firefly-linux-kernel-4.4.55.git
13 years agoleds: cpcap: Convert to global non-reentrant workqueue
Colin Cross [Wed, 12 Jan 2011 20:03:05 +0000 (12:03 -0800)]
leds: cpcap: Convert to global non-reentrant workqueue

schedule_work puts work items into a global workqueue that is
not single threaded.  If the work item is running when a
second schedule_work is called, the second one could be
immediately scheduled onto the second cpu while the first
one is running, breaking the synchronization expected in
cpcap_brightness_work.

Fix by converting schedule_work to queue_work(system_nrt_wq),
which puts the work on a global single-threaded workqueue.

Change-Id: Ieba89e0353b86f11350cb37552afbce5abe87088
Signed-off-by: Colin Cross <ccross@android.com>
13 years agoMerge remote branch 'tegra/android-tegra-2.6.36' into android-tegra-moto-2.6.36
Rebecca Schultz Zavin [Wed, 12 Jan 2011 19:39:11 +0000 (11:39 -0800)]
Merge remote branch 'tegra/android-tegra-2.6.36' into android-tegra-moto-2.6.36

13 years agoMerge remote branch 'tegra/linux-tegra-2.6.36' into android-tegra-2.6.36
Rebecca Schultz Zavin [Wed, 12 Jan 2011 19:38:21 +0000 (11:38 -0800)]
Merge remote branch 'tegra/linux-tegra-2.6.36' into android-tegra-2.6.36

13 years agovideo: tegra: nvmap: Don't store task struct for kernel threads
Rebecca Schultz Zavin [Wed, 12 Jan 2011 19:10:34 +0000 (11:10 -0800)]
video: tegra: nvmap: Don't store task struct for kernel threads

No need to maintain a reference to the task struct if the client
is a kernel thread.  In this case just set the task to NULL.

Change-Id: Ica4785388932f6b298eeb0da04b78b0e1cdc3a44
Signed-off-by: Rebecca Schultz Zavin <rebecca@android.com>
13 years agocrypto: tegra-aes: call INIT_COMPLETION before setting the transfer
Varun Wadekar [Tue, 11 Jan 2011 08:33:54 +0000 (14:03 +0530)]
crypto: tegra-aes: call INIT_COMPLETION before setting the transfer

Change-Id: I368a9bf8aa1c311532e54d18a813b556576b8da1
Signed-off-by: Varun Wadekar <vwadekar@nvidia.com>
13 years agocrypto: tegra-aes: do not reset bsev/vde per operation
Varun Wadekar [Fri, 7 Jan 2011 13:02:00 +0000 (18:32 +0530)]
crypto: tegra-aes: do not reset bsev/vde per operation

avp does not use the crypto engine inside the vde block,
so it is ok if we do not reset the entire block before
any crypto operation. this would save more than 100us per crypto
operation.

Change-Id: I4e6f4940c5b6e95498b6f7a62498fbf956e25254
Signed-off-by: Varun Wadekar <vwadekar@nvidia.com>
13 years agoMerge branch 'android-tegra-2.6.36' into android-tegra-moto-2.6.36
Colin Cross [Wed, 12 Jan 2011 08:04:17 +0000 (00:04 -0800)]
Merge branch 'android-tegra-2.6.36' into android-tegra-moto-2.6.36

13 years agoMerge branch 'linux-tegra-2.6.36' into android-tegra-2.6.36
Colin Cross [Wed, 12 Jan 2011 08:04:01 +0000 (00:04 -0800)]
Merge branch 'linux-tegra-2.6.36' into android-tegra-2.6.36

13 years ago[ARM] tegra_i2s_audio: fixed several issues
Eric Laurent [Wed, 12 Jan 2011 03:39:58 +0000 (19:39 -0800)]
[ARM] tegra_i2s_audio: fixed several issues

- Acquire the DMA request spinlock in dma_tx_complete_callback()
(this was omitted)
- allow_suspend() was called by mistake in start_recording_if_necessary()
every time it was executed which could cause the wake lock to be
released during audio capture.

Signed-off-by: Eric Laurent <elaurent@google.com>
Signed-off-by: Iliyan Malchev <malchev@google.com>
13 years ago[ARM] mfd: cpcap-audio: CPCAP CODEC Capture pop
Chris Fries [Tue, 11 Jan 2011 05:01:57 +0000 (23:01 -0600)]
[ARM] mfd: cpcap-audio: CPCAP CODEC Capture pop

There appears to be a warm-up time for the CODEC to stabilize of
25 milliseconds (by measurement).  Avoid noise at the start of
audio capture by holding off the CODEC enable for this duration.

Change-Id: I078af9272875cb77388fbbbd2796c4ba69301cb4
Signed-off-by: Iliyan Malchev <malchev@google.com>
13 years ago[ARM] mfd: cpcap-audio: Deadlock in audio capture after LP0
Chris Fries [Tue, 11 Jan 2011 23:10:01 +0000 (17:10 -0600)]
[ARM] mfd: cpcap-audio: Deadlock in audio capture after LP0

LP0 was clearing out the DAC/DAP configuration registers.
Currently the CPCAP audio driver owns their configuration at
runtime, so fix the port config during system resume.

Change-Id: Ib7f26085d55a9184649b79ec8f647830d135489b
Signed-off-by: Iliyan Malchev <malchev@google.com>
13 years ago[ARM] mfd: cpcap-audio: fixed mic inversion.
Eric Laurent [Mon, 10 Jan 2011 18:48:19 +0000 (10:48 -0800)]
[ARM] mfd: cpcap-audio: fixed mic inversion.

There was an inversion between built-in mic and headset mic
in cpcap_audio_ctl_ioctl() which made that the wrong mic was
activated when exiting standby.

Signed-off-by: Eric Laurent <elaurent@google.com>
13 years agoRevert "ARM: tegra: stingray: Disable panic on softlockup for debugging"
Colin Cross [Wed, 12 Jan 2011 01:48:42 +0000 (17:48 -0800)]
Revert "ARM: tegra: stingray: Disable panic on softlockup for debugging"

This reverts commit 05b22a4f5420263e66c330ad7d58da1c598b5334.

13 years agoARM: tegra: clock: Round frequency up in clock dividers
Colin Cross [Wed, 12 Jan 2011 00:59:55 +0000 (16:59 -0800)]
ARM: tegra: clock: Round frequency up in clock dividers

When picking clock divider values, the clock framework picks
the closest frequency that is lower than the requested
frequency.  If the value from a clock divider rounds down,
and then the new rounded down frequency is requested, it
will get rounded down again, resulting in a frequency two
steps lower than the original requested frequency.

Fix the problem by rounding up when calculating the frequency
coming out of a clock divider, so if that frequency is
requested again, the same divider value will be picked.

