Commit graph

850 commits

Author SHA1 Message Date
Thibault Saunier
27ba8d24ec doc: Port to hotdoc 2019-05-13 11:34:08 -04:00
Jan Alexander Steffens (heftig)
0bec36d5fa oggmux: Fix a maybe-uninitialized warning
Arch Linux, GCC 8.2.1+20181127, CFLAGS='-Og -g3'
2018-12-05 21:08:33 +00:00
Sebastian Dröge
e0dee7f4c1 oggdemux: Answer POSITION query 2018-10-17 19:44:22 +03:00
Sebastian Dröge
6024b8e46b oggdemux: Ensure that no pad values are set when setting up the mapper
Otherwise we might have arbitrary values set that are used later and can
cause undefined behaviour, as found by ossfuzz.
2018-08-27 11:07:47 +03:00
Nirbheek Chauhan
eadedc68f8 meson: Add feature options for all plugins
GL dependency detection is still automagic.

https://bugzilla.gnome.org/show_bug.cgi?id=795107
2018-07-27 18:42:11 +05:30
Sebastian Dröge
3dd95b1311 oggdemux: Make sure that events are writable before changing their seqnum 2018-06-18 13:28:15 +03:00
Edward Hervey
5c118e5924 oggdemux: Properly relay seqnum of segments
Not all cases were handled regarding properly propagating the
seqnum of SEGMENT events on all downstream segment-related events
2018-06-05 17:24:05 +02:00
Edward Hervey
63e8900c47 ogg: Avoid undefined granule shift
A granule is a 64bit signed integer, shifting by 63 or more is
undefined and most likely an indication that the stream is
corrupted or invalid.

Detected by oss-fuzz
2018-05-30 10:57:14 +02:00
Xavier Claessens
201e7c7803 Meson: Generate pc file for all plugins in base
https://bugzilla.gnome.org/show_bug.cgi?id=794568
2018-04-25 11:05:56 +01:00
Edward Hervey
17d5f08316 oggstream: protect against out-of-bounds read
We need at least 17 bytes of data for a valid flac header

oss-fuzz #6974
2018-03-17 06:34:47 +01:00
Edward Hervey
0d14819ef2 oggdemux: Handle invalid-sized packets
On invalid packets there is the possibility we might end up wanting
to trim/offset more than what is available.

oss-fuzz issue #5866
2018-02-01 10:51:21 +01:00
Edward Hervey
a28880b458 oggdemux: Check encoder name is valid
Encoder names should be valid utf-8, if not just ignore them
2017-12-08 10:38:20 +01:00
Edward Hervey
427df2d5d9 ogmparse: Make sure we set valid string on caps
the fourcc might not contain printable characters, use the convenience
macro to turn it into a printable fourcc, like it's done in
gstogmparse
2017-12-06 14:23:49 +01:00
Edward Hervey
51eeb4534a ogmparse: Set caps after sending pending events
Avoids sending STREAM_START before CAPS
2017-11-15 10:55:55 +01:00
Edward Hervey
6c6d8b0b95 oggstream: Demote error to warning 2017-11-15 10:55:55 +01:00
Edward Hervey
25fa4802fe oggdemux: Move mutex/cond initialization/release
We only need to initialize the mutex/cond once when creating the
element and then release them when we are done with the element.

Avoids weird "mutex_clear called when still locked" issues
2017-11-15 10:55:55 +01:00
Edward Hervey
895d884701 oggdemux: Solidify gst_ogg_demux_loop_push() some more
There were still some races going on where seeking events wouldn't
be properly intercepted/executed by this thread.

* Instead of always waiting for the GCond to be emitted, first just
  check if there is an event available
* Take ownership of the event *while* the lock is taken and not
  after releasing/reacquiring it
* Finally acquire lock at the very top and release it at the end
  to make it a bit more streamlined

This removes the remaining issues with seeks not being executed
2017-11-08 17:51:52 +01:00
Edward Hervey
9f678bb27f oggdemux: Don't double-unlock
The previous branch will release the lock in the call to
gst_ogg_demux_seek_back_after_push_duration_check_unlock()

Only unlock it if we didn't call that function
2017-11-08 17:51:51 +01:00
Edward Hervey
c86df789ed oggdemux: Drop data before new segment
When calculating duration in push-mode we seek to a certain position
and discard any data until we get data from that requested position.

