power cut happens after we have
invalidated the VID header, but before we have invalidated the EC
header, we end up with a PEB which satisfies the above 3 conditions,
and the scanning code will treat it as corrupted, and will print
scary warnings, wrongly.

This patch fixes the issue by firt invalidating the EC header, then
invalidating the VID header. In case of power cut inbetween, we still
just lose the EC header, and UBI can deal with this situation gracefully.

Thanks to Anatolij Gustschin <agust@denx.de> for tracking this down.

Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
Reported-by: Anatolij Gustschin <agust@denx.de>
Tested-by: Anatolij Gustschin <agust@denx.de>
13 years agoMerge branch 'sh/urgent' into sh-fixes-for-linus
Paul Mundt [Fri, 3 Dec 2010 05:42:29 +0000 (14:42 +0900)]
Merge branch 'sh/urgent' into sh-fixes-for-linus

13 years agosh: se/7724: Remove FSI/B of GPIO init code
Nobuhiro Iwamatsu [Fri, 3 Dec 2010 05:37:54 +0000 (05:37 +0000)]
sh: se/7724: Remove FSI/B of GPIO init code

se7724 board does not have FSI/B.

Signed-off-by: Nobuhiro Iwamatsu <nobuhiro.iwamatsu.yj@renesas.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
13 years agosh: se/7724: Update clock framework of FSI clock to non-legacy
Nobuhiro Iwamatsu [Fri, 3 Dec 2010 05:37:53 +0000 (05:37 +0000)]
sh: se/7724: Update clock framework of FSI clock to non-legacy

Signed-off-by: Nobuhiro Iwamatsu <nobuhiro.iwamatsu.yj@renesas.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
13 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid
Linus Torvalds [Fri, 3 Dec 2010 01:40:04 +0000 (17:40 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/jikos/hid

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid:
  HID: length resolution should be reported units/mm
  HID: add support for F430 Force Feedback Wheel
  HID: egalax: Use kzalloc
  HID: Remove KERN_DEBUG from dbg_hid use

Manually fixed trivial conflict in drivers/hid/hid-input.c (due to
removal of KERN_DEBUG from dbg_hid use clashing with new keycode
interface switch)

13 years agoxen: allocate irq descs on any NUMA node
Jeremy Fitzhardinge [Fri, 3 Dec 2010 00:14:27 +0000 (16:14 -0800)]
xen: allocate irq descs on any NUMA node

Allocate irq descs on any NUMA node (we don't care) rather than
specifically node 0, which may not exist.

(At the moment NUMA is meaningless within a domain, so any info
the kernel has is just from an SRAT table we haven't suppressed/disabled.)

Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
13 years agoxen: prevent crashes with non-HIGHMEM 32-bit kernels with largeish memory
Jeremy Fitzhardinge [Thu, 2 Dec 2010 23:30:06 +0000 (15:30 -0800)]
xen: prevent crashes with non-HIGHMEM 32-bit kernels with largeish memory

If this is a non-HIGHMEM 32-bit kernel, then the page structures only go
up to the limit of addressable memory, even if more memory is physically
present.  Don't try to add that extra memory to the balloon.

Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
13 years agodo_exit(): make sure that we run with get_fs() == USER_DS
Nelson Elhage [Thu, 2 Dec 2010 22:31:21 +0000 (14:31 -0800)]
do_exit(): make sure that we run with get_fs() == USER_DS

If a user manages to trigger an oops with fs set to KERNEL_DS, fs is not
otherwise reset before do_exit().  do_exit may later (via mm_release in
fork.c) do a put_user to a user-controlled address, potentially allowing
a user to leverage an oops into a controlled write into kernel memory.

This is only triggerable in the presence of another bug, but this
potentially turns a lot of DoS bugs into privilege escalations, so it's
worth fixing.  I have proof-of-concept code which uses this bug along
with CVE-2010-3849 to write a zero to an arbitrary kernel address, so
I've tested that this is not theoretical.

A more logical place to put this fix might be when we know an oops has
occurred, before we call do_exit(), but that would involve changing
every architecture, in multiple places.

Let's just stick it in do_exit instead.

[akpm@linux-foundation.org: update code comment]
Signed-off-by: Nelson Elhage <nelhage@ksplice.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agoksm: annotate ksm_thread_mutex is no deadlock source
KOSAKI Motohiro [Thu, 2 Dec 2010 22:31:20 +0000 (14:31 -0800)]
ksm: annotate ksm_thread_mutex is no deadlock source

commit 62b61f611e ("ksm: memory hotremove migration only") caused the
following new lockdep warning.

  =======================================================
  [ INFO: possible circular locking dependency detected ]
  -------------------------------------------------------
  bash/1621 is trying to acquire lock:
   ((memory_chain).rwsem){.+.+.+}, at: [<ffffffff81079339>]
  __blocking_notifier_call_chain+0x69/0xc0

  but task is already holding lock:
   (ksm_thread_mutex){+.+.+.}, at: [<ffffffff8113a3aa>]
  ksm_memory_callback+0x3a/0xc0

  which lock already depends on the new lock.

  the existing dependency chain (in reverse order) is:

  -> #1 (ksm_thread_mutex){+.+.+.}:
       [<ffffffff8108b70a>] lock_acquire+0xaa/0x140
       [<ffffffff81505d74>] __mutex_lock_common+0x44/0x3f0
       [<ffffffff81506228>] mutex_lock_nested+0x48/0x60
       [<ffffffff8113a3aa>] ksm_memory_callback+0x3a/0xc0
       [<ffffffff8150c21c>] notifier_call_chain+0x8c/0xe0
       [<ffffffff8107934e>] __blocking_notifier_call_chain+0x7e/0xc0
       [<ffffffff810793a6>] blocking_notifier_call_chain+0x16/0x20
       [<ffffffff813afbfb>] memory_notify+0x1b/0x20
       [<ffffffff81141b7c>] remove_memory+0x1cc/0x5f0
       [<ffffffff813af53d>] memory_block_change_state+0xfd/0x1a0
       [<ffffffff813afd62>] store_mem_state+0xe2/0xf0
       [<ffffffff813a0bb0>] sysdev_store+0x20/0x30
       [<ffffffff811bc116>] sysfs_write_file+0xe6/0x170
       [<ffffffff8114f398>] vfs_write+0xc8/0x190
       [<ffffffff8114fc14>] sys_write+0x54/0x90
       [<ffffffff810028b2>] system_call_fastpath+0x16/0x1b

