openwrt/staging/blogic.git
12 years agoASoC: wm0010: Allow slow GPIO for reset
Mark Brown [Tue, 25 Sep 2012 15:35:26 +0000 (16:35 +0100)]
ASoC: wm0010: Allow slow GPIO for reset

We never set the GPIO from atomic context so there's no reason why we
can't support a GPIO that needs to sleep when configuring.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: wm5110: Enable bypass mode for MICVDD
Mark Brown [Fri, 7 Sep 2012 04:58:54 +0000 (12:58 +0800)]
ASoC: wm5110: Enable bypass mode for MICVDD

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: wm5102: Enable bypass mode for MICVDD
Mark Brown [Fri, 7 Sep 2012 04:58:43 +0000 (12:58 +0800)]
ASoC: wm5102: Enable bypass mode for MICVDD

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: dapm: Allow regulators to bypass as well as disable when idle
Mark Brown [Fri, 7 Sep 2012 04:57:11 +0000 (12:57 +0800)]
ASoC: dapm: Allow regulators to bypass as well as disable when idle

Allow regulators managed via DAPM to make use of the bypass support that
has recently been added to the regulator API by setting a flag
SND_SOC_DAPM_REGULATOR_BYPASS. When this flag is set the regulator will
be put into bypass mode before being disabled, allowing the regulator to
fall into bypass mode if it can't be disabled due to other users.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoMerge tag 'bypass' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator...
Mark Brown [Wed, 26 Sep 2012 11:29:31 +0000 (12:29 +0100)]
Merge tag 'bypass' of git://git./linux/kernel/git/broonie/regulator into for-3.7

regulator: Bypass mode support

Allow regulators to be put into a non-regulating mode bypassing the
input straight to the output, mostly used by low power retention modes.

12 years agoASoC: cs4270: Remove mono support
Fabio Estevam [Tue, 18 Sep 2012 16:03:54 +0000 (13:03 -0300)]
ASoC: cs4270: Remove mono support

According to cs4270 datasheet, there is no reference to mono mode.

Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Acked-by: Timur Tabi <timur@freescale.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoMerge remote-tracking branch 'asoc/topic/ux500' into for-3.7
Mark Brown [Sat, 22 Sep 2012 22:47:58 +0000 (18:47 -0400)]
Merge remote-tracking branch 'asoc/topic/ux500' into for-3.7

12 years agoASoC: wm2000: Add regulator support
Mark Brown [Sat, 22 Sep 2012 22:33:23 +0000 (18:33 -0400)]
ASoC: wm2000: Add regulator support

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: wm2000: Convert to devm_regmap_init_i2c()
Mark Brown [Sat, 22 Sep 2012 22:32:08 +0000 (18:32 -0400)]
ASoC: wm2000: Convert to devm_regmap_init_i2c()

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoMerge tag 'v3.6-rc6' into for-3.7
Mark Brown [Sat, 22 Sep 2012 15:26:27 +0000 (11:26 -0400)]
Merge tag 'v3.6-rc6' into for-3.7

Linux 3.6-rc6 has all our bug fixes.

Conflicts (trivial overlap):
sound/soc/omap/am3517evm.c

12 years agoASoC: twl4030: Support for DT booted kernel
Peter Ujfalusi [Mon, 10 Sep 2012 10:46:32 +0000 (13:46 +0300)]
ASoC: twl4030: Support for DT booted kernel

When the kernel has been booted with DT blob the platform data is NULL for
the driver.
We need to construct the pdata based on the DT information for runtime use.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: twl4030: Add pointer to pdata within the private data
Peter Ujfalusi [Mon, 10 Sep 2012 10:46:31 +0000 (13:46 +0300)]
ASoC: twl4030: Add pointer to pdata within the private data

Access the pdata via a pointer within the twl4030_priv structure.
In preparation for DeviceTree support.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: twl4030: Convert to use devm_kzalloc
Peter Ujfalusi [Mon, 10 Sep 2012 10:46:30 +0000 (13:46 +0300)]
ASoC: twl4030: Convert to use devm_kzalloc

Allocate the private data with devm_kzalloc.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC/mfd: twl4030: Remove set_hs_extmute callback from platform data
Peter Ujfalusi [Mon, 10 Sep 2012 10:46:29 +0000 (13:46 +0300)]
ASoC/mfd: twl4030: Remove set_hs_extmute callback from platform data

We no longer have users for the set_hs_extmute callback which has been
replaced by hs_extmute_gpio so the codec driver can handle the external
mute if it is needed by the board.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Acked-by: Samuel Ortiz <sameo@linux.intel.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoARM: OMAP/ASoC: Zoom2: Let the codec to handle the hs_extmute GPIO
Peter Ujfalusi [Mon, 10 Sep 2012 10:46:28 +0000 (13:46 +0300)]
ARM: OMAP/ASoC: Zoom2: Let the codec to handle the hs_extmute GPIO

Remove the use of set_hs_extmute callback and let the codec driver to
handle the extmute GPIO.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Acked-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: twl4030: Move hs_extmute GPIO handling to driver
Peter Ujfalusi [Mon, 10 Sep 2012 10:46:27 +0000 (13:46 +0300)]
ASoC: twl4030: Move hs_extmute GPIO handling to driver

The external mute (if it is in use) is handled by a GPIO line. Prepare to
remove the set_hs_extmute callback and replace it with:
hs_extmute_gpio: the GPIO number to use for external mute

When the users of set_hs_extmute has been converted the callback can be removed.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoInput: twl4030-vibra: Support for DT booted kernel
Peter Ujfalusi [Mon, 10 Sep 2012 10:46:26 +0000 (13:46 +0300)]
Input: twl4030-vibra: Support for DT booted kernel

Add support when the kernel has been booted with DT blob. In this case the
pdata is NULL, we need to reach up to the core node and check if the codec
part has been enabled to determine if we need to coexist with the codec or
not.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Acked-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agomfd: twl4030-audio: Add DT support
Peter Ujfalusi [Mon, 10 Sep 2012 10:46:25 +0000 (13:46 +0300)]
mfd: twl4030-audio: Add DT support

Support for loading the twl4030 audio module via devicetree.
Sub devices for codec and vibra will be created as mfd devices once the
core MFD driver is loaded when the kernel is booted with a DT blob.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Acked-by: Samuel Ortiz <sameo@linux.intel.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agodt: Add empty of_find_node_by_name() function
Peter Ujfalusi [Mon, 10 Sep 2012 10:46:24 +0000 (13:46 +0300)]
dt: Add empty of_find_node_by_name() function

This commit adds an empty of_find_node_by_name() function for !CONFIG_OF
builds.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agomfd: twl4030-audio: Get audio MCLK via twl-core API instead of pdata
Peter Ujfalusi [Mon, 10 Sep 2012 10:46:23 +0000 (13:46 +0300)]
mfd: twl4030-audio: Get audio MCLK via twl-core API instead of pdata

twl-core has API to get the boot time configured HFCLK rate which has the
same rate as the audio MCLK.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Acked-by: Samuel Ortiz <sameo@linux.intel.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agomfd: twl-core: Add API to query the HFCLK rate
Peter Ujfalusi [Mon, 10 Sep 2012 10:46:22 +0000 (13:46 +0300)]
mfd: twl-core: Add API to query the HFCLK rate

CFG_BOOT register's HFCLK_FREQ field hold information about the used HFCLK
frequency.
Add possibility for users to get the configured rate based on this
register.
This register was configured during boot, without it the chip would not
operate correctly, so we can trust on this information.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Acked-by: Samuel Ortiz <sameo@linux.intel.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agomfd: twl4030-audio: Rearange and clean-up the probe function
Peter Ujfalusi [Mon, 10 Sep 2012 10:46:21 +0000 (13:46 +0300)]
mfd: twl4030-audio: Rearange and clean-up the probe function

