USB: Fix PS3 USB shutdown problems

Add ehci_shutdown() or ohci_shutdown() calls to the USB
PS3 bus glue.  ehci_shutdown() and ohci_shutdown() do some
controller specific cleanups not done by usb_remove_hcd().

Fixes errors on shutdown or reboot similar to these:

  ps3-ehci-driver sb_07: HC died; cleaning up
  irq 51: nobody cared (try booting with the "irqpoll" option)

Related bugzilla reports:

  http://bugzilla.kernel.org/show_bug.cgi?id=11819
  http://bugzilla.terrasoftsolutions.com/show_bug.cgi?id=317

Signed-off-by: Geoff Levand <geoffrey.levand@am.sony.com>
Cc: stable <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

USB: unusual_devs entry for Argosy USB mass-storage interface

This patch (as1162) adds an unusual_devs entry for Argosy's USB-IDE
interface.  This fixes Bugzilla #11843.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Tested-by: Luciano Rocha <luciano@eurotux.com>
Cc: stable <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

USB: cdc-acm.c: fix recursive lock in acm_start_wb error path

Fixes an obvious bug in cdc-acm by avoiding a recursive lock on
acm_start_wb()'s error path. Should apply towards 2.6.27 stable and
2.6.28.

=============================================
[ INFO: possible recursive locking detected ]
2.6.27-2-pae #109
---------------------------------------------
python/31449 is trying to acquire lock:
 (&acm->write_lock){++..}, at: [<f89a0348>] acm_start_wb+0x5c/0x7b [cdc_acm]

but task is already holding lock:
 (&acm->write_lock){++..}, at: [<f89a04fb>] acm_tty_write+0xe1/0x167 [cdc_acm]

other info that might help us debug this:
2 locks held by python/31449:
 #0:  (&tty->atomic_write_lock){--..}, at: [<c0260fae>] tty_write_lock+0x14/0x3b
 #1:  (&acm->write_lock){++..}, at: [<f89a04fb>] acm_tty_write+0xe1/0x167 [cdc_acm]

stack backtrace:
Pid: 31449, comm: python Not tainted 2.6.27-2-pae #109
 [<c030f42f>] ? printk+0xf/0x18
 [<c0149f33>] __lock_acquire+0xc7b/0x1316
 [<c014a63e>] lock_acquire+0x70/0x97
 [<f89a0348>] ? acm_start_wb+0x5c/0x7b [cdc_acm]
 [<c0312109>] _spin_lock_irqsave+0x37/0x47
 [<f89a0348>] ? acm_start_wb+0x5c/0x7b [cdc_acm]
 [<f89a0348>] acm_start_wb+0x5c/0x7b [cdc_acm]
 [<f89a055d>] acm_tty_write+0x143/0x167 [cdc_acm]
 [<c0262a98>] write_chan+0x1cd/0x297
 [<c012527e>] ? default_wake_function+0x0/0xd
 [<c026111e>] tty_write+0x149/0x1b9
 [<c02628cb>] ? write_chan+0x0/0x297
 [<c01912c5>] ? rw_verify_area+0x76/0x98
 [<c0260fd5>] ? tty_write+0x0/0x1b9
 [<c01919ba>] vfs_write+0x8c/0x136
 [<c0191afd>] sys_write+0x3b/0x60
 [<c0103beb>] sysenter_do_call+0x12/0x3f
 =======================

Signed-off-by: Brandon Philips <bphilips@suse.de>
Cc: Oliver Neukum <oliver@neukum.org>
Cc: stable <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

USB: CP2101 Add device ID for AMB2560

This patch adds the device vendor and product IDs for Amber Wireless AMB2560

Signed-off-by: Craig Shelley <craig@microtron.org.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

USB: mention URB_FREE_BUFFER in usb_free_urb documentation

The usb_free_urb comment says that the transfer buffer will not be
freed, but this is not the case when URB_FREE_BUFFER is set.

Signed-off-by: Rabin Vincent <rabin@rab.in>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

USB: Add YISO u893 usb modem vendor and product IDs to option driver

This patch adds YISO u893 usb modem vendor and product ID to option.c.

I had a better experience using this modification and the same system.

Signed-off-by: Leslie Harlley Watter <leslie@watter.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

usb: musb: fix BULK request on different available endpoints

Fixes co-working issue of usb serial device with usb/net devices while
oter endpoints are free and can be used.This patch implements the policy
that if endpoint resources are available then different BULK request goes
to different endpoint otherwise they are multiplexed to one reserved
endpoint as currently done.

Switch statement case is reordered in musb_giveback() to take care of
bulk request both in multiplex scenario and otherwise.

NAK limit scheme has to be added for multiplexed BULK request scenario
to avoid endpoint starvation due to usb/net devices.

Signed-off-by: Ajay Kumar Gupta <ajay.gupta@ti.com>
Signed-off-by: Felipe Balbi <felipe.balbi@nokia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

usb: musb: fix debug global variable name

In order to avoid namespace conflicts, add a prefix
to our kernel-wise symbol.

Signed-off-by: Felipe Balbi <felipe.balbi@nokia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

usb: musb: Removes compilation warning in gadget mode

Fixes compilation warning when musb is configured in gadget mode.

Signed-off-by: Ajay Kumar Gupta <ajay.gupta@ti.com>
Signed-off-by: Felipe Balbi <felipe.balbi@nokia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

usb: musb: tusb6010: kill compile warning

Add an errno to failing case.

Signed-off-by: Felipe Balbi <felipe.balbi@nokia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

usb: musb: Fix for isochronous IN transfer

Fixes blurred capture images in dma mode. Isochronous error field in
urb and source data buffer pointer were not updated properly in dma
mode.

Signed-off-by: Ajay Kumar Gupta <ajay.gupta@ti.com>
Signed-off-by: Felipe Balbi <felipe.balbi@nokia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

USB: don't register endpoints for interfaces that are going away

This patch (as1155) fixes a bug in usbcore.  When interfaces are
deleted, either because the device was disconnected or because of a
configuration change, the extra attribute files and child endpoint
devices may get left behind.  This is because the core removes them
before calling device_del().  But during device_del(), after the
driver is unbound the core will reinstall altsetting 0 and recreate
those extra attributes and children.