  -> #0 ((memory_chain).rwsem){.+.+.+}:
       [<ffffffff8108b5ba>] __lock_acquire+0x155a/0x1600
       [<ffffffff8108b70a>] lock_acquire+0xaa/0x140
       [<ffffffff81506601>] down_read+0x51/0xa0
       [<ffffffff81079339>] __blocking_notifier_call_chain+0x69/0xc0
       [<ffffffff810793a6>] blocking_notifier_call_chain+0x16/0x20
       [<ffffffff813afbfb>] memory_notify+0x1b/0x20
       [<ffffffff81141f1e>] remove_memory+0x56e/0x5f0
       [<ffffffff813af53d>] memory_block_change_state+0xfd/0x1a0
       [<ffffffff813afd62>] store_mem_state+0xe2/0xf0
       [<ffffffff813a0bb0>] sysdev_store+0x20/0x30
       [<ffffffff811bc116>] sysfs_write_file+0xe6/0x170
       [<ffffffff8114f398>] vfs_write+0xc8/0x190
       [<ffffffff8114fc14>] sys_write+0x54/0x90
       [<ffffffff810028b2>] system_call_fastpath+0x16/0x1b

But it's a false positive.  Both memory_chain.rwsem and ksm_thread_mutex
have an outer lock (mem_hotplug_mutex).  So they cannot deadlock.

Thus, This patch annotate ksm_thread_mutex is not deadlock source.

[akpm@linux-foundation.org: update comment, from Hugh]
Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Acked-by: Hugh Dickins <hughd@google.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agomem-hotplug: introduce {un}lock_memory_hotplug()
KOSAKI Motohiro [Thu, 2 Dec 2010 22:31:19 +0000 (14:31 -0800)]
mem-hotplug: introduce {un}lock_memory_hotplug()

Presently hwpoison is using lock_system_sleep() to prevent a race with
memory hotplug.  However lock_system_sleep() is a no-op if
CONFIG_HIBERNATION=n.  Therefore we need a new lock.

Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Suggested-by: Hugh Dickins <hughd@google.com>
Acked-by: Hugh Dickins <hughd@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agoDocumentation/filesystems/vfs.txt: fix ->repeasepage() description
Andrew Morton [Thu, 2 Dec 2010 22:31:19 +0000 (14:31 -0800)]
Documentation/filesystems/vfs.txt: fix ->repeasepage() description

->releasepage() does not remove the page from the mapping.

Acked-by: Neil Brown <neilb@suse.de>
Cc: Trond Myklebust <Trond.Myklebust@netapp.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agovmalloc: eagerly clear ptes on vunmap
Jeremy Fitzhardinge [Thu, 2 Dec 2010 22:31:18 +0000 (14:31 -0800)]
vmalloc: eagerly clear ptes on vunmap

On stock 2.6.37-rc4, running:

  # mount lilith:/export /mnt/lilith
  # find  /mnt/lilith/ -type f -print0 | xargs -0 file

crashes the machine fairly quickly under Xen.  Often it results in oops
messages, but the couple of times I tried just now, it just hung quietly
and made Xen print some rude messages:

    (XEN) mm.c:2389:d80 Bad type (saw 7400000000000001 != exp
    3000000000000000) for mfn 1d7058 (pfn 18fa7)
    (XEN) mm.c:964:d80 Attempt to create linear p.t. with write perms
    (XEN) mm.c:2389:d80 Bad type (saw 7400000000000010 != exp
    1000000000000000) for mfn 1d2e04 (pfn 1d1fb)
    (XEN) mm.c:2965:d80 Error while pinning mfn 1d2e04

Which means the domain tried to map a pagetable page RW, which would
allow it to map arbitrary memory, so Xen stopped it.  This is because
vm_unmap_ram() left some pages mapped in the vmalloc area after NFS had
finished with them, and those pages got recycled as pagetable pages
while still having these RW aliases.

Removing those mappings immediately removes the Xen-visible aliases, and
so it has no problem with those pages being reused as pagetable pages.
Deferring the TLB flush doesn't upset Xen because it can flush the TLB
itself as needed to maintain its invariants.

When unmapping a region in the vmalloc space, clear the ptes
immediately.  There's no point in deferring this because there's no
amortization benefit.

The TLBs are left dirty, and they are flushed lazily to amortize the
cost of the IPIs.

This specific motivation for this patch is an oops-causing regression
since 2.6.36 when using NFS under Xen, triggered by the NFS client's use
of vm_map_ram() introduced in 56e4ebf877b60 ("NFS: readdir with vmapped
pages") .  XFS also uses vm_map_ram() and could cause similar problems.

Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Nick Piggin <npiggin@kernel.dk>
Cc: Bryan Schumaker <bjschuma@netapp.com>
Cc: Trond Myklebust <Trond.Myklebust@netapp.com>
Cc: Alex Elder <aelder@sgi.com>
Cc: Dave Chinner <david@fromorbit.com>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agocs5535-gpio: apply CS5536 errata workaround for GPIOs
Andres Salomon [Thu, 2 Dec 2010 22:31:17 +0000 (14:31 -0800)]
cs5535-gpio: apply CS5536 errata workaround for GPIOs

The AMD Geode CS5536 Companion Device Silicon Revision B1 Specification
Update mentions the follow as issue #36:

 "Atomic write transactions to the atomic GPIO High Bank Feature Bit
  registers should only affect the bits selected [...]"