Change-Id: Ieaf74448f67d91aeb7ba08226e48c092d8afaa2b
Signed-off-by: Colin Cross <ccross@android.com>
13 years agoARM: tegra: stingray: Add watchdog timer to stingray board file
Rebecca Schultz Zavin [Tue, 11 Jan 2011 23:40:50 +0000 (15:40 -0800)]
ARM: tegra: stingray: Add watchdog timer to stingray board file

Change-Id: Ia20586b42fb95601e45e4fbb9623b8e3b0db6b86
Signed-off-by: Rebecca Schultz Zavin <rebecca@android.com>
13 years agoARM: configs: Add watchdog to stingray defconfig
Rebecca Schultz Zavin [Tue, 11 Jan 2011 23:41:58 +0000 (15:41 -0800)]
ARM: configs: Add watchdog to stingray defconfig

Change-Id: I909c61b9068c5ba53832fff1712e157e04d21163
Signed-off-by: Rebecca Schultz Zavin <rebecca@android.com>
13 years agoMerge remote branch 'tegra/android-tegra-2.6.36' into android-tegra-moto-2.6.36
Rebecca Schultz Zavin [Wed, 12 Jan 2011 00:43:37 +0000 (16:43 -0800)]
Merge remote branch 'tegra/android-tegra-2.6.36' into android-tegra-moto-2.6.36

13 years agoMerge remote branch 'tegra/linux-tegra-2.6.36' into android-tegra-2.6.36
Rebecca Schultz Zavin [Wed, 12 Jan 2011 00:42:30 +0000 (16:42 -0800)]
Merge remote branch 'tegra/linux-tegra-2.6.36' into android-tegra-2.6.36

13 years agovideo: tegra: work around overlay corruption on underflows
Erik Gilling [Wed, 12 Jan 2011 00:32:20 +0000 (16:32 -0800)]
video: tegra: work around overlay corruption on underflows

Overlays can get their internal state corrupted during and underflow
condition.  The only way to fix this state is to reset the DC.  If we get
4 consecutive frames with underflows, assume we're hosed and reset.

Change-Id: Icdf61517837c8570b8de35f585075de08aa35fe7
Signed-off-by: Erik Gilling <konkers@android.com>
Cc: Michael I. Gold <gold@nvidia.com>
13 years agowatchdog: tegra_wdt: Several changes to watchdog driver
Rebecca Schultz Zavin [Tue, 11 Jan 2011 23:35:47 +0000 (15:35 -0800)]
watchdog: tegra_wdt: Several changes to watchdog driver

Add a config option to enable the watchdog driver from probe
Reduce the default timeout to 30 seconds
Read the reset reason and print a messge if the last reboot was
due to watchdog reset

Change-Id: I7500ef236089b06ffcff3d8e7d86a5b9060b59cb
Signed-off-by: Rebecca Schultz Zavin <rebecca@android.com>
13 years agotegra: enable filters in backlight controller
makarand.karvekar [Tue, 11 Jan 2011 22:33:33 +0000 (16:33 -0600)]
tegra: enable filters in backlight controller

updated hysteresis and enabled slope, advanced slope
to smooth brightness transitions.

Signed-off-by: makarand.karvekar <makarand.karvekar@motorola.com>
13 years agoARM: tegra: clock: Save and restore audio clock during suspend
Colin Cross [Tue, 11 Jan 2011 20:15:01 +0000 (12:15 -0800)]
ARM: tegra: clock: Save and restore audio clock during suspend

Change-Id: Id1c7a6c8e8679995ad462f17309c2f3550c00da1
Signed-off-by: Colin Cross <ccross@android.com>
13 years agoMerge branch 'android-tegra-2.6.36' into android-tegra-moto-2.6.36
Colin Cross [Tue, 11 Jan 2011 03:09:49 +0000 (19:09 -0800)]
Merge branch 'android-tegra-2.6.36' into android-tegra-moto-2.6.36

13 years agoMerge branch 'linux-tegra-2.6.36' into android-tegra-2.6.36
Colin Cross [Tue, 11 Jan 2011 03:09:22 +0000 (19:09 -0800)]
Merge branch 'linux-tegra-2.6.36' into android-tegra-2.6.36

13 years agoARM: tegra: clock: Save and restore plld, plls, and pllu in suspend
Colin Cross [Tue, 11 Jan 2011 02:39:43 +0000 (18:39 -0800)]
ARM: tegra: clock: Save and restore plld, plls, and pllu in suspend

HDMI was not working after LP0 because the plld were being reset
during suspend.  plls and pllu were also not being saved.  Add
all three to tegra_clk_suspend.

The lock time for plld is 1000 us, so increase the delay after
setting the PLLs.

Add a BUG_ON to ensure the size of the suspend context area is
correct.

Originally fixed by Mayuresh Kulkarni.

Original-author: Mayuresh Kulkarni <mkulkarni@nvidia.com>
Change-Id: I50a3e994c6e3cab5989aa7a8e26e7a2eb66b6dfb
Signed-off-by: Colin Cross <ccross@android.com>
13 years agoARM: configs: Add NVMAP_CARVEOUT_KILLER to stingray_defconfig
Rebecca Schultz Zavin [Tue, 11 Jan 2011 01:34:51 +0000 (17:34 -0800)]
ARM: configs: Add NVMAP_CARVEOUT_KILLER to stingray_defconfig

Change-Id: I745308dd74d45d10bc11dd1f46a0462fa7984ab1
Signed-off-by: Rebecca Schultz Zavin <rebecca@android.com>
13 years agofirmware: cpcap: Improvements to reset procedures
Greg Meiste [Tue, 11 Jan 2011 00:44:03 +0000 (18:44 -0600)]
firmware: cpcap: Improvements to reset procedures

1) Fix an issue where soft resets would only occur every other
   boot.

2) On a soft reset, put the CPCAP back into hard reset mode.  This
   allows the user to issue a second reset command to get a hard
   reset.  (Of course, when the kernel re-loads, CPCAP will be
   put back into soft reset mode.)

Change-Id: Icc46a14fa795971acc4d6de60e4d2db749ab489c
Signed-off-by: Greg Meiste <w30289@motorola.com>
13 years agoMerge branch 'android-tegra-2.6.36' into android-tegra-moto-2.6.36
Colin Cross [Tue, 11 Jan 2011 01:19:05 +0000 (17:19 -0800)]
Merge branch 'android-tegra-2.6.36' into android-tegra-moto-2.6.36

13 years agoMerge branch 'android-2.6.36' into android-tegra-2.6.36
Colin Cross [Tue, 11 Jan 2011 01:18:52 +0000 (17:18 -0800)]
Merge branch 'android-2.6.36' into android-tegra-2.6.36

13 years agoMerge branch 'linux-tegra-2.6.36' into android-tegra-2.6.36
Colin Cross [Tue, 11 Jan 2011 01:18:42 +0000 (17:18 -0800)]
Merge branch 'linux-tegra-2.6.36' into android-tegra-2.6.36

13 years agocgroup: leave cg_list valid upon cgroup_exit
Simon Wilson [Sat, 8 Jan 2011 02:46:34 +0000 (18:46 -0800)]
cgroup: leave cg_list valid upon cgroup_exit

A thread/process in cgroup_attach_task() could have called
list_del(&tsk->cg_list) after cgroup_exit() had already called
list_del() on the same list. Since it only checked for
!list_empty(&tsk->cg_list) before doing this, the list_del()
call would thus be made twice.

The solution is to leave tsk->cg_list in a valid state in
cgroup_exit() with list_del_init(&tsk->cg_list), which leaves
an empty list.

Change-Id: I4e7c1d0665fced629f5ca033c18dd98afe080e0c
Signed-off-by: Simon Wilson <simonwilson@google.com>
13 years agortc: alarm: Update hrtimer if alarm at the head of the queue is reprogrammed
Arve Hjønnevåg [Sat, 8 Jan 2011 03:00:01 +0000 (19:00 -0800)]
rtc: alarm: Update hrtimer if alarm at the head of the queue is reprogrammed

If an alarm was restarted with a value that moved it away from the head
of a queue, the hrtimer would not be updated. This would cause unnecessary
wakeups.

Change-Id: If379f8dd92b0bdb3173bd8d057adfe0dc1d15259
Signed-off-by: Arve Hjønnevåg <arve@android.com>
13 years ago[ARM] tegra: Add ioctls to tegra2_fuse driver
Greg Meiste [Tue, 4 Jan 2011 20:51:46 +0000 (14:51 -0600)]
[ARM] tegra: Add ioctls to tegra2_fuse driver

Exposing the fuse API to userspace for factory.