To facilitate the device tree support the probe function need to be rearanged.
Small cleanup in the APLL frequency selection part as well.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Acked-by: Samuel Ortiz <sameo@linux.intel.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agomfd: twl4030-audio: Convert to use devm_kzalloc
Peter Ujfalusi [Mon, 10 Sep 2012 10:46:20 +0000 (13:46 +0300)]
mfd: twl4030-audio: Convert to use devm_kzalloc

To clean up the module probe and remove functions.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Acked-by: Samuel Ortiz <sameo@linux.intel.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agomfd: twl4030-audio: Clean up MODULE_* and platform_driver part
Peter Ujfalusi [Mon, 10 Sep 2012 10:46:19 +0000 (13:46 +0300)]
mfd: twl4030-audio: Clean up MODULE_* and platform_driver part

Place the MODULE_* lines in the same block and add MODULE_DESCRIPTION.
Rearange the platform_driver structure at the same time.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Acked-by: Samuel Ortiz <sameo@linux.intel.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: omap-pcm: Convert to use dmaengine
Peter Ujfalusi [Fri, 14 Sep 2012 12:05:58 +0000 (15:05 +0300)]
ASoC: omap-pcm: Convert to use dmaengine

Original author: Russell King <rmk+kernel@arm.linux.org.uk>

Switch the omap-pcm to use dmaengine.
Certain features are not supported by after dmaengine conversion:
1. No period wakeup mode
   DMA engine has no way to communicate this information through
   standard channels.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: Janusz Krzysztofik <jkrzyszt@tis.icnet.pl>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: OMAP: mcbsp, mcpdm, dmic, hdmi: Set dma_data at startup time
Peter Ujfalusi [Fri, 14 Sep 2012 12:05:57 +0000 (15:05 +0300)]
ASoC: OMAP: mcbsp, mcpdm, dmic, hdmi: Set dma_data at startup time

Set the dma_data for the stream (snd_soc_dai_set_dma_data) at dai_startup
time so omap-pcm will have access to the needed information regarding to
the DMA channel earlier.
This is needed for the clean dmaengine support.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: Janusz Krzysztofik <jkrzyszt@tis.icnet.pl>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: omap-pcm, omap-dmic: Change the use of omap_pcm_dma_data->data_type
Peter Ujfalusi [Fri, 14 Sep 2012 12:05:56 +0000 (15:05 +0300)]
ASoC: omap-pcm, omap-dmic: Change the use of omap_pcm_dma_data->data_type

Instead of the OMAP DMA data type definition the data_type will be used to
specify the number of bits the DMA word should be configured or 0 in case
when based on the stream's format the omap-pcm can decide the needed DMA
word size.
This feature is needed for the omap-hdmi where the sDMA need to be
configured for 32bit word type regardless of the audio format used.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: Janusz Krzysztofik <jkrzyszt@tis.icnet.pl>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: OMAP: mcbsp, mcpdm, dmic: Let omap-pcm to pick the dma_type
Peter Ujfalusi [Fri, 14 Sep 2012 12:05:55 +0000 (15:05 +0300)]
ASoC: OMAP: mcbsp, mcpdm, dmic: Let omap-pcm to pick the dma_type

omap-pcm can figure out the correct dma_type based on the stream's format.
In this way we can get rid of the plat/dma.h include from these drivers.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: Janusz Krzysztofik <jkrzyszt@tis.icnet.pl>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: omap-mcpdm: Use platform_get_resource_* to get resources
Peter Ujfalusi [Fri, 14 Sep 2012 12:05:54 +0000 (15:05 +0300)]
ASoC: omap-mcpdm: Use platform_get_resource_* to get resources

Get the needed resources in a correct way and avoid using defines for them.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: Janusz Krzysztofik <jkrzyszt@tis.icnet.pl>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoARM: OMAP4: hwmod_data: Add resource names to McPDM memory ranges
Peter Ujfalusi [Fri, 14 Sep 2012 12:05:53 +0000 (15:05 +0300)]
ARM: OMAP4: hwmod_data: Add resource names to McPDM memory ranges

To help the driver to get the correct memory range to access McPDM
registers.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: Janusz Krzysztofik <jkrzyszt@tis.icnet.pl>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: omap-pcm: Prepare to configure the DMA data_type based on stream properties
Peter Ujfalusi [Fri, 14 Sep 2012 12:05:52 +0000 (15:05 +0300)]
ASoC: omap-pcm: Prepare to configure the DMA data_type based on stream properties

Based on the format of the stream the omap-pcm can decide alone what data
type should be used with by the sDMA.
Keep the possibility for OMAP dai drivers to tell omap-pcm if they want to
use different data type. This is needed for the omap-hdmi for example which
needs 32bit data type even if the stream format is S16_LE.

The check if (dma_data->data_type) is safe at the moment since omap-pcm
does not support 8bit samples (OMAP_DMA_DATA_TYPE_S8 == 0x00).

The next step is to redefine the meaning of dma_data->data_type to unblock
this limitation.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: Janusz Krzysztofik <jkrzyszt@tis.icnet.pl>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: OMAP: Remove sync_mode from omap_pcm_dma_data struct
Peter Ujfalusi [Fri, 14 Sep 2012 12:05:51 +0000 (15:05 +0300)]
ASoC: OMAP: Remove sync_mode from omap_pcm_dma_data struct

The omap-pcm platform driver no longer needs this parameter to select
between ELEMENT and PACKET mode. The selection is based on the configured
packet_size.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: Janusz Krzysztofik <jkrzyszt@tis.icnet.pl>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: omap-pcm: Select sDMA synchronization based on packet_size
Peter Ujfalusi [Fri, 14 Sep 2012 12:05:50 +0000 (15:05 +0300)]
ASoC: omap-pcm: Select sDMA synchronization based on packet_size

Since we only have element or packet synchronization we can use the
dma_data->packet_size to select the desired mode:
if packet_size is 0 we use ELEMENT mode
if packet_size is not 0 we use PACKET mode for sDMA synchronization.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: Janusz Krzysztofik <jkrzyszt@tis.icnet.pl>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: omap-mcbsp: Use sDMA packet mode instead of frame mode
Peter Ujfalusi [Fri, 14 Sep 2012 12:05:49 +0000 (15:05 +0300)]
ASoC: omap-mcbsp: Use sDMA packet mode instead of frame mode

When McBSP is configured in threshold mode we can use sDMA packet mode in
all cases.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: Janusz Krzysztofik <jkrzyszt@tis.icnet.pl>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agodmaengine: omap-dma: Add support to suppress interrupts in cyclic mode
Peter Ujfalusi [Fri, 14 Sep 2012 12:05:48 +0000 (15:05 +0300)]
dmaengine: omap-dma: Add support to suppress interrupts in cyclic mode

When requested (DMA_PREP_INTERRUPT is cleared in flags) disable all DMA
interrupts for the channel. In this mode user space does not expect
periodic reports from kernel about the progress of the audio stream.
PulseAudio for example support this type of mode.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Acked-by: Vinod Koul <vinod.koul@linux.intel.com>
Tested-by: Janusz Krzysztofik <jkrzyszt@tis.icnet.pl>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agodmaengine: Pass flags via device_prep_dma_cyclic() callback
Peter Ujfalusi [Fri, 14 Sep 2012 12:05:47 +0000 (15:05 +0300)]
dmaengine: Pass flags via device_prep_dma_cyclic() callback

Change the parameter list of device_prep_dma_cyclic() so the DMA drivers
can receive the flags coming from clients.
This feature can be used during audio operation to disable all audio
related interrupts when the DMA_PREP_INTERRUPT is cleared from the flags.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Acked-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Acked-by: Shawn Guo <shawn.guo@linaro.org>
Acked-by: Vinod Koul <vinod.koul@linux.intel.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agodmaengine: omap: Add support for pause/resume in cyclic dma mode
Peter Ujfalusi [Fri, 14 Sep 2012 12:05:45 +0000 (15:05 +0300)]
dmaengine: omap: Add support for pause/resume in cyclic dma mode

