Jonas Holmberg [Tue, 9 Jul 2013 09:17:59 +0000 (11:17 +0200)]
matroskademux: Unlock stream lock after use
Stream lock of sink pad was not unlocked after non-updating seek.
Ognyan Tonchev [Thu, 27 Jun 2013 11:26:31 +0000 (13:26 +0200)]
multipartmux: Re-set need_segment flag after FLUSH_STOP
https://bugzilla.gnome.org/show_bug.cgi?id=703182
Michael Olbrich [Fri, 5 Jul 2013 09:51:04 +0000 (11:51 +0200)]
v4l2: bufferpool: don't forget to release buffer on error
If the pool is stopped while gst_v4l2_buffer_pool_dqbuf() waits for a
buffer then the return value is GST_FLOW_FLUSHING. In this case the buffer
to queue must also be released. Otherwise is will never be deleted or
returned to its pool.
https://bugzilla.gnome.org/show_bug.cgi?id=703764
Sebastian Dröge [Mon, 8 Jul 2013 12:15:10 +0000 (14:15 +0200)]
rtp: Fail payloading unit test if an error message is received
Sebastian Dröge [Mon, 8 Jul 2013 12:09:37 +0000 (14:09 +0200)]
rtph263ppay: Don't pass upstream filter caps to downstream
Downstream usually can't accept video/x-h263 but only application/x-rtp,
so we would always get an empty intersection here.
https://bugzilla.gnome.org/show_bug.cgi?id=702632
Piotr Drąg [Fri, 5 Jul 2013 20:00:37 +0000 (22:00 +0200)]
po: update POTFILES.in
https://bugzilla.gnome.org/show_bug.cgi?id=703685
Wim Taymans [Tue, 2 Jul 2013 09:13:25 +0000 (11:13 +0200)]
rtspsrc: avoid some strdup
Wim Taymans [Tue, 2 Jul 2013 08:37:50 +0000 (10:37 +0200)]
rtspsrc: add select-stream signal
Add a signal to let the app select what streams will be selected.
See https://bugzilla.gnome.org/show_bug.cgi?id=634419
Wim Taymans [Tue, 2 Jul 2013 08:37:35 +0000 (10:37 +0200)]
rtspsrc: avoid strdup
J. Rick Ramstetter [Tue, 2 Jul 2013 08:12:17 +0000 (10:12 +0200)]
rtp: Fix documentation and comments to use rtpbin instead of old gstrtpbin
https://bugzilla.gnome.org/show_bug.cgi?id=703426
Michael Olbrich [Mon, 1 Jul 2013 14:55:01 +0000 (16:55 +0200)]
v4l2: don't extract data from caps twice
gst_video_info_from_caps() always extract width, height, interlace mode and
framerate now. It is no longer necessary to do it again for encoded
formats.
https://bugzilla.gnome.org/show_bug.cgi?id=703399
Andoni Morales Alastruey [Thu, 20 Jun 2013 12:41:48 +0000 (09:41 -0300)]
souphttpsrc: also consider stop positions in seeks
Use seek stop position as range end for requests
https://bugzilla.gnome.org/show_bug.cgi?id=702206
Thiago Santos [Wed, 19 Jun 2013 17:06:40 +0000 (14:06 -0300)]
souphttpsrc: allow seeks in ready
On is_seekable, check if the server's headers have already been
received. If not, do a HEAD request to get them before responding
to basesrc.
https://bugzilla.gnome.org/show_bug.cgi?id=702206
Wim Taymans [Mon, 1 Jul 2013 15:28:55 +0000 (17:28 +0200)]
rtspsrc: add signal to notify of the SDP
This way, the app can look and modify the SDP.
