linux-stable.git/drivers/media, branch v3.7.9

media: pwc-if: must check vb2_queue_init() success

2013-02-14T18:48:01+00:00

commit eda94710d6502672c5ee7de198fa78a63ddfae3a upstream.

drivers/media/usb/pwc/pwc-if.c: In function 'usb_pwc_probe':
drivers/media/usb/pwc/pwc-if.c:1003:16: warning: ignoring return value of 'vb2_queue_init', declared with attribute warn_unused_result [-Wunused-result]
In the past, it used to have a logic there at queue init that would
BUG() on errors. This logic got removed. Drivers are now required
to explicitly handle the queue initialization errors, or very bad
things may happen.

Signed-off-by: Mauro Carvalho Chehab 
Signed-off-by: Paul Bolle

media: gspca_kinect: add Kinect for Windows USB id

2013-01-28T04:49:01+00:00

commit 98fd485795db064d0885150e2c0c7f296d8fe06e upstream.

Add the USB ID for the Kinect for Windows RGB camera so it can be used
with the gspca_kinect driver.

Signed-off-by: Jacob Schloss 
Signed-off-by: Antonio Ospite 
Signed-off-by: Hans de Goede 
Signed-off-by: Mauro Carvalho Chehab 
Signed-off-by: Greg Kroah-Hartman

[media] s5p-mfc: Handle multi-frame input buffer

2012-11-26T20:43:27+00:00

When one input buffer has multiple frames, it should be fed
again to the hardware with the remaining bytes. Removed the
check for P frame in this scenario as this condition can come with
all frame types.

Signed-off-by: Arun Kumar K 
Signed-off-by: ARUN MANKUZHI 
Acked-by: Kamil Debski 
Signed-off-by: Sylwester Nawrocki 
Signed-off-by: Mauro Carvalho Chehab

[media] s5p-mfc: Bug fix of timestamp/timecode copy mechanism

2012-11-26T20:43:24+00:00

Modified the function s5p_mfc_get_dec_y_adr_v6 to access the
decode Y address register instead of display Y address.

Signed-off-by: Sunil Mazhavanchery 
Signed-off-by: Arun Kumar K 
Acked-by: Kamil Debski 
Signed-off-by: Sylwester Nawrocki 
Signed-off-by: Mauro Carvalho Chehab

[media] exynos-gsc: Add missing video device vfl_dir flag initialization

2012-11-26T20:43:22+00:00

vfl_dir should be set to VFL_DIR_M2M so valid ioctls for this
mem-to-mem device can be properly determined in the v4l2 core.

Signed-off-by: Sylwester Nawrocki 
Signed-off-by: Mauro Carvalho Chehab

[media] exynos-gsc: Fix settings for input and output image RGB type

2012-11-26T20:43:14+00:00

Macros used to set input and output RGB type aren't correct.
Updating the macros as per register manual.

Signed-off-by: Shaik Ameer Basha 
Signed-off-by: Sylwester Nawrocki 
Signed-off-by: Mauro Carvalho Chehab

[media] exynos-gsc: Don't use mutex_lock_interruptible() in device release()

2012-11-26T20:43:08+00:00

Use uninterruptible mutex_lock in the release() file op to make
sure all resources are properly freed when a process is being
terminated. Returning -ERESTARTSYS has no effect for a terminating
process and this may cause driver resources not to be released.

Reported-by: Marek Szyprowski 
Signed-off-by: Sylwester Nawrocki 
Signed-off-by: Kyungmin Park 
Signed-off-by: Mauro Carvalho Chehab

[media] fimc-lite: Don't use mutex_lock_interruptible() in device release()

2012-11-26T20:31:20+00:00

Use uninterruptible mutex_lock in the release() file op to make
sure all resources are properly freed when a process is being
terminated. Returning -ERESTARTSYS has no effect for a terminating
process and this may cause driver resources not to be released.
This patch is required for stable kernels v3.5+.

Reported-by: Marek Szyprowski 
Signed-off-by: Sylwester Nawrocki 
Signed-off-by: Kyungmin Park 
Signed-off-by: Mauro Carvalho Chehab

[media] s5p-fimc: Don't use mutex_lock_interruptible() in device release()

2012-11-26T20:30:19+00:00

Use uninterruptible mutex_lock in the release() file op to make
sure all resources are properly freed when a process is being
terminated. Returning -ERESTARTSYS has no effect for a terminating
process and this caused driver resources not to be released. Not
releasing the buffer queue also prevented other drivers to free
memory, e.g. in MMAP -> USERPTR scenario.
This patch is required for stable kernels v3.6+.

Reported-by: Kamil Debski 
Reported-by: Marek Szyprowski 
Signed-off-by: Sylwester Nawrocki 
Signed-off-by: Kyungmin Park 
Signed-off-by: Mauro Carvalho Chehab

[media] s5p-fimc: Prevent race conditions during subdevs registration

2012-11-26T20:29:31+00:00

Make sure when fimc and fimc-lite capture video node is registered
it has valid pipeline_ops assigned to it. Otherwise when a video
node is opened right after is was registered there, might be an
attempt to use ops that are just being assigned, after function
v4l2_device_register_subdev() returns.

Signed-off-by: Sylwester Nawrocki 
Signed-off-by: Kyungmin Park 
Signed-off-by: Mauro Carvalho Chehab