The audio stack used omap_stop_dma/omap_start_dma to pause/resume the DMA.
This method has been used for years on OMAP based products.
We only allow pause/resume when the DMA has been configured in cyclic mode
which is used by the audio stack.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Acked-by: Vinod Koul <vinod.koul@linux.intel.com>
Tested-by: Janusz Krzysztofik <jkrzyszt@tis.icnet.pl>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agodmaengine: omap: Support for element mode in cyclic DMA
Peter Ujfalusi [Fri, 14 Sep 2012 12:05:44 +0000 (15:05 +0300)]
dmaengine: omap: Support for element mode in cyclic DMA

When src_maxburst/dst_maxburst is set to 0 by the users of cyclic DMA
(mostly audio) indicates that we should configure the omap DMA to element
sync mode instead of packet mode.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Acked-by: Vinod Koul <vinod.koul@linux.intel.com>
Tested-by: Janusz Krzysztofik <jkrzyszt@tis.icnet.pl>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: fsl: register the wm9712-codec
Eric Millbrandt [Thu, 20 Sep 2012 14:36:45 +0000 (10:36 -0400)]
ASoC: fsl: register the wm9712-codec

The mpc5200-psc-ac97 driver does not enumerate attached ac97 devices, so
register the device here.

Signed-off-by: Eric Millbrandt <emillbrandt@dekaresearch.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: fsl: pcm030-audio-fabric use snd_soc_register_card
Eric Millbrandt [Thu, 20 Sep 2012 14:36:44 +0000 (10:36 -0400)]
ASoC: fsl: pcm030-audio-fabric use snd_soc_register_card

Convert pcm030-audio-fabric to use the new snd_soc_register_card api
instead of the older method of registering a separate platform device.
Create the dai_link to the mpc5200_psc_ac97 platform using the device tree.
Convert the pcm030-audio-fabric driver to a platform-driver and add a
remove function.

Signed-off-by: Eric Millbrandt <emillbrandt@dekaresearch.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: fsl: add PPC_MPC52xx dependency to SND_POWERPC_SOC
Eric Millbrandt [Thu, 20 Sep 2012 14:36:43 +0000 (10:36 -0400)]
ASoC: fsl: add PPC_MPC52xx dependency to SND_POWERPC_SOC

mpc52xx socs do not define FSL_SOC but need SND_POWERPC_SOC defined to build
ASoC drivers.

Signed-off-by: Eric Millbrandt <emillbrandt@dekaresearch.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: twl6040: Convert to use DAI DAPM widgets
Peter Ujfalusi [Thu, 20 Sep 2012 13:32:15 +0000 (16:32 +0300)]
ASoC: twl6040: Convert to use DAI DAPM widgets

Use DAPM mapping for stream events and give unique names for the streams.
This change also fixes the following warning:
twl6040-codec twl6040-codec: Failed to create Capture debugfs file

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: twl4030: Convert to use DAI DAPM widgets
Peter Ujfalusi [Thu, 20 Sep 2012 13:32:02 +0000 (16:32 +0300)]
ASoC: twl4030: Convert to use DAI DAPM widgets

Use DAPM mapping for stream events and give unique names for the streams.
This change also fixes the following warning:
twl4030-codec twl4030-codec: Failed to create Capture debugfs file

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: wm8960: remove 'dres' field from platform data structure
Timur Tabi [Thu, 20 Sep 2012 18:57:27 +0000 (13:57 -0500)]
ASoC: wm8960: remove 'dres' field from platform data structure

The 'dres' field (discharge resistance for headphone outputs) is no longer
used in the driver, so remove it.

It was used in the original version of the driver when entering standby
from off, but we stopped using it when we switched from having a single
startup sequence to having separate cap and capless sequences.

Signed-off-by: Timur Tabi <timur@freescale.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoDocumentation: Define the MSP Driver Device Tree bindings
Lee Jones [Tue, 31 Jul 2012 10:59:21 +0000 (11:59 +0100)]
Documentation: Define the MSP Driver Device Tree bindings

Here we add the required documentation for the new Device Tree
bindings pertaining to the MSP CPU-side DAI Driver.

Acked-by: Ola Lilja <ola.o.lilja@stericsson.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
12 years agoDocumentation: Define the MOP500 Audio Machine Driver Device Tree bindings
Lee Jones [Tue, 31 Jul 2012 10:56:51 +0000 (11:56 +0100)]
Documentation: Define the MOP500 Audio Machine Driver Device Tree bindings

Here we add the required documentation for the new Device Tree
bindings pertaining to the MOP500 Audio Machine driver.

Acked-by: Ola Lilja <ola.o.lilja@stericsson.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
12 years agoASoC: Ux500: Minor coding layout changes
Lee Jones [Tue, 21 Aug 2012 09:06:44 +0000 (10:06 +0100)]
ASoC: Ux500: Minor coding layout changes

Includes removal of duplicate debug print affirming entry into
the probe function, an unnecessary line break of a coding line
<80 chars and a white space change (unintentional tab).

Acked-by: Ola Lilja <ola.o.lilja@stericsson.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
12 years agoASoC: codecs: Enable AB8500 CODEC for Device Tree
Lee Jones [Fri, 27 Jul 2012 07:50:05 +0000 (08:50 +0100)]
ASoC: codecs: Enable AB8500 CODEC for Device Tree

We continue to allow the AB8500 CODEC to be registered via the AB8500
Multi Functional Device API, only this time we extract its configuration
from the Device Tree binary.

Acked-by: Ola Lilja <ola.o.lilja@stericsson.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
12 years agoASoC: Ux500: Enable ux500 MSP driver for Device Tree
Lee Jones [Thu, 26 Jul 2012 16:07:26 +0000 (17:07 +0100)]
ASoC: Ux500: Enable ux500 MSP driver for Device Tree

Register both parts of the MSP driver from Device Tree so that they
are probed when Device Tree is enabled. Also, as there is platform
data involved, we ensure that there is allocated memory to place the
configuration into and that the correct information is extracted from
the DT binary.

Acked-by: Ola Lilja <ola.o.lilja@stericsson.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
12 years agoASoC: Ux500: Enable MOP500 driver for Device Tree
Lee Jones [Thu, 26 Jul 2012 15:48:34 +0000 (16:48 +0100)]
ASoC: Ux500: Enable MOP500 driver for Device Tree

Here we ensure that the MOP500 audio driver will be probed during a
Device Tree boot. We also parse the sound node to link together the
codec, dma and the CPU-side Digital Audio Interface.

Acked-by: Ola Lilja <ola.o.lilja@stericsson.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
12 years agoASoC: Ux500: Move MSP pinctrl setup into the MSP driver
Lee Jones [Fri, 14 Sep 2012 15:16:08 +0000 (16:16 +0100)]
ASoC: Ux500: Move MSP pinctrl setup into the MSP driver

In the initial submission of the MSP driver msp1 and msp3's associated
pinctrl mechanism was passed back to platform code using a plat_init()
call-back routine, but it has no place in platform code. The MSP driver
should set this up for the appropriate ports. Instead we use a use_pinctrl
identifier which is passed from platform_data/Device Tree which indicates
which ports should use pinctrl.

