Commit graph

124 commits

Author SHA1 Message Date
Thiago Santos
613b110b25 mssdemux: seeks in the manifest never fail
Set void as the return and remove error handling as it could
never fail
2014-11-07 19:30:18 -03:00
Tim-Philipp Müller
efbae700a6 smoothstreaming: initialise new GstH264NalUnit fields
CID #1251107.
2014-11-07 00:16:12 +00:00
Tim-Philipp Müller
f216b7bb11 Sprinkle some G_PARAM_DEPRECATED and #ifndef GST_REMOVE_DEPRECATED 2014-11-02 17:19:34 +00:00
Thiago Santos
a38ada61e3 mssdemux: add type of stream to debug log
Makes it faster to debug why certain streams are not showing
2014-10-31 17:33:55 -03:00
Thiago Santos
e644b5e6ae mssdemux: do not unlock not locked mutex 2014-10-31 17:33:55 -03:00
George Kiagiadakis
e289ab07c1 mssdemux: synchronize with the download loop thread to signal it to continue
If EOS or ERROR happens before the download loop thread has reached its
g_cond_wait() call, then the g_cond_signal doesn't have any effect and
the download loop thread stucks later.

https://bugzilla.gnome.org/show_bug.cgi?id=735663
2014-09-18 13:38:29 -03:00
Thiago Santos
01ccac24fa mssdemux: fix clearing of eos state in pads
The internal pad still keeps its EOS flag and event as it can be assigned
after the flush-start/stop pair is sent. The EOS is assigned from the streaming
thread so this is racy.

To be sure to clear it, it has to be done after setting the source to READY to
be sure that its streaming thread isn't running.

https://bugzilla.gnome.org/show_bug.cgi?id=736012
2014-09-18 12:14:30 -03:00
Matthieu Bouron
bce21e7398 mssdemux: improve position detection when a download restart
Query other src pads before falling back to the position of the last
known pushed segment (which can be far ahead in time).

https://bugzilla.gnome.org/show_bug.cgi?id=736421
2014-09-11 18:25:19 -03:00
Sebastian Dröge
bda4eae1cc mssdemux: Don't send flush events to deactivated pads
https://bugzilla.gnome.org/show_bug.cgi?id=736012
2014-09-04 18:21:38 +03:00
Sebastian Dröge
546feaef5c mssdemux: Always use the redirect target to resolve relative URIs
But redownload the playlists from the original URI if it's not
a permanent redirect.
2014-05-28 12:56:24 +02:00
Thiago Santos
ce7319c62d mssdemux: avoid deadlocking when erroring
It might get stuck when posting the message while
a reconfigure event is being pushed upstream
2014-05-12 18:28:45 -03:00
Thiago Santos
a1f87d0b4f mssdemux: fix caps leak 2014-05-12 18:08:12 -03:00
Thiago Santos
6b8448c245 mssdemux: handle error messages from internal source
set up a bin message handler to get notified of error messages from
the internal source element
2014-05-12 18:08:12 -03:00
Thiago Santos
b627d7df83 mssdemux: Always accept the latency configure event
Upstream or internal elements are irrelevant and we should
always succeed.
2014-05-12 16:27:37 -03:00
Thiago Santos
b54367e211 mssdemux: reenable segment tracking for streams
Store the latest pushed timestamp so that mssdemux streams know
where they are in time
2014-05-12 16:27:37 -03:00
Thiago Santos
45d5859131 mssdemux: deprecate unused max-queue-size-buffers
There is no internal storage of buffers anymore
2014-05-12 16:27:37 -03:00
Thiago Santos
99a2e425de mssdemux: measure bitrate similarly to hlsdemux
new bitrate = (old bitrate + (last fragment bitrate * 3)) / 4
2014-05-12 16:27:37 -03:00
Thiago Santos
90577c3ae7 mssdemux: remove uridownloader in favor of internal source
Refactor mssdemux to remove uridownloader to use an internal
source element which reduces startup latency and provides smaller
buffers for better buffering management downstream
2014-05-12 16:27:37 -03:00
Sebastian Dröge
3c2fe80ab6 mssdemux: Update for uri downloader API changes 2014-05-05 09:41:51 +02:00
Sebastian Dröge
5d34b071d9 mssdemux: Update for URI downloader API changes
And include the manifest file as referer in requests, and make sure
to set refresh=TRUE when updating the manifest.
2014-05-02 10:43:06 +02:00
Vincent Penquerc'h
14f8172d2b smoothstreaming: avoid using an uninitialized manifest
If we did not set the manifest yet, early out in the seeking query.

