platform/upstream/gstreamer.git
10 years agoRelease 1.3.2
Sebastian Dröge [Wed, 21 May 2014 11:06:35 +0000 (13:06 +0200)]
Release 1.3.2

10 years agoUpdate .po files
Sebastian Dröge [Wed, 21 May 2014 10:19:39 +0000 (12:19 +0200)]
Update .po files

10 years agoAutomatic update of common submodule
Sebastian Dröge [Wed, 21 May 2014 08:51:10 +0000 (10:51 +0200)]
Automatic update of common submodule

From 211fa5f to 1f5d3c3

10 years agovp8enc: Don't dereference NULL variable
Edward Hervey [Tue, 20 May 2014 06:23:06 +0000 (08:23 +0200)]
vp8enc: Don't dereference NULL variable

CID #1139838

10 years agovp9enc: Don't dereference NULL checks
Edward Hervey [Tue, 20 May 2014 06:20:42 +0000 (08:20 +0200)]
vp9enc: Don't dereference NULL checks

CID #1197703

10 years agov4l2bufferpool: Explicitly cast enum "subtype" to its "supertype"
Sebastian Dröge [Mon, 19 May 2014 09:26:46 +0000 (11:26 +0200)]
v4l2bufferpool: Explicitly cast enum "subtype" to its "supertype"

gstv4l2bufferpool.c:608:18: error: implicit conversion from enumeration type
      'enum _GstV4l2BufferPoolAcquireFlags' to different enumeration type
      'GstBufferPoolAcquireFlags' [-Werror,-Wenum-conversion]
  params.flags = GST_V4L2_POOL_ACQUIRE_FLAG_RESURECT;
               ~ ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

10 years agogoom: Use fabs() instead of abs() to calculate the floating point absolute value
Sebastian Dröge [Mon, 19 May 2014 09:24:06 +0000 (11:24 +0200)]
goom: Use fabs() instead of abs() to calculate the floating point absolute value

tentacle3d.c:268:7: error: using integer absolute value function 'abs' when
      argument is of floating point type [-Werror,-Wabsolute-value]
  if (abs (tmp - fx_data->rot) > abs (tmp - (fx_data->rot + 2.0 * G_PI))) {
      ^

10 years agodebugutils: Properly calculate the difference with unsigned types
Sebastian Dröge [Mon, 19 May 2014 09:21:36 +0000 (11:21 +0200)]
debugutils: Properly calculate the difference with unsigned types

tests.c:161:16: error: taking the absolute value of unsigned type
      'unsigned long' has no effect [-Werror,-Wabsolute-value]
    t->diff += labs (GST_BUFFER_TIMESTAMP (buffer) - t->expected);

10 years agov4l2videodec: Handle flush while in start_streaming
Nicolas Dufresne [Fri, 16 May 2014 21:46:30 +0000 (17:46 -0400)]
v4l2videodec: Handle flush while in start_streaming

We need to handle the case where a flush occure while the streaming
thread is being brought up. In this case, the flushing state of the poll
object is cleared. To solve this, we simply set the capture poll to flushing
again, this way we know the thread will exit. The decoder streamlock
is used to synchronize with handle frame.

10 years agov4l2allocator: Don't trace twice the same message
Nicolas Dufresne [Fri, 16 May 2014 20:44:37 +0000 (16:44 -0400)]
v4l2allocator: Don't trace twice the same message

10 years agortspsrc: always use a random ssrc for the internal session
Aleix Conchillo Flaqué [Thu, 15 May 2014 18:25:50 +0000 (11:25 -0700)]
rtspsrc: always use a random ssrc for the internal session

Use a random SSRC different than 0 for the internal session SSRC.

https://bugzilla.gnome.org/show_bug.cgi?id=730212

10 years agortpsession: update last_activity when sending RTP
Wim Taymans [Fri, 16 May 2014 14:52:25 +0000 (16:52 +0200)]
rtpsession: update last_activity when sending RTP

Also update last_activity when doing something with the internal
source to make sure don't timeout early.