 "after Suspend, an atomic write transaction [...] will clear all
  non-selected bits of the accessed register."

In other words, writing to the high bank for a single GPIO bit will
clear every other GPIO bit (but only sometimes after a suspend).

The workaround described is obvious and simple; do a read-modify-write.
This patch does that, and documents why we're doing it.

Signed-off-by: Andres Salomon <dilinger@queued.net>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agoreiserfs: don't acquire lock recursively in reiserfs_acl_chmod
Frederic Weisbecker [Thu, 2 Dec 2010 22:31:16 +0000 (14:31 -0800)]
reiserfs: don't acquire lock recursively in reiserfs_acl_chmod

reiserfs_acl_chmod() can be called by reiserfs_set_attr() and then take
the reiserfs lock a second time.  Thereafter it may call journal_begin()
that definitely requires the lock not to be nested in order to release
it before taking the journal mutex because the reiserfs lock depends on
the journal mutex already.

So, aviod nesting the lock in reiserfs_acl_chmod().

Reported-by: Pawel Zawora <pzawora@gmail.com>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Tested-by: Pawel Zawora <pzawora@gmail.com>
Cc: Jeff Mahoney <jeffm@suse.com>
Cc: <stable@kernel.org> [2.6.32.x+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agoleds: fix up dependencies
Johannes Berg [Thu, 2 Dec 2010 22:31:14 +0000 (14:31 -0800)]
leds: fix up dependencies

It's not useful to build LED triggers when there's no LEDs that can be
triggered by them.  Therefore, fix up the dependencies so that this
cannot happen, and fix a few users that select triggers to depend on
LEDS_CLASS as well (there is also one user that also selects LEDS_CLASS,
which is OK).

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Reported-by: Randy Dunlap <randy.dunlap@oracle.com>
Acked-by: Randy Dunlap <randy.dunlap@oracle.com>
Tested-by: Ingo Molnar <mingo@elte.hu>
Cc: Arnd Hannemann <arnd@arndnet.de>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Richard Purdie <rpurdie@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agovmstat: fix dirty threshold ordering
Wu Fengguang [Thu, 2 Dec 2010 22:31:13 +0000 (14:31 -0800)]
vmstat: fix dirty threshold ordering

The nr_dirty_[background_]threshold fields are misplaced before the
numa_* fields, and users will read strange values.

This is the right order.  Before patch, nr_dirty_background_threshold
will read as 0 (the value from numa_miss).

numa_hit 128501
numa_miss 0
numa_foreign 0
numa_interleave 7388
numa_local 128501
numa_other 0
nr_dirty_threshold 144291
nr_dirty_background_threshold 72145

Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
Cc: Michael Rubin <mrubin@google.com>
Reviewed-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Reviewed-by: Minchan Kim <minchan.kim@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agomm/mempolicy.c: add rcu read lock to protect pid structure
Zeng Zhaoming [Thu, 2 Dec 2010 22:31:13 +0000 (14:31 -0800)]
mm/mempolicy.c: add rcu read lock to protect pid structure

find_task_by_vpid() should be protected by rcu_read_lock(), to prevent
free_pid() reclaiming pid.

Signed-off-by: Zeng Zhaoming <zengzm.kernel@gmail.com>
Cc: "Paul E. McKenney" <paulmck@us.ibm.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Christoph Lameter <cl@linux-foundation.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agomm/hugetlb.c: avoid double unlock_page() in hugetlb_fault()
Dean Nelson [Thu, 2 Dec 2010 22:31:12 +0000 (14:31 -0800)]
mm/hugetlb.c: avoid double unlock_page() in hugetlb_fault()

Have hugetlb_fault() call unlock_page(page) only if it had previously
called lock_page(page).

Setting CONFIG_DEBUG_VM=y and then running the libhugetlbfs test suite,
resulted in the tripping of VM_BUG_ON(!PageLocked(page)) in
unlock_page() having been called by hugetlb_fault() when page ==
pagecache_page.  This patch remedied the problem.

Signed-off-by: Dean Nelson <dnelson@redhat.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agoMerge branch 'staging-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh...
Linus Torvalds [Thu, 2 Dec 2010 20:59:11 +0000 (12:59 -0800)]
Merge branch 'staging-linus' of git://git./linux/kernel/git/gregkh/staging-2.6

* 'staging-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging-2.6: (27 commits)
  Staging: rt2870: Add USB ID for Buffalo Airstation WLI-UC-GN
  staging: easycap needs smp_lock.h, fixes build error
  Staging: batman-adv: ensure that eth_type_trans gets linear memory
  Staging: batman-adv: Don't remove interface with spinlock held
  staging: brcm80211: updated maintainers contact information
  staging: fix winbond build, needs delay.h
  Staging: line6: fix up my fixup for some sysfs attribute permissions
  Staging: zram: fix up my fixup for some sysfs attribute permissions
  Staging: udlfb: fix up my fixup for some sysfs attribute permissions
  Staging: samsung-laptop: fix up my fixup for some sysfs attribute permissions
  Staging: iio: adis16220: fix up my fixup for some sysfs attribute permissions
  Staging: frontier: fix up my fixup for some sysfs attribute permissions
  Staging: asus_oled: fix up my fixup for some sysfs attribute permissions
  staging: spectra: fix build error
  Staging: intel_sst: fix memory leak
  Staging: rtl8712: signedness bug in init
  staging: rtl8187se: Change panic to warn when RF switch turned off
  staging: comedi: fix memory leak
  Staging: quickstart: free after input_unregister_device()
  Staging: speakup: free after input_unregister_device()
  ...