Kishore Arepalli [Fri, 21 Jun 2013 16:10:28 +0000 (18:10 +0200)]
gdkpixbufoverlay: Allow negative offsets to specify offset from bottom/right
https://bugzilla.gnome.org/show_bug.cgi?id=702826
Matej Knopp [Sun, 30 Jun 2013 19:01:20 +0000 (21:01 +0200)]
qtdemux: compute framerate from average sample duration
https://bugzilla.gnome.org/show_bug.cgi?id=703350
Alban Browaeys [Tue, 25 Jun 2013 19:16:38 +0000 (21:16 +0200)]
flvdemux: Add flvversion 1 to the flash-video caps
This allows using avdec_flv which requires this field to be
present in the caps. FLV only supports flash-video version 1
right now.
https://bugzilla.gnome.org/show_bug.cgi?id=703076
Sebastian Dröge [Mon, 1 Jul 2013 09:37:00 +0000 (11:37 +0200)]
deinterleave: Don't hold object lock while sending events downstream
Based on a patch by Kishore Arepalli <kishore.arepalli@gmail.com>
https://bugzilla.gnome.org/show_bug.cgi?id=703114
Sebastian Dröge [Mon, 1 Jul 2013 08:59:07 +0000 (10:59 +0200)]
matroskademux: Add MPEG4 video profile/level to the caps
Sebastian Dröge [Mon, 1 Jul 2013 08:56:28 +0000 (10:56 +0200)]
matroskademux: Add AAC profile/level to the caps
https://bugzilla.gnome.org/show_bug.cgi?id=703312
Wim Taymans [Fri, 28 Jun 2013 13:21:56 +0000 (15:21 +0200)]
vorbispay: add support for config-interval
Align code with the theora payloader and add support for the config-interval to
periodically send out the config headers.
Wim Taymans [Fri, 28 Jun 2013 13:21:12 +0000 (15:21 +0200)]
theorapay: small cleanups
Wim Taymans [Fri, 28 Jun 2013 10:08:19 +0000 (12:08 +0200)]
theorapay: handle streamheaders as well
Wim Taymans [Fri, 28 Jun 2013 10:06:21 +0000 (12:06 +0200)]
vorbispay: always collect headers on data
When we see a data packet, always check if we need to collect any previous
headers.
Wim Taymans [Fri, 28 Jun 2013 09:43:17 +0000 (11:43 +0200)]
vorbispay: handle streamheader as well
Take config strings from the streamheader when we can
Fixes https://bugzilla.gnome.org/show_bug.cgi?id=664312
David Svensson Fors [Thu, 27 Jun 2013 05:40:29 +0000 (07:40 +0200)]
rtph264pay: avoid double buffer unmap on error
Fixes https://bugzilla.gnome.org/show_bug.cgi?id=703171
Wim Taymans [Thu, 27 Jun 2013 15:02:14 +0000 (17:02 +0200)]
rtspsrc: reset-sync before play
Call reset-sync on the rtpbin before we go to playing. This makes us require SR
packets for all streams again before we attempt to sync them. If we don't reset,
it might be that we combine SR packets from before and after the PAUSE/PLAYING
state change and end up with huge bogus offsets.
Wim Taymans [Thu, 27 Jun 2013 14:23:20 +0000 (16:23 +0200)]
jitterbuffer: improve sync on first packets
Don't throw away the first RTCP packet if it arrives before the first
RTP packet but remember and use it to signal sync once we get the
RTP packet.
See https://bugzilla.gnome.org/show_bug.cgi?id=691400
Wim Taymans [Thu, 27 Jun 2013 14:15:45 +0000 (16:15 +0200)]
jitterbuffer: only signal loop when active
Only signal the loop function when it is active.
Wim Taymans [Thu, 27 Jun 2013 14:13:37 +0000 (16:13 +0200)]
jitterbuffer: signal timestamp discont
We can now use the RESYNC buffer flag to mark a timestamp discont when we update
the ts-offset property.
Wim Taymans [Wed, 26 Jun 2013 18:49:41 +0000 (20:49 +0200)]
jpegpay: turn some errors into warnings
Turn some errors into warnings, we can continue processing so this should
not be fatal.
Fixes https://bugzilla.gnome.org/show_bug.cgi?id=657079
Wim Taymans [Wed, 26 Jun 2013 12:58:53 +0000 (14:58 +0200)]
rtspsrc: avoid some flushes
Wim Taymans [Wed, 26 Jun 2013 12:41:00 +0000 (14:41 +0200)]
rtspsrc: handle data message when waiting for reply
When we are waiting for a server reply, handle data messages instead of
ignoring them.