See https://bugzilla.gnome.org/show_bug.cgi?id=730217

10 years agov4l2: Cleanup M2M properties
Nicolas Dufresne [Thu, 15 May 2014 22:08:53 +0000 (18:08 -0400)]
v4l2: Cleanup M2M properties

M2M devices were sharing the same properties as src and sink. Most of
these made no sense. This patch reduces the number of propeties and
makes io-mode clearer by having capture-io-mode and output-io-mode. This
also accidently fixed a bug in gstv4l2transform io-mode code, where the
capture io-mode could not be set.

https://bugzilla.gnome.org/show_bug.cgi?id=729591

10 years agov4l2bufferpool: Update pool limit with hardware requiremenst
Benjamin Gaignard [Thu, 15 May 2014 15:39:39 +0000 (17:39 +0200)]
v4l2bufferpool: Update pool limit with hardware requiremenst

If the driver need more buffers than requested by the config,
update the pool min/max values. The minimum value for the pool
could be provided either by the driver or by the pool. This is
best effort for drivers that don't support
CID V4L2_CID_MIN_BUFFERS_FOR_CAPTURE.

https://bugzilla.gnome.org/show_bug.cgi?id=730200

10 years agov4l2videodec: Handle start_streaming error
Nicolas Dufresne [Thu, 15 May 2014 14:44:29 +0000 (10:44 -0400)]
v4l2videodec: Handle start_streaming error

https://bugzilla.gnome.org/show_bug.cgi?id=730207

10 years agov4l2videodec: Print the flow return causing the loop to leave
Nicolas Dufresne [Thu, 15 May 2014 14:39:40 +0000 (10:39 -0400)]
v4l2videodec: Print the flow return causing the loop to leave

https://bugzilla.gnome.org/show_bug.cgi?id=730207

10 years agov4l2videodec: Don't lock the decoder when stopping task
Nicolas Dufresne [Thu, 15 May 2014 14:31:40 +0000 (10:31 -0400)]
v4l2videodec: Don't lock the decoder when stopping task

That src pad task may need to take the lock when being pulled
down. takeing that lock can lead to a deadlock.

https://bugzilla.gnome.org/show_bug.cgi?id=730207

10 years agov4l2transform: Don't leak pool if activation failed
Nicolas Dufresne [Wed, 14 May 2014 21:18:52 +0000 (17:18 -0400)]
v4l2transform: Don't leak pool if activation failed

https://bugzilla.gnome.org/show_bug.cgi?id=730207

10 years agov4l2: Split flush in start/stop_streaming
Nicolas Dufresne [Wed, 14 May 2014 21:18:35 +0000 (17:18 -0400)]
v4l2: Split flush in start/stop_streaming

This allow calling start streaming later for capture device. Currently it breaks
in dmabuf-import because downstream is holding a buffer that will only be
released after stream-start.

https://bugzilla.gnome.org/show_bug.cgi?id=730207

10 years agov4l2transform: Flush buffer pools on flush stop
Nicolas Dufresne [Wed, 14 May 2014 19:12:26 +0000 (15:12 -0400)]
v4l2transform: Flush buffer pools on flush stop

https://bugzilla.gnome.org/show_bug.cgi?id=730207

10 years agov4l2allocator: Fix use of atomic active marker
Nicolas Dufresne [Wed, 14 May 2014 17:28:31 +0000 (13:28 -0400)]
v4l2allocator: Fix use of atomic active marker

https://bugzilla.gnome.org/show_bug.cgi?id=730207

10 years agov4l2bufferpool: Don't deactivate otherpool
Nicolas Dufresne [Wed, 14 May 2014 17:05:42 +0000 (13:05 -0400)]
v4l2bufferpool: Don't deactivate otherpool

We should not stop the otherpool unless we also stop our own
pool, otherwise it will never get restarted.

https://bugzilla.gnome.org/show_bug.cgi?id=730207

10 years agov4l2bufferpool: Also update num_buffers for import cases
Nicolas Dufresne [Wed, 14 May 2014 16:33:58 +0000 (12:33 -0400)]
v4l2bufferpool: Also update num_buffers for import cases

https://bugzilla.gnome.org/show_bug.cgi?id=730207

10 years agortpbin: update rtp encoder/decoder docs
Aleix Conchillo Flaqué [Wed, 14 May 2014 20:42:25 +0000 (13:42 -0700)]
rtpbin: update rtp encoder/decoder docs

Use %u in RTP encoder/decoder pads to match other rtpbin pads.

https://bugzilla.gnome.org/show_bug.cgi?id=730146

10 years agotests/check: rtpsession: test internal sources timing out
George Kiagiadakis [Fri, 27 Dec 2013 09:55:18 +0000 (11:55 +0200)]
tests/check: rtpsession: test internal sources timing out