13 years agoMerge branch 'driver-core-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Thu, 2 Dec 2010 20:58:36 +0000 (12:58 -0800)]
Merge branch 'driver-core-linus' of git://git./linux/kernel/git/gregkh/driver-core-2.6

* 'driver-core-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core-2.6:
  uio: Change mail address of Hans J. Koch
  driver core: prune docs about device_interface
  driver core: the development tree has switched to git

13 years agoMerge branch 'tty-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh...
Linus Torvalds [Thu, 2 Dec 2010 20:58:16 +0000 (12:58 -0800)]
Merge branch 'tty-linus' of git://git./linux/kernel/git/gregkh/tty-2.6

* 'tty-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty-2.6:
  serial: mfd: adjust the baud rate setting
  TTY: open/hangup race fixup
  TTY: don't allow reopen when ldisc is changing
  NET: wan/x25, fix ldisc->open retval
  TTY: ldisc, fix open flag handling
  serial8250: Mark console as CON_ANYTIME

13 years agoMerge branch 'usb-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh...
Linus Torvalds [Thu, 2 Dec 2010 20:57:35 +0000 (12:57 -0800)]
Merge branch 'usb-linus' of git://git./linux/kernel/git/gregkh/usb-2.6

* 'usb-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6:
  USB: fix autosuspend bug in usb-serial
  USB: ehci: disable LPM and PPCD for nVidia MCP89 chips
  USB: serial: ftdi_sio: Vardaan USB RS422/485 converter PID added
  USB: yurex: add .llseek fop to file_operations
  USB: ftdi_sio: Add ID for RT Systems USB-29B radio cable
  usb: musb: do not use dma for control transfers
  usb: musb: gadget: fix compilation warning
  usb: musb: clear RXCSR_AUTOCLEAR before PIO read
  usb: musb: unmap dma buffer when switching to PIO
  xhci: Don't let the USB core disable SuperSpeed ports.
  xhci: Setup array of USB 2.0 and USB 3.0 ports.
  xhci: Fix reset-device and configure-endpoint commands

13 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/wim/linux-2.6-watchdog
Linus Torvalds [Thu, 2 Dec 2010 20:11:31 +0000 (12:11 -0800)]
Merge git://git./linux/kernel/git/wim/linux-2.6-watchdog

* git://git.kernel.org/pub/scm/linux/kernel/git/wim/linux-2.6-watchdog:
  watchdog: it8712f_wdt: add note to Kconfig
  watchdog: gef_wdt: include fs.h
  watchdog: bcm63xx_wdt: improve platform part.
  watchdog: iTCO_wdt: TCO Watchdog patch for Intel Patsburg PCH

13 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland...
Linus Torvalds [Thu, 2 Dec 2010 20:10:56 +0000 (12:10 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/roland/infiniband

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband:
  IB: Fix information leak in marshalling code
  IB/pack: Remove some unused code added by the IBoE patches
  IB/mlx4: Fix IBoE link state
  IB/mlx4: Fix IBoE reported link rate
  mlx4_core: Workaround firmware bug in query dev cap
  IB/mlx4: Fix memory ordering of VLAN insertion control bits
  MAINTAINERS: Update NetEffect entry

13 years agoMerge branch 'for-linus' of git://oss.sgi.com/xfs/xfs
Linus Torvalds [Thu, 2 Dec 2010 17:13:36 +0000 (09:13 -0800)]
Merge branch 'for-linus' of git://oss.sgi.com/xfs/xfs

* 'for-linus' of git://oss.sgi.com/xfs/xfs:
  xfs: only run xfs_error_test if error injection is active
  xfs: avoid moving stale inodes in the AIL
  xfs: delayed alloc blocks beyond EOF are valid after writeback
  xfs: push stale, pinned buffers on trylock failures
  xfs: fix failed write truncation handling.

13 years agoMerge branch 'fix/asoc' into for-linus
Takashi Iwai [Thu, 2 Dec 2010 16:33:53 +0000 (17:33 +0100)]
Merge branch 'fix/asoc' into for-linus

13 years agoMerge branch 'for-2.6.37' of git://git.kernel.org/pub/scm/linux/kernel/git/lrg/asoc...
Takashi Iwai [Thu, 2 Dec 2010 16:31:18 +0000 (17:31 +0100)]
Merge branch 'for-2.6.37' of git://git./linux/kernel/git/lrg/asoc-2.6 into fix/asoc

13 years agoMerge branch 'fix/hda' into for-linus
Takashi Iwai [Thu, 2 Dec 2010 16:14:50 +0000 (17:14 +0100)]
Merge branch 'fix/hda' into for-linus

13 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/lrg/voltage-2.6
Linus Torvalds [Thu, 2 Dec 2010 16:06:16 +0000 (08:06 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/lrg/voltage-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/lrg/voltage-2.6:
  regulator: fix kernel-doc for set_consumer_device_supply
  regulator: enable supply regulator only when use count is zero
  regulator: twl-regulator - fix twlreg_set_mode
  regulator: lock supply in regulator enable
  regulator: Return proper error for regulator_register()
  regulator: Ensure enough delay time for enabling regulator
  regulator: Remove a redundant device_remove_file call in create_regulator
  regulator: Staticise mc13783_powermisc_rmw()
  regulator: regulator disable supply fix

13 years agoMerge branch 'v4l_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab...
Linus Torvalds [Thu, 2 Dec 2010 16:05:56 +0000 (08:05 -0800)]
Merge branch 'v4l_for_linus' of git://git./linux/kernel/git/mchehab/linux-2.6

* 'v4l_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-2.6:
  [media] v4l: Remove module_name argument to the v4l2_i2c_new_subdev* functions
  [media] v4l: Remove hardcoded module names passed to v4l2_i2c_new_subdev* (2)

13 years agoMerge branch 'rbd-sysfs' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph...
Linus Torvalds [Thu, 2 Dec 2010 16:05:22 +0000 (08:05 -0800)]
Merge branch 'rbd-sysfs' of git://git./linux/kernel/git/sage/ceph-client

* 'rbd-sysfs' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client:
  rbd: replace the rbd sysfs interface