The patch prevents this by adding a flag to record when the interface
is in the midst of being unregistered.  When the flag is set, the
attribute files and child devices will not be created.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Cc: stable <stable@kernel.org> [2.6.27, 2.6.26, 2.6.25]
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

USB: add Nikon D300 camera to unusual_devs

Addresses http://bugzilla.kernel.org/show_bug.cgi?id=11685

When A Nikon D300 camera is connected to a system it is seen in
/proc/bus/pci/devices but is not accessible.

This is seen in the above file:

T:  Bus=01 Lev=01 Prnt=01 Port=05 Cnt=03 Dev#= 11 Spd=480 MxCh= 0
D:  Ver= 2.00 Cls=00(>ifc ) Sub=00 Prot=00 MxPS=64 #Cfgs=  1
P:  Vendor=04b0 ProdID=041a Rev= 1.03
S:  Manufacturer=NIKON
S:  Product=NIKON DSC D300
S:  SerialNumber=000008014379
C:* #Ifs= 1 Cfg#= 1 Atr=c0 MxPwr=  2mA
I:* If#= 0 Alt= 0 #EPs= 3 Cls=06(still) Sub=01 Prot=01 Driver=usbfs
E:  Ad=81(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=02(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=83(I) Atr=03(Int.) MxPS=   8 Ivl=32ms

Cc: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

USB: vstusb: fix compiler warning on x86-64

This fixes a reported compiler warning.

Reported-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Merge branch 'for-linus' of git://git./linux/kernel/git/jbarnes/pci-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jbarnes/pci-2.6:
  PCI: ignore bit0 of _OSC return code

Merge branch 'for-linus' of git://git./linux/kernel/git/teigland/dlm

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/teigland/dlm:
  dlm: fix shutdown cleanup

Merge branch 'for-linus' of git://git./linux/kernel/git/penberg/slab-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/penberg/slab-2.6:
  slab: document SLAB_DESTROY_BY_RCU
  Kconfig: SLUB is the default slab allocator

Merge master.kernel.org:/home/rmk/linux-2.6-arm

* master.kernel.org:/home/rmk/linux-2.6-arm:
  [ARM] cdb89712: avoid namespace clashes with SRAM_ and BOOTROM_ constants
  [ARM] cdb89712,clps7500,h720x: avoid namespace clash for FLASH_* constants
  [ARM] integrator,realview,versatile: remove FLASH_* and EPROM_* constants
  [ARM] dma-mapping: fix compiler warning
  [ARM] iop: iop3xx needs registers mapped uncached+unbuffered
  [ARM] versatile: correct MMC clock rate
  [ARM] realview: correct MMC clock rate
  [ARM] 5329/1: Feroceon: fix feroceon_l2_inv_range

fix cpumask build breakage on parisc

Commit 2d3854a37e8b767a51aba38ed6d22817b0631e33 ("cpumask: introduce new
API, without changing anything") introduced a build breakage on parisc.

This trivial patch fixes it.

Signed-off-by: Helge Deller <deller@gmx.de>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Kyle Mc Martin <kyle@hera.kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

dlm: fix shutdown cleanup

Fixes a regression from commit 0f8e0d9a317406612700426fad3efab0b7bbc467,
"dlm: allow multiple lockspace creates".

An extraneous 'else' slipped into a code fragment being moved from
release_lockspace() to dlm_release_lockspace().  The result of the
unwanted 'else' is that dlm threads and structures are not stopped
and cleaned up when the final dlm lockspace is removed.  Trying to
create a new lockspace again afterward will fail with
"kmem_cache_create: duplicate cache dlm_conn" because the cache
was not previously destroyed.

Signed-off-by: David Teigland <teigland@redhat.com>

slab: document SLAB_DESTROY_BY_RCU

Explain this SLAB_DESTROY_BY_RCU thing...

[hugh@veritas.com: add a pointer to comment in mm/slab.c]
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Jens Axboe <jens.axboe@oracle.com>
Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Acked-by: Christoph Lameter <cl@linux-foundation.org>
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Pekka Enberg <penberg@cs.helsinki.fi>

Kconfig: SLUB is the default slab allocator

In 2007, a0acd820807680d2ccc4ef3448387fcdbf152c73 changed the default
slab allocator to SLUB, but the SLAB help text still says SLAB is the
default. This change fixes that.

Signed-off-by: Simon Arlott <simon@fire.lp0.eu>
Signed-off-by: Pekka Enberg <penberg@cs.helsinki.fi>

[ARM] cdb89712: avoid namespace clashes with SRAM_ and BOOTROM_ constants

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

[ARM] cdb89712,clps7500,h720x: avoid namespace clash for FLASH_* constants

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

[ARM] integrator,realview,versatile: remove FLASH_* and EPROM_* constants

FLASH_* and EPROM_* constants are unused, and clash with drivers:

drivers/atm/ambassador.h:257:1: warning: "FLASH_BASE" redefined
drivers/atm/ambassador.h:258:1: warning: "FLASH_SIZE" redefined
drivers/atm/iphase.h:332:1: warning: "EPROM_SIZE" redefined

so remove them.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

[ARM] dma-mapping: fix compiler warning

arch/arm/mm/dma-mapping.c: In function `dma_sync_sg_for_cpu':
arch/arm/mm/dma-mapping.c:588: warning: statement with no effect

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

Merge branch 'sched-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'sched-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  sched: fix init_idle()'s use of sched_clock()
  sched: fix stale value in average load per task

c2port: add c2port support for Eurotech Duramar 2150

Signed-off-by: Rodolfo Giometti <giometti@linux.it>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Add c2 port support

C2port implements a two wire serial communication protocol (bit
banging) designed to enable in-system programming, debugging, and
boundary-scan testing on low pin-count Silicon Labs devices.

Currently this code supports only flash programming through sysfs
interface but extensions shoud be easy to add.

Signed-off-by: Rodolfo Giometti <giometti@linux.it>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

hdq: documentation for OMAP HDQ

Add a brief document about omap2430/3430 HDQ/1-wire driver.

Signed-off-by: Madhusudhan Chikkature<madhu.cr@ti.com>
Acked-by: Evgeniy Polyakov <johnpol@2ka.mipt.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

hdq: bQ27000 HDQ Slave Interface Driver

Provide the BQ27000 slave interface driver.

Signed-off-by: Madhusudhan Chikkature<madhu.cr@ti.com>
Acked-by: Evgeniy Polyakov <johnpol@2ka.mipt.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

hdq driver for OMAP2430/3430

The HDQ/1-Wire module of TI OMAP2430/3430 platforms implement the hardware
protocol of the master functions of the Benchmark HDQ and the Dallas
Semiconductor 1-Wire protocols.  These protocols use a single wire for
communication between the master (HDQ/1-Wire controller) and the slave
(HDQ/1-Wire external compliant device).