Wim Taymans [Wed, 26 Jun 2013 12:27:34 +0000 (14:27 +0200)]
rtspsrc: handle data messages in separate method
Refactor and make a method to handle a data message.
Wim Taymans [Tue, 25 Jun 2013 18:36:18 +0000 (20:36 +0200)]
rtspsrc: add some more docs to handle-request signal
See https://bugzilla.gnome.org/show_bug.cgi?id=702705
Youness Alaoui [Mon, 10 Jun 2013 21:20:30 +0000 (17:20 -0400)]
Send a clock_provide message on the bus when we get a netclock
Youness Alaoui [Mon, 10 Jun 2013 21:20:14 +0000 (17:20 -0400)]
rtspsrc: Expose use-pipeline-clock property
Wim Taymans [Mon, 24 Jun 2013 15:11:35 +0000 (17:11 +0200)]
udpsink: bind to the given interface
Actually call BINDTODEVICE to bind to the interface as given by the
property.
Fixes https://bugzilla.gnome.org/show_bug.cgi?id=702819
Sebastian Dröge [Sat, 22 Jun 2013 08:59:17 +0000 (10:59 +0200)]
vp8dec: Error out gracefully if we get an unsupported color format
In theory we can only get I420 though, just to be on the safe side.
Sebastian Dröge [Sat, 22 Jun 2013 08:57:41 +0000 (10:57 +0200)]
vp9: Add support for YV12, Y42B and Y444 color formats
The encoder does not work with Y42B and Y444 yet it seems.
Sebastian Dröge [Sat, 22 Jun 2013 08:26:18 +0000 (10:26 +0200)]
vp9dec: Update default postproc settings from vp9_dx_iface.c
Sebastian Dröge [Fri, 21 Jun 2013 11:11:32 +0000 (13:11 +0200)]
matroska: Add initial VP9 support
Sebastian Dröge [Fri, 21 Jun 2013 11:07:30 +0000 (13:07 +0200)]
vpx: Add initial, experimental VP9 support
Youness Alaoui [Fri, 21 Jun 2013 08:32:30 +0000 (10:32 +0200)]
rtsp: go back into the loop after doing pause
After we do a pause request, go back to loop mode so that we can listen
for server messages again.
See https://bugzilla.gnome.org/show_bug.cgi?id=702705
Olivier Crête [Fri, 21 Jun 2013 03:16:17 +0000 (23:16 -0400)]
rtpptdemux: Wait after the caps to forward the other events
First forward the stream-start, then the caps, then the rest
Tim-Philipp Müller [Thu, 20 Jun 2013 23:42:02 +0000 (00:42 +0100)]
ximagesrc: clear dts on buffer acquired from pool
When setting timestamps on outgoing buffers, clear the
dts explicitly, otherwise it may end up being set to a
bogus value from last time it was used. Avoids every
second or so buffer's dts being set to 0. Not that it
should matter for raw video.
Wim Taymans [Thu, 20 Jun 2013 13:35:11 +0000 (15:35 +0200)]
v4l2: don't redefine the PERFORMANCE debug variable
It is already defined in core.
fixes https://bugzilla.gnome.org/show_bug.cgi?id=702732
Wim Taymans [Thu, 20 Jun 2013 12:43:47 +0000 (14:43 +0200)]
rtspsrc: fix race in state change to paused
When we go to paused, we first flush the connection and then send the pause
command. As a result of the flushing, the scheduled paused command can get
lost. Wait until the connection is completely flushed and the rtsp task is
waiting before issuing the paused or playing request.
Fixes https://bugzilla.gnome.org/show_bug.cgi?id=702705
Wim Taymans [Thu, 20 Jun 2013 09:31:22 +0000 (11:31 +0200)]
qtdemux: handle SEGMENT query
Sjoerd Simons [Wed, 19 Jun 2013 10:37:31 +0000 (12:37 +0200)]
v4l2: Optimize negotiation by removing the query filter
As cameras tend to have a quite specific set of capabilities (specific
framerates for each resolution), getting the peer caps filtered by our
probed caps can cause a big increase in the caps size which slows down
things quire a bit.
