Commit graph

15622 commits

Author SHA1 Message Date
Jan Schmidt
d178f7626a id3v2: Add missing overrun check for frame sizes
When frames claim to have a footer, ensure they
are large enough to contain one to avoid an invalid
read overrun.

Spotted by Joshua Yabut
2016-12-09 18:06:35 +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
Tim-Philipp Müller
6bb574ba89 docs: design: remove outdated draft docs (hw-acceleration, va) 2016-12-08 23:15:31 +00:00
Tim-Philipp Müller
46138b1b1d docs: design: move most design docs to gst-docs module 2016-12-08 23:00:45 +00:00
Seungha Yang
49653b058a decodebin3: Remove unused variable
https://bugzilla.gnome.org/show_bug.cgi?id=773341
2016-12-08 16:59:07 +01:00
Seungha Yang
582c7cef18 decodebin3: More cleanup DecodebinOutputStream and MultiQueueSlot
When removing DecodebinInputStream, cleanup DecodebinOutputStream and
MultiQueueSlot also if they were drained.

https://bugzilla.gnome.org/show_bug.cgi?id=773341
2016-12-08 16:59:07 +01:00
Seungha Yang
6bd7a5602c decodebin3: Drop duration query during _input_pad_unlink ()
Playbin3 takes lock when querying duration and handling
stream-collection message. So,to post stream-collection message,
duration query should be dropped when input pad is being unlinked.

https://bugzilla.gnome.org/show_bug.cgi?id=773341
2016-12-08 16:59:07 +01:00
Seungha Yang
7e76405895 decodebin3: Update stream-collection with _input_pad_unlink()
Since parsebin does not post new stream-collection message when
it was being removed, decodebin3 should update it itself.

https://bugzilla.gnome.org/show_bug.cgi?id=773341
2016-12-08 16:59:07 +01:00
Seungha Yang
f08e4592a4 decodebin3: Cleanup no more used DecodebinInput
Remove DecodebinInput using gst_element_call_async() API.

https://bugzilla.gnome.org/show_bug.cgi?id=773341
2016-12-08 16:59:07 +01:00
Seungha Yang
bd6ec225b9 decodebin3: Cleanup no more used MultiQueueSlot
Since MultiQueueSlot cannot be removed inside of streaming thread,
use gst_element_call_async() API.

https://bugzilla.gnome.org/show_bug.cgi?id=773341
2016-12-08 16:59:07 +01:00
Seungha Yang
4fcbcf4e48 decodebin3: Send custom-eos event to notify drained state
Likewise how urisourcebin is doing, use custom event if other streams
are still alive.

https://bugzilla.gnome.org/show_bug.cgi?id=773341
2016-12-08 16:59:07 +01:00
Seungha Yang
f4d6aa7167 playbin3: Reconfigure playsink again with pad-removed
If selected streams and actived streams are matched,
do reconfigure of playsink again with pad-removed signal

https://bugzilla.gnome.org/show_bug.cgi?id=773341
2016-12-08 16:59:07 +01:00
Seungha Yang
858ee3d913 playback: Remove trailing whitespace
https://bugzilla.gnome.org/show_bug.cgi?id=773341
2016-12-08 16:59:07 +01:00
Seungha Yang
cbd4bcd7a4 urisourcebin: Try to link output slot before cleanup
Before cleaning up output slot, check pending pads first, if available.
Then, cleanup it only if linking was failed.

https://bugzilla.gnome.org/show_bug.cgi?id=773341
2016-12-08 16:59:07 +01:00
Seungha Yang
5760bd4543 urisourcebin: Cleanup unused output slot
Since urisourcebin cannot cleanup unused output slot
in streaming thread, it will be handled in thread pool
with gst_element_call_async ().

https://bugzilla.gnome.org/show_bug.cgi?id=773341
2016-12-08 16:59:07 +01:00
Sebastian Dröge
7c1a32e28b tagdemux: Fix crash when shutting down element during getrange()
Ensure that nothing is in any of the streaming thread functions
anymore when going from PAUSED to READY. While the parent's state change
function has deactivated all pads, there is nothing preventing
downstream from activating our srcpad again and calling the getrange()
function. Although we're in READY!

https://bugzilla.gnome.org/show_bug.cgi?id=775687
2016-12-06 16:29:23 +02:00
Edward Hervey
9fd90258e6 Automatic update of common submodule
From f980fd9 to 39ac2f5
2016-12-03 08:19:15 +01:00
Reynaldo H. Verdejo Pinochet
13a491fb6c typefind: add another test to itc typefinder
Report certainty after every test passes.