This patch provides the HDQ driver to suppport TI OMAP2430/3430 platforms.

Signed-off-by: Madhusudhan Chikkature<madhu.cr@ti.com>
Acked-by: Felipe Balbi <felipe.balbi@nokia.com>
Acked-by: Evgeniy Polyakov <johnpol@2ka.mipt.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

w1: export w1_read_8 function

Export the w1_read_8 function for use of drivers.  The OMAP HDQ
driver(drivers/w1/masters/omap_hdq.c) uses this function.

[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Madhusudhan Chikkature<madhu.cr@ti.com>
Acked-by: Evgeniy Polyakov <johnpol@2ka.mipt.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

rtc: add Dallas DS1390/93/94 RTC chips

Add support for the Dallas DS1390/93/94 SPI RTC chip.

Signed-off-by: Mark Jackson <mpfj@mimc.co.uk>
Acked-by: Alessandro Zummo <a.zummo@towertech.it>
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>

rtc: basic implementation of Epson RX-8581 I2C Real Time Clock

Provide the basic "get" and "set" functionality for the Epson RX-8581 I2C
RTC.  It currently does not support the RTC's Alarm or Fixed-cycle timer.

[akpm@linux-foundation.org: need log2.h for ilog2(), remove unneeded initialisation]
Signed-off-by: Martyn Welch <martyn.welch@gefanuc.com>
Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Cc: David Brownell <david-b@pacbell.net>
Cc: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

rtc: rtc-wm8350: add support for WM8350 RTC

This adds support for the RTC provided by the Wolfson Microelectronics
WM8350.

This driver was originally written by Graeme Gregory and Liam Girdwood,
though it has been modified since then to update it to current mainline
coding standards and for API completeness.

[akpm@linux-foundation.org: s/schedule_timeout_interruptible/schedule_timeout_uninterruptible/ to prevent bogus timeout when signal_pending()]
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Cc: Alessandro Zummo <a.zummo@towertech.it>
Cc: David Brownell <david-b@pacbell.net>
Cc: Liam Girdwood <linux@wolfsonmicro.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

ics932s401: new clock generator chip driver

The ics932s401 is a clock generator chip.  This driver allows users to
read the current clock outputs.

Signed-off-by: Darrick J. Wong <djwong@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

adt7462: new hwmon driver

New driver to play with.  As Jean mentioned a couple of years ago, this
chip is a beast with odd combinations of 8 fans, 4 temperatures, and 13
voltage sensors.  This driver has been tested on an IntelliStation Z30.

Signed-off-by: Darrick J. Wong <djwong@us.ibm.com>
Cc: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

LIS3LV02Dx Accelerometer driver

This adds a driver to the accelerometer sensor found in several HP
laptops (under the commercial names of "HP Mobile Data Protection System
3D" and "HP 3D driveguard").  It tries to have more or less the same
interfaces as the hdaps and other accelerometer drivers: in sysfs and as
a joystick.

This driver was first written by Yan Burman.  Eric Piel has updated it
and slimed it up (including the removal of an interface to access to the
free-fall feature of the sensor because it is not reliable enough for
now).  Pavel Machek removed few more features and switched locking from
semaphore to mutex.

Several people have contributed to the database of the axes.

[eric.piel@tremplin-utc.net: LIS3LV02D: Conform to the new ACPI API]
Signed-off-by: Eric Piel <eric.piel@tremplin-utc.net>
Signed-off-by: Yan Burman <burman.yan@gmail.com>
Signed-off-by: Pavel Machek <pavel@suse.cz>
Cc: "Mark M. Hoffman" <mhoffman@lightlink.com>
Signed-off-by: Eric Piel <eric.piel@tremplin-utc.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

memcg: bugfix for memory hotplug

The start pfn calculation in page_cgroup's memory hotplug notifier chain
is wrong.

Tested-by: Badari Pulavarty <pbadari@us.ibm.com>
Signed-off-by: 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>

ext3: Clean up outdated and incorrect comment for ext3_write_super()

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Cc: <linux-ext4@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

atmel_spi: work-around required for new HW bug in AT91SAM9263 Rev.B SPI controller

We're working with an AT91SAM9263 Rev B in our design and I experienced
some inconsistency in spi-based touchscreen usage between our board and
the Atmel evaluation kit we have that runs on a Rev A chip.

The data was apparently delayed by 1 byte and got ridiculous data out of
the touchscreen driver, very strange.  As everything looked normal in
the spi, touchscreen and dma logs, I contacted the Atmel support and
they triggered me on a new HW bug that appeared in the Rev B SPI
controller.

The problem is that the SPI controller on the Rev B needs that the
software reset is performed two times so that it's performed correctly.

Applying the patch below solves the issue on my Rev B board.  I've tested
it as well on my Rev A evaluation kit and it has apparently no unwanted
side effect, things continue to work as expected.

Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com>
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>

lm85: support adt7468 chips

The adt7468 is a follow-on to the adt7463, so plumb in adt7468 support
along the same code paths.

Signed-off-by: Darrick J. Wong <djwong@us.ibm.com>
Cc: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

adt7473: check inputs from sysfs writes

Implement correct range checking for adt7470 to prevent userland from
writing impossible values into the chip, and cap out-of-range values per
standard hwmon conventions.

Implement correct rounding of input values per standard hwmon conventions.

Signed-off-by: Darrick J. Wong <djwong@us.ibm.com>
Cc: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

adt7470: check input range when sysfs files are written

Implement correct range checking for adt7470 to prevent userland from
writing impossible values into the chip, and cap out-of-range values per
standard hwmon conventions.

Implement correct rounding of input values per standard hwmon conventions.

Signed-off-by: Darrick J. Wong <djwong@us.ibm.com>
Cc: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

kernel/kprobes.c: don't pad kretprobe_table_locks[] on uniprocessor builds

We only need the cacheline padding on SMP kernels.  Saves 6k:

   text    data     bss     dec     hex filename
   5713     388    8840   14941    3a5d kernel/kprobes.o
   5713     388    2632    8733    221d kernel/kprobes.o

Acked-by: Masami Hiramatsu <mhiramat@redhat.com>
Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

DOC: update xip method info

xip documentation updated:
- change "get_xip_page" to "get_xip_mem";
- explain changed function parameters

Signed-off-by: Marco Stornelli <marco.stornelli@gmail.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Cc: Carsten Otte <cotte@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

DOC: add printk-formats.txt

Add printk-formats.txt so that we don't have to keep fixing the
same things over and over again.  <wishful thinking>

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

remove ratelimt()

It mistakenly assumes that a static local in an inlined function is a
kernel-wide singleton.  It also has no callers, so let's remove it.