Acked-by: Ola Lilja <ola.o.lilja@stericsson.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
12 years agoASoC: wm8961: Move device identification and reset to I2C probe
Mark Brown [Thu, 13 Sep 2012 05:31:38 +0000 (13:31 +0800)]
ASoC: wm8961: Move device identification and reset to I2C probe

This is more idiomatic as it means we verify that the device is there
prior to trying to do the card probe.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: wm8961: Convert to direct regmap API usage
Mark Brown [Thu, 13 Sep 2012 04:53:59 +0000 (12:53 +0800)]
ASoC: wm8961: Convert to direct regmap API usage

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: wm8960: Support shared LRCLK
Mark Brown [Thu, 13 Sep 2012 03:46:58 +0000 (11:46 +0800)]
ASoC: wm8960: Support shared LRCLK

If the LRCLK is shared and the WM8960 is clock master then we should
enable the LRCM bit to tell the device that it should drive LRCLK when
either ADC or DAC is enabled rather than separately driving the two
LRCLKs.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: wm8776: Convert to direct regmap API usage
Mark Brown [Wed, 12 Sep 2012 12:15:22 +0000 (20:15 +0800)]
ASoC: wm8776: Convert to direct regmap API usage

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: fsl: cleanup headers in pcm030-audio-fabric
Eric Millbrandt [Thu, 13 Sep 2012 21:43:13 +0000 (17:43 -0400)]
ASoC: fsl: cleanup headers in pcm030-audio-fabric

Remove unreferenced header files.

Signed-off-by: Eric Millbrandt <emillbrandt@dekaresearch.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: fsl: mpc5200 add missing information to snd_soc_dai_driver
Eric Millbrandt [Thu, 13 Sep 2012 21:43:12 +0000 (17:43 -0400)]
ASoC: fsl: mpc5200 add missing information to snd_soc_dai_driver

Add missing dai_driver information to avoid these runtime errors

[   16.433788] asoc: error - multiple DAI f0002c00.i2s registered with no name
[   16.453551] Failed to register DAI
[   16.461222] mpc5200-psc-i2s: probe of f0002c00.i2s failed with error -22
[   16.475242] asoc: error - multiple DAI f0002000.ac97 registered with no name
[   16.488087] mpc5200-psc-ac97 f0002000.ac97: Failed to register DAI
[   16.502222] mpc5200-psc-ac97: probe of f0002000.ac97 failed with error -22

Signed-off-by: Eric Millbrandt <emillbrandt@dekaresearch.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: fsl: mpc5200 combine psc_dma platform data
Eric Millbrandt [Thu, 13 Sep 2012 21:43:11 +0000 (17:43 -0400)]
ASoC: fsl: mpc5200 combine psc_dma platform data

The mpc5200_psc_ac97 and mpc5200_psc_i2s modules rely on shared platform data
with mpc5200_dma.

Signed-off-by: Eric Millbrandt <emillbrandt@dekaresearch.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: imx-audmux: remove null check of audmux_base in audmux_read_file
Richard Zhao [Tue, 18 Sep 2012 08:38:30 +0000 (16:38 +0800)]
ASoC: imx-audmux: remove null check of audmux_base in audmux_read_file

When audmux_read_file is called, it means the driver is already
initialised successfully, so we don't need to check audmux_base.

It also fix smatch warning:
sound/soc/fsl/imx-audmux.c:78 audmux_read_file() warn: possible memory leak of 'buf'

Signed-off-by: Richard Zhao <richard.zhao@freescale.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: imx-pcm-dma: open function failed when snd_dmaengine_pcm_open fail
Richard Zhao [Tue, 18 Sep 2012 09:20:06 +0000 (17:20 +0800)]
ASoC: imx-pcm-dma: open function failed when snd_dmaengine_pcm_open fail

snd_imx_open should return error code returned by snd_dmaengine_pcm_open.

Signed-off-by: Richard Zhao <richard.zhao@freescale.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: imx-pcm-dma: check kzalloc return value in function snd_imx_open
Richard Zhao [Tue, 18 Sep 2012 09:20:05 +0000 (17:20 +0800)]
ASoC: imx-pcm-dma: check kzalloc return value in function snd_imx_open

It fixed smatch warning:
sound/soc/fsl/imx-pcm-dma.c:112 snd_imx_open() error: potential null dereference 'dma_data'.  (kzalloc returns null)

Signed-off-by: Richard Zhao <richard.zhao@freescale.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: Avoid recalculating the bitmask for SOC_ENUM controls
Lars-Peter Clausen [Fri, 14 Sep 2012 11:57:27 +0000 (13:57 +0200)]
ASoC: Avoid recalculating the bitmask for SOC_ENUM controls

For ENUM controls the bitmask is calculated based on the number of items.
Currently this is done each time the control is accessed. And while the
performance impact of this should be negligible we can easily do better. The
roundup_pow_of_two macro performs the same calculation which is currently done
manually, but it is also possible to use this macro with compile time constants
and so it can be used to initialize static data. So we can use it to initialize
the mask field of a ENUM control during its declaration.

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: wm8960: add support for big-endian audio samples
Timur Tabi [Fri, 14 Sep 2012 21:14:37 +0000 (16:14 -0500)]
ASoC: wm8960: add support for big-endian audio samples

PowerPC ASoC drivers frequently use the _BE variants of the SNDRV_PCM_FORMAT
macros, so we need to look for those as well.

Signed-off-by: Timur Tabi <timur@freescale.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: fsl: remove unnecessary call to dma_unmap_single
Timur Tabi [Fri, 14 Sep 2012 21:14:36 +0000 (16:14 -0500)]
ASoC: fsl: remove unnecessary call to dma_unmap_single

Remove a call to dma_unmap_single() from the PowerPC ASoC DMA driver.  The
buffer is allocated and not actually mapped, so the unmap call doesn't
make sense.  It was probably left over from some early version of the driver.

This bug was unnoticed for so long because the DMA mapping functions normally
don't do anything on PowerPC.

Signed-off-by: Timur Tabi <timur@freescale.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: fsl: use snd_soc_register_card to register the card
Timur Tabi [Fri, 14 Sep 2012 21:14:34 +0000 (16:14 -0500)]
ASoC: fsl: use snd_soc_register_card to register the card

Use snd_soc_register_card() instead of platform_device_alloc("soc-audio")
to register the sound card from the machine drivers.  The use of
platform_device_alloc is deprecated.

Although several other drivers still use platform_device_alloc(), the
Freescale drivers were not using it to pass driver data.  Instead of fixing
the driver data usage, it's better to replace the deprecated code.

Signed-off-by: Timur Tabi <timur@freescale.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoASoC: mx27vis: retrieve gpio numbers from platform_data
Shawn Guo [Mon, 17 Sep 2012 05:34:31 +0000 (13:34 +0800)]
ASoC: mx27vis: retrieve gpio numbers from platform_data

Rather than including mach/iomux-mx27.h to define gpio numbers and set
up the pins, the patch moves all these into machine code and has the
gpio numbers passed to driver via platform_data.  As the result, we
can remove the mach/iomux-mx27.h inclusion from driver.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
Acked-by: Javier Martin <javier.martin@vista-silicon.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoLinux 3.6-rc6
Linus Torvalds [Sun, 16 Sep 2012 21:58:51 +0000 (14:58 -0700)]
Linux 3.6-rc6

12 years agoMerge tag 'mfd-for-linus-3.6-2' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 16 Sep 2012 20:22:21 +0000 (13:22 -0700)]
Merge tag 'mfd-for-linus-3.6-2' of git://git./linux/kernel/git/sameo/mfd-2.6

Pull mfd fixes from Samuel Ortiz:
 "This is the remaining MFD fixes for 3.6, with 5 pending fixes:

   - A tps65217 build error fix.
   - A lcp_ich regression fix caused by the MFD driver failing to
     initialize the watchdog sub device due to ACPI conflicts.
   - 2 MAX77693 interrupt handling bug fixes.
   - An MFD core fix, adding an IRQ domain argument to the MFD device
     addition API in order to prevent silent and potentially harmful
     remapping behaviour changes for drivers supporting non-DT
     platforms."