13 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/sfrench/cifs-2.6
Linus Torvalds [Thu, 2 Dec 2010 16:04:21 +0000 (08:04 -0800)]
Merge git://git./linux/kernel/git/sfrench/cifs-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/sfrench/cifs-2.6:
  cifs: fix parsing of hostname in dfs referrals
  cifs: display fsc in /proc/mounts
  cifs: enable fscache iff fsc mount option is used explicitly
  cifs: allow fsc mount option only if CONFIG_CIFS_FSCACHE is set
  cifs: Handle extended attribute name cifs_acl to generate cifs acl blob (try #4)
  cifs: Misc. cleanup in cifsacl handling [try #4]
  cifs: trivial comment fix for cifs_invalidate_mapping
  [CIFS] fs/cifs/Kconfig: CIFS depends on CRYPTO_HMAC
  cifs: don't take extra tlink reference in initiate_cifs_search
  cifs: Percolate error up to the caller during get/set acls [try #4]
  cifs: fix another memleak, in cifs_root_iget
  cifs: fix potential use-after-free in cifs_oplock_break_put

13 years agoxen: unplug the emulated devices at resume time
Stefano Stabellini [Wed, 1 Dec 2010 14:51:44 +0000 (14:51 +0000)]
xen: unplug the emulated devices at resume time

Early after being resumed we need to unplug again the emulated devices.

Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
13 years agoxen: fix save/restore for PV on HVM guests with pirq remapping
Stefano Stabellini [Wed, 1 Dec 2010 14:51:44 +0000 (14:51 +0000)]
xen: fix save/restore for PV on HVM guests with pirq remapping

Re-map and re-bind all the pirqs at resume time.

Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
13 years agoxen: resume the pv console for hvm guests too
Stefano Stabellini [Wed, 1 Dec 2010 14:51:44 +0000 (14:51 +0000)]
xen: resume the pv console for hvm guests too

Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
13 years agoxen: fix MSI setup and teardown for PV on HVM guests
Stefano Stabellini [Wed, 1 Dec 2010 14:51:44 +0000 (14:51 +0000)]
xen: fix MSI setup and teardown for PV on HVM guests

When remapping MSIs into pirqs for PV on HVM guests, qemu is responsible
for doing the actual mapping and unmapping.
We only give qemu the desired pirq number when we ask to do the mapping
the first time, after that we should be reading back the pirq number
from qemu every time we want to re-enable the MSI.

This fixes a bug in xen_hvm_setup_msi_irqs that manifests itself when
trying to enable the same MSI for the second time: the old MSI to pirq
mapping is still valid at this point but xen_hvm_setup_msi_irqs would
try to assign a new pirq anyway.
A simple way to reproduce this bug is to assign an MSI capable network
card to a PV on HVM guest, if the user brings down the corresponding
ethernet interface and up again, Linux would fail to enable MSIs on the
device.

Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
13 years agoxen: use PHYSDEVOP_get_free_pirq to implement find_unbound_pirq
Stefano Stabellini [Wed, 1 Dec 2010 14:51:44 +0000 (14:51 +0000)]
xen: use PHYSDEVOP_get_free_pirq to implement find_unbound_pirq

Use the new hypercall PHYSDEVOP_get_free_pirq to ask Xen to allocate a
pirq. Remove the unsupported PHYSDEVOP_get_nr_pirqs hypercall to get the
amount of pirq available.

This fixes find_unbound_pirq that otherwise would return a number
starting from nr_irqs that might very well be out of range in Xen.

The symptom of this bug is that when you passthrough an MSI capable pci
device to a PV on HVM guest, Linux would fail to enable MSIs on the
device.

Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
13 years agowatchdog: it8712f_wdt: add note to Kconfig
Wim Van Sebroeck [Thu, 2 Dec 2010 14:03:29 +0000 (14:03 +0000)]
watchdog: it8712f_wdt: add note to Kconfig

On some motherboards the it8712f watchdog does not work unless
the game port was enabled. see Bug 13140. We therefor add a note
to Kconfig.

Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
13 years agowatchdog: gef_wdt: include fs.h
Wolfram Sang & Martyn Welch [Wed, 1 Dec 2010 23:11:16 +0000 (00:11 +0100)]
watchdog: gef_wdt: include fs.h

Add missing include "linux/fs.h".
This fixes compile failure.

Signed-off-by: Wolfram Sang <w.sang@pengutronix.de>
Signed-off-by: Martyn Welch <martyn.welch@ge.com>
Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
13 years agowatchdog: bcm63xx_wdt: improve platform part.
Wim Van Sebroeck [Sat, 23 Oct 2010 20:59:42 +0000 (20:59 +0000)]
watchdog: bcm63xx_wdt: improve platform part.

* fix devinit and devexit sections
* fix platform removal code so that the iounmap happens after the removal of the timer.
* changes the reboot_notifier by a platform shutdown method.

Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
13 years agowatchdog: iTCO_wdt: TCO Watchdog patch for Intel Patsburg PCH
Seth Heasley [Wed, 17 Nov 2010 19:15:08 +0000 (12:15 -0700)]
watchdog: iTCO_wdt: TCO Watchdog patch for Intel Patsburg PCH

This patch adds an additional LPC Controller DeviceID for the Intel Patsburg PCH for TCO Watchdog.

Signed-off-by: Seth Heasley <seth.heasley@intel.com>
Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
13 years agoASoC: omap: N810: Don't select CONFIG_OMAP_MUX but make it as dependency
Jarkko Nikula [Wed, 1 Dec 2010 09:01:20 +0000 (11:01 +0200)]
ASoC: omap: N810: Don't select CONFIG_OMAP_MUX but make it as dependency

Not all omap boards use kernel based pin multiplexing so
CONFIG_SND_OMAP_SOC_N810 should not select it by default as it can make
harm to other boards in multi-board kernels.

Therefore put CONFIG_OMAP_MUX as a dependency to N810 ASoC machine driver.

Thanks to Tony Lindgren <tony@atomide.com> for noticing.