Cc: Dave Young <hidave.darkstar@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

vfs: fix shrink_submounts

In the last refactoring of shrink_submounts a variable was not completely
renamed.  So finish the renaming of mnt to m now.

Without this if you attempt to mount an nfs mount that has both automatic
nfs sub mounts on it, and has normal mounts on it.  The unmount will
succeed when it should not.

Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Al Viro <viro@ZenIV.linux.org.uk
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm: remove lru_add_drain_all() from the munlock path

lockdep warns about following message at boot time on one of my test
machine.  Then, schedule_on_each_cpu() sholdn't be called when the task
have mmap_sem.

Actually, lru_add_drain_all() exist to prevent the unevictalble pages
stay on reclaimable lru list.  but currenct unevictable code can rescue
unevictable pages although it stay on reclaimable list.

So removing is better.

In addition, this patch add lru_add_drain_all() to sys_mlock() and
sys_mlockall().  it isn't must.  but it reduce the failure of moving to
unevictable list.  its failure can rescue in vmscan later.  but reducing
is better.

Note, if above rescuing happend, the Mlocked and the Unevictable field
mismatching happend in /proc/meminfo.  but it doesn't cause any real
trouble.

=======================================================
[ INFO: possible circular locking dependency detected ]
2.6.28-rc2-mm1 #2
-------------------------------------------------------
lvm/1103 is trying to acquire lock:
 (&cpu_hotplug.lock){--..}, at: [<c0130789>] get_online_cpus+0x29/0x50

but task is already holding lock:
 (&mm->mmap_sem){----}, at: [<c01878ae>] sys_mlockall+0x4e/0xb0

which lock already depends on the new lock.

the existing dependency chain (in reverse order) is:

-> #3 (&mm->mmap_sem){----}:
       [<c0153da2>] check_noncircular+0x82/0x110
       [<c0185e6a>] might_fault+0x4a/0xa0
       [<c0156161>] validate_chain+0xb11/0x1070
       [<c0185e6a>] might_fault+0x4a/0xa0
       [<c0156923>] __lock_acquire+0x263/0xa10
       [<c015714c>] lock_acquire+0x7c/0xb0 (*) grab mmap_sem
       [<c0185e6a>] might_fault+0x4a/0xa0
       [<c0185e9b>] might_fault+0x7b/0xa0
       [<c0185e6a>] might_fault+0x4a/0xa0
       [<c0294dd0>] copy_to_user+0x30/0x60
       [<c01ae3ec>] filldir+0x7c/0xd0
       [<c01e3a6a>] sysfs_readdir+0x11a/0x1f0 (*) grab sysfs_mutex
       [<c01ae370>] filldir+0x0/0xd0
       [<c01ae370>] filldir+0x0/0xd0
       [<c01ae4c6>] vfs_readdir+0x86/0xa0 (*) grab i_mutex
       [<c01ae75b>] sys_getdents+0x6b/0xc0
       [<c010355a>] syscall_call+0x7/0xb
       [<ffffffff>] 0xffffffff

-> #2 (sysfs_mutex){--..}:
       [<c0153da2>] check_noncircular+0x82/0x110
       [<c01e3d2c>] sysfs_addrm_start+0x2c/0xc0
       [<c0156161>] validate_chain+0xb11/0x1070
       [<c01e3d2c>] sysfs_addrm_start+0x2c/0xc0
       [<c0156923>] __lock_acquire+0x263/0xa10
       [<c015714c>] lock_acquire+0x7c/0xb0 (*) grab sysfs_mutex
       [<c01e3d2c>] sysfs_addrm_start+0x2c/0xc0
       [<c04f8b55>] mutex_lock_nested+0xa5/0x2f0
       [<c01e3d2c>] sysfs_addrm_start+0x2c/0xc0
       [<c01e3d2c>] sysfs_addrm_start+0x2c/0xc0
       [<c01e3d2c>] sysfs_addrm_start+0x2c/0xc0
       [<c01e422f>] create_dir+0x3f/0x90
       [<c01e42a9>] sysfs_create_dir+0x29/0x50
       [<c04faaf5>] _spin_unlock+0x25/0x40
       [<c028f21d>] kobject_add_internal+0xcd/0x1a0
       [<c028f37a>] kobject_set_name_vargs+0x3a/0x50
       [<c028f41d>] kobject_init_and_add+0x2d/0x40
       [<c019d4d2>] sysfs_slab_add+0xd2/0x180
       [<c019d580>] sysfs_add_func+0x0/0x70
       [<c019d5dc>] sysfs_add_func+0x5c/0x70 (*) grab slub_lock
       [<c01400f2>] run_workqueue+0x172/0x200
       [<c014008f>] run_workqueue+0x10f/0x200
       [<c0140bd0>] worker_thread+0x0/0xf0
       [<c0140c6c>] worker_thread+0x9c/0xf0
       [<c0143c80>] autoremove_wake_function+0x0/0x50
       [<c0140bd0>] worker_thread+0x0/0xf0
       [<c0143972>] kthread+0x42/0x70
       [<c0143930>] kthread+0x0/0x70
       [<c01042db>] kernel_thread_helper+0x7/0x1c
       [<ffffffff>] 0xffffffff

-> #1 (slub_lock){----}:
       [<c0153d2d>] check_noncircular+0xd/0x110
       [<c04f650f>] slab_cpuup_callback+0x11f/0x1d0
       [<c0156161>] validate_chain+0xb11/0x1070
       [<c04f650f>] slab_cpuup_callback+0x11f/0x1d0
       [<c015433d>] mark_lock+0x35d/0xd00
       [<c0156923>] __lock_acquire+0x263/0xa10
       [<c015714c>] lock_acquire+0x7c/0xb0
       [<c04f650f>] slab_cpuup_callback+0x11f/0x1d0
       [<c04f93a3>] down_read+0x43/0x80
       [<c04f650f>] slab_cpuup_callback+0x11f/0x1d0 (*) grab slub_lock
       [<c04f650f>] slab_cpuup_callback+0x11f/0x1d0
       [<c04fd9ac>] notifier_call_chain+0x3c/0x70
       [<c04f5454>] _cpu_up+0x84/0x110
       [<c04f552b>] cpu_up+0x4b/0x70 (*) grab cpu_hotplug.lock
       [<c06d1530>] kernel_init+0x0/0x170
       [<c06d15e5>] kernel_init+0xb5/0x170
       [<c06d1530>] kernel_init+0x0/0x170
       [<c01042db>] kernel_thread_helper+0x7/0x1c
       [<ffffffff>] 0xffffffff