As for negotiation v4l2 iterates through the caps of the peer to find the
first intersection with the probed caps, getting the fully expanded
intersection of capabilities is not useful.
Using the same testcase as for bug #702632, adding this patch on top of
the patches suggested there speeds up getting the inital frame from
around ~14-15 seconds to around ~3-4 seconds.
https://bugzilla.gnome.org/show_bug.cgi?id=702638
Kishore Arepalli [Wed, 19 Jun 2013 08:30:56 +0000 (10:30 +0200)]
avidemux: duration query returns zero for DV video in avi
https://bugzilla.gnome.org/show_bug.cgi?id=702625
Sebastian Dröge [Wed, 19 Jun 2013 09:06:37 +0000 (11:06 +0200)]
qtdemux: Disable usage of allocation queries
This can only reliably work if demuxers have a
separate streaming thread per srcpad. This should be
done in a demuxer base class, which integrates parts
of multiqueue
https://bugzilla.gnome.org/show_bug.cgi?id=701856
Alex Ashley [Tue, 11 Jun 2013 14:02:21 +0000 (15:02 +0100)]
Avoid skipping moov atoms for fragmented MP4 files.
bug #700505
Following a representation change that causes a resolution change,
the video decoder fails to decode correctly. Dashdemux detects the
representation change and pushes a new caps event and an
initialization segment (a new moov atom) to the downstream qtdemux,
but it doesn't handle this new moov yet, it will only parse the
first one it receives.
This commit changes qtdemux to accept a new moov in a dash bitstream
switching scenario.
Thiago Santos [Wed, 19 Jun 2013 03:42:54 +0000 (00:42 -0300)]
qtdemux: send stream-start only once for each stream
Do not send stream start again when reconfiguring a pad for new caps.
That is common for adaptive streams
Andoni Morales Alastruey [Wed, 5 Jun 2013 15:02:49 +0000 (17:02 +0200)]
osxvideosink: fix support in VM's without hardware acceleration
Jens Georg [Sat, 15 Jun 2013 10:29:31 +0000 (12:29 +0200)]
rtpmp2tdepay: accept mislabelled streams from GStreamer 0.10 as well
The mp2t payloader in 0.10 mislabelled the streams as MP2T-ES
instead of MP2T, so accept that as well for compatibility reasons.
https://bugzilla.gnome.org/show_bug.cgi?id=702457
Wim Taymans [Sun, 16 Jun 2013 03:40:13 +0000 (05:40 +0200)]
rtspsrc: manage element state ourselves
Lock the state of the all our elements and manage their states
outselves. Because we are working async, we can't rely on the state
change function to set the state at the right time or to return the
right return value from the state change function.
Fixes https://bugzilla.gnome.org/show_bug.cgi?id=702046
Bruno Gonzalez [Fri, 14 Jun 2013 12:09:50 +0000 (14:09 +0200)]
matroskademux: Don't unlock stream lock without locking it first
https://bugzilla.gnome.org/show_bug.cgi?id=702167
Wim Taymans [Thu, 13 Jun 2013 14:00:33 +0000 (16:00 +0200)]
rtpsession: Use the right hashtable to calculate bandwidth
Don't use an unused hashtable to iterate source to calculate bandwidth.
Remove unused code.
Brendan Long [Wed, 12 Jun 2013 22:27:24 +0000 (16:27 -0600)]
pulsesink: Require PulseAudio >= 2.0
This is needed for pa_format_info_get_prop_* functions.
https://bugzilla.gnome.org/show_bug.cgi?id=686459
Arun Raghavan [Thu, 13 Jun 2013 08:53:08 +0000 (14:23 +0530)]
Revert "pulsesink: Make 2.0 dependency optional"
This reverts commit
01457027e0d384aca3e551ae684e0aa074ee5498.
We'll just depend on PulseAudio 2.0 or above instead of having the bug
partially fixed based on the installed libpulse version.
Arun Raghavan [Thu, 13 Jun 2013 07:10:15 +0000 (12:40 +0530)]
pulsesink: Make 2.0 dependency optional
The getcaps function we added uses some pa_format_info_get_prop...
accessor functions that were only added in 2.0, so we only have our
getcaps implementation exist if we're compiling against libpulse 2.0 or
above.