Signed-off-by: Jarkko Nikula <jhnikula@gmail.com>
Cc: Tony Lindgren <tony@atomide.com>
Acked-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Liam Girdwood <lrg@slimlogic.co.uk>
13 years agoHID: length resolution should be reported units/mm
Dmitry Torokhov [Thu, 2 Dec 2010 10:49:06 +0000 (11:49 +0100)]
HID: length resolution should be reported units/mm

Input ABI requires reporting resolution on main axes in units per
millimeter, not units per inch, so we need to convert accordingly.

Tested-by: Nikolai Kondrashov <spbnick@gmail.com>
Acked-by: Nikolai Kondrashov <spbnick@gmail.com>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
13 years agoALSA: hda: Use "alienware" model quirk for another SSID
Daniel T Chen [Thu, 2 Dec 2010 00:16:07 +0000 (19:16 -0500)]
ALSA: hda: Use "alienware" model quirk for another SSID

BugLink: https://launchpad.net/bugs/683695
The original reporter states that headphone jacks do not appear to
work.  Upon inspecting his codec dump, and upon further testing, it is
confirmed that the "alienware" model quirk is correct.

Reported-and-tested-by: Cody Thierauf
Cc: <stable@kernel.org> [2.6.32+]
Signed-off-by: Daniel T Chen <crimsun@ubuntu.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
13 years agoMerge branches 'misc', 'mlx4' and 'nes' into for-next
Roland Dreier [Thu, 2 Dec 2010 00:33:47 +0000 (16:33 -0800)]
Merge branches 'misc', 'mlx4' and 'nes' into for-next

13 years agoIB: Fix information leak in marshalling code
Vasiliy Kulikov [Sun, 14 Nov 2010 09:22:52 +0000 (09:22 +0000)]
IB: Fix information leak in marshalling code

ib_ucm_init_qp_attr() and ucma_init_qp_attr() pass struct ib_uverbs_qp_attr
with reserved, qp_state, {ah_attr,alt_ah_attr}{reserved,->grh.reserved}
fields uninitialized to copy_to_user().  This leads to leaking of
contents of kernel stack memory to userspace.

Signed-off-by: Vasiliy Kulikov <segoon@openwall.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
13 years agoIB/pack: Remove some unused code added by the IBoE patches
Or Gerlitz [Mon, 29 Nov 2010 07:20:50 +0000 (07:20 +0000)]
IB/pack: Remove some unused code added by the IBoE patches

Remove unused functions added by commit ff7f5aab354d ("IB/pack: IBoE UD
packet packing support").

Signed-off-by: Or Gerlitz <ogerlitz@voltaire.com>
13 years agoIB/mlx4: Fix IBoE link state
Eli Cohen [Thu, 11 Nov 2010 21:05:58 +0000 (21:05 +0000)]
IB/mlx4: Fix IBoE link state

Use netif_running() and netif_carrier_ok() to report link state,
exactly as is done to report Ethernet link state in sysfs.

Signed-off-by: Eli Cohen <eli@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
13 years agoIB/mlx4: Fix IBoE reported link rate
Eli Cohen [Mon, 15 Nov 2010 14:04:39 +0000 (14:04 +0000)]
IB/mlx4: Fix IBoE reported link rate

The link rate is the product of the link speed in the link width. For
Etherent ports the rate is 10G, so we use 1 for the width and 4 for
speed to get the correct rate.

Signed-off-by: Eli Cohen <eli@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
13 years agomlx4_core: Workaround firmware bug in query dev cap
Eli Cohen [Wed, 10 Nov 2010 12:52:37 +0000 (12:52 +0000)]
mlx4_core: Workaround firmware bug in query dev cap

ConnectX firmware is supposed to report the number blue flame
registers per page as log2 of the value.  However, due to a firmware
bug, it reports actual number.  This patch works around this by
checking if the number of registers calculated fits within a page.  If
it does not, we use 8 registers per page.

Signed-off-by: Eli Cohen <eli@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
13 years agorbd: replace the rbd sysfs interface
Yehuda Sadeh [Fri, 19 Nov 2010 22:51:04 +0000 (14:51 -0800)]
rbd: replace the rbd sysfs interface

The new interface creates directories per mapped image
and under each it creates a subdir per available snapshot.
This allows keeping a cleaner interface within the sysfs
guidelines. The ABI documentation was updated too.

Acked-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net>
Signed-off-by: Sage Weil <sage@newdream.net>
13 years agoIB/mlx4: Fix memory ordering of VLAN insertion control bits
Eli Cohen [Wed, 17 Nov 2010 13:39:43 +0000 (13:39 +0000)]
IB/mlx4: Fix memory ordering of VLAN insertion control bits

We must fully update the control segment before marking it as valid,
so that hardware doesn't start executing it before we're ready.

Signed-off-by: Eli Cohen <eli@mellanox.co.il>
[ Move VLAN control bit setting to before wmb().  - Roland ]

Signed-off-by: Roland Dreier <rolandd@cisco.com>
13 years agoMAINTAINERS: Update NetEffect entry
Chien Tung [Tue, 2 Nov 2010 16:29:54 +0000 (16:29 +0000)]
MAINTAINERS: Update NetEffect entry

Correct web link as www.neteffect.com is no longer valid.  Remove
Chien Tung as maintainer.  I am moving on to other responsibilities at
Intel.  Thanks for all the fish.

Signed-off-by: Chien Tung <chien.tin.tung@intel.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
13 years agoxfs: only run xfs_error_test if error injection is active
Dave Chinner [Tue, 30 Nov 2010 04:15:31 +0000 (15:15 +1100)]
xfs: only run xfs_error_test if error injection is active

Recent tests writing lots of small files showed the flusher thread
being CPU bound and taking a long time to do allocations on a debug
kernel. perf showed this as the prime reason:

             samples  pcnt function                    DSO
             _______ _____ ___________________________ _________________

           224648.00 36.8% xfs_error_test              [kernel.kallsyms]
            86045.00 14.1% xfs_btree_check_sblock      [kernel.kallsyms]
            39778.00  6.5% prandom32                   [kernel.kallsyms]
            37436.00  6.1% xfs_btree_increment         [kernel.kallsyms]
            29278.00  4.8% xfs_btree_get_rec           [kernel.kallsyms]
            27717.00  4.5% random32                    [kernel.kallsyms]

Walking btree blocks during allocation checking them requires each
block (a cache hit, so no I/O) call xfs_error_test(), which then
does a random32() call as the first operation.  IOWs, ~50% of the
CPU is being consumed just testing whether we need to inject an
error, even though error injection is not active.