-> #0 (&cpu_hotplug.lock){--..}:
       [<c0155bff>] validate_chain+0x5af/0x1070
       [<c040f7e0>] dev_status+0x0/0x50
       [<c0156923>] __lock_acquire+0x263/0xa10
       [<c015714c>] lock_acquire+0x7c/0xb0
       [<c0130789>] get_online_cpus+0x29/0x50
       [<c04f8b55>] mutex_lock_nested+0xa5/0x2f0
       [<c0130789>] get_online_cpus+0x29/0x50
       [<c0130789>] get_online_cpus+0x29/0x50
       [<c017bc30>] lru_add_drain_per_cpu+0x0/0x10
       [<c0130789>] get_online_cpus+0x29/0x50 (*) grab cpu_hotplug.lock
       [<c0140cf2>] schedule_on_each_cpu+0x32/0xe0
       [<c0187095>] __mlock_vma_pages_range+0x85/0x2c0
       [<c0156945>] __lock_acquire+0x285/0xa10
       [<c0188f09>] vma_merge+0xa9/0x1d0
       [<c0187450>] mlock_fixup+0x180/0x200
       [<c0187548>] do_mlockall+0x78/0x90 (*) grab mmap_sem
       [<c01878e1>] sys_mlockall+0x81/0xb0
       [<c010355a>] syscall_call+0x7/0xb
       [<ffffffff>] 0xffffffff

other info that might help us debug this:

1 lock held by lvm/1103:
 #0:  (&mm->mmap_sem){----}, at: [<c01878ae>] sys_mlockall+0x4e/0xb0

stack backtrace:
Pid: 1103, comm: lvm Not tainted 2.6.28-rc2-mm1 #2
Call Trace:
 [<c01555fc>] print_circular_bug_tail+0x7c/0xd0
 [<c0155bff>] validate_chain+0x5af/0x1070
 [<c040f7e0>] dev_status+0x0/0x50
 [<c0156923>] __lock_acquire+0x263/0xa10
 [<c015714c>] lock_acquire+0x7c/0xb0
 [<c0130789>] get_online_cpus+0x29/0x50
 [<c04f8b55>] mutex_lock_nested+0xa5/0x2f0
 [<c0130789>] get_online_cpus+0x29/0x50
 [<c0130789>] get_online_cpus+0x29/0x50
 [<c017bc30>] lru_add_drain_per_cpu+0x0/0x10
 [<c0130789>] get_online_cpus+0x29/0x50
 [<c0140cf2>] schedule_on_each_cpu+0x32/0xe0
 [<c0187095>] __mlock_vma_pages_range+0x85/0x2c0
 [<c0156945>] __lock_acquire+0x285/0xa10
 [<c0188f09>] vma_merge+0xa9/0x1d0
 [<c0187450>] mlock_fixup+0x180/0x200
 [<c0187548>] do_mlockall+0x78/0x90
 [<c01878e1>] sys_mlockall+0x81/0xb0
 [<c010355a>] syscall_call+0x7/0xb

Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Tested-by: Kamalesh Babulal <kamalesh@linux.vnet.ibm.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
Cc: Christoph Lameter <cl@linux-foundation.org>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Cc: Hugh Dickins <hugh@veritas.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

kprobes: disable preempt for module_text_address() and kernel_text_address()

__register_kprobe() can be preempted after checking probing address but
before module_text_address() or try_module_get(), and in this interval
the module can be unloaded.  In that case, try_module_get(probed_mod)
will access to invalid address, or kprobe will probe invalid address.

This patch uses preempt_disable() to protect it and uses
__module_text_address() and __kernel_text_address().

Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Signed-off-by: Masami Hiramatsu <mhiramat@redhat.com>
Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com>
Cc: Hiroshi Shimamoto <h-shimamoto@ct.jp.nec.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

freezer_cg: disable writing freezer.state of root cgroup

With this change, control file 'freezer.state' doesn't exist in root
cgroup, making root cgroup unfreezable.

I think it's reasonable to disallow freeze tasks in the root cgroup.  And
then we can avoid fork overhead when freezer subsystem is compiled but not
used.

Also make writing invalid value to freezer.state returns EINVAL rather
than EIO.  This is more consistent with other cgroup subsystem.

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Acked-by: Paul Menage <menage@google.com>
Cc: Cedric Le Goater <clg@fr.ibm.com>
Cc: Paul Menage <menage@google.com>
Cc: Matt Helsley <matthltc@us.ibm.com>
Cc: "Serge E. Hallyn" <serue@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

freezer_cg: remove task_lock from freezer_fork()

In theory the task can be moved to another cgroup and the freezer will be
freed right after task_lock is dropped, so the lock results in zero
protection.

But in the case of freezer_fork() no lock is needed, since the task is not
in tasklist yet so it won't be moved to another cgroup, so task->cgroups
won't be changed or invalidated.

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Cc: Matt Helsley <matthltc@us.ibm.com>
Cc: Cedric Le Goater <clg@fr.ibm.com>
Cc: "Serge E. Hallyn" <serue@us.ibm.com>
Cc: Paul Menage <menage@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

GPIO: add new Xilinx driver for powerpc

This driver supports the Xilinx XPS GPIO IP core which has the typical
GPIO features.

Signed-off-by: Kiran Sutariya <kirans@xilinx.com>
Signed-off-by: John Linn <john.linn@xilinx.com>
Cc: David Brownell <david-b@pacbell.net>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Kumar Gala <galak@gate.crashing.org>
Cc: "Grant Likely" <grant.likely@secretlab.ca>
Cc: <avorontsov@ru.mvista.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

cpusets: update mems allowed in page allocator

If all allowable memory is unreclaimable, it is possible to loop forever
in the page allocator for ~__GFP_NORETRY allocations.