The problem is that basing ourselves solely on offset to determine
whether we reached the target offset is wrong since the source might
be fast enough  to send us that target position *before* it processed
the requested seek.

This would end up in a situation where:
* We think we're done with duration estimate
* We fire a seek back to "0" in the loop thread
* We resume normal processing
* ... except that we're still getting data from too far ahead which
  we decide to process.
* And we start doing totally wrong granule/time/duration calculation
  and pushing wrong data.

Instead of this confusion, wait until we receive data from the requested
seek. We do that by using the fact that the seqnum in
seek_event_drop_til will be non-zero until the SEGMENT corresponding
to the requested SEEK has been received.

Bonus: makes startup slightly faster
2017-11-07 15:16:52 +01:00
Edward Hervey
0a49b93ade oggdemux: Wait for push loop to be started
Code using the push_loop_thread (using for sending seeks) assumes
that the thread was properly started, except that this isn't always
true and the thread might not have completely started.

Instead wait for the thread to properly start before doing anything
else.
2017-11-07 15:16:52 +01:00
Edward Hervey
5e373936bb oggstream: Ensure enough bytes for fishead header parsing 2017-11-06 14:29:26 +01:00
Edward Hervey
af48682d44 oggstream: Use proper type for sample calculation
If we are going to return a (potentially) 64bit integer, don't use
a 32bit one for calculation, otherwise we could end up exceeding
the maximum size of a 32bit int.
2017-11-05 12:16:13 +01:00
Edward Hervey
7f5ffcaa49 oggstream: More fixes for invalid granuleshift
Don't use granuleshift if it wasn't set
2017-11-05 12:15:33 +01:00
Edward Hervey
2cf5d596ee oggstream: Fix default granuleshift usage
For stream mappers that don't set a specific granuleshift, it will
have the default value of -1.

Protect the code for that and return the granule value as-is
2017-11-04 11:50:13 +01:00
Edward Hervey
1172e5efc7 oggdemux: Protect against invalid granule positions
Only valid values are -1, 0 or positive values. Anything else is
most likely corrupted data streams
2017-11-04 11:50:10 +01:00
Edward Hervey
33dfed5af8 oggdemux: Only track time for initialized streams
in push-mode we only can track time (or most operations on streams
for that matter) if the underlying GstOggMap was properly initialized.
2017-11-01 18:39:26 +01:00
Edward Hervey
c3006b17d1 oggstream: Add a default GstOggMap
Since the default value of a GstOggPad.map.map was 0 ... we would
end up using wrong functions from mappers() if the stream wasn't
initialized yet.

Instead of that, use a default blank/empty first entry.
2017-11-01 18:39:26 +01:00
Edward Hervey
f1c44332d5 ogmparse: Ensure we don't create bogus fractions
The clamping of the fraction denominator was bogus (it needs to
be >0)
2017-11-01 11:19:58 +01:00
Edward Hervey
eb8bf0f162 oggdemux: Fix chain leak in push mode
In some corner cases we end up with the building chain not being
properly tracked (and therefore not properly freed).

Add a FIXME so it can later be fixed, but for now just fix the leak
2017-11-01 11:18:12 +01:00
Edward Hervey
0297bc2a4f ogmparse: Error out if we can't figure out the format
In some cases we might not be able to figure out the fixed format
from the header.

Properly error out in those cases.
2017-11-01 10:32:54 +01:00
Edward Hervey
e638f81165 ogg: Don't add tags to empty taglist 2017-10-31 11:57:19 +01:00
Edward Hervey
90106b6bf2 oggdemux: Don't forget to reacquire lock when needed
Fixup to ef93130cf0

I overlooked the issue. There is a case when the lock is released
and we need to reacquire it
2017-10-26 10:07:15 +02:00
Edward Hervey
6fd8d78d8b oggdemux: Don't drop sticky events
Previous commit was wrong. We should still send all events to the
pad (so that sticky events get attached to it and sent when pad
gets added).
2017-10-24 11:05:20 +02:00
Edward Hervey
e8a60b3de9 oggdemux: Improve handling of EOS without source pads
We might have a chain to use, but it might not have any active pads