10 years agortpsession: remove unused if branch
George Kiagiadakis [Thu, 26 Dec 2013 15:30:42 +0000 (17:30 +0200)]
rtpsession: remove unused if branch

1) sources that have sent BYE in the past cannot be senders, since
they would have timed out to being receivers in the meantime...
2) sources that have sent BYE are now being removed earlier inside
this function

10 years agortpsession: cleanup sources that have sent BYE
George Kiagiadakis [Thu, 26 Dec 2013 15:29:42 +0000 (17:29 +0200)]
rtpsession: cleanup sources that have sent BYE

10 years agortpsession: unify nested if clauses
George Kiagiadakis [Thu, 26 Dec 2013 15:24:51 +0000 (17:24 +0200)]
rtpsession: unify nested if clauses

10 years agortpsession: timeout internal sources that are inactive for a long time and send BYE
George Kiagiadakis [Thu, 26 Dec 2013 15:21:44 +0000 (17:21 +0200)]
rtpsession: timeout internal sources that are inactive for a long time and send BYE

10 years agortpjitterbuffer: don't stop looping if event found in the queue
Aleix Conchillo Flaqué [Tue, 13 May 2014 19:25:04 +0000 (12:25 -0700)]
rtpjitterbuffer: don't stop looping if event found in the queue

If we are inserting a packet into the jitter queue we need to keep
looping through the items until the right position is found. Currently,
the code stops as soon as an event is found in the queue.

Regarding events, we should only move packets before an event if there
is another packet before the event that has a larger seqnum.

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=730078

10 years agomatroskamux: fix the memory leak of language attribute
Adrien SCH [Thu, 17 Apr 2014 13:04:00 +0000 (13:04 +0000)]
matroskamux: fix the memory leak of language attribute

https://bugzilla.gnome.org/show_bug.cgi?id=728418

10 years agov4l2object: Fix regression in offset extrapolation
Nicolas Dufresne [Tue, 13 May 2014 17:44:20 +0000 (13:44 -0400)]
v4l2object: Fix regression in offset extrapolation

When extrapolating the offset, we need to use the extrapolate
stride rather then the base stride. This should fix support for format
with more then two planes (I420, Y42B, etc).

10 years agov4l2bufferpool: Use default VideoInfo for frame operation
Nicolas Dufresne [Mon, 12 May 2014 22:03:18 +0000 (18:03 -0400)]
v4l2bufferpool: Use default VideoInfo for frame operation

When doing frame operation, we need to use the default VideoInfo
and let the frame API read the video meta in order to get the stride
and offset right. Currently we where using the specialized VideoInfo
which reflects what the HW is setup to.

10 years agov4l2src: minor GValue handling optimisation in probing code
Tim-Philipp Müller [Mon, 12 May 2014 16:23:19 +0000 (17:23 +0100)]
v4l2src: minor GValue handling optimisation in probing code

10 years agov4l2src: avoid lists with one single framerate in probed caps
Tim-Philipp Müller [Mon, 12 May 2014 16:20:14 +0000 (17:20 +0100)]
v4l2src: avoid lists with one single framerate in probed caps

Simplify framerate field if possible, so we don't end up with
e.g. framerate = (fraction) { 30/1 }. Maybe the helper function
should be moved to core, but we can do this later.

10 years agoqtdemux: Fix leak of palette_data in error cases
Edward Hervey [Mon, 12 May 2014 14:56:35 +0000 (16:56 +0200)]
qtdemux: Fix leak of palette_data in error cases

CID #1212151

10 years agoqtmux: Free node_header in error cases
Edward Hervey [Mon, 12 May 2014 14:53:32 +0000 (16:53 +0200)]
qtmux: Free node_header in error cases

CID #1212134

10 years agoflvdemux: Don't use WARNING for not-linked flow return
Edward Hervey [Mon, 12 May 2014 11:46:01 +0000 (13:46 +0200)]
flvdemux: Don't use WARNING for not-linked flow return

Pollutes debug logs for no reason. It's only an error if all pads
return not-linked

10 years agoflvdemux: Skip unknown tags in push-mode
Edward Hervey [Mon, 12 May 2014 11:45:06 +0000 (13:45 +0200)]
flvdemux: Skip unknown tags in push-mode