During this time, it is also possible for a task's cpuset to expand its
set of allowable nodes so that it now includes free memory.  The cached
copy of this set, current->mems_allowed, is stale, however, since there
has not been a subsequent call to cpuset_update_task_memory_state().

The cached copy of the set of allowable nodes is now updated in the page
allocator's slow path so the additional memory is available to
get_page_from_freelist().

[akpm@linux-foundation.org: add comment]
Signed-off-by: David Rientjes <rientjes@google.com>
Cc: Paul Menage <menage@google.com>
Cc: Christoph Lameter <cl@linux-foundation.org>
Cc: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

i5000-edac: hold reference to mci kobject

It turns out that edac_mc_del_mc will kobject_put the last kref on the
mci object.

If the timing is just right, that means that the mci object is freed
before before i5000_remove_one has a chance to free the resources
associated with it, causing a null pointer exceptions when unloading the
driver.  Insert a kobject_{get,put} pair so that this doesn't happen.

Signed-off-by: Darrick J. Wong <djwong@us.ibm.com>
Cc: Doug Thompson <norsk5@yahoo.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Fix platform drivers that crash on suspend/resume

It turns out that if one registers a struct platform_device, the
platform device code expects that platform_device.device->driver points
to a struct driver inside a struct platform_driver.

This is not the case with the ipmi-si, ipmi-msghandler and ibmaem
drivers, which causes the suspend/resume hook functions to jump off into
nowhere, causing a crash.  Make this assumption hold true for these
three drivers.

Signed-off-by: Darrick J. Wong <djwong@us.ibm.com>
Acked-by: Corey Minyard <cminyard@mvista.com>
Cc: Jean Delvare <khali@linux-fr.org>
Cc: Kay Sievers <kay.sievers@vrfy.org>
Cc: Greg KH <greg@kroah.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

atmel_lcdfb: change irq_base definition to allow error reporting

Changed because old the definition of unsigned long cannot be negative.

Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Reported-by: Roel Kluin <roel.kluin@gmail.com>
Cc: Haavard Skinnemoen <hskinnemoen@atmel.com>
Cc: Andrew Victor <linux@maxim.org.za>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

hwmon: applesmc: add support for Macbook 4

This patch adds accelerometer and temperature sensor support for Macbook 4.

Signed-off-by: Henrik Rydberg <rydberg@euromail.se>
Cc: Nicolas Boichat <nicolas@boichat.ch
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

hugetlb: make unmap_ref_private multi-size-aware

Oops.  Part of the hugetlb private reservation code was not fully
converted to use hstates.

When a huge page must be unmapped from VMAs due to a failed COW,
HPAGE_SIZE is used in the call to unmap_hugepage_range() regardless of
the page size being used.  This works if the VMA is using the default
huge page size.  Otherwise we might unmap too much, too little, or
trigger a BUG_ON.  Rare but serious -- fix it.

Signed-off-by: Adam Litke <agl@us.ibm.com>
Cc: Jon Tollefson <kniht@linux.vnet.ibm.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Merge git://git./linux/kernel/git/davem/net-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6:
  bnx2: fix poll_controller to pass proper structures and check all rx queues
  niu: Fix readq implementation when architecture does not provide one.
  hostap: pad the skb->cb usage in lieu of a proper fix
  rtl8187 : support for Sitecom WL-168 0001 v4
  mac80211: fix notify_mac function
  rtl8187: Add Abocom USB ID
  net: put_cmsg_compat + SO_TIMESTAMP[NS]: use same name for value as caller
  tcp_htcp: last_cong bug fix
  [netdrvr] smc911x: fix for driver resume (and compilation warning)
  RDMA/cxgb3: deadlock in iw_cxgb3 can cause hang when configuring interface.
  cxgb3 - Limit multiqueue setting to msi-x
  cxgb3 - eeprom read fixes
  myri10ge: fix stop/go ordering even more

bnx2: fix poll_controller to pass proper structures and check all rx queues

Fix bnx2 so that netpoll works properly.  Specifically:

1) Fix parameters to bnx2_interrupt to be a struct bnx2_napi rather than a
struct net_device

2) Fix poll_controller method to check every queue in the rx case so frames
aren't missed

Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge branch 'master' of git://git./linux/kernel/git/linville/wireless-2.6

niu: Fix readq implementation when architecture does not provide one.

This fixes a TX hang reported by Jesper Dangaard Brouer.

When an architecutre cannot provide a fully functional
64-bit atomic readq/writeq, the driver must implement
it's own.  This is because only the driver can say whether
doing something like using two 32-bit reads to implement
the full 64-bit read will actually work properly.

In particular one of the issues is whether the top 32-bits
or the bottom 32-bits of the 64-bit register should be read
first.  There could be side effects, and in fact that is
exactly the problem here.

The TX_CS register has counters in the upper 32-bits and
state bits in the lower 32-bits.  A read clears the state
bits.

We would read the counter half before the state bit half.
That first read would clear the state bits, and then the
driver thinks that no interrupts are pending because the
interrupt indication state bits are seen clear every time.

Fix this by reading the bottom half before the upper half.

Tested-by: Jesper Dangaard Brouer <jdb@comx.dk>
Signed-off-by: David S. Miller <davem@davemloft.net>

hostap: pad the skb->cb usage in lieu of a proper fix

Like mac80211 did, this driver makes 'clever' use of skb->cb to pass
information along with an skb as it is requeued from the virtual device
to the physical wireless device.  Unfortunately, that trick no longer
works...

Unlike mac80211, code complexity and driver apathy makes this hack
the best option we have in the short run.  Hopefully someone will
eventually be motivated to code a proper fix before all the effected
hardware dies.

(Above text by me.  Johannes officially disavows all knowledge of this
hack. -- JWL)

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Cc: stable@kernel.org
Signed-off-by: John W. Linville <linville@tuxdriver.com>

rtl8187 : support for Sitecom WL-168 0001 v4

the Sitecom 0001 v4 with product id 0x0df6:0028, uses Realtek's
RTL8187B and work fine with new 2.6.27 driver.