Submitted on behalf of Eyob Tesfu <eyob.tesfu@motorola.com>

Change-Id: Ib2a03c49bde7990d89165db24db41f1ea832a5b7
Signed-off-by: Greg Meiste <w30289@motorola.com>
13 years agoARM: tegra: Fix powergate_debugfs_init return value
Benoit Goby [Mon, 10 Jan 2011 22:58:00 +0000 (14:58 -0800)]
ARM: tegra: Fix powergate_debugfs_init return value

Change-Id: Iff865e0f74e1667aa6e998a63d33d3b4cd09694c
Signed-off-by: Benoit Goby <benoit@android.com>
13 years agomdm6600_ctrl: Hold down pwron pin until acknowledged by modem
James Wylder [Mon, 10 Jan 2011 21:23:26 +0000 (15:23 -0600)]
mdm6600_ctrl: Hold down pwron pin until acknowledged by modem

This ensures that the modem powers down correctly and is
part of the fix required to ensure that the BP powers up
correctly after reflash / over the air update.

Change-Id: Ied6a1fe2213127c944547ac2ed86abaf0eb63bd2
Signed-off-by: James Wylder <james.wylder@motorola.com>
13 years agoinpus: touchscreen: qtouch: Fix permissions on device attributes
makarand.karvekar [Fri, 7 Jan 2011 16:55:05 +0000 (10:55 -0600)]
inpus: touchscreen: qtouch: Fix permissions on device attributes

Change-Id: I7aa9004ee42375e0bd7cf4d79cc0ef4580918e70
Signed-off-by: makarand.karvekar <makarand.karvekar@motorola.com>
13 years agovideo: tegra: nvmap: Add carveout killer
Rebecca Schultz Zavin [Fri, 3 Dec 2010 23:43:33 +0000 (15:43 -0800)]
video: tegra: nvmap: Add carveout killer

This change attempts to reclaim carveout memory by killing
other carveout users when an allocation fails.  Processes
are killed in order of priority from lowest to highest, and then
from largest to smallest users.

Change-Id: Iee8a6f36269bc8165d691000a153dbf9f4337775
Signed-off-by: Rebecca Schultz Zavin <rebecca@android.com>
13 years agoMerge branch 'android-tegra-2.6.36' into android-tegra-moto-2.6.36
Colin Cross [Mon, 10 Jan 2011 03:20:26 +0000 (19:20 -0800)]
Merge branch 'android-tegra-2.6.36' into android-tegra-moto-2.6.36

13 years agoMerge branch 'linux-tegra-2.6.36' into android-tegra-2.6.36
Colin Cross [Mon, 10 Jan 2011 03:20:14 +0000 (19:20 -0800)]
Merge branch 'linux-tegra-2.6.36' into android-tegra-2.6.36

13 years agoARM: tegra: pll_a clock fixes
Stephen Warren [Wed, 5 Jan 2011 00:47:55 +0000 (17:47 -0700)]
ARM: tegra: pll_a clock fixes

Increase the max_frequency entries for clocks that can be driven from pll_a
to match the fastest pll_a table entry.

Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Colin Cross <ccross@android.com>
13 years agoARM: tegra: Prevent requeuing in-progress DMA requests
Stephen Warren [Wed, 5 Jan 2011 21:24:12 +0000 (14:24 -0700)]
ARM: tegra: Prevent requeuing in-progress DMA requests

If a request already in the queue is passed to tegra_dma_enqueue_req,
tegra_dma_req.node->{next,prev} will end up pointing to itself instead
of at tegra_dma_channel.list, which is the way a the end-of-list
should be set up. When the DMA request completes and is list_del'd,
the list head will still point at it, yet the node's next/prev will
contain the list poison values. When the next DMA request completes,
a kernel panic will occur when those poison values are dereferenced.

This makes the DMA driver more robust in the face of buggy clients.

Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Colin Cross <ccross@android.com>
13 years agortc: tps6586x: Fix error in RTC tick calculations
Lowell Dennis [Wed, 15 Dec 2010 00:35:17 +0000 (16:35 -0800)]
rtc: tps6586x: Fix error in RTC tick calculations

In the TPS6586x PMU/PMIC RTC support code, when
converting from seconds to ticks using a shift
operator, the most significant bits were being lost
due to seconds being a 32-bit value and ticks being
a 64-bit value.  A hard cast was added to avoid this
loss.

Reviewed-by: Lowell Dennis <ldennis@nvidia.com>
Tested-by: Lowell Dennis <ldennis@nvidia.com>
Reviewed-by: Jonathan Mayo <jmayo@nvidia.com>
Reviewed-by: Peter Zu <pzu@nvidia.com>
Reviewed-by: Bharat Nihalani <bnihalani@nvidia.com>
Signed-off-by: Robert Morell <rmorell@nvidia.com>
Signed-off-by: Colin Cross <ccross@android.com>
13 years agousb: ehci: tegra: continues driving FS-J during resume
Jay Cheng [Fri, 7 Jan 2011 07:03:46 +0000 (02:03 -0500)]
usb: ehci: tegra: continues driving FS-J during resume

To prevent USB glitch.
Also only program PTC bits when resume from LP0

Change-Id: Iced668e33f986828d3a483b411055948b5b257e1
Signed-off-by: Jay Cheng <jacheng@nvidia.com>
13 years agoARM: tegra: usb_phy: continues driving FS-J during resume
Jay Cheng [Fri, 7 Jan 2011 07:03:46 +0000 (02:03 -0500)]
ARM: tegra: usb_phy: continues driving FS-J during resume

to prevent USB glitch.

Change-Id: Iced668e33f986828d3a483b411055948b5b257e1
Signed-off-by: Jay Cheng <jacheng@nvidia.com>
13 years agousb: ehci: tegra: remove reschedule rh_timer
Jay Cheng [Thu, 6 Jan 2011 05:16:18 +0000 (00:16 -0500)]
usb: ehci: tegra: remove reschedule rh_timer

Reschedule rh_timer may cause usb device resume fail, as rh_timer may be
timeout and send USB_REQ_GET_STATUS SETUP control transfer by the time when
the device is handling clear suspend feature, which in turn the device may
drop clear suspend feature request.

Actually on port resume case, the host driver don't need to reschedule
rh_timer to check port status. The host driver will check port status right
after suspend feature is cleared.

Change-Id: I6205e97af49ed4349b6215b851f6b5f1394258d8
Signed-off-by: Jay Cheng <jacheng@nvidia.com>
13 years agoMerge branch 'android-tegra-2.6.36' into android-tegra-moto-2.6.36
Colin Cross [Sat, 8 Jan 2011 22:48:26 +0000 (14:48 -0800)]
Merge branch 'android-tegra-2.6.36' into android-tegra-moto-2.6.36

Conflicts:
drivers/misc/Kconfig
drivers/misc/Makefile

Change-Id: Ic136562830dcd821eaae66f4c1fc66b4e0b825bd

13 years agoMerge branch 'linux-tegra-2.6.36' into android-tegra-2.6.36
Colin Cross [Sat, 8 Jan 2011 22:47:29 +0000 (14:47 -0800)]
Merge branch 'linux-tegra-2.6.36' into android-tegra-2.6.36

Conflicts:
drivers/misc/Kconfig
drivers/misc/Makefile

Change-Id: I826244513f799fdd2fb0269abad3271c96661293

13 years agoARM: tegra: powergate: Allow calling power up when already powered
Colin Cross [Sat, 8 Jan 2011 21:59:15 +0000 (13:59 -0800)]
ARM: tegra: powergate: Allow calling power up when already powered

Allow calling tegra_powergate_sequence_power_up on a partition
that is already powered.  Reset the partition, and return success
with the clock enabled.