Kill this overhead when error injection is not active by adding a
global counter of active error traps and only calling into
xfs_error_test when fault injection is active.

Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
13 years agoxfs: avoid moving stale inodes in the AIL
Dave Chinner [Tue, 30 Nov 2010 04:15:46 +0000 (15:15 +1100)]
xfs: avoid moving stale inodes in the AIL

When an inode has been marked stale because the cluster is being
freed, we don't want to (re-)insert this inode into the AIL. There
is a race condition where the cluster buffer may be unpinned before
the inode is inserted into the AIL during transaction committed
processing. If the buffer is unpinned before the inode item has been
committed and inserted, then it is possible for the buffer to be
released and hence processthe stale inode callbacks before the inode
is inserted into the AIL.

In this case, we then insert a clean, stale inode into the AIL which
will never get removed by an IO completion. It will, however, get
reclaimed and that triggers an assert in xfs_inode_free()
complaining about freeing an inode still in the AIL.

This race can be avoided by not moving stale inodes forward in the AIL
during transaction commit completion processing. This closes the
race condition by ensuring we never insert clean stale inodes into
the AIL. It is safe to do this because a dirty stale inode, by
definition, must already be in the AIL.

Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
13 years agoxfs: delayed alloc blocks beyond EOF are valid after writeback
Dave Chinner [Tue, 30 Nov 2010 04:16:02 +0000 (15:16 +1100)]
xfs: delayed alloc blocks beyond EOF are valid after writeback

There is an assumption in the parts of XFS that flushing a dirty
file will make all the delayed allocation blocks disappear from an
inode. That is, that after calling xfs_flush_pages() then
ip->i_delayed_blks will be zero.

This is an invalid assumption as we may have specualtive
preallocation beyond EOF and they are recorded in
ip->i_delayed_blks. A flush of the dirty pages of an inode will not
change the state of these blocks beyond EOF, so a non-zero
deeelalloc block count after a flush is valid.

The bmap code has an invalid ASSERT() that needs to be removed, and
the swapext code has a bug in that while it swaps the data forks
around, it fails to swap the i_delayed_blks counter associated with
the fork and hence can get the block accounting wrong.

Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
13 years agoxfs: push stale, pinned buffers on trylock failures
Dave Chinner [Tue, 30 Nov 2010 04:16:16 +0000 (15:16 +1100)]
xfs: push stale, pinned buffers on trylock failures

As reported by Nick Piggin, XFS is suffering from long pauses under
highly concurrent workloads when hosted on ramdisks. The problem is
that an inode buffer is stuck in the pinned state in memory and as a
result either the inode buffer or one of the inodes within the
buffer is stopping the tail of the log from being moved forward.

The system remains in this state until a periodic log force issued
by xfssyncd causes the buffer to be unpinned. The main problem is
that these are stale buffers, and are hence held locked until the
transaction/checkpoint that marked them state has been committed to
disk. When the filesystem gets into this state, only the xfssyncd
can cause the async transactions to be committed to disk and hence
unpin the inode buffer.

This problem was encountered when scaling the busy extent list, but
only the blocking lock interface was fixed to solve the problem.
Extend the same fix to the buffer trylock operations - if we fail to
lock a pinned, stale buffer, then force the log immediately so that
when the next attempt to lock it comes around, it will have been
unpinned.

Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
13 years agoxfs: fix failed write truncation handling.
Dave Chinner [Tue, 30 Nov 2010 04:14:39 +0000 (15:14 +1100)]
xfs: fix failed write truncation handling.

Since the move to the new truncate sequence we call xfs_setattr to
truncate down excessively instanciated blocks.  As shown by the testcase
in kernel.org BZ #22452 that doesn't work too well.  Due to the confusion
of the internal inode size, and the VFS inode i_size it zeroes data that
it shouldn't.

But full blown truncate seems like overkill here.  We only instanciate
delayed allocations in the write path, and given that we never released
the iolock we can't have converted them to real allocations yet either.

The only nasty case is pre-existing preallocation which we need to skip.
We already do this for page discard during writeback, so make the delayed
allocation block punching a generic function and call it from the failed
write path as well as xfs_aops_discard_page. The callers are
responsible for ensuring that partial blocks are not truncated away,
and that they hold the ilock.

Based on a fix originally from Christoph Hellwig. This version used
filesystem blocks as the range unit.

Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
13 years agoMerge branch 'kconfig_fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab...
Michal Marek [Wed, 1 Dec 2010 12:13:56 +0000 (13:13 +0100)]
Merge branch 'kconfig_fixes' of git://git./linux/kernel/git/mchehab/linux-next into kbuild/rc-fixes

13 years agoinitramfs: Really fix build break on symbol-prefixed archs
Hendrik Brueckner [Mon, 29 Nov 2010 22:38:50 +0000 (14:38 -0800)]
initramfs: Really fix build break on symbol-prefixed archs

Define the __initramfs_size variable using VMLINUX_SYMBOL() to take care
of symbol-prefixed architectures, for example, blackfin.

Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Cc: Mike Frysinger <vapier.adi@gmail.com>
Cc: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>,
Cc: Michal Marek <mmarek@suse.cz>
Cc: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
[mmarek: leave out Makefile change, since d63f6d1 already takes care of the
 SYMBOL_PREFIX define]
Signed-off-by: Michal Marek <mmarek@suse.cz>
13 years agoASoC: WM8731: Fix incorrect mask for bypass path disable
Dimitris Papastamos [Wed, 1 Dec 2010 09:38:55 +0000 (09:38 +0000)]
ASoC: WM8731: Fix incorrect mask for bypass path disable

According to the datasheet the bypass path enable/disable is
bit 3 therefore we need 0x8 and not 0x4.