Signed-off-by: John W. Linville <linville@tuxdriver.com>

mac80211: fix notify_mac function

The ieee80211_notify_mac() function uses ieee80211_sta_req_auth() which
in turn calls ieee80211_set_disassoc() which calls a few functions that
need to be able to sleep, so ieee80211_notify_mac() cannot use RCU
locking for the interface list and must use rtnl locking instead.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

rtl8187: Add Abocom USB ID

Signed-off-by: Ivan Kuten <ivan.kuten@promwad.com>
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

Merge branch 'timers-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'timers-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  hrtimer: clean up unused callback modes

Merge branch 'fixes' of git://git./linux/kernel/git/djbw/async_tx

* 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/djbw/async_tx:
  dmaengine: struct device - replace bus_id with dev_name(), dev_set_name()
  iop-adma: use iop_paranoia() for debug BUG_ONs
  iop-adma: add a dummy read to flush next descriptor update

sched: fix init_idle()'s use of sched_clock()

Maciej Rutecki reported:

> I have this bug during suspend to disk:
>
> [  188.592151] Enabling non-boot CPUs ...
> [  188.592151] SMP alternatives: switching to SMP code
> [  188.666058] BUG: using smp_processor_id() in preemptible
> [00000000]
> code: suspend_to_disk/2934
> [  188.666064] caller is native_sched_clock+0x2b/0x80

Which, as noted by Linus, was caused by me, via:

  7cbaef9c "sched: optimize sched_clock() a bit"

Move the rq locking a bit earlier in the initialization sequence,
that will make the sched_clock() call in init_idle() non-preemptible.

Reported-by: Maciej Rutecki <maciej.rutecki@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

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:
  IPoIB: Fix crash in path_rec_completion()
  IPoIB: Fix hang in ipoib_flush_paths()
  IPoIB: Don't enable NAPI when it's already enabled
  RDMA/cxgb3: Fix deadlock in iw_cxgb3 (hang when configuring interface)
  IB/ehca: Remove reference to special QP in case of port activation failure
  IB/mlx4: Set umem field to NULL in mlx4_ib_alloc_fast_reg_mr()
  mlx4_core: Fix unused variable warning
  RDMA/nes: Mitigate compatibility issue regarding PCIe write credits
  RDMA/nes: Fix CQ allocation scheme for multicast receive queue apps
  RDMA/nes: Correct handling of PBL resources
  RDMA/nes: Reindent mis-indented spinlocks
  RDMA/cxgb3: Fix too-big reserved field zeroing in iwch_post_zb_read()
  IB/ipath: Fix RDMA write with immediate copy of last packet

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:
  ALSA: hda - Fix IDT/STAC multiple HP detection
  ALSA: hda - Fix input pin initialization for STAC/IDT codecs
  ALSA: hda - Add missing analog-mux mixer creation for STAC9200

MN10300: Don't do misalignment handling for userspace

Don't do misalignment handling for userspace misalignment faults: just
generate an appropriate SIGBUS instead.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

MN10300: Don't handle misaligned loading and storing of SP

Don't handle the misaligned loading and storing of the SP register as in C code
that's most certainly a compiler bug.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

MN10300: Handle misaligned SP-based operands

Support misalignment handling for instructions that have kernel SP-based
address operands, including fixing those that include IMM8 or IMM16
displacements.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

MN10300: Fix misaligned index-register addressing handling

Fix misalignment handling for an address calculated from the sum of two
registers.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

MN10300: Handle misaligned postinc-with-imm addressing mode correctly

Correctly handle misalignment in MOV instructions with postinc-with-immediate
addressing mode operands.  In these, the immediate value is the increment to
be applied the address register, not the displacement to the address.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

MN10300: Perform misalignment fixups of MOV_Lcc

Perform misalignment fixups of the MOV_Lcc instructions (move postinc memory
to register and conditionally loop).

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

MN10300: Allow misalignment fixup in interrupt handling

Allow misalignment fixup in interrupt handling in the MN10300 arch.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

MN10300: Fix register-postinc addressing misalignment handling

Fix misalignment handling of operands with register postincrement addressing.
The flag to indicate that postincrement is required should not be interpreted
as an specification of a value to be added to the address.

Also add BUGs to catch unimplemented parameter markings in the opcodes table.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

MN10300: Extract the displacement from an insn correctly in misalignment fixup

Extract the displacement from an MN10300 instruction correctly in the
misalignment fixup handler.

The code should extract the displacement in LSB order, not MSB order.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

MN10300: Add further misalignment fixups

Add further misalignment fixup support to the MN10300 arch, notably for ABS32
and SP+disp addressing.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

MN10300: Add built-in testing for misalignment handler

Add configurable built-in testing for the MN10300 misalignment handler.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

MN10300: Clean up the misalignment handler a little

Clean up the MN10300 misalignment handler a little by:

 (1) Use ilog2() rather than doing implementing log2() locally.

 (2) Make format_tbl[] const and static.

 (3) Making the debugging prints more consistent.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Merge branch 'kvm-updates/2.6.28' of git://git./linux/kernel/git/avi/kvm

* 'kvm-updates/2.6.28' of git://git.kernel.org/pub/scm/linux/kernel/git/avi/kvm:
  KVM: Fix pit memory leak if unable to allocate irq source id
  KVM: ia64: fix vmm_spin_{un}lock for !CONFIG_SMP
  KVM: VMX: Set IGMT bit in EPT entry
  KVM: Require the PCI subsystem
  x86: KVM guest: fix section mismatch warning in kvmclock.c
  KVM: ia64: Use guest signal mask when blocking
  KVM: MMU: increase per-vcpu rmap cache alloc size

parisc: fix find_extend_vma() breakage

The STACK_GROWSUP case of stack expansion was missing a test for 'prev',
which got removed by commit cb8f488c33539f096580e202f5438a809195008f
("mmap.c: deinline a few functions") by mistake.

I found my original email in "sent" folder. The patch in that mail
does NOT remove !prev. That change had beed added by someone else.

Ok, I think we are not much interested in who did it, let's
fix it for good.