Change-Id: I776c6a84091f0bb8faca22d87b3fabf0cfede564
Signed-off-by: Colin Cross <ccross@android.com>
13 years agoNCT1008 temperature sensor driver
Todd Poynor [Mon, 20 Dec 2010 23:05:52 +0000 (15:05 -0800)]
NCT1008 temperature sensor driver

Replace the PM-only driver for NCT1008 with a new version written by
Varun Wadekar and Dmitriy Gruzman.  Add a callback to an alarm
function specified in the board platform data.

Change-Id: Ib429533930ee75af3402d24b0bc286da9f6ee67b
Signed-off-by: Todd Poynor <toddpoynor@google.com>
13 years agomisc: Initial NCT1008 driver
Greg Meiste [Tue, 16 Nov 2010 19:33:06 +0000 (13:33 -0600)]
misc: Initial NCT1008 driver

Initial version of the NCT1008 driver to turn off the sensor when the
device is suspended.  This improves standby current drain.

Change-Id: Ia64613c33c0052434d5e304c434605611e5ef789
Signed-off-by: Greg Meiste <w30289@motorola.com>
13 years agovideo: tegra: fix debug regs file hang
Erik Gilling [Sat, 8 Jan 2011 02:45:43 +0000 (18:45 -0800)]
video: tegra: fix debug regs file hang

Change-Id: I2bdfe5ec9f69b7b7960e59bec76fb9b02b61f3fb
Signed-off-by: Erik Gilling <konkers@android.com>
13 years agoARM: tegra: stingray: Enable NCT1008 temperature monitor
Dmitriy Gruzman [Fri, 17 Dec 2010 05:15:53 +0000 (23:15 -0600)]
ARM: tegra: stingray: Enable NCT1008 temperature monitor

From Varun Wadekar and Dmitriy Gruzman, plus changes to pass
NCT1008 alarm function from platform data.

Change-Id: I4b625e9278a9c214c71ee8859316bb1a9a9bfd91
Signed-off-by: Todd Poynor <toddpoynor@google.com>
13 years agoNCT1008 temperature sensor driver
Todd Poynor [Mon, 20 Dec 2010 23:05:52 +0000 (15:05 -0800)]
NCT1008 temperature sensor driver

Replace the PM-only driver for NCT1008 with a new version written by
Varun Wadekar and Dmitriy Gruzman.  Add a callback to an alarm
function specified in the board platform data.

Change-Id: Ib429533930ee75af3402d24b0bc286da9f6ee67b
Signed-off-by: Todd Poynor <toddpoynor@google.com>
13 years agoARM: tegra: stingray: Enable keyreset driver in defconfig
Colin Cross [Sat, 8 Jan 2011 01:33:43 +0000 (17:33 -0800)]
ARM: tegra: stingray: Enable keyreset driver in defconfig

Change-Id: I2beb427a263ff54d6ae0da2c694241d1c8015cf6
Signed-off-by: Colin Cross <ccross@android.com>
13 years agoARM: tegra: stingray: Log when hard reset buttons are pushed
Colin Cross [Fri, 7 Jan 2011 22:52:26 +0000 (14:52 -0800)]
ARM: tegra: stingray: Log when hard reset buttons are pushed

Change-Id: Ie8b7cccbf616b2e72a01a88afafc06820f9a3efb
Signed-off-by: Colin Cross <ccross@android.com>
13 years agoMerge branch 'android-tegra-2.6.36' into android-tegra-moto-2.6.36
Colin Cross [Sat, 8 Jan 2011 01:21:20 +0000 (17:21 -0800)]
Merge branch 'android-tegra-2.6.36' into android-tegra-moto-2.6.36

13 years agoMerge branch 'linux-tegra-2.6.36' into android-tegra-2.6.36
Colin Cross [Sat, 8 Jan 2011 01:20:53 +0000 (17:20 -0800)]
Merge branch 'linux-tegra-2.6.36' into android-tegra-2.6.36

Conflicts:
arch/arm/mm/cache-v6.S

Change-Id: I1a2063218dd705a762a40f4a9dfe504ce1a1d491

13 years agoMerge branch 'android-2.6.36' into android-tegra-2.6.36
Colin Cross [Sat, 8 Jan 2011 01:16:59 +0000 (17:16 -0800)]
Merge branch 'android-2.6.36' into android-tegra-2.6.36

13 years agoMerge commit 'v2.6.36.3' into linux-tegra-2.6.36
Colin Cross [Sat, 8 Jan 2011 01:14:48 +0000 (17:14 -0800)]
Merge commit 'v2.6.36.3' into linux-tegra-2.6.36

13 years agocrytpo: tegra-aes: make aes_handle_req aynschronous
Varun Wadekar [Tue, 28 Dec 2010 09:31:18 +0000 (15:01 +0530)]
crytpo: tegra-aes: make aes_handle_req aynschronous

the encrypt/decrypt callbacks have to return with -EINPROGRESS
error code and the request complete callback needs to be
called from handle_req for aynchronous block ciphers. use
work queue to make the driver asynchronous.

Change-Id: I0dec1185c31e5de7ba039c39d6bd87c8b3487b2a
Signed-off-by: Varun Wadekar <vwadekar@nvidia.com>
13 years agoinput: keyreset: Allow reset function to be overriden
Colin Cross [Fri, 7 Jan 2011 22:52:56 +0000 (14:52 -0800)]
input: keyreset: Allow reset function to be overriden

Change-Id: Ibb3dda05772b2e89d7b2646689944d309cb1f74e
Signed-off-by: Colin Cross <ccross@android.com>
13 years agoLinux 2.6.36.3
Greg Kroah-Hartman [Fri, 7 Jan 2011 22:39:58 +0000 (14:39 -0800)]
Linux 2.6.36.3

13 years agoARM: 6490/1: MM: bugfix: initialize spinlock for init_mm.context
MyungJoo Ham [Tue, 23 Nov 2010 10:39:23 +0000 (11:39 +0100)]
ARM: 6490/1: MM: bugfix: initialize spinlock for init_mm.context

init_mm used at kernel/sched.c:idle_task_exit() has spin_lock
(init_mm.context.id_lock) that is not initialized when spin_lock/unlock
is called at an ARM machine. Note that mm_struct.context.id_lock is
usually initialized except for the instance of init_mm at
linux/arch/arm/mm/context.c

Not initializing this spinlock incurs "BUG: pinlock bad magic"
warning when spinlock debug is enabled. We have observed such
instances when testing PM in S5PC210 machines.

Change-Id: I52c6c6d035c220822a5675faaf8d163a3535985d
Signed-off-by: MyungJoo Ham <myungjoo.ham@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Colin Cross <ccross@android.com>
13 years agogspca - sonixj: Add a flag in the driver_info table
Jean-Francois Moine [Tue, 14 Dec 2010 19:15:37 +0000 (16:15 -0300)]
gspca - sonixj: Add a flag in the driver_info table

commit c6c14330717f9850b4b4c054b81424b9979cd07d upstream.

Signed-off-by: Jean-François Moine <moinejf@free.fr>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agogspca - sonixj: Set the flag for some devices
Jean-Francois Moine [Tue, 14 Dec 2010 19:16:16 +0000 (16:16 -0300)]
gspca - sonixj: Set the flag for some devices

commit b2272a49e7df37732d73988f00468ce31e1ebc92 upstream.

The flag PDN_INV indicates that the sensor pin S_PWR_DN has not the same
value as other webcams with the same sensor. For now, only two webcams have
been so detected: the Microsoft's VX1000 and VX3000.

Signed-off-by: Jean-François Moine <moinejf@free.fr>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agoorinoco: initialise priv->hw before assigning the interrupt
David Kilroy [Tue, 7 Dec 2010 18:50:42 +0000 (18:50 +0000)]
orinoco: initialise priv->hw before assigning the interrupt

commit 229bd792be0bad245b78ed8f119952733a4752e5 upstream.