* tag 'mfd-for-linus-3.6-2' of git://git.kernel.org/pub/scm/linux/kernel/git/sameo/mfd-2.6:
  mfd: MAX77693: Fix NULL pointer error when initializing irqs
  mfd: MAX77693: Fix interrupt handling bug
  mfd: core: Push irqdomain mapping out into devices
  mfd: lpc_ich: Fix a 3.5 kernel regression for iTCO_wdt driver
  mfd: Move tps65217 regulator plat data handling to regulator

12 years agoMerge tag 'for-3.6-rc6' of git://gitorious.org/linux-pwm/linux-pwm
Linus Torvalds [Sun, 16 Sep 2012 20:20:43 +0000 (13:20 -0700)]
Merge tag 'for-3.6-rc6' of git://gitorious.org/linux-pwm/linux-pwm

Pull pwm fixes from Thierry Reding:
 "While this comes a bit later than I had wished, both patches are
  rather minor and touch only new drivers so I think these are still
  safe for merging."

* tag 'for-3.6-rc6' of git://gitorious.org/linux-pwm/linux-pwm:
  pwm: pwm-tiehrpwm: Fix conflicting channel period setting
  pwm: pwm-tiecap: Disable APWM mode after configure

12 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending
Linus Torvalds [Sun, 16 Sep 2012 20:00:36 +0000 (13:00 -0700)]
Merge git://git./linux/kernel/git/nab/target-pending

Pull scsi target fixes from Nicholas Bellinger:
 "Here is the current set of target-pending fixes headed for v3.6-final

  The main parts of this series include bug-fixes from Paolo Bonzini to
  address an use-after-free bug in pSCSI sense exception handling, along
  with addressing some long-standing bugs wrt the handling of zero-
  length SCSI CDB payloads also specific to pSCSI pass-through device
  backends."

* git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending:
  target: go through normal processing for zero-length REQUEST_SENSE
  target: support zero allocation length in REQUEST SENSE
  target: support zero-size allocation lengths in transport_kmap_data_sg
  target: fail REPORT LUNS with less than 16 bytes of payload
  target: report too-small parameter lists everywhere
  target: go through normal processing for zero-length PSCSI commands
  target: fix use-after-free with PSCSI sense data
  target: simplify code around transport_get_sense_data
  target: move transport_get_sense_data
  target: Check idr_get_new return value in iscsi_login_zero_tsih_s1
  target: Fix ->data_length re-assignment bug with SCSI overflow

12 years agoMerge tag 'pm-for-3.6-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael...
Linus Torvalds [Sun, 16 Sep 2012 19:59:42 +0000 (12:59 -0700)]
Merge tag 'pm-for-3.6-rc6' of git://git./linux/kernel/git/rafael/linux-pm

Pull power management fixes from Rafael J. Wysocki:
 "Three ACPI device power management fixes related to checking and
  setting device power states."

* tag 'pm-for-3.6-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  ACPI / PM: Use KERN_DEBUG when no power resources are found
  ACPI / PM: Fix resource_lock dead lock in acpi_power_on_device
  ACPI / PM: Infer parent power state from child if unknown, v2

12 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux...
Linus Torvalds [Sun, 16 Sep 2012 19:58:44 +0000 (12:58 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/mason/linux-btrfs

Pull a btrfs revert from Chris Mason:
 "My for-linus branch has one revert in the new quota code.

  We're building up more fixes at etc for the next merge window, but I'm
  keeping them out unless they are bigger regressions or have a huge
  impact."

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs:
  Revert "Btrfs: fix some error codes in btrfs_qgroup_inherit()"

12 years agoMerge tag 'sound-3.6' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound
Linus Torvalds [Sun, 16 Sep 2012 19:57:59 +0000 (12:57 -0700)]
Merge tag 'sound-3.6' of git://git./linux/kernel/git/tiwai/sound

Pull more sound fixes from Takashi Iwai:
 "Yet more (a bunch of) small fixes that slipped from the previous pull
  request.  Most of commits are pending ASoC fixes, all of which are
  fairly trivial commits."

* tag 'sound-3.6' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
  ASoC: wm8904: correct the index
  ALSA: hda - Yet another position_fix quirk for ASUS machines
  ASoC: tegra: fix maxburst settings in dmaengine code
  ASoC: samsung dma - Don't indicate support for pause/resume.
  ASoC: mc13783: Remove mono support
  ASoC: arizona: Fix typo in 44.1kHz rates
  ASoC: spear: correct the check for NULL dma_buffer pointer
  sound: tegra_alc5632: remove HP detect GPIO inversion
  ASoC: atmel-ssc: include linux/io.h for raw io
  ASoC: dapm: Don't force card bias level to be updated
  ASoC: dapm: Make sure we update the bias level for CODECs with no op
  ASoC: am3517evm: fix error return code
  ASoC: ux500_msp_i2s: better use devm functions and fix error return code
  ASoC: imx-sgtl5000: fix error return code

12 years agoRevert "sched: Improve scalability via 'CPU buddies', which withstand random perturba...
Linus Torvalds [Sun, 16 Sep 2012 19:29:43 +0000 (12:29 -0700)]
Revert "sched: Improve scalability via 'CPU buddies', which withstand random perturbations"

This reverts commit 970e178985cadbca660feb02f4d2ee3a09f7fdda.

Nikolay Ulyanitsky reported thatthe 3.6-rc5 kernel has a 15-20%
performance drop on PostgreSQL 9.2 on his machine (running "pgbench").

Borislav Petkov was able to reproduce this, and bisected it to this
commit 970e178985ca ("sched: Improve scalability via 'CPU buddies' ...")
apparently because the new single-idle-buddy model simply doesn't find
idle CPU's to reschedule on aggressively enough.

Mike Galbraith suspects that it is likely due to the user-mode spinlocks
in PostgreSQL not reacting well to preemption, but we don't really know
the details - I'll just revert the commit for now.

There are hopefully other approaches to improve scheduler scalability
without it causing these kinds of downsides.

Reported-by: Nikolay Ulyanitsky <lystor@gmail.com>
Bisected-by: Borislav Petkov <bp@alien8.de>
Acked-by: Mike Galbraith <efault@gmx.de>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agomfd: MAX77693: Fix NULL pointer error when initializing irqs
Chanwoo Choi [Tue, 21 Aug 2012 06:16:23 +0000 (15:16 +0900)]
mfd: MAX77693: Fix NULL pointer error when initializing irqs

This patch initialize register map of MUIC device because mfd driver
of Maxim MAX77693 use regmap-muic instance of MUIC device when irqs of
Maxim MAX77693 is initialized before call max77693-muic probe() function.

Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Signed-off-by: Myungjoo Ham <myungjoo.ham@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Reported-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
12 years agomfd: MAX77693: Fix interrupt handling bug
Chanwoo Choi [Tue, 21 Aug 2012 06:15:52 +0000 (15:15 +0900)]
mfd: MAX77693: Fix interrupt handling bug

This patch fix bug related to interrupt handling for MAX77693 devices.
- Unmask interrupt masking bit for charger/flash/muic to revolve
that interrupt isn't happened when external connector is attached.
- Fix wrong regmap instance when muic interrupt is happened.

This patch were discussed and confirm discussion about this patch on below url:
http://lkml.org/lkml/2012/7/16/118

Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Signed-off-by: Myungjoo Ham <myungjoo.ham@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
12 years agomfd: core: Push irqdomain mapping out into devices
Mark Brown [Tue, 11 Sep 2012 07:16:36 +0000 (15:16 +0800)]
mfd: core: Push irqdomain mapping out into devices

Currently the MFD core supports remapping MFD cell interrupts using an
irqdomain but only if the MFD is being instantiated using device tree
and only if the device tree bindings use the pattern of registering IPs
in the device tree with compatible properties.  This will be actively
harmful for drivers which support non-DT platforms and use this pattern
for their DT bindings as it will mean that the core will silently change
remapping behaviour and it is also limiting for drivers which don't do
DT with this particular pattern.  There is also a potential fragility if
there are interrupts not associated with MFD cells and all the cells are
omitted from the device tree for some reason.