Eventually, we could bump the minimum requirement to 2.0 or above.
https://bugzilla.gnome.org/show_bug.cgi?id=686459
Sebastian Dröge [Wed, 12 Jun 2013 16:23:46 +0000 (18:23 +0200)]
Revert "videomixer: When all sinkpads are eos, update output segment stop and forward it"
This reverts commit
2d3910fc7901b5f29e16c0fdd4e9067a6d7f66fe.
It's not solving any problem and instead causes code to fall apart.
https://bugzilla.gnome.org/show_bug.cgi?id=701519
Tim-Philipp Müller [Wed, 9 Jan 2013 09:39:33 +0000 (09:39 +0000)]
matroskademux: mark subtitle streams as sparse in stream-start event
And also mark the streams that should be selected by default if
marked so in the headers.
https://bugzilla.gnome.org/show_bug.cgi?id=600648
Stefan Sauer [Tue, 11 Jun 2013 20:12:58 +0000 (22:12 +0200)]
audiopanorama: add prebuilt files
Stefan Sauer [Tue, 11 Jun 2013 18:27:51 +0000 (20:27 +0200)]
audiopanorama: cleanup and expand the tests
Split out two more tests. Extract more common code into helpers. Add coverage for float.
Stefan Sauer [Mon, 10 Jun 2013 19:15:20 +0000 (21:15 +0200)]
audiopanorama: cleanup of transform()
Only map input if we are reading it. Cleanup the logging and the comments a bit.
Stefan Sauer [Sun, 9 Jun 2013 18:35:18 +0000 (20:35 +0200)]
audiopanorama: use orc to speedup processing
Use special variants for the case when we don't change the panorama (pan=0.0).
Simplify the processing functions by passing the panorama value directy instead
of the instance. Use orc for clearing buffers too.
Mathieu Duponchelle [Tue, 11 Jun 2013 17:24:49 +0000 (19:24 +0200)]
videomixer: check last end_time after conversion to running segment
The last end_time was saved after conversion, so the comparison
had to be made after conversion for it to make sense.
https://bugzilla.gnome.org/show_bug.cgi?id=701385
Mathieu Duponchelle [Tue, 11 Jun 2013 17:22:20 +0000 (19:22 +0200)]
videomixer: add mix->segment.start to output_end_time
When the segment start is not 0, this created a situation where
the output_end_time is inferior to output_start_time, and the duration
of the next buffer ended up underflowing.
https://bugzilla.gnome.org/show_bug.cgi?id=701385
Sebastian Dröge [Tue, 11 Jun 2013 11:54:53 +0000 (13:54 +0200)]
matroskademux: Send stream headers after the segment event
https://bugzilla.gnome.org/show_bug.cgi?id=700799
Sebastian Dröge [Tue, 11 Jun 2013 10:26:24 +0000 (12:26 +0200)]
qtdemux: Do allocation query after exposing all pads and no-more-pads
Also configure video streams as early as possible.
Related https://bugzilla.gnome.org/show_bug.cgi?id=701856
but not fixing that.
Sebastian Dröge [Tue, 11 Jun 2013 10:25:46 +0000 (12:25 +0200)]
flvdemux: Don't forward CAPS events from upstream
Just use the default pad event handler.
https://bugzilla.gnome.org/show_bug.cgi?id=701976
Arun Raghavan [Sun, 26 May 2013 02:48:04 +0000 (08:18 +0530)]
pulsesink: Cache the getcaps/acceptcaps probe stream
getcaps is called frequently during stream setup, and creating a new
stream each time is very inefficient. There's some more room for
optimisation by caching the queried sink formats as well, but this needs
some more changes to listen for format changes on the sink (for when
supported formats change between probe stream creation and sink
querying).
https://bugzilla.gnome.org/show_bug.cgi?id=686459
Arun Raghavan [Thu, 23 May 2013 16:09:08 +0000 (21:39 +0530)]
pulsesink: Add a getcaps function
This allows us to have more fine-tuned caps in READY or above. However,
this is _really_ inefficient since we create a new stream and query sink
for every getcaps in READY, which on a simple gst-launch line happens
about 35 times. The next step is to cache getcaps results.