Coverity 1139737
2014-04-10 09:51:07 +01:00
Vincent Penquerc'h
88879654df smoothstreaming: remove unnecessary check
The function is static, the only call site passes a non NULL
pointer, and the pointer is dereferenced before anyway.

Coverity 1139839
2014-04-09 17:10:06 +01:00
Sebastian Dröge
b47a4faf5f ext: Use Codec/Demuxer/Adaptive for the adaptive streaming demuxers 2014-02-23 00:11:04 +01:00
Sebastian Dröge
a11c6640bf mssdemux: Download the manifest with compression, the fragments without 2014-02-12 13:11:38 +01:00
Sebastian Dröge
3a5abc233d ext: Update for GstUriDownloader API changes 2014-02-11 14:18:32 +01:00
Thiago Santos
6348113777 mssdemux: updating docs explaining how it works
mssdemux changed a lot and the docs were not correct anymore.

Also adds the 'Adaptive' category to its details
2013-12-20 00:40:42 -03:00
Thiago Santos
ff1e88ce93 mssdemux: do not push event holding object lock
Can lead to deadlocks as the push might block downstream in
serialized event cases.
2013-12-20 00:40:42 -03:00
Thiago Santos
aa61251371 mssdemux: lock around 'cancelled' flag
Prevents race conditions when pipeline is seeking near eos
2013-12-20 00:40:42 -03:00
Thiago Santos
628461b4d1 mssdemux: fix segfault when trying to access codec_data
Check if codec_data was obtained before trying to use it to
prevent a segfault
2013-12-18 18:52:19 -03:00
Thiago Santos
5948a07324 mssdemux: also generate caps from audiotag attribute
Some audio will have its audio format defined by the AudioTag
instead of the Fourcc. Fallback to using AudioTag if Fourcc
isn't available
2013-12-18 18:52:19 -03:00
Thiago Santos
26acbbcca9 mssdemux: make sure streams stop properly on cancelled
specially when the gst_task_join is waiting for the task
to go stop and the task itself sets its back to paused, causing
a deadlock on exit.
2013-12-18 18:52:19 -03:00
Thiago Santos
99c8fb7330 mssdemux: track segments per stream
As streams now flow independently, the GstSegment needs to be put
on each stream so they can track the position of each one correctly
instead of being mixed in a single segment
2013-12-18 18:52:19 -03:00
Guillaume Desmottes
f44c6ed526 mssdemux: push language code tag
https://bugzilla.gnome.org/show_bug.cgi?id=712358
2013-12-18 18:52:19 -03:00
Guillaume Desmottes
d6a9f17d8a mssmanifest: store the language of the stream
https://bugzilla.gnome.org/show_bug.cgi?id=712358
2013-12-18 18:52:19 -03:00
Thiago Santos
67f3190301 mssdemux: remove the stream loop task
Download and push from the same task, makes code a lot simpler
to maintain. Also pushing from separate threads avoids deadlocking
when gst_pad_push blocks due to downstream queues being full
2013-12-18 18:52:19 -03:00
Thiago Santos
4e6e1315da mssdemux: Improve logging
Show the stream's pad on log messages to make easier to debug
issues in the multiple threads
2013-12-18 18:52:18 -03:00
Thiago Santos
056420940e mssdemux: improve flow return handling
Handle different flow returns both from the streaming and the
downloading loops
2013-12-18 18:52:18 -03:00
Thiago Santos
e847bea1e1 mssdemux: remove stream locks
Simplify the locking by using a single lock instead of having one
lock per stream. This still works and is simpler to maintain.
2013-12-18 18:52:18 -03:00
Thiago Santos
2ce4f6a8e4 mssdemux: avoid downloading not-linked streams
When a stream gets a not-linked return, it will be marked as so and
won't download any more new fragments until a reconfigure event
is received. This will make mssdemux expose all pads, but only download
fragments for the streams that are actually being used.

Relying on the pads being linked/unlinked isn't enough in this scenario
as there might be an input-selector downstream that is actually discarding
buffers for a given linked pad.