The interrupt handler takes a lock - but since commit bcad6e80f3f this
lock goes through an indirection specified in the hermes_t structure.
We must therefore initialise the structure before setting up the
interrupt handler.

Fix orinoco_cs and spectrum_cs

<https://bugzilla.kernel.org/show_bug.cgi?id=23932>

Bisected by: Matt Domsch <Matt_Domsch@dell.com>
Signed-off by: David Kilroy <kilroyd@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agoDon't export format_by_forcc on two different drivers
Mauro Carvalho Chehab [Thu, 2 Dec 2010 03:57:03 +0000 (01:57 -0200)]
Don't export format_by_forcc on two different drivers

commit a757ee2216211278680dd8ac869aabe7b4a9970d upstream.

Drivers should append their name on exported symbols, to avoid
conflicts with allyesconfig:

drivers/staging/built-in.o: In function `format_by_fourcc':
/home/v4l/work_trees/linus/drivers/staging/cx25821/cx25821-video.c:96: multiple definition of `format_by_fourcc'
drivers/media/built-in.o:/home/v4l/work_trees/linus/drivers/media/common/saa7146_video.c:88: first defined here

Let's rename both occurences with a small shellscript:

for i in drivers/staging/cx25821/*.[ch]; do sed s,format_by_fourcc,cx25821_format_by_fourcc,g <$i >a && mv a $i; done
for i in drivers/media/common/saa7146*.[ch]; do sed s,format_by_fourcc,saa7146_format_by_fourcc,g <$i >a && mv a $i; done
for i in include/media/saa7146*.[ch]; do sed s,format_by_fourcc,saa7146_format_by_fourcc,g <$i >a && mv a $i; done

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agomx2_camera: fix pixel clock polarity configuration
Baruch Siach [Wed, 27 Oct 2010 07:03:52 +0000 (04:03 -0300)]
mx2_camera: fix pixel clock polarity configuration

commit 42cc37fe20cc680fb58fe12ae5ba718d683b8ca2 upstream.

When SOCAM_PCLK_SAMPLE_FALLING, just leave CSICR1_REDGE unset, otherwise we get
the inverted behaviour.

Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agoposix-cpu-timers: workaround to suppress the problems with mt exec
Oleg Nesterov [Fri, 5 Nov 2010 15:53:42 +0000 (16:53 +0100)]
posix-cpu-timers: workaround to suppress the problems with mt exec

commit e0a70217107e6f9844628120412cb27bb4cea194 upstream.

posix-cpu-timers.c correctly assumes that the dying process does
posix_cpu_timers_exit_group() and removes all !CPUCLOCK_PERTHREAD
timers from signal->cpu_timers list.

But, it also assumes that timer->it.cpu.task is always the group
leader, and thus the dead ->task means the dead thread group.

This is obviously not true after de_thread() changes the leader.
After that almost every posix_cpu_timer_ method has problems.

It is not simple to fix this bug correctly. First of all, I think
that timer->it.cpu should use struct pid instead of task_struct.
Also, the locking should be reworked completely. In particular,
tasklist_lock should not be used at all. This all needs a lot of
nontrivial and hard-to-test changes.

Change __exit_signal() to do posix_cpu_timers_exit_group() when
the old leader dies during exec. This is not the fix, just the
temporary hack to hide the problem for 2.6.37 and stable. IOW,
this is obviously wrong but this is what we currently have anyway:
cpu timers do not work after mt exec.

In theory this change adds another race. The exiting leader can
detach the timers which were attached to the new leader. However,
the window between de_thread() and release_task() is small, we
can pretend that sys_timer_create() was called before de_thread().

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agox86/microcode: Fix double vfree() and remove redundant pointer checks before vfree()
Jesper Juhl [Sat, 25 Dec 2010 18:57:41 +0000 (19:57 +0100)]
x86/microcode: Fix double vfree() and remove redundant pointer checks before vfree()

commit 5cdd2de0a76d0ac47f107c8a7b32d75d25768dc1 upstream.

In arch/x86/kernel/microcode_intel.c::generic_load_microcode()
we have  this:

while (leftover) {
...
if (get_ucode_data(mc, ucode_ptr, mc_size) ||
    microcode_sanity_check(mc) < 0) {
vfree(mc);
break;
}
...
}

if (mc)
vfree(mc);

This will cause a double free of 'mc'. This patch fixes that by
just  removing the vfree() call in the loop since 'mc' will be
freed nicely just  after we break out of the loop.

There's also a second change in the patch. I noticed a lot of
checks for  pointers being NULL before passing them to vfree().
That's completely  redundant since vfree() deals gracefully with
being passed a NULL pointer.  Removing the redundant checks
yields a nice size decrease for the object  file.

Size before the patch:
   text    data     bss     dec     hex filename
   4578     240    1032    5850    16da arch/x86/kernel/microcode_intel.o
Size after the patch:
   text    data     bss     dec     hex filename
   4489     240     984    5713    1651 arch/x86/kernel/microcode_intel.o

Signed-off-by: Jesper Juhl <jj@chaosbits.net>
Acked-by: Tigran Aivazian <tigran@aivazian.fsnet.co.uk>
Cc: Shaohua Li <shaohua.li@intel.com>
LKML-Reference: <alpine.LNX.2.00.1012251946100.10759@swampdragon.chaosbits.net>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agoblock: Deprecate QUEUE_FLAG_CLUSTER and use queue_limits instead
Martin K. Petersen [Wed, 1 Dec 2010 18:41:49 +0000 (19:41 +0100)]
block: Deprecate QUEUE_FLAG_CLUSTER and use queue_limits instead

commit e692cb668fdd5a712c6ed2a2d6f2a36ee83997b4 upstream.

When stacking devices, a request_queue is not always available. This
forced us to have a no_cluster flag in the queue_limits that could be
used as a carrier until the request_queue had been set up for a
metadevice.

There were several problems with that approach. First of all it was up
to the stacking device to remember to set queue flag after stacking had
completed. Also, the queue flag and the queue limits had to be kept in
sync at all times. We got that wrong, which could lead to us issuing
commands that went beyond the max scatterlist limit set by the driver.

The proper fix is to avoid having two flags for tracking the same thing.
We deprecate QUEUE_FLAG_CLUSTER and use the queue limit directly in the
block layer merging functions. The queue_limit 'no_cluster' is turned
into 'cluster' to avoid double negatives and to ease stacking.
Clustering defaults to being enabled as before. The queue flag logic is
removed from the stacking function, and explicitly setting the cluster
flag is no longer necessary in DM and MD.

Reported-by: Ed Lin <ed.lin@promise.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Acked-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agoSched: fix skip_clock_update optimization
Mike Galbraith [Wed, 8 Dec 2010 10:05:42 +0000 (11:05 +0100)]
Sched: fix skip_clock_update optimization

commit f26f9aff6aaf67e9a430d16c266f91b13a5bff64 upstream.

idle_balance() drops/retakes rq->lock, leaving the previous task
vulnerable to set_tsk_need_resched().  Clear it after we return
from balancing instead, and in setup_thread_stack() as well, so
no successfully descheduled or never scheduled task has it set.

Need resched confused the skip_clock_update logic, which assumes
that the next call to update_rq_clock() will come nearly immediately
after being set.  Make the optimization robust against the waking
a sleeper before it sucessfully deschedules case by checking that
the current task has not been dequeued before setting the flag,
since it is that useless clock update we're trying to save, and
clear unconditionally in schedule() proper instead of conditionally
in put_prev_task().

Signed-off-by: Mike Galbraith <efault@gmx.de>
Reported-by: Bjoern B. Brandenburg <bbb.lst@gmail.com>
Tested-by: Yong Zhang <yong.zhang0@gmail.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <1291802742.1417.9.camel@marge.simson.net>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agoALSA: hda: Use LPIB quirk for Dell Inspiron m101z/1120
Daniel T Chen [Tue, 28 Dec 2010 22:20:02 +0000 (17:20 -0500)]
ALSA: hda: Use LPIB quirk for Dell Inspiron m101z/1120

commit e03fa055bc126e536c7f65862e08a9b143138ea9 upstream.