Instead change the code to take an IRQ domain as an optional argument,
allowing drivers to take the decision about the parent domain for their
interrupts.  The one current user of this feature is ab8500-core, it has
the domain lookup pushed out into the driver.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
12 years agoMerge tag 'asoc-3.6' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound...
Takashi Iwai [Sat, 15 Sep 2012 06:24:42 +0000 (08:24 +0200)]
Merge tag 'asoc-3.6' of git://git./linux/kernel/git/broonie/sound into for-linus

ASoC: Updates for 3.6

A bigger set of updates than I'm entirely comfortable with - things
backed up a bit due to travel.  As ever the majority of these are small,
focused updates for specific drivers though there are a couple of core
changes.  There's been good exposure in -next.

The AT91 patch fixes a build break.

12 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/steve/gfs2-3.0-fixes
Linus Torvalds [Sat, 15 Sep 2012 01:05:14 +0000 (18:05 -0700)]
Merge git://git./linux/kernel/git/steve/gfs2-3.0-fixes

Pull GFS2 fixes from Steven Whitehouse:
 "Here are three GFS2 fixes for the current kernel tree.  These are all
  related to the block reservation code which was added at the merge
  window.  That code will be getting an update at the forthcoming merge
  window too.  In the mean time though there are a few smaller issues
  which should be fixed.

  The first patch resolves an issue with write sizes of greater than 32
  bits with the size hinting code.  The second ensures that the
  allocation data structure is initialised when using xattrs and the
  third takes into account allocations which may have been made by other
  nodes which affect a reservation on the local node."

* git://git.kernel.org/pub/scm/linux/kernel/git/steve/gfs2-3.0-fixes:
  GFS2: Take account of blockages when using reserved blocks
  GFS2: Fix missing allocation data for set/remove xattr
  GFS2: Make write size hinting code common

12 years agoMerge branch 'for_linus' of git://cavan.codon.org.uk/platform-drivers-x86
Linus Torvalds [Sat, 15 Sep 2012 00:59:35 +0000 (17:59 -0700)]
Merge branch 'for_linus' of git://cavan.codon.org.uk/platform-drivers-x86

Pull x86 platform driver updates from Matthew Garrett:
 "A few small updates for 3.6 - a trivial regression fix and a couple of
  conformance updates for the gmux driver, plus some tiny fixes for
  asus-wmi, eeepc-laptop and thinkpad_acpi."

* 'for_linus' of git://cavan.codon.org.uk/platform-drivers-x86:
  thinkpad_acpi: buffer overflow in fan_get_status()
  eeepc-laptop: fix device reference count leakage in eeepc_rfkill_hotplug()
  platform/x86: fix asus_laptop.wled_type description
  asus-laptop: HRWS/HWRS typo
  drivers-platform-x86: remove useless #ifdef CONFIG_ACPI_VIDEO
  apple-gmux: Fix port address calculation in gmux_pio_write32()
  apple-gmux: Fix index read functions
  apple-gmux: Obtain version info from indexed gmux

12 years agoMerge branch 'i2c-embedded/for-current' of git://git.pengutronix.de/git/wsa/linux
Linus Torvalds [Sat, 15 Sep 2012 00:55:57 +0000 (17:55 -0700)]
Merge branch 'i2c-embedded/for-current' of git://git.pengutronix.de/git/wsa/linux

Pull i2c embedded fixes from Wolfram Sang:
 "The last bunch of (typical) i2c-embedded driver fixes for 3.6.

  Also update the MAINTAINERS file to point to my tree since people keep
  asking where to find their patches."

* 'i2c-embedded/for-current' of git://git.pengutronix.de/git/wsa/linux:
  i2c: algo: pca: Fix mode selection for PCA9665
  MAINTAINERS: fix tree for current i2c-embedded development
  i2c: mxs: correctly setup speed for non devicetree
  i2c: pnx: Fix read transactions of >= 2 bytes
  i2c: pnx: Fix bit definitions

12 years agoMerge tag 'ecryptfs-3.6-rc6-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 15 Sep 2012 00:53:55 +0000 (17:53 -0700)]
Merge tag 'ecryptfs-3.6-rc6-fixes' of git://git./linux/kernel/git/tyhicks/ecryptfs

Pull ecryptfs fixes from Tyler Hicks:

 - Fixes a regression, introduced in 3.6-rc1, when a file is closed
   before its shared memory mapping is dirtied and unmapped.  The lower
   file was being released when the eCryptfs file was closed and the
   dirtied pages could not be written out.
 - Adds a call to the lower filesystem's ->flush() from
   ecryptfs_flush().
 - Fixes a regression, introduced in 2.6.39, when a file is renamed on
   top of another file.  The target file's inode was not being evicted
   and the space taken by the file was not reclaimed until eCryptfs was
   unmounted.

* tag 'ecryptfs-3.6-rc6-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tyhicks/ecryptfs:
  eCryptfs: Copy up attributes of the lower target inode after rename
  eCryptfs: Call lower ->flush() from ecryptfs_flush()
  eCryptfs: Write out all dirty pages just before releasing the lower file

12 years agoMerge branch 'fixes-for-3.6' of git://git.linaro.org/people/mszyprowski/linux-dma...
Linus Torvalds [Sat, 15 Sep 2012 00:53:11 +0000 (17:53 -0700)]
Merge branch 'fixes-for-3.6' of git://git.linaro.org/people/mszyprowski/linux-dma-mapping

Pull one more DMA-mapping fix from Marek Szyprowski:
 "This patch fixes very subtle bug (typical off-by-one error) which
  might appear in very rare circumstances."

* 'fixes-for-3.6' of git://git.linaro.org/people/mszyprowski/linux-dma-mapping:
  arm: mm: fix DMA pool affiliation check

12 years agoMerge tag 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck...
Linus Torvalds [Sat, 15 Sep 2012 00:52:29 +0000 (17:52 -0700)]
Merge tag 'hwmon-for-linus' of git://git./linux/kernel/git/groeck/linux-staging

Pull hwmon fixes from Guenter Roeck:
 "Fix word size register read and write operations in ina2xx driver, and
  initialize uninitialized structure elements in twl4030-madc-hwmon
  driver."

* tag 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging:
  hwmon: (ina2xx) Fix word size register read and write operations
  hwmon: (twl4030-madc-hwmon) Initialize uninitialized structure elements

12 years agoMerge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux
Linus Torvalds [Sat, 15 Sep 2012 00:51:10 +0000 (17:51 -0700)]
Merge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux

Pull drm fixes from Dave Airlie:
 "I realise this a bit bigger than I would want at this point.

  Exynos is a large chunk, I got them to half what they wanted already,
  and hey its ARM based, so not going to hurt many people.

  Radeon has only two fixes, but the PLL fixes were a bit bigger, but
  required for a lot of scenarios, the fence fix is really urgent.

  vmwgfx: I've pulled in a dumb ioctl support patch that I was going to
  shove in later and cc stable, but we need it asap, its mainly to stop
  mesa growing a really ugly dependency in userspace to run stuff on
  vmware, and if I don't stick it in the kernel now, everyone will have
  to ship ugly userspace libs to workaround it.

  nouveau: single urgent fix found in F18 testing, causes X to not start
  properly when f18 plymouth is used

  i915: smattering of fixes and debug quieting

  gma500: single regression fix

  So as I said a bit large, but its fairly well scattered and its all
  stuff I'll be shipping in F18's 3.6 kernel."