We add a new mode (SKIP) in push-mode to skip tags that we don't known about

Partially fixes https://bugzilla.gnome.org/show_bug.cgi?id=670712

10 years agoflacdec: Add support for variable block size files and remove dead code
Sebastian Dröge [Sat, 10 May 2014 07:14:33 +0000 (09:14 +0200)]
flacdec: Add support for variable block size files and remove dead code

This dead code wasn't used since the 1.0 port and would need to
be modified heavily for variable block size support.

https://bugzilla.gnome.org/show_bug.cgi?id=729894

10 years agov4l2transform: Fix NULL check copy paste error
Nicolas Dufresne [Fri, 9 May 2014 16:14:23 +0000 (12:14 -0400)]
v4l2transform: Fix NULL check copy paste error

CID 1212129

10 years agov4l2transform: Fix potential deadlock due to missing break
Nicolas Dufresne [Fri, 9 May 2014 16:11:54 +0000 (12:11 -0400)]
v4l2transform: Fix potential deadlock due to missing break

CID 1212131

10 years agortpjitterbuffer: increment accepted packets after loss
Wim Taymans [Fri, 9 May 2014 16:01:28 +0000 (18:01 +0200)]
rtpjitterbuffer: increment accepted packets after loss

When we detect a lost packet, expect packets with higher
seqnum on the input.

Also update the unit test.

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=729524

10 years agoAdd new test case.
Jason Litzinger [Sun, 4 May 2014 17:12:54 +0000 (11:12 -0600)]
Add new test case.

10 years agoshapewipe: no need to activate pads
Wim Taymans [Fri, 9 May 2014 14:14:21 +0000 (16:14 +0200)]
shapewipe: no need to activate pads

Activation will happen in the state change

10 years agov4l2object: Don't leak config structure
Nicolas Dufresne [Fri, 9 May 2014 16:10:04 +0000 (12:10 -0400)]
v4l2object: Don't leak config structure

this fixes a leak of the config structure and take care of making sure
caps can't reach ref 0 before we are done doing our check.

CID 1212144

10 years agov4l2object: Remove uneeded cast for code clarity
Nicolas Dufresne [Fri, 9 May 2014 16:08:11 +0000 (12:08 -0400)]
v4l2object: Remove uneeded cast for code clarity

10 years agov4l2pool: Fix leak of config structure in error case
Nicolas Dufresne [Fri, 9 May 2014 15:56:52 +0000 (11:56 -0400)]
v4l2pool: Fix leak of config structure in error case

CIDs 1212167 and  1212167

10 years agov4l2object: Fix use of unitilized pool pointer
Nicolas Dufresne [Fri, 9 May 2014 15:51:26 +0000 (11:51 -0400)]
v4l2object: Fix use of unitilized pool pointer

CID #1212173

10 years agoqtdemux: don't respond to a position query in BYTE format with a TIME position
Eric Trousset [Fri, 9 May 2014 14:48:58 +0000 (16:48 +0200)]
qtdemux: don't respond to a position query in BYTE format with a TIME position

https://bugzilla.gnome.org/show_bug.cgi?id=729553

10 years agomatroskademux: don't leak doctype string in error code path
Tim-Philipp Müller [Fri, 9 May 2014 13:22:42 +0000 (14:22 +0100)]
matroskademux: don't leak doctype string in error code path

CID 1212145.

10 years agov4l2object: Readback pool config if used within the baseclass
Nicolas Dufresne [Tue, 6 May 2014 17:37:47 +0000 (13:37 -0400)]
v4l2object: Readback pool config if used within the baseclass

10 years agov4l2: Replace miss-use of crop meta in favour of proper offset
Nicolas Dufresne [Tue, 6 May 2014 16:58:59 +0000 (12:58 -0400)]
v4l2: Replace miss-use of crop meta in favour of proper offset

This moves away from copying information and store everything inside
the GstVideoInfo structure. The alignement exposed by v4l2 api
is now handled using proper offset.

10 years agov4l2object: Style fix
Nicolas Dufresne [Tue, 6 May 2014 16:55:30 +0000 (12:55 -0400)]
v4l2object: Style fix

10 years agov4l2allocator: Reset imported buffer size with expected size
Nicolas Dufresne [Mon, 5 May 2014 16:38:33 +0000 (12:38 -0400)]
v4l2allocator: Reset imported buffer size with expected size

This ensure that the buffer pool won't always discard buffer with these
memory when they are released.