Sjoerd Simons reports that, without using position_fix=1, recording
experiences overruns. Work around that by applying the LPIB quirk
for his hardware.

Reported-and-tested-by: Sjoerd Simons <sjoerd@debian.org>
Signed-off-by: Daniel T Chen <crimsun@ubuntu.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agomac80211: fix mesh forwarding
Johannes Berg [Wed, 22 Dec 2010 09:15:07 +0000 (10:15 +0100)]
mac80211: fix mesh forwarding

commit b51aff057c9d0ef6c529dc25fd9f775faf7b6c63 upstream.

Under memory pressure, the mac80211 mesh code
may helpfully print a message that it failed
to clone a mesh frame and then will proceed
to crash trying to use it anyway. Fix that.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Acked-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agolibata-sff: fix HSM_ST_ERR handling in __ata_sff_port_intr()
Tejun Heo [Fri, 3 Dec 2010 14:19:13 +0000 (15:19 +0100)]
libata-sff: fix HSM_ST_ERR handling in __ata_sff_port_intr()

commit 687a993339c4f3a63654746230da3aab8bbdbffd upstream.

While separating out BMDMA irq handler from SFF, commit c3b28894
(libata-sff: separate out BMDMA irq handler) incorrectly made
__ata_sff_port_intr() consider an IRQ to be an idle one if the host
state was transitioned to HSM_ST_ERR by ata_bmdma_port_intr().

This makes BMDMA drivers ignore IRQs reporting host bus error which
leads to timeouts instead of triggering EH immediately.  Fix it by
making __ata_sff_port_intr() consider the IRQ to be an idle one iff
the state is HSM_ST_IDLE.  This is equivalent to adding HSM_ST_ERR to
the "break"ing case but less error-prone.

Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Antonio Toma <antonio.toma@gmail.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agoima: fix add LSM rule bug
Mimi Zohar [Mon, 3 Jan 2011 22:59:10 +0000 (14:59 -0800)]
ima: fix add LSM rule bug

commit 867c20265459d30a01b021a9c1e81fb4c5832aa9 upstream.

If security_filter_rule_init() doesn't return a rule, then not everything
is as fine as the return code implies.

This bug only occurs when the LSM (eg. SELinux) is disabled at runtime.

Adding an empty LSM rule causes ima_match_rules() to always succeed,
ignoring any remaining rules.

 default IMA TCB policy:
  # PROC_SUPER_MAGIC
  dont_measure fsmagic=0x9fa0
  # SYSFS_MAGIC
  dont_measure fsmagic=0x62656572
  # DEBUGFS_MAGIC
  dont_measure fsmagic=0x64626720
  # TMPFS_MAGIC
  dont_measure fsmagic=0x01021994
  # SECURITYFS_MAGIC
  dont_measure fsmagic=0x73636673

  < LSM specific rule >
  dont_measure obj_type=var_log_t

  measure func=BPRM_CHECK
  measure func=FILE_MMAP mask=MAY_EXEC
  measure func=FILE_CHECK mask=MAY_READ uid=0

Thus without the patch, with the boot parameters 'tcb selinux=0', adding
the above 'dont_measure obj_type=var_log_t' rule to the default IMA TCB
measurement policy, would result in nothing being measured.  The patch
prevents the default TCB policy from being replaced.

Signed-off-by: Mimi Zohar <zohar@us.ibm.com>
Cc: James Morris <jmorris@namei.org>
Acked-by: Serge Hallyn <serge.hallyn@canonical.com>
Cc: David Safford <safford@watson.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agomv_xor: fix race in tasklet function
Saeed Bishara [Tue, 21 Dec 2010 14:53:39 +0000 (16:53 +0200)]
mv_xor: fix race in tasklet function

commit 8333f65ef094e47020cd01452b4637e7daf5a77f upstream.

use mv_xor_slot_cleanup() instead of __mv_xor_slot_cleanup() as the former function
aquires the spin lock that needed to protect the drivers data.

Signed-off-by: Saeed Bishara <saeed@marvell.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agoKVM: Fix OSXSAVE after migration
Sheng Yang [Wed, 8 Dec 2010 02:49:43 +0000 (10:49 +0800)]
KVM: Fix OSXSAVE after migration

commit 3ea3aa8cf67d3bbe00a19b6a4013d19efa7d0f41 upstream.

CPUID's OSXSAVE is a mirror of CR4.OSXSAVE bit. We need to update the CPUID
after migration.

KVM-Stable-Tag.
Signed-off-by: Sheng Yang <sheng@linux.intel.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agoKVM: SVM: Do not report xsave in supported cpuid
Joerg Roedel [Tue, 7 Dec 2010 16:15:05 +0000 (17:15 +0100)]
KVM: SVM: Do not report xsave in supported cpuid

commit 24d1b15f72abe3465e871d11cfc9dc34d1aab8b2 upstream.

To support xsave properly for the guest the SVM module need
software support for it. As long as this is not present do
not report the xsave as supported feature in cpuid.
As a side-effect this patch moves the bit() helper function
into the x86.h file so that it can be used in svm.c too.

KVM-Stable-Tag.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agoKVM: enlarge number of possible CPUID leaves
Andre Przywara [Wed, 1 Dec 2010 11:17:44 +0000 (12:17 +0100)]
KVM: enlarge number of possible CPUID leaves

commit 73c1160ce377d8fc6d84cb630ebf9658808bec49 upstream.

Currently the number of CPUID leaves KVM handles is limited to 40.
My desktop machine (AthlonII) already has 35 and future CPUs will
expand this well beyond the limit. Extend the limit to 80 to make
room for future processors.

KVM-Stable-Tag.
Signed-off-by: Andre Przywara <andre.przywara@amd.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agosound: Prevent buffer overflow in OSS load_mixer_volumes
Dan Rosenberg [Sat, 25 Dec 2010 21:23:40 +0000 (16:23 -0500)]
sound: Prevent buffer overflow in OSS load_mixer_volumes

commit d81a12bc29ae4038770e05dce4ab7f26fd5880fb upstream.

The load_mixer_volumes() function, which can be triggered by
unprivileged users via the SOUND_MIXER_SETLEVELS ioctl, is vulnerable to
a buffer overflow.  Because the provided "name" argument isn't
guaranteed to be NULL terminated at the expected 32 bytes, it's possible
to overflow past the end of the last element in the mixer_vols array.
Further exploitation can result in an arbitrary kernel write (via
subsequent calls to load_mixer_volumes()) leading to privilege
escalation, or arbitrary kernel reads via get_mixer_levels().  In
addition, the strcmp() may leak bytes beyond the mixer_vols array.

Signed-off-by: Dan Rosenberg <drosenberg@vsecurity.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agomemcg: fix wrong VM_BUG_ON() in try_charge()'s mm->owner check
KAMEZAWA Hiroyuki [Wed, 29 Dec 2010 22:07:11 +0000 (14:07 -0800)]
memcg: fix wrong VM_BUG_ON() in try_charge()'s mm->owner check

commit ebb76ce16daf6908dc030dec1c00827d37129fe5 upstream.

At __mem_cgroup_try_charge(), VM_BUG_ON(!mm->owner) is checked.
But as commented in mem_cgroup_from_task(), mm->owner can be NULL
in some racy case. This check of VM_BUG_ON() is bad.

A possible story to hit this is at swapoff()->try_to_unuse(). It passes
mm_struct to mem_cgroup_try_charge_swapin() while mm->owner is NULL. If we
can't get proper mem_cgroup from swap_cgroup information, mm->owner is used
as charge target and we see NULL.

Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Reported-by: Hugh Dickins <hughd@google.com>
Reported-by: Thomas Meyer <thomas@m3y3r.de>
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Reviewed-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Signed-off-by: Hugh Dickins <hughd@google.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agodrm/i915, intel_ips: When i915 loads after IPS, make IPS relink to i915.
Eric Anholt [Tue, 21 Dec 2010 02:40:06 +0000 (18:40 -0800)]
drm/i915, intel_ips: When i915 loads after IPS, make IPS relink to i915.

commit 63ee41d794d9c555f84205517a68509848988760 upstream.

The IPS driver is designed to be able to run detached from i915 and
just not enable GPU turbo in that case, in order to avoid module
dependencies between the two drivers.  This means that we don't know
what the load order between the two is going to be, and we had
previously only supported IPS after (optionally) i915, but not i915
after IPS.  If the wrong order was chosen, you'd get no GPU turbo, and
something like half the possible graphics performance.

Signed-off-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agodrm/i915/dp: Fix I2C/EDID handling with active DisplayPort to DVI converter
David Flynn [Wed, 8 Dec 2010 16:10:21 +0000 (16:10 +0000)]
drm/i915/dp: Fix I2C/EDID handling with active DisplayPort to DVI converter

commit 8316f33766a82907c694267ff911e45e256f09f9 upstream.

The DisplayPort standard (1.1a) states that:
  The I2C-over-AUX Reply field is valid only when Native AUX CH Reply
  field is AUX_ACK (00). When Native AUX CH Reply field is not 00, then,
  I2C-over-AUX Reply field must be 00 and be ignored.

This fixes broken EDID reading when using an active DisplayPort to
duallink DVI converter.  If the AUX CH replier chooses to defer the
transaction, a short read occurs and erroneous data is returned as
the i2c reply due to a lack of length checking and failure to check
for AUX ACK.

As a result, broken EDIDs can look like:
     0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f    0123456789abcdef
00: bc bc bc ff bc bc bc ff bc bc bc ac bc bc bc 45    ???.???.???????E
10: bc bc bc 10 bc bc bc 34 bc bc bc ee bc bc bc 4c    ???????4???????L
20: bc bc bc 50 bc bc bc 00 bc bc bc 40 bc bc bc 00    ???P???.???@???.
30: bc bc bc 01 bc bc bc 01 bc bc bc a0 bc bc bc 40    ???????????????@
40: bc bc bc 00 bc bc bc 00 bc bc bc 00 bc bc bc 55    ???.???.???.???U
50: bc bc bc 35 bc bc bc 31 bc bc bc 20 bc bc bc fc    ???5???1??? ????
60: bc bc bc 4c bc bc bc 34 bc bc bc 46 bc bc bc 00    ???L???4???F???.
70: bc bc bc 38 bc bc bc 11 bc bc bc 20 bc bc bc 20    ???8??????? ???
80: bc bc bc ff bc bc bc ff bc bc bc ff bc bc bc ff    ???.???.???.???.
...

which can lead to:
[drm:drm_edid_block_valid] *ERROR* EDID checksum is invalid, remainder
[drm:drm_edid_block_valid] *ERROR* Raw EDID:
<3>30 30 30 30 30 30 30 32 38 32 30 32 63 63 31 61  000000028202cc1a
<3>28 00 02 8c 00 00 00 00 18 00 00 00 00 00 00 00  (...............
<3>20 4c 61 73 74 20 62 65 61 63 6f 6e 3a 20 33 32   Last beacon: 32
<3>32 30 6d 73 20 61 67 6f 46 00 05 8c 00 00 00 00  20ms agoF.......
<3>36 00 00 00 00 00 00 00 00 0c 57 69 2d 46 69 20  6.........Wi-Fi
<3>52 6f 75 74 65 72 01 08 82 84 8b 96 24 30 48 6c  Router......$0Hl
<3>03 01 01 06 02 00 00 2a 01 00 2f 01 00 32 04 0c  .......*../..2..
<3>12 18 60 dd 09 00 10 18 02 00 00 01 00 00 18 00  ..`.............

Signed-off-by: David Flynn <davidf@rd.bbc.co.uk>
[ickle: fix up some surrounding checkpatch warnings]
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agodrm/radeon/kms: reorder display resume to avoid problems
Alex Deucher [Mon, 20 Dec 2010 16:22:29 +0000 (11:22 -0500)]
drm/radeon/kms: reorder display resume to avoid problems

commit a93f344d3c04e4b84490c65f2a574387c593be40 upstream.

On resume, we were attemping to unblank the displays before the
timing and plls had be reprogrammed which led to atom timeouts
waiting for things that are not yet programmed.  Re-program
the mode first, then reset the dpms state.

This fixes the infamous atombios timeouts on resume.

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agodrm/radeon/kms: fix evergreen asic reset
Alex Deucher [Mon, 20 Dec 2010 17:35:03 +0000 (12:35 -0500)]
drm/radeon/kms: fix evergreen asic reset

commit 9f0c4f9c2f835eee1bbb93f96bf9483d56f1892b upstream.

Only reset the grbm blocks, srbm tends to lock the GPU
if not done properly and in most cases is not necessary.
Also, no need to call asic init after reset the grbm blocks.

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agodrm/radeon/kms/evergreen: reset the grbm blocks at resume and init
Alex Deucher [Mon, 20 Dec 2010 17:35:04 +0000 (12:35 -0500)]
drm/radeon/kms/evergreen: reset the grbm blocks at resume and init

commit 86f5c9edbb3bac37cc8cee6528a929005ba72aad upstream.

This fixes module reloading and resume as the gfx block seems to
be left in a bad state in some cases.

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agomfd: Supply IRQ base for WM832x devices
Mark Brown [Wed, 24 Nov 2010 18:01:39 +0000 (18:01 +0000)]
mfd: Supply IRQ base for WM832x devices

commit bd7c72ed18d719c1fb0fdf6ff9042d8ab78fdf71 upstream.

Without this the IRQ base will not be correctly configured for the
subdevices.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agomfd: Support additional parent IDs for wm831x
Mark Brown [Thu, 2 Dec 2010 16:25:43 +0000 (16:25 +0000)]
mfd: Support additional parent IDs for wm831x

commit b93cef556162b0f33399bfe5f307c54f51554e09 upstream.

Some newer device revisions add a second parent ID. Support this in
the device validity checks done at startup.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agoarch/x86/oprofile/op_model_amd.c: Perform initialisation on a single CPU
Robert Richter [Mon, 3 Jan 2011 11:15:14 +0000 (12:15 +0100)]
arch/x86/oprofile/op_model_amd.c: Perform initialisation on a single CPU

commit c7c25802b39c443b3745cfa973dc49a97a3491f8 upstream.

Disable preemption in init_ibs(). The function only checks the
ibs capabilities and sets up pci devices (if necessary). It runs
only on one cpu but operates with the local APIC and some MSRs,
thus it is better to disable preemption.

[    7.034377] BUG: using smp_processor_id() in preemptible [00000000] code: modprobe/483
[    7.034385] caller is setup_APIC_eilvt+0x155/0x180
[    7.034389] Pid: 483, comm: modprobe Not tainted 2.6.37-rc1-20101110+ #1
[    7.034392] Call Trace:
[    7.034400]  [<ffffffff812a2b72>] debug_smp_processor_id+0xd2/0xf0
[    7.034404]  [<ffffffff8101e985>] setup_APIC_eilvt+0x155/0x180
[ ... ]

Addresses https://bugzilla.kernel.org/show_bug.cgi?id=22812

Reported-by: <atswartz@gmail.com>
Signed-off-by: Robert Richter <robert.richter@amd.com>
Cc: oprofile-list@lists.sourceforge.net <oprofile-list@lists.sourceforge.net>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Rafael J. Wysocki <rjw@sisk.pl>
Cc: Dan Carpenter <error27@gmail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
LKML-Reference: <20110103111514.GM4739@erda.amd.com>
[ small cleanups ]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agowatchdog: Improve initialisation error message and documentation
Ben Hutchings [Sun, 2 Jan 2011 23:02:42 +0000 (23:02 +0000)]
watchdog: Improve initialisation error message and documentation

commit 551423748a4eba55f2eb0fc250d757986471f187 upstream.