* 'drm-fixes' of git://people.freedesktop.org/~airlied/linux: (26 commits)
  drm/nouveau: fix booting with plymouth + dumb support
  drm/radeon: make 64bit fences more robust v3
  drm/radeon: rework pll selection (v3)
  drm: Drop the NV12M and YUV420M formats
  drm/exynos: remove DRM_FORMAT_NV12M from plane module
  drm/exynos: fix double call of drm_prime_(init/destroy)_file_private
  drm/exynos: add dummy support for dmabuf-mmap
  drm/exynos: Add missing braces around sizeof in exynos_mixer.c
  drm/exynos: Add missing braces around sizeof in exynos_hdmi.c
  drm/exynos: Make g2d_pm_ops static
  drm/exynos: Add dependency for G2D in Kconfig
  drm/exynos: fixed page align bug.
  drm/exynos: Use ERR_CAST inlined function instead of ERR_PTR(PTR_ERR(.. [1]
  drm/exynos: Use devm_* functions in exynos_drm_g2d.c file
  drm/exynos: Use devm_kzalloc in exynos_drm_hdmi.c file
  drm/exynos: Use devm_kzalloc in exynos_drm_vidi.c file
  drm/exynos: Remove redundant check in exynos_drm_fimd.c file
  drm/exynos: Remove redundant check in exynos_hdmi.c file
  vmwgfx: add dumb ioctl support
  gma500: Fix regression on Oaktrail devices
  ...

12 years agoMerge branch 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 15 Sep 2012 00:44:52 +0000 (17:44 -0700)]
Merge branch 'sched-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull scheduler fixes from Ingo Molnar:
 "Smaller fixlets"

* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  sched: Fix kernel-doc warnings in kernel/sched/fair.c
  sched: Unthrottle rt runqueues in __disable_runtime()
  sched: Add missing call to calc_load_exit_idle()
  sched: Fix load avg vs cpu-hotplug

12 years agoMerge branch 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 15 Sep 2012 00:43:45 +0000 (17:43 -0700)]
Merge branch 'perf-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull perf fixes from Ingo Molnar:
 "This tree includes various fixes"

Ingo really needs to improve on the whole "explain git pull" part.
"Various fixes" indeed.

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf/hwpb: Invoke __perf_event_disable() if interrupts are already disabled
  perf/x86: Enable Intel Cedarview Atom suppport
  perf_event: Switch to internal refcount, fix race with close()
  oprofile, s390: Fix uninitialized memory access when writing to oprofilefs
  perf/x86: Fix microcode revision check for SNB-PEBS

12 years agoMerge branch 'core-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 15 Sep 2012 00:43:14 +0000 (17:43 -0700)]
Merge branch 'core-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull a core sparse warning fix from Ingo Molnar

* 'core-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  mm/memblock: Use NULL instead of 0 for pointers

12 years agoRevert "Btrfs: fix some error codes in btrfs_qgroup_inherit()"
Chris Mason [Sat, 15 Sep 2012 00:06:30 +0000 (20:06 -0400)]
Revert "Btrfs: fix some error codes in btrfs_qgroup_inherit()"

This reverts commit 5986802c2fcc754040bb7ed95f30bb16c4a843b7.

Both paths are not error paths but regular cases where non-qgroup
subvols are involved.

Signed-off-by: Chris Mason <chris.mason@fusionio.com>
12 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
Linus Torvalds [Fri, 14 Sep 2012 22:34:07 +0000 (15:34 -0700)]
Merge git://git./linux/kernel/git/davem/net

Pull networking fixes from David Miller:

 1) Use after free and new device IDs in bluetooth from Andre Guedes,
    Yevgeniy Melnichuk, Gustavo Padovan, and Henrik Rydberg.

 2) Fix crashes with short packet lengths and VLAN in pktgen, from
    Nishank Trivedi.

 3) mISDN calls flush_work_sync() with locks held, fix from Karsten
    Keil.

 4) Packet scheduler gred parameters are reported to userspace
    improperly scaled, and WRED idling is not performed correctly.  All
    from David Ward.

 5) Fix TCP socket refcount problem in ipv6, from Julian Anastasov.

 6) ibmveth device has RX queue alignment requirements which are not
    being explicitly met resulting in sporadic failures, fix from
    Santiago Leon.

 7) Netfilter needs to take care when interpreting sockets attached to
    socket buffers, they could be time-wait minisockets.  Fix from Eric
    Dumazet.

 8) sock_edemux() has the same issue as netfilter did in #7 above, fix
    from Eric Dumazet.

 9) Avoid infinite loops in CBQ scheduler with some configurations, from
    Eric Dumazet.

10) Deal with "Reflection scan: an Off-Path Attack on TCP", from Jozsef
    Kadlecsik.

11) SCTP overcharges socket for TX packets, fix from Thomas Graf.

12) CODEL packet scheduler should not reset it's state every time it
    builds a new flow, fix from Eric Dumazet.

13) Fix memory leak in nl80211, from Wei Yongjun.

14) NETROM doesn't check skb_copy_datagram_iovec() return values, from
    Alan Cox.

15) l2tp ethernet was using sizeof(ETH_HLEN) instead of plain ETH_HLEN,
    oops.  From Eric Dumazet.

16) Fix selection of ath9k chips on which PA linearization and AM2PM
    predistoration are used, from Felix Fietkau.

17) Flow steering settings in mlx4 driver need to be validated properly,
    from Hadar Hen Zion.

18) bnx2x doesn't show the correct link duplex setting, from Yaniv
    Rosner.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (75 commits)
  pktgen: fix crash with vlan and packet size less than 46
  bnx2x: Add missing afex code
  bnx2x: fix registers dumped
  bnx2x: correct advertisement of pause capabilities
  bnx2x: display the correct duplex value
  bnx2x: prevent timeouts when using PFC
  bnx2x: fix stats copying logic
  bnx2x: Avoid sending multiple statistics queries
  net: qmi_wwan: call subdriver with control intf only
  net_sched: gred: actually perform idling in WRED mode
  net_sched: gred: fix qave reporting via netlink
  net_sched: gred: eliminate redundant DP prio comparisons
  net_sched: gred: correct comment about qavg calculation in RIO mode
  mISDN: Fix wrong usage of flush_work_sync while holding locks
  netfilter: log: Fix log-level processing
  net-sched: sch_cbq: avoid infinite loop
  net: qmi_wwan: fix Gobi device probing for un2430
  net: fix net/core/sock.c build error
  ixp4xx_hss: fix build failure due to missing linux/module.h inclusion
  caif: move the dereference below the NULL test
  ...

12 years agoMerge tag 'usb-3.6-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb
Linus Torvalds [Fri, 14 Sep 2012 21:54:57 +0000 (14:54 -0700)]
Merge tag 'usb-3.6-rc6' of git://git./linux/kernel/git/gregkh/usb

Pull USB patches from Greg Kroah-Hartman:
 "Here are a number of USB patches, a bit more than I normally like this
  late in the -rc series, but given people's vacations (myself
  included), and the kernel summit, it seems to have happened this way.

  All are tiny, but they add up.  A number of gadget and xhci fixes, and
  a few new device ids.  All have been tested in linux-next.

Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>"
* tag 'usb-3.6-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb: (33 commits)
  usb: chipidea: udc: don't stall endpoint if request list is empty in isr_tr_complete_low
  usb: chipidea: cleanup dma_pool if udc_start() fails
  usb: chipidea: udc: fix error path in udc_start()
  usb: chipidea: udc: add pullup fuction, needed by the uvc gadget
  usb: chipidea: udc: fix setup of endpoint maxpacket size
  USB: option: replace ZTE K5006-Z entry with vendor class rule
  EHCI: Update qTD next pointer in QH overlay region during unlink
  USB: cdc-wdm: fix wdm_find_device* return value
  USB: ftdi_sio: do not claim CDC ACM function
  usb: dwc3: gadget: fix pending isoc handling
  usb: renesas_usbhs: fixup DMA transport data alignment
  usb: gadget: at91udc: Don't check for ep->ep.desc
  usb: gadget: at91udc: don't overwrite driver data
  usb: dwc3: core: fix incorrect usage of resource pointer
  usb: musb: musbhsdma: fix IRQ check
  usb: musb: tusb6010: fix error path in tusb_probe()
  usb: musb: host: fix for musb_start_urb Oops
  usb: gadget: dummy_hcd: add support for USB_DT_BOS on rh
  usb: gadget: dummy_hcd: fixup error probe path
  usb: gadget: s3c-hsotg.c: fix error return code
  ...