https://bugzilla.gnome.org/show_bug.cgi?id=686459
Arun Raghavan [Fri, 10 May 2013 06:02:44 +0000 (11:32 +0530)]
pulsesink: Take a lock on the ringbuffer in acceptcaps
This is needed as a concurrent state change could pull the context or
stream out from under our feet.
https://bugzilla.gnome.org/show_bug.cgi?id=686459
Stefan Sauer [Sun, 9 Jun 2013 18:29:09 +0000 (20:29 +0200)]
audiopanorama: move the enum to the header and use instead of gint
Move the enum for the processing method to the header so that we can use the
type for the instance struct.
Stefan Sauer [Sun, 9 Jun 2013 18:32:22 +0000 (20:32 +0200)]
level: rework the tests to cover other formats too
Michael Olbrich [Wed, 5 Jun 2013 14:32:30 +0000 (16:32 +0200)]
v4l2: make sure the element is not deleted before the pool
The pool accesses data from the v4l2object so it must exist at least
as long as the pool. Refcount the element which controls the object
live-time.
https://bugzilla.gnome.org/show_bug.cgi?id=701650
Sebastian Dröge [Fri, 7 Jun 2013 13:38:25 +0000 (15:38 +0200)]
png: Link with libgstbase for GstByteReader and GstAdapter
Sebastian Dröge [Fri, 7 Jun 2013 13:15:15 +0000 (15:15 +0200)]
wavenc: Link with libgstbase for GstByteWriter
Sebastian Dröge [Fri, 7 Jun 2013 11:26:35 +0000 (13:26 +0200)]
wavparse: Push stream-start event in pull mode before anything else
Arun Raghavan [Fri, 10 May 2013 06:39:19 +0000 (12:09 +0530)]
pulsesink: Get rid of acceptcaps side-effects
The sink info callback should not have side-effects on the GstPulseSink
object since we are sometimes using with a dummy stream in acceptcaps.
https://bugzilla.gnome.org/show_bug.cgi?id=686459
Sebastian Dröge [Wed, 5 Jun 2013 16:36:40 +0000 (18:36 +0200)]
Back to development
Sebastian Dröge [Wed, 5 Jun 2013 15:58:51 +0000 (17:58 +0200)]
Release 1.1.1
Sebastian Dröge [Wed, 5 Jun 2013 14:35:19 +0000 (16:35 +0200)]
Update .po files
Sebastian Dröge [Wed, 5 Jun 2013 13:50:04 +0000 (15:50 +0200)]
wavenc: Fix taglist ref handling that made the unit test fail
Sebastian Dröge [Wed, 5 Jun 2013 13:14:54 +0000 (15:14 +0200)]
Automatic update of common submodule
From 098c0d7 to 01a7a46
Michael Olbrich [Mon, 3 Jun 2013 07:17:43 +0000 (09:17 +0200)]
v4l2: iterate controls with V4L2_CTRL_FLAG_NEXT_CTRL if possible
In v2.6.18 control classes where added to the v4l2 API.
Iterating over CIDs starting with V4L2_CID_BASE will only find controls for
the first control class.
By iterating with V4L2_CTRL_FLAG_NEXT_CTRL all controls are found.
This is necessary to make controls from other control classes available in
the extra-controls property.
If V4L2_CTRL_FLAG_NEXT_CTRL is not defined at compile time or not supported
at runtime then the old mechanism for iterating is used.
https://bugzilla.gnome.org/show_bug.cgi?id=701540
Wim Taymans [Wed, 5 Jun 2013 10:12:53 +0000 (12:12 +0200)]
udpsink: avoid leaking the host
Fixes https://bugzilla.gnome.org/show_bug.cgi?id=701586
Michael Olbrich [Tue, 4 Jun 2013 06:26:33 +0000 (08:26 +0200)]
v4l2: improve pixel aspect ratio handling
Instead of just assuming a aspect ratio of 1/1 use VIDIOC_CROPCAP to ask
the device.