10 years agov4l2allocator: Reset flushed group
Nicolas Dufresne [Mon, 5 May 2014 16:37:43 +0000 (12:37 -0400)]
v4l2allocator: Reset flushed group

This ensure that a flushed group memory are the same size as when they
where originally allocated / imported.

10 years agov4l2bufferpool: Get number of allocated buffers from allocator
Nicolas Dufresne [Mon, 5 May 2014 16:07:31 +0000 (12:07 -0400)]
v4l2bufferpool: Get number of allocated buffers from allocator

The value of num_allocated buffer would get confused when
buffer are being discarded.

10 years agov4l2allocator: Add a method to read number of allocated group
Nicolas Dufresne [Mon, 5 May 2014 16:06:44 +0000 (12:06 -0400)]
v4l2allocator: Add a method to read number of allocated group

10 years agov4l2bufferpool: Improve debugging
Nicolas Dufresne [Mon, 5 May 2014 00:23:42 +0000 (20:23 -0400)]
v4l2bufferpool: Improve debugging

10 years agov4l2bufferpool: Ensure we don't re-enqueue buffer during flush
Nicolas Dufresne [Sun, 4 May 2014 23:51:48 +0000 (19:51 -0400)]
v4l2bufferpool: Ensure we don't re-enqueue buffer during flush

10 years agov4l2transform: Initilialize debug category
Nicolas Dufresne [Sun, 4 May 2014 23:13:37 +0000 (19:13 -0400)]
v4l2transform: Initilialize debug category

10 years agov4l2allocator: Fix libv4l2 support
Nicolas Dufresne [Sun, 4 May 2014 20:11:09 +0000 (16:11 -0400)]
v4l2allocator: Fix libv4l2 support

Need to include config.h, otherwise we endup directly using the
ioct/mmap/munmap calls and need to vall v4l2_munmap.

10 years agov4l2allocator: Set the flags on the object
Nicolas Dufresne [Thu, 1 May 2014 17:04:08 +0000 (13:04 -0400)]
v4l2allocator: Set the flags on the object

We where not setting the probed flags on the allocator, which mean even if
CREATE_BUFS was supported on some driver, it would endup being ignored.

10 years agov4l2bufferpool: Re-enqueue buffer at stream start
Nicolas Dufresne [Tue, 29 Apr 2014 20:49:52 +0000 (16:49 -0400)]
v4l2bufferpool: Re-enqueue buffer at stream start

10 years agov4l2allocator: There is not group on error
Nicolas Dufresne [Tue, 29 Apr 2014 20:06:00 +0000 (16:06 -0400)]
v4l2allocator: There is not group on error

10 years agov4l2transform: Handle FLUSH_STOP event
Nicolas Dufresne [Tue, 29 Apr 2014 18:56:31 +0000 (14:56 -0400)]
v4l2transform: Handle FLUSH_STOP event

10 years agov4l2bufferpool: Acquire cannot return a buffer from another pool
Nicolas Dufresne [Tue, 29 Apr 2014 17:05:41 +0000 (13:05 -0400)]
v4l2bufferpool: Acquire cannot return a buffer from another pool

Return a buffer from an otherpool has unwanted side effects that lead to leaks and
prevents deactivating the pool. Instead, we change the _process() API so it can
replace the internal buffer with the buffer from the downstream pool. This implied
moving from _fill() to _create() method in the src.

10 years agov4l2bufferpool: Remove unreached acquire code
Nicolas Dufresne [Tue, 29 Apr 2014 17:00:32 +0000 (13:00 -0400)]
v4l2bufferpool: Remove unreached acquire code

The acquire is done in _prepare now.

10 years agov4l2bufferpool: Sanetize buffer refount handling
Nicolas Dufresne [Tue, 29 Apr 2014 16:57:08 +0000 (12:57 -0400)]
v4l2bufferpool: Sanetize buffer refount handling

Buffer refcounting is a bit hard, because of the duality between CAPTURE and
OUTPUT mode. In the long term, we should consider having two seperate pool
instead of this mess. At least state should be better kept this way.