12 years agoMerge tag 'tty-3.6-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty
Linus Torvalds [Fri, 14 Sep 2012 21:54:29 +0000 (14:54 -0700)]
Merge tag 'tty-3.6-rc6' of git://git./linux/kernel/git/gregkh/tty

Pull TTY fixes from Greg Kroah-Hartman:
 "Here are 2 tiny patches for a serial driver to resolve issues that
  people have reported with the 3.6-rc tree.

  Both of these have been in the linux-next tree for a while now.

Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>"
* tag 'tty-3.6-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty:
  tty: serial: imx: don't reinit clock in imx_setup_ufcr()
  tty: serial: imx: console write routing is unsafe on SMP

12 years agoMerge tag 'staging-3.6-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh...
Linus Torvalds [Fri, 14 Sep 2012 21:53:51 +0000 (14:53 -0700)]
Merge tag 'staging-3.6-rc6' of git://git./linux/kernel/git/gregkh/staging

Pull staging tree fixes from Greg Kroah-Hartman:
 "Here are a few staging tree fixes for problems that have been
  reported.

  Nothing major, just a number of tiny driver fixes.  All of these have
  been in the linux-next tree for a while.

Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>"
* tag 'staging-3.6-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging:
  drm/omap: add more new timings fields
  drm/omap: update for interlaced
  staging: r8712u: fix bug in r8712_recv_indicatepkt()
  staging: zcache: fix cleancache race condition with shrinker
  Staging: Android alarm: IOCTL command encoding fix
  staging: vt6656: [BUG] - Failed connection, incorrect endian.
  staging: ozwpan: fix memcmp() test in oz_set_active_pd()
  staging: wlan-ng: Fix problem with wrong arguments
  staging: comedi: das08: Correct AO output for das08jr-16-ao
  staging: comedi: das08: Correct AI encoding for das08jr-16-ao
  staging: comedi: das08: Fix PCI ref count
  staging: comedi: amplc_pci230: Fix PCI ref count
  staging: comedi: amplc_pc263: Fix PCI ref count
  staging: comedi: amplc_pc236: Fix PCI ref count
  staging: comedi: amplc_dio200: Fix PCI ref count
  staging: comedi: amplc_pci224: Fix PCI ref count
  drivers/iio/adc/at91_adc.c: adjust inconsistent IS_ERR and PTR_ERR
  staging iio: fix potential memory leak in lis3l02dq_ring.c
  staging:iio: prevent divide by zero bugs

12 years agoMerge tag 'driver-core-3.6-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Fri, 14 Sep 2012 21:53:22 +0000 (14:53 -0700)]
Merge tag 'driver-core-3.6-rc6' of git://git./linux/kernel/git/gregkh/driver-core

Pull driver core fix from Greg Kroah-Hartman:
 "Here is one fix for 3.6-rc6 for the kobject.h file.

  It fixes a reported oops if CONFIG_HOTPLUG is disabled.  It's been in
  the linux-next tree for a while now.

Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>"
* tag 'driver-core-3.6-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core:
  kobject: fix oops with "input0: bad kobj_uevent_env content in show_uevent()"

12 years agovfs: make O_PATH file descriptors usable for 'fstat()'
Linus Torvalds [Fri, 14 Sep 2012 21:48:21 +0000 (14:48 -0700)]
vfs: make O_PATH file descriptors usable for 'fstat()'

We already use them for openat() and friends, but fstat() also wants to
be able to use O_PATH file descriptors.  This should make it more
directly comparable to the O_SEARCH of Solaris.

Note that you could already do the same thing with "fstatat()" and an
empty path, but just doing "fstat()" directly is simpler and faster, so
there is no reason not to just allow it directly.

See also commit 332a2e1244bd, which did the same thing for fchdir, for
the same reasons.

Reported-by: ольга крыжановская <olga.kryzhanovska@gmail.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: stable@kernel.org # O_PATH introduced in 3.0+
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agoACPI / PM: Use KERN_DEBUG when no power resources are found
Aaron Lu [Fri, 14 Sep 2012 18:54:44 +0000 (20:54 +0200)]
ACPI / PM: Use KERN_DEBUG when no power resources are found

commit a606dac368eed5696fb38e16b1394f1d049c09e9 adds support to link
devices which have _PRx, if a device does not have _PRx, a warning
message will be printed.

This commit is for ZPODD on Intel ZPODD capable platforms, on other
platforms, it has no problem if there is no power resource for this
device, so a warning here is not appropriate, change it to debug.

Reported-by: Borislav Petkov <bp@amd64.org>
Signed-off-by: Aaron Lu <aaron.lu@intel.com>
Cc: stable@vger.kernel.org
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
12 years agoASoC: wm8904: correct the index
Bo Shen [Fri, 14 Sep 2012 08:09:09 +0000 (16:09 +0800)]
ASoC: wm8904: correct the index

Signed-off-by: Bo Shen <voice.shen@atmel.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
12 years agoeCryptfs: Copy up attributes of the lower target inode after rename
Tyler Hicks [Thu, 13 Sep 2012 19:00:56 +0000 (12:00 -0700)]
eCryptfs: Copy up attributes of the lower target inode after rename

After calling into the lower filesystem to do a rename, the lower target
inode's attributes were not copied up to the eCryptfs target inode. This
resulted in the eCryptfs target inode staying around, rather than being
evicted, because i_nlink was not updated for the eCryptfs inode. This
also meant that eCryptfs didn't do the final iput() on the lower target
inode so it stayed around, as well. This would result in a failure to
free up space occupied by the target file in the rename() operation.
Both target inodes would eventually be evicted when the eCryptfs
filesystem was unmounted.

This patch calls fsstack_copy_attr_all() after the lower filesystem
does its ->rename() so that important inode attributes, such as i_nlink,
are updated at the eCryptfs layer. ecryptfs_evict_inode() is now called
and eCryptfs can drop its final reference on the lower inode.

http://launchpad.net/bugs/561129

Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Tested-by: Colin Ian King <colin.king@canonical.com>
Cc: <stable@vger.kernel.org> [2.6.39+]
12 years agoeCryptfs: Call lower ->flush() from ecryptfs_flush()
Tyler Hicks [Thu, 13 Sep 2012 01:38:00 +0000 (18:38 -0700)]
eCryptfs: Call lower ->flush() from ecryptfs_flush()

Since eCryptfs only calls fput() on the lower file in
ecryptfs_release(), eCryptfs should call the lower filesystem's
->flush() from ecryptfs_flush().

If the lower filesystem implements ->flush(), then eCryptfs should try
to flush out any dirty pages prior to calling the lower ->flush(). If
the lower filesystem does not implement ->flush(), then eCryptfs has no
need to do anything in ecryptfs_flush() since dirty pages are now
written out to the lower filesystem in ecryptfs_release().

Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
12 years agoeCryptfs: Write out all dirty pages just before releasing the lower file
Tyler Hicks [Thu, 13 Sep 2012 01:02:46 +0000 (18:02 -0700)]
eCryptfs: Write out all dirty pages just before releasing the lower file

Fixes a regression caused by:

821f749 eCryptfs: Revert to a writethrough cache model

That patch reverted some code (specifically, 32001d6f) that was
necessary to properly handle open() -> mmap() -> close() -> dirty pages
-> munmap(), because the lower file could be closed before the dirty
pages are written out.

Rather than reapplying 32001d6f, this approach is a better way of
ensuring that the lower file is still open in order to handle writing
out the dirty pages. It is called from ecryptfs_release(), while we have
a lock on the lower file pointer, just before the lower file gets the
final fput() and we overwrite the pointer.

https://launchpad.net/bugs/1047261

Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Reported-by: Artemy Tregubenko <me@arty.name>
Tested-by: Artemy Tregubenko <me@arty.name>
Tested-by: Colin Ian King <colin.king@canonical.com>