linux-stable.git/drivers/media, branch v3.12

[media] saa7134: Fix crash when device is closed before streamoff

2013-10-14T09:37:00+00:00

pm_qos_remove_request was not called on video_release, resulting in the PM
core's list of requests being corrupted when the file handle was freed.
This has no immediate symptoms, but later in operation, the kernel will
panic as the PM core dereferences a dangling pointer.

Signed-off-by: Simon Farnsworth 
Acked-by: Hans Verkuil 
Signed-off-by: Hans Verkuil 
Signed-off-by: Mauro Carvalho Chehab

[media] adv7511: fix error return code in adv7511_probe()

2013-10-14T09:36:44+00:00

Fix to return -ENOMEM in the new i2c client and create workqueue error
handling case instead of 0, as done elsewhere in this function.

Signed-off-by: Wei Yongjun 
Signed-off-by: Hans Verkuil 
Signed-off-by: Mauro Carvalho Chehab

[media] ths8200: fix compilation with GCC < 4.4.6

2013-10-14T09:36:27+00:00

Signed-off-by: Gianluca Gennari 
Signed-off-by: Hans Verkuil 
Signed-off-by: Mauro Carvalho Chehab

[media] ad9389b: fix compilation with GCC < 4.4.6

2013-10-14T09:36:19+00:00

Signed-off-by: Gianluca Gennari 
Signed-off-by: Hans Verkuil 
Signed-off-by: Mauro Carvalho Chehab

[media] adv7511: fix compilation with GCC < 4.4.6

2013-10-14T09:36:05+00:00

Signed-off-by: Gianluca Gennari 
Signed-off-by: Hans Verkuil 
Signed-off-by: Mauro Carvalho Chehab

[media] adv7842: fix compilation with GCC < 4.4.6

2013-10-14T09:35:50+00:00

Signed-off-by: Gianluca Gennari 
Signed-off-by: Hans Verkuil 
Signed-off-by: Mauro Carvalho Chehab

[media] s5p-jpeg: Initialize vfd_decoder->vfl_dir field

2013-09-24T16:41:53+00:00

This patch fixes regression introduced in commit 5c77879ff9ab9e7
"[media] v4l2-dev: add new VFL_DIR_ defines" caused by not initializing
the vfl_dir field of the vfd_decoder instance of struct video_device,
after the field was introduced. It precluded calling the driver's ioctls
which require vfl_dir not to be equal to VFL_DIR_RX which is defined as
0 and uninitialized vfl_dir field is interpreted as such. In effect the
test in the v4l_s_fmt function failed for the ioctls that expect is_tx
to be false, which prevented the ioctl callbacks registered by the driver
from being called.

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

[media] videobuf2-dc: Fix support for mappings without struct page in userptr mode

2013-09-24T16:41:38+00:00

Earlier version of dma-contig allocator in user ptr mode assumed that in
all cases DMA address equals physical address. This was just a special case.
Commit e15dab752d4c588544ccabdbe020a7cc092e23c8 introduced correct support
for converting userpage to dma address, but unfortunately it broke the
support for simple dma address = physical address for the case, when given
physical frame has no struct page associated with it (this happens if one
use for example dma_declare_coherent api or other reserved memory approach).
This commit restores support for such cases.

Signed-off-by: Marek Szyprowski 
[s.nawrocki@samsung.com: replaced #elsif with #elif]
Signed-off-by: Sylwester Nawrocki 

Signed-off-by: Mauro Carvalho Chehab

[media] vb2: Allow queuing OUTPUT buffers with zeroed 'bytesused'

2013-09-24T16:40:56+00:00

Modify the bytesused/data_offset check to not fail if both bytesused
and data_offset is set to 0. This should minimize possible issues in
existing applications which worked before we enforced the plane lengths
for output buffers checks introduced in commit 8023ed09cb278004a2
"videobuf2-core: Verify planes lengths for output buffers"

Signed-off-by: Sylwester Nawrocki 
Signed-off-by: Kyungmin Park 
Acked-by: Laurent Pinchart 
Acked-by: Marek Szyprowski 
Signed-off-by: Mauro Carvalho Chehab

[media] mx3-camera: locking cleanup in mx3_videobuf_queue()

2013-09-24T16:33:55+00:00

Smatch complains about the locking here because we mix spin_lock_irq()
with spin_lock_irqsave() in an unusual way.  According to Smatch, it's
not always clear if the IRQs are enabled or disabled when we return.  It
turns out this function is always called with IRQs enabled and we can
just use spin_lock_irq().
It's called from __enqueue_in_driver().

Signed-off-by: Dan Carpenter 
Signed-off-by: Guennadi Liakhovetski 
Signed-off-by: Mauro Carvalho Chehab