When streams are switching, the old active stream can be blocked because
input-selector will block not-linked streams. In case the mssdemux's
stream loop is blocked pushing a buffer to a full queue downstream it will
never unblock as the queue will not drain (input-selector is blocking).

In this scenario, stream switching will deadlock as input-selector is
waiting for the newly active stream data and the stream_loop that would
push this data is blocked waiting for input-selector.

To solve this issue, whenever an stream is reactivated on a reconfigure
it will enter into the 'catch up mode', in this mode it can push buffers
from its download thread until it reaches the currrent GstSegment's position.
This works because this timestamp will always be behind or equal to the maximum
timestamp pushed for all streams, after pushing data for this timestamp,
the stream will go back to default and be pushed sequentially from the main
streaming thread. By this time, the input-selector should have already
released the thread.

https://bugzilla.gnome.org/show_bug.cgi?id=711849
2013-12-18 18:52:18 -03:00
Antonio Ospite
c4fbff50df mssdemux: fix setting codec_data when using WaveFormatEx
When using WaveFormatEx the actual codec private data is at the end of
the wav header structure, after the cbSize field, see for example the
notes at http://wiki.multimedia.cx/index.php?title=WAVEFORMATEX

The previous code was erroneously passing the whole WAVEFORMATEX
structure as codec_data.

The smoothstreaming specifications[1,2,3] always mention WAVEFORMATEX
and never smaller structures like WAVEFORMAT or PCMWAVEFORMAT, so the
buffer can assumed to be at least 18 bytes and always consuming the wav
header gives the (possibly empty) codec private data.

[1] http://web.archive.org/web/20120907004742/http://www.iis.net/community/files/media/smoothspecs/%5BMS-SMTH%5D.pdf
[2] http://download.microsoft.com/download/B/0/B/B0B199DB-41E6-400F-90CD-C350D0C14A53/%5BMS-SSTR%5D.pdf
[3] http://download.microsoft.com/download/9/5/E/95EF66AF-9026-4BB0-A41D-A4F81802D92C/%5BMS-SSTR%5D.pdf

https://bugzilla.gnome.org/show_bug.cgi?id=699924
2013-11-27 09:26:36 -03:00
Antonio Ospite
f02e7a2439 mssdemux: use mssdemux_debug debug category for gstmssmanifest.c too 2013-11-27 09:26:36 -03:00
Antonio Ospite
8345c5acfb mssdemux: fix coding style 2013-11-27 09:26:36 -03:00
Thiago Santos
27b0086e4a mssdemux: do not try to deinit the downloadrate while it is used
This can cause an assertion or deadlocks.
2013-11-15 12:01:55 -03:00
Thiago Santos
fe77ee47c9 mssdemux: also consider /manifest
The spec says /Manifest, but /manifest is easily found on some
services, so handle that, too.

https://bugzilla.gnome.org/show_bug.cgi?id=710223
2013-11-15 12:01:54 -03:00
Sebastian Dröge
fef9ef6a8e dash/smoothstreaming: Remove unused debug category with the same name
It caused static linking to fail.
2013-09-19 20:09:59 +02:00
Thiago Santos
1b1332cb6d mssdemux: use gst_data_queue_push_force for adding events to queue
This prevents locking on startup when a stream only has a single buffer
for one of the streams and mssdemux decides to push an EOS event right
after it.
2013-08-13 12:44:41 -03:00
Edward Hervey
97426a1caa all: Fix for GST_DISABLE_GST_DEBUG
Where applicable, remove methods that don't do anything different than
the default implementation.
2013-07-29 09:42:12 +02:00
Sebastian Dröge
5b702b894e mssdemux: Add support for group-id in the stream-start event 2013-07-23 10:33:31 +02:00
Thiago Santos
750137c4be mssdemux: fix wma pro caps creation
WmaPro is actually wmaversion 3, and can also be found by the
WMAP fourcc.

Some manifests also contain the block_align field as "PacketSize"
in the audio track description, the libav decoders require it
to be present in caps.

Fixes #699921
2013-05-17 12:08:10 -03:00
Thiago Santos
504142ea93 mssdemux: fix assertions when a stream caps can't be created
Fixes two assertions related to cleanup for a pad that couldn't
be exposed because the caps couldn't be identified from the
Manifest

Fixes #699921
2013-05-17 11:37:30 -03:00