10 years agov4l2transform: Add more traces
Nicolas Dufresne [Tue, 29 Apr 2014 16:48:04 +0000 (12:48 -0400)]
v4l2transform: Add more traces

10 years agov4l2-allocator: Add S to REQBUFS/CREATE_BUFS enum
Nicolas Dufresne [Mon, 28 Apr 2014 12:48:26 +0000 (08:48 -0400)]
v4l2-allocator: Add S to REQBUFS/CREATE_BUFS enum

All enum that has REQBUFS and CREATE_BUFS where missing S, which was
confusing since they are supposed to match with associcated ioctl name. This
also fixes the yet unused CAN_REQUEST flag check.

10 years agov4l2transform: Enabled QoS
Nicolas Dufresne [Fri, 18 Apr 2014 21:51:07 +0000 (17:51 -0400)]
v4l2transform: Enabled QoS

10 years agov4l2: Fixup USERPTR/DMABUF capture support
Nicolas Dufresne [Fri, 18 Apr 2014 21:02:50 +0000 (17:02 -0400)]
v4l2: Fixup USERPTR/DMABUF capture support

10 years agov4l2object: Improve selecton of min/max in decide allocation
Nicolas Dufresne [Fri, 18 Apr 2014 18:45:00 +0000 (14:45 -0400)]
v4l2object: Improve selecton of min/max in decide allocation

10 years agov4l2bufferpool: Update config if meta is missing
Nicolas Dufresne [Fri, 18 Apr 2014 17:09:00 +0000 (13:09 -0400)]
v4l2bufferpool: Update config if meta is missing

Rather then hard failure, we should update the config with the meta option we
need and return false.

10 years agov4l2: Add DMABUF and USERPTR importation
Nicolas Dufresne [Fri, 11 Apr 2014 21:10:11 +0000 (17:10 -0400)]
v4l2: Add DMABUF and USERPTR importation

10 years agov4l2allocator: Valid FD are bigger or equal to zero
Nicolas Dufresne [Fri, 18 Apr 2014 01:45:58 +0000 (21:45 -0400)]
v4l2allocator: Valid FD are bigger or equal to zero

10 years agov4l2object: Don't leak downstream pool in propose_allocation
Nicolas Dufresne [Wed, 16 Apr 2014 21:04:42 +0000 (17:04 -0400)]
v4l2object: Don't leak downstream pool in propose_allocation

parse_nth_allocation_pool() give a ref on the pool, we need to unref it
when done.

10 years agov4l2: Introduce DMABUF_IMPORT IO mode
Nicolas Dufresne [Mon, 14 Apr 2014 16:19:39 +0000 (12:19 -0400)]
v4l2: Introduce DMABUF_IMPORT IO mode

10 years agov4l2: Add dmabuf export support
Nicolas Dufresne [Thu, 10 Apr 2014 20:26:34 +0000 (16:26 -0400)]
v4l2: Add dmabuf export support

This can be enabled sing io-mode=dmabuf. This will enabled mmap base
drivers to export the buffers as dmabuf.

10 years agov4l2allocator: Guaranty queued state integrety
Nicolas Dufresne [Wed, 16 Apr 2014 19:51:03 +0000 (15:51 -0400)]
v4l2allocator: Guaranty queued state integrety

Because of the buf in videobuf2, dqbuf may leave the DONE flag being,
which would implied that the buffer is queued. As this has been broken
for 4 years, simply guaranty the state flags integrity when doing
qbuf/dqbuf.

See https://patchwork.linuxtv.org/patch/23641/

10 years agov4l2transform: Implement open/close
Nicolas Dufresne [Tue, 15 Apr 2014 21:31:42 +0000 (17:31 -0400)]
v4l2transform: Implement open/close

10 years agov4l2transform: Ensure output pool is configured
Nicolas Dufresne [Tue, 15 Apr 2014 20:43:41 +0000 (16:43 -0400)]
v4l2transform: Ensure output pool is configured

10 years agov4l2transform: Check if caps have changes before asserting
Nicolas Dufresne [Tue, 15 Apr 2014 20:43:15 +0000 (16:43 -0400)]
v4l2transform: Check if caps have changes before asserting

In set_caps, now checks if caps actually changed and succeed if they didn't
change.