[ "It looks like this was caused by me fixing rejects.  That was the
  fancy include-lots-of-context-so-it-wont-apply patch." - akpm ]

Reported-and-bisected-by: Helge Deller <deller@gmx.de>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Jiri Kosina <jkosina@suse.cz>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Merge branch 'release' of git://git./linux/kernel/git/lenb/linux-acpi-2.6

* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux-acpi-2.6: (47 commits)
  ACPI: pci_link: remove acpi_irq_balance_set() interface
  fujitsu-laptop: Add DMI callback for Lifebook S6420
  ACPI: EC: Don't do transaction from GPE handler in poll mode.
  ACPI: EC: lower interrupt storm treshold
  ACPICA: Use spinlock for acpi_{en|dis}able_gpe
  ACPI: EC: restart failed command
  ACPI: EC: wait for last write gpe
  ACPI: EC: make kernel messages more useful when GPE storm is detected
  ACPI: EC: revert msleep patch
  thinkpad_acpi: fingers off backlight if video.ko is serving this functionality
  sony-laptop: fingers off backlight if video.ko is serving this functionality
  msi-laptop: fingers off backlight if video.ko is serving this functionality
  fujitsu-laptop: fingers off backlight if video.ko is serving this functionality
  eeepc-laptop: fingers off backlight if video.ko is serving this functionality
  compal: fingers off backlight if video.ko is serving this functionality
  asus-acpi: fingers off backlight if video.ko is serving this functionality
  Acer-WMI: fingers off backlight if video.ko is serving this functionality
  ACPI video: if no ACPI backlight support, use vendor drivers
  ACPI: video: Ignore devices that aren't present in hardware
  Delete an unwanted return statement at evgpe.c
  ...

Merge branches 'cxgb3', 'ehca', 'ipath', 'ipoib', 'mlx4' and 'nes' into for-next

IPoIB: Fix crash in path_rec_completion()

Fix a crash in path_rec_completion() during an SM up/down loop.  If
more than one path record request is issued, the first completion
releases path->done, allowing ipoib_flush_paths() to free the path,
and thus corrupting it for the second completion.

Commit ee1e2c82 ("IPoIB: Refresh paths instead of flushing them on SM
change events") added the field path->valid and changed the test "if
(!path)" to "if (!path || !path->valid)".  This change made it
possible for a path with an outstanding query to pass the test and
issue another query on the same path.  Having two queries on the same
path leads to a crash.

This fixes <https://bugs.openfabrics.org/show_bug.cgi?id=1325>.

Signed-off-by: Yossi Etigin <yosefe@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>

IPoIB: Fix hang in ipoib_flush_paths()

ipoib_flush_paths() can hang during an SM up/down loop: if
path_rec_start() fails (for instance, because there is no sm_ah), the
path is still added to the path list by neigh_add_path().  Then,
ipoib_flush_paths() will wait for path->done, but it will never
complete because the request was not issued at all.  Fix this by
completing path->done if issuing the query fails.

This fixes <https://bugs.openfabrics.org/show_bug.cgi?id=1329>.

Signed-off-by: Yossi Etigin <yosefe@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>

IPoIB: Don't enable NAPI when it's already enabled

If a P_Key is not present when an interface is created, ipoib_open()
will return after doing napi_enable().  ipoib_open() will be called
again from ipoib_pkey_poll() when the P_Key appears, after NAPI has
already been enabled, and try to enable it again. This triggers a
BUG_ON() in napi_enable().

Fix this by moving the call to napi_enable() to after the test for
P_Key presence.

Signed-off-by: Yossi Etigin <yosefe@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>

RDMA/cxgb3: Fix deadlock in iw_cxgb3 (hang when configuring interface)

When the iw_cxgb3 module's cxgb3_client "add" func gets called by the
cxgb3 module, the iwarp driver ends up calling the ethtool ops
get_drvinfo function in cxgb3 to get the fw version and other info.
Currently the iwarp driver grabs the rtnl lock around this down call
to serialize.  As of 2.6.27 or so, things changed such that the rtnl
lock is held around the call to the netdev driver open function.  Also
the cxgb3_client "add" function doesn't get called if the device is
down.

So, if you load cxgb3, then load iw_cxgb3, then ifconfig up the
device, the iw_cxgb3 add func gets called with the rtnl_lock held.  If
you load cxgb3, ifconfig up the device, then load iw_cxgb3, the add
func gets called without the rtnl_lock held.  The former causes the
deadlock, the latter does not.

In addition, there are iw_cxgb3 sysfs handlers that also can call down
into cxgb3 to gather the fw and hw versions.  These can be called
concurrently on different processors and at any time.  Thus we need to
push this serialization down in the cxgb3 driver get_drvinfo func.

The fix is to remove rtnl lock usage, and use a per-device lock in cxgb3.

Signed-off-by: Steve Wise <swise@opengridcomputing.com>
Acked-by: Divy Le Ray <divy@chelsio.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>

sched: fix stale value in average load per task

Impact: fix load balancer load average calculation accuracy

cpu_avg_load_per_task() returns a stale value when nr_running is 0.
It returns an older stale (caculated when nr_running was non zero) value.

This patch returns and sets rq->avg_load_per_task to zero when nr_running
is 0.

Compile and boot tested on a x86_64 box.

Signed-off-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

net: put_cmsg_compat + SO_TIMESTAMP[NS]: use same name for value as caller

In __sock_recv_timestamp() the additional SCM_TIMESTAMP[NS] is used. This
has the same value as SO_TIMESTAMP[NS], so this is a purely cosmetic change.

Signed-off-by: Patrick Ohly <patrick.ohly@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

tcp_htcp: last_cong bug fix

This patch fixes a minor bug in tcp_htcp.c which has been
highlighted by Lachlan Andrew and Lawrence Stewart.  Currently, the
time since the last congestion event, which is stored in variable
last_cong, is reset whenever there is a state change into
TCP_CA_Open.  This includes transitions of the type
TCP_CA_Open->TCP_CA_Disorder->TCP_CA_Open which are not associated
with backoff of cwnd.  The patch changes last_cong to be updated
only on transitions into TCP_CA_Open that occur after experiencing
the congestion-related states TCP_CA_Loss, TCP_CA_Recovery,
TCP_CA_CWR.

Signed-off-by: Doug Leith <doug.leith@nuim.ie>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge branch 'topic/fix/hda' into for-linus

ALSA: hda - Fix IDT/STAC multiple HP detection

Due to the recent change for multiple HP as line-out switch, only
one of the multiple headphons (usually a wrong one) is toggled
and the other pins are still disabled.  This causes the silent output
problem on some Dell laptops.

Also, the hp_switch check is screwed up when a line-in or a mic-in
jack exists.  This is added as an additional output, but hp_switch
check doesn't take it into account.

This patch fixes these issues: simplify hp_switch check by using
the NID instead of bool, and clean up / fix the toggle of HP pins
in unsol event handler code.

Reference: Novell bnc#443267
https://bugzilla.novell.com/show_bug.cgi?id=443267

Signed-off-by: Takashi Iwai <tiwai@suse.de>

hrtimer: clean up unused callback modes

Impact: cleanup

git grep HRTIMER_CB_IRQSAFE revealed half the callback modes are actually
unused.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>