The error message 'NMI watchdog failed to create perf event...'
does not make it clear that this is a fatal error for the
watchdog.  It also currently prints the error value as a
pointer, rather than extracting the error code with PTR_ERR().
Fix that.

Add a note to the description of the 'nowatchdog' kernel
parameter to associate it with this message.

Reported-by: Cesare Leonardi <celeonar@gmail.com>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Cc: 599368@bugs.debian.org
Cc: 608138@bugs.debian.org
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
LKML-Reference: <1294009362.3167.126.camel@localhost>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agowatchdog: Fix null pointer dereference while accessing rdc321x platform_data
Florian Fainelli [Fri, 26 Nov 2010 09:39:55 +0000 (10:39 +0100)]
watchdog: Fix null pointer dereference while accessing rdc321x platform_data

commit 3b3c1f24e96c411a95daabb6af9e09c5381f713b upstream.

rdc321x-wdt currently fetches its driver specific data by using the
platform_device->platform_data pointer, this is wrong because the mfd
device which registers our platform_device has been added using
mfd_add_device() which sets the platform_device->driver_data pointer
instead.

Signed-off-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agoRAMOOPS: Don't overflow over non-allocated regions
Ahmed S. Darwish [Sat, 25 Dec 2010 09:57:09 +0000 (11:57 +0200)]
RAMOOPS: Don't overflow over non-allocated regions

commit 1873bb8115e678ad9fd0aac9dbbc68383bc36e06 upstream.

The current code mis-calculates the ramoops header size, leading to an
overflow over the next record at best, or over a non-allocated region at
worst.  Fix that calculation.

Signed-off-by: Ahmed S. Darwish <darwish.07@gmail.com>
Acked-by: Marco Stornelli <marco.stornelli@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agortc: rs5c372: fix buffer size
Wolfram Sang [Wed, 22 Dec 2010 01:24:24 +0000 (17:24 -0800)]
rtc: rs5c372: fix buffer size

commit 118364948fad7b6c0469ef2d3ddaee447d7a0b5f upstream.

Match the buffer size to the amount of initialized values.  Before, it was
one too big and thus destroyed the neighbouring register causing the clock
to run at false speeds.

Reported-by: Andre van Rooyen <a.v.rooyen@sercom.nl>
Signed-off-by: Wolfram Sang <w.sang@pengutronix.de>
Cc: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agofix freeing user_struct in user cache
Hillf Danton [Wed, 29 Dec 2010 13:55:28 +0000 (21:55 +0800)]
fix freeing user_struct in user cache

commit 4ef9e11d6867f88951e30db910fa015300e31871 upstream.

When racing on adding into user cache, the new allocated from mm slab
is freed without putting user namespace.

Since the user namespace is already operated by getting, putting has
to be issued.

Signed-off-by: Hillf Danton <dhillf@gmail.com>
Acked-by: Serge Hallyn <serge@hallyn.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agommc: Fix re-probing with PM_POST_RESTORE notification
Takashi Iwai [Fri, 10 Dec 2010 07:40:31 +0000 (08:40 +0100)]
mmc: Fix re-probing with PM_POST_RESTORE notification

commit 274476f8fe0b6ac9bac542cc39de12c3dd0f43f6 upstream.

In the error-path where PM notifies PM_POST_RESTORE, the rescan-blockage
should be cleared as well.  Otherwise it'll be never re-probed.

Also, as a bonus, this fixes a bug in S4 with user-mode suspend in the
current code, as it sends PM_POST_RESTORE instead of
PM_POST_HIBERNATION wrongly.

Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Chris Ball <cjb@laptop.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agommc: atmel-mci: fix multiblock SDIO transfers
Nicolas Ferre [Fri, 10 Dec 2010 18:14:32 +0000 (19:14 +0100)]
mmc: atmel-mci: fix multiblock SDIO transfers

commit 2f1d791882d21a4002a719fb016a1ac21c8bd6b7 upstream.

Based on report made by Yauhen in:
"MMC: Fix multiblock SDIO transfers in AT91 MCI" patch,
I report those changes to the brother driver: atmel-mci.

So, this patch sets SDIO transfer types: SDIO block and SDIO byte
transfers instead of using ordinary MMC block transfers.
It is checking opcode for SDIO CMD53 and setting transfer
type in MCI_CMDR register properly.

Reported-by: Yauhen Kharuzhy <yauhen.kharuzhy@promwad.com>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Acked-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agommc: at91_mci: fix multiblock SDIO transfers
Yauhen Kharuzhy [Thu, 25 Nov 2010 10:11:51 +0000 (12:11 +0200)]
mmc: at91_mci: fix multiblock SDIO transfers

commit a2255ff45143001fecbc5e5a4b58fcb999d393ae upstream.

The AT91 MCI has special SDIO transfer types: SDIO block and SDIO byte
transfers, but at91_mci driver doesn't use them and handles all SDIO
transfers as ordinary MMC block transfers. This causes problems for
multiple-block SDIO transfers (in particular for 256-bytes blocks).

Fix this situation by checking the opcode for SDIO CMD53 and setting
the transfer type in the AT91_MCI_CMDR register properly.

This patch was tested with libertas SDIO driver: problem with TX
timeouts on big packets was eliminated.

Signed-off-by: Yauhen Kharuzhy <yauhen.kharuzhy@promwad.com>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Acked-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agocs5535-gpio: handle GPIO regs where higher (clear) bits are set
Andres Salomon [Tue, 21 Dec 2010 21:04:52 +0000 (13:04 -0800)]
cs5535-gpio: handle GPIO regs where higher (clear) bits are set

commit 44658a11f312fb9217674cb90b1a11cbe17fd18d upstream.

The default for non-READ_BACK GPIO regs is to have the clear bits set;
this means that our original errata fix was too simplistic.  This
changes it to the following behavior:

 - when setting GPIOs, ignore the higher order bits (they're for
   clearing, we don't need to care about them).

 - when clearing GPIOs, keep all the bits, but unset (via XOR) the
   lower order bit that negates the clear bit that we care about.  That
   is, if we're clearing GPIO 26 (val = 0x04000000), we first XOR what's
   currently in the register with 0x0400 (GPIO 26's SET bit), and then
   OR that with the GPIO 26's CLEAR bit.

Tested-by: Daniel Drake <dsd@laptop.org>
Signed-off-by: Andres Salomon <dilinger@queued.net>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
13 years agocs5535-gpio: don't apply errata #36 to edge detect GPIOs
Andres Salomon [Tue, 21 Dec 2010 21:04:42 +0000 (13:04 -0800)]
cs5535-gpio: don't apply errata #36 to edge detect GPIOs

commit 001851659354cce436b749a793f3512a53394d80 upstream.

The edge detect status GPIOs function differently from the other atomic
model CS5536 GPIO registers; writing 1 to the high bits clears the GPIO,
but writing 1 to the lower bits also clears the bit.

This means that read-modify-write doesn't actually work for it, so don't
apply the errata here.  If a negative edge status gets lost after
resume..  well, we tried our best!

Tested-by: Daniel Drake <dsd@laptop.org>
Signed-off-by: Andres Salomon <dilinger@queued.net>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>