10 years agov4l2videodec: Ensure pool is configured
Nicolas Dufresne [Tue, 15 Apr 2014 20:41:46 +0000 (16:41 -0400)]
v4l2videodec: Ensure pool is configured

10 years agov4l2object: Always set a size when deciding allocation
Nicolas Dufresne [Tue, 8 Apr 2014 22:54:09 +0000 (18:54 -0400)]
v4l2object: Always set a size when deciding allocation

10 years agov4l2object: Improved decide allocation
Nicolas Dufresne [Tue, 8 Apr 2014 22:20:25 +0000 (18:20 -0400)]
v4l2object: Improved decide allocation

Improve decide allocation so it properly configure both local and downstream
buffer pools. Also read back the pool config if it was changed to to driver
limitations.

10 years agov4l2bufferpool: Do not pre-configure the pool
Nicolas Dufresne [Tue, 15 Apr 2014 17:30:02 +0000 (13:30 -0400)]
v4l2bufferpool: Do not pre-configure the pool

Pre-configuring the pool is error prone, since it may hide a configuration failure and
endup with a pool that is not configured the way it should (e.g. no video meta, wrong
queue size, etc.)

10 years agov4l2bufferpool: Preserve downstream minimum even in RW
Nicolas Dufresne [Tue, 15 Apr 2014 17:23:33 +0000 (13:23 -0400)]
v4l2bufferpool: Preserve downstream minimum even in RW

10 years agov4l2bufferpool: Turn cropmeta into a custom option
Nicolas Dufresne [Tue, 15 Apr 2014 17:20:12 +0000 (13:20 -0400)]
v4l2bufferpool: Turn cropmeta into a custom option

Turn crop meta into a custom option and make sure it's there is needed.

10 years agov4l2bufferpool: Early catch short allocation
Nicolas Dufresne [Wed, 9 Apr 2014 16:53:19 +0000 (12:53 -0400)]
v4l2bufferpool: Early catch short allocation

Catch short allocation after saving the format. This is not a catch all, but should catch
most of the miss-behaving drivers when doing S_FMT/G_FMT and avoid potential crash.

10 years agov4l2bufferpool: Port to use GstV4l2Allocator
Nicolas Dufresne [Sat, 5 Apr 2014 02:46:40 +0000 (22:46 -0400)]
v4l2bufferpool: Port to use GstV4l2Allocator

10 years agoImplement V4l2 Allocator
Nicolas Dufresne [Sat, 5 Apr 2014 02:35:48 +0000 (22:35 -0400)]
Implement V4l2 Allocator

This goal of this allocator is mainly to allow tracking the memory.
Currently, when a buffer memory has been modified, the buffer and it's
memory is disposed and lost until the stream is restarted.

10 years agov4l2object: Don't advertise crop meta
Nicolas Dufresne [Wed, 16 Apr 2014 20:35:49 +0000 (16:35 -0400)]
v4l2object: Don't advertise crop meta

Currently we advertise crop meta, but not element handle support this meta.

10 years agov4l2object: Setup pool already send element error
Nicolas Dufresne [Tue, 8 Apr 2014 22:18:57 +0000 (18:18 -0400)]
v4l2object: Setup pool already send element error

10 years agov4l2object: Workaround decoder that set num_planes to 0 in the format
Nicolas Dufresne [Tue, 8 Apr 2014 22:17:31 +0000 (18:17 -0400)]
v4l2object: Workaround decoder that set num_planes to 0 in the format

Some well known decoder wrongly set num_planes to 0 in their format instead of
one. In this case we would endup with no size when deciding buffer allocation.

10 years agov4l2object: Ensure size before configuring the pool
Nicolas Dufresne [Tue, 8 Apr 2014 21:34:19 +0000 (17:34 -0400)]
v4l2object: Ensure size before configuring the pool

10 years agov4l2object: Set minimum buffers to 2
Nicolas Dufresne [Sat, 5 Apr 2014 02:38:05 +0000 (22:38 -0400)]
v4l2object: Set minimum buffers to 2

All the element requires at least two buffers. This is not used for RW mode.

10 years agov4l2object: Remove unused MAX_BUFFERS define
Nicolas Dufresne [Sat, 5 Apr 2014 02:37:14 +0000 (22:37 -0400)]
v4l2object: Remove unused MAX_BUFFERS define

10 years agov4l2object: Don't hardcode min/max use default instead
Nicolas Dufresne [Sat, 5 Apr 2014 02:36:37 +0000 (22:36 -0400)]
v4l2object: Don't hardcode min/max use default instead