Properly detect that and send an error message on EOS
2017-10-24 10:56:00 +02:00
Edward Hervey
f902286a26 oggdemux: Error out on EOS if we have no chains to use
There are not active and pending chains, if we get EOS we need to
inform the user via an error message
2017-10-20 18:41:52 +02:00
Edward Hervey
ef93130cf0 oggdemux: Don't double lock
The lock was already taken just before this block and is released after
2017-10-20 18:41:46 +02:00
Vincent Penquerc'h
bb0abf8558 oggdemux: fix artifacts at chain boundaries
https://bugzilla.gnome.org/show_bug.cgi?id=782132
2017-05-29 16:22:04 +01:00
Vincent Penquerc'h
523e98396f oggdemux: fix clipping more samples than exist in the first packet
This can happen in Opus (and maybe other codecs ?), and would cause
failure to play.

https://bugzilla.gnome.org/show_bug.cgi?id=782157
2017-05-26 14:41:47 +01:00
Nicolas Dufresne
8e6c6266d7 Remove plugin specific static build option
Static and dynamic plugins now have the same interface. The standard
--enable-static/--enable-shared toggle are sufficient.
2017-05-16 13:42:07 -04:00
Thibault Saunier
099ac9faf2 docs: Convert gtkdoc comments to markdown
Modernizing the documentation, making it simpler to read an
modify and allowing us to possibly switch to hotdoc in the
future.
2017-03-10 18:19:17 -03:00
Jan Schmidt
8596ec23cb oggdemux: Fix reverse playback
Fix various issues with reverse playback by clearing tracking
vars when working in reverse, and where possible using the
timestamp interpolation code to generate timestamps for
outgoing buffers. Make sure to mark things as discontinuous
only when looping backward to a new position and fix seeking
to the next page when starting.
2017-03-04 00:30:37 +11:00
Jan Schmidt
fe1f47aa17 oggdemux: Timestamp tracking fixes
In gst_ogg_demux_do_seek() when calculating the
keyframe time, account for a non-zero start-time

Handle a discontinuous first packet in
gst_ogg_demux_setup_first_granule() because that's pretty
normal after a seek. Also differentiate between a genuinely
truncated first packet and just bailing out early, by not using
granule = -1 as an error code.

Make the debug output logs clearer about which timestamps
are stream times (PTS) and which are ogg timestamps.
2017-03-04 00:30:37 +11:00
Jan Schmidt
342132a700 oggdemux: Don't arbitrarily guess a timestamp of 0
When we haven't managed to manufacture a timestamp for
a packet, don't just guess '0', leave it at none and
let downstream decide
2017-03-04 00:30:37 +11:00
Jan Schmidt
226dfc3f32 ogg: Fix element factory klass for OGM parsers
They're parsers, not decoders, so fix the klass info
accordingly.
2016-12-09 18:06:35 +11:00
Sebastian Dröge
732ecf0925 oggdemux: Don't end up ignoring caps just because there are no headers for this stream
https://bugzilla.gnome.org/show_bug.cgi?id=775459
2016-12-01 15:12:59 +02:00
Scott D Phillips
ab3c8caeed Remove 'return' from void functions
https://bugzilla.gnome.org/show_bug.cgi?id=774293
2016-11-12 10:51:30 +02:00
Vincent Penquerc'h
a30c713e43 oggmux: take audio clip meta into account for buffer duration
https://bugzilla.gnome.org/show_bug.cgi?id=768763
2016-10-05 14:49:52 +01:00
Vincent Penquerc'h
31438ef49b oggdemux: safety for failing to determine time length in push mode
If we can't find a valid granule near the end of the file, we
disable seeking. This guards against the whole file being then
read and never going to PLAYING.

https://bugzilla.gnome.org/show_bug.cgi?id=770314
2016-09-05 11:42:44 +01:00
Vincent Penquerc'h
6f856cb54d oggdemux: increase EOS granpos detection chunk size
This can be too small on some files to find a valid granule.

https://bugzilla.gnome.org/show_bug.cgi?id=770314
2016-09-05 11:41:43 +01:00
Thibault Saunier
bc6aae6ca7 Use the new API to post flow ERROR messages on the bus
https://bugzilla.gnome.org/show_bug.cgi?id=770158
2016-08-26 19:23:24 -03:00