This also add a pixel-aspect-ratio property to overwrite the value from the
driver and a force-aspect-ratio property to ignore it.
https://bugzilla.gnome.org/show_bug.cgi?id=700285
Stirling Westrup [Tue, 4 Jun 2013 15:04:11 +0000 (17:04 +0200)]
v4l2: Fix compilation with older kernels
https://bugzilla.gnome.org/show_bug.cgi?id=701595
Michael Olbrich [Mon, 3 Jun 2013 15:07:10 +0000 (17:07 +0200)]
v4l2: call VIDIOC_REQBUFS with count = 0 in pool_finalize
Without this the following sequence fails:
- set_caps()
- object_stop() (does nothing)
- set_format() -> VIDIOC_S_FMT
- set_config() -> VIDIOC_REQBUFS with count = N
- set_caps()
- object_stop()
- pool_finalize()
- set_format() -> VIDIOC_S_FMT => EBUSY
Usually the pool is started after set_config(), in which case object_stop()
will result in a pool_stop and therefore VIDIOC_REQBUFS with count = 0 but
that is not guaranteed.
Also calling VIDIOC_REQBUFS with count = 0 in pool_finalize() if necessary
fixes this problem.
Fixes https://bugzilla.gnome.org/show_bug.cgi?id=701543
Michael Olbrich [Tue, 28 May 2013 17:14:15 +0000 (19:14 +0200)]
v4l2: rework sink buffer refcounting
This is a followup patch for #700781, which is not quite correct.
The buffer handling is quite complicated here.
The original code intended to the the following:
- gst_v4l2_buffer_pool_process() calls QBUF and adds the buffer to the
local list.
- The sink calls gst_buffer_unref() which returns the buffer to the pool
but not the 'free list'.
- Some time later DQBUF returns the buffer and
gst_v4l2_buffer_pool_release_buffer() puts in on the 'free list'.
If the buffer must be copied then (parent_class)->acquire_buffer() is
called directly to keep the buffer in the pool.
This has two problems:
1. If gst_v4l2_buffer_pool_release_buffer() is called before the buffer is
returned to the pool, then the buffer is put on the 'free list' twice.
This can happen if a reference to the buffer is kept outside the sink,
of if DQBUF returns the buffer, that was just queued with QBUF.
2. If buffers are copied, then all buffers are in the pool at all times. As
a result gst_v4l2_buffer_pool_stop() and gst_v4l2_buffer_pool_dqbuf()
can access pool->buffers at the same time, which can lead to memory
corruption.
The patch for #700781 fixes those problems, but with the side effect that
there are always buffers outside the pool (because they are queued) and
the pool is never stopped.
This patch fixes this by releasing the reference to the buffer after
handling it (to avoid problem 2.) so it can be returned to the pool.
gst_v4l2_buffer_pool_release_buffer() is only called if the buffer is
already in the pool (to avoid problem 1.).
Fixes https://bugzilla.gnome.org/show_bug.cgi?id=701375
Thiago Santos [Sun, 2 Jun 2013 18:24:38 +0000 (15:24 -0300)]
qtdemux: make sure taglist is writable before adding tags
Avoids assertions
Thiago Santos [Thu, 30 May 2013 22:24:13 +0000 (19:24 -0300)]
qtdemux: effectively skip tracks that weren't listed on the 1st moov
Without this, stream is NULL and the code will try to access it, leading
to segfaults.
Thiago Santos [Thu, 30 May 2013 22:23:50 +0000 (19:23 -0300)]
qtdemux: skip redundant check
!got_moov is already checked the line above
Stefan Sauer [Sun, 2 Jun 2013 11:03:40 +0000 (13:03 +0200)]
tests: cleanup level tests
Split out a few more tests to avoid checking the same stuff over and over again.
Stefan Sauer [Sat, 1 Jun 2013 19:33:46 +0000 (21:33 +0200)]
level: remove unused variables in instance struct
Stefan Sauer [Fri, 31 May 2013 16:13:02 +0000 (18:13 +0200)]
level: add a test for continous timestamps
A test that checks that msg[n].ts + msg[n].dur == msg[n+1].ts.