Signed-off-by: Dimitris Papastamos <dp@opensource.wolfsonmicro.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
13 years agolxfb: Maintain video processor palette through suspend/resume
Daniel Drake [Tue, 30 Nov 2010 20:34:52 +0000 (20:34 +0000)]
lxfb: Maintain video processor palette through suspend/resume

The Geode X driver uses both of the LX's palettes, one for gamma
correction and one for colormaps.

The kernel driver currently only backs up the one used for colormaps
during suspend/resume. If you mess with gamma settings and do a
suspend/resume, colors go funny.

Fix this by backing up the video proc palette during suspend/resume,
alongside the display controller one which is already handled.

Signed-off-by: Daniel Drake <dsd@laptop.org>
Acked-by: Andres Salomon <dilinger@queued.net>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
13 years agovideo: da8xx: Register IRQ as last thing in driver probing.
Caglar Akyuz [Tue, 30 Nov 2010 20:04:14 +0000 (20:04 +0000)]
video: da8xx: Register IRQ as last thing in driver probing.

Following commit exposed a bug in driver:

"fbdev: da8xx/omap-l1xx: implement double buffering"

Bug is, if interrupt handler is called before initialization is
finished, raster controller is enabled and following register
modifications causes hardware to stay in a broken state.

By looking at this one may say that proper locking is missing in
this driver, and a more proper fix should be prepared. However,
aformentioned commit causes a regression in the driver and some
fix to current one should be applied first.

Signed-off-by: Caglar Akyuz <caglar@bilkon-kontrol.com.tr>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
13 years agoMerge branch 'sh/cachetlb' into sh-fixes-for-linus
Paul Mundt [Wed, 1 Dec 2010 07:39:08 +0000 (16:39 +0900)]
Merge branch 'sh/cachetlb' into sh-fixes-for-linus

13 years agosh: Assume new page cache pages have dirty dcache lines.
Paul Mundt [Wed, 1 Dec 2010 06:39:51 +0000 (15:39 +0900)]
sh: Assume new page cache pages have dirty dcache lines.

This follows the ARM change c01778001a4f5ad9c62d882776235f3f31922fdd
("ARM: 6379/1: Assume new page cache pages have dirty D-cache") for the
same rationale:

    There are places in Linux where writes to newly allocated page
    cache pages happen without a subsequent call to flush_dcache_page()
    (several PIO drivers including USB HCD). This patch changes the
    meaning of PG_arch_1 to be PG_dcache_clean and always flush the
    D-cache for a newly mapped page in update_mmu_cache().

This addresses issues seen with executing binaries from MMC, in
addition to some of the other HCDs that don't explicitly do cache
management for their pipe-in buffers.

Requested-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
13 years agoMerge branch 'drm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied...
Linus Torvalds [Wed, 1 Dec 2010 04:13:35 +0000 (20:13 -0800)]
Merge branch 'drm-fixes' of git://git./linux/kernel/git/airlied/drm-2.6

* 'drm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6:
  drm/radeon/kms: add workaround for dce3 ddc line vbios bug
  drm/radeon/kms: fix interlaced and doublescan handling
  drm/radeon/kms: fix typos in disabled vbios code
  Revert "drm/i915/dp: use VBT provided eDP params if available"
  drm/i915: Clear pfit registers when not used by any outputs
  drm: record monitor status in output_poll_execute
  drm: Set connector DPMS status to ON in drm_crtc_helper_set_config
  drm/i915: fix regression due to ba3d8d749b01548b9
  Revert "drm/radeon/kms: fix typo in r600 cs checker"
  drm/i915/sdvo: Always add a 30ms delay to make SDVO TV detection reliable
  MAINTAINERS: INTEL DRM DRIVERS list (intel-gfx) is subscribers-only
  drm/i915/sdvo: Always fallback to querying the shared DDC line
  drm/i915: Handle pagefaults in execbuffer user relocations
  drm/i915/sdvo: Only enable HDMI encodings only if the commandset is supported
  drm/radeon/kms: fix resume regression for some r5xx laptops
  drm/radeon/kms: fix regression in rs4xx i2c setup
  drm/i915: Only save/restore cursor regs if !KMS
  drm/i915: Prevent integer overflow when validating the execbuffer

13 years agodrm/radeon/kms: add workaround for dce3 ddc line vbios bug
Alex Deucher [Tue, 30 Nov 2010 05:15:10 +0000 (00:15 -0500)]
drm/radeon/kms: add workaround for dce3 ddc line vbios bug

fixes:
https://bugzilla.kernel.org/show_bug.cgi?id=23752

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Cc:stable@kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm/radeon/kms: fix interlaced and doublescan handling
Alex Deucher [Tue, 30 Nov 2010 20:46:47 +0000 (15:46 -0500)]
drm/radeon/kms: fix interlaced and doublescan handling

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Cc: stable@kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm/radeon/kms: fix typos in disabled vbios code
Alex Deucher [Wed, 1 Dec 2010 00:11:45 +0000 (19:11 -0500)]
drm/radeon/kms: fix typos in disabled vbios code

6xx/7xx was hitting the wrong BUS_CNTL reg and bits.

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Cc: stable@kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agoMerge remote branch 'intel/drm-intel-fixes' of /ssd/git/drm-next into drm-fixes
Dave Airlie [Wed, 1 Dec 2010 02:10:34 +0000 (12:10 +1000)]
Merge remote branch 'intel/drm-intel-fixes' of /ssd/git/drm-next into drm-fixes

* 'intel/drm-intel-fixes' of /ssd/git/drm-next:
  Revert "drm/i915/dp: use VBT provided eDP params if available"
  drm/i915: Clear pfit registers when not used by any outputs
  drm/i915: fix regression due to ba3d8d749b01548b9