Additionally:

 - Remove self-explanatory comment.
2016-12-02 16:13:24 -08:00
Sebastian Dröge
d452ad0eca ssaparse: Free initialization section before storing the next one
If getting multiple caps events.

https://bugzilla.gnome.org/show_bug.cgi?id=775480
2016-12-01 19:57:47 +02:00
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
Sebastian Dröge
52d7441942 audioconvert: Error out if mapping input/output buffer failed 2016-11-30 10:55:16 +02:00
Sebastian Dröge
1e64800278 audioconvert: Don't map the input buffer in in-place mode
Input and output buffer are the same, let's not do unnecessary work.

https://bugzilla.gnome.org/show_bug.cgi?id=775369
2016-11-30 10:49:58 +02:00
Sebastian Dröge
71e819ae7d audio-converter: In passthrough, also don't copy if in and out block are the same
In and out array are usually different, they are stack allocated arrays.
However the blocks inside them still can be the same.

https://bugzilla.gnome.org/show_bug.cgi?id=775369
2016-11-30 10:43:50 +02:00
Sebastian Dröge
1631a38640 audioconvert: Don't call transform_ip() in passthrough mode
https://bugzilla.gnome.org/show_bug.cgi?id=775369
2016-11-30 10:36:14 +02:00
Jan Alexander Steffens (heftig)
58be309344 multihandlesink: Fix buffers-queued being off by one
max_buffer_usage is the index of the oldest buffer in the queue,
starting at zero, not the number of buffers queued.

find_limits returns the index of the oldest buffer that satisfies the
limits in its min_idx parameter, not the number of buffers needed. Fix
this use too in order to keep passing the tests that read
buffers-queued.

https://bugzilla.gnome.org/show_bug.cgi?id=775351
2016-11-29 19:14:10 +02:00
Jan Alexander Steffens (heftig)
9bdf7ff6d0 multifdsink: Add a test involving a slow client
https://bugzilla.gnome.org/show_bug.cgi?id=774908
2016-11-29 19:13:00 +02:00
Jan Alexander Steffens (heftig)
8b9ebd8f88 multihandlesink: Update bufpos in a separate pass
If a client gets dropped and the iteration gets restarted, bufpos is
incremented again for all clients that preceded the dropped one, causing
havoc.

Adjust the bufpos for all clients first before trying to drop any.

https://bugzilla.gnome.org/show_bug.cgi?id=774908
2016-11-29 19:13:00 +02:00
Garima Gaur
6c5381eda6 playbin: Fix caps memory leak in usage of gst_static_caps_get() API
https://bugzilla.gnome.org/show_bug.cgi?id=775310
2016-11-29 13:42:28 +02:00
Tim-Philipp Müller
88469fdb76 win32: update .def file for new audioconverter API
Fixes distcheck.
2016-11-28 20:25:54 +00:00
Scott D Phillips
56c1b41c6a meson: Add headers and libm to has_function checks
The functions from math.h may be implemented in libm.

https://bugzilla.gnome.org/show_bug.cgi?id=774876
2016-11-29 00:28:08 +05:30
Sebastian Dröge
7e24acc67f discoverer: Handle NULL/ANY/EMPTY caps without crashing 2016-11-28 19:46:10 +02:00
Edward Hervey
647b183adb check/videorate: Avoid leaking extra buffers 2016-11-28 16:54:55 +01:00
Edward Hervey
216ee52090 video-info: Properly initialize/set extra fields
The flags and field order weren't properly initialized in the
gst_video_info_init().

Furthermore in gst_video_info_from_caps() we might set unitiliazed
values previously, this only sets them if valid.
2016-11-28 16:53:10 +01:00
Edward Hervey
e3216d1648 sdp: Fix attribute leak
We need to free the strdup'd string (to_free) in all cases
2016-11-28 16:51:43 +01:00
Edward Hervey
fb4856c839 rtsp: Don't leak authorization string 2016-11-28 16:51:23 +01:00
Petr Kulhavy
010b9547d3 audio-converter: optimize endian conversion
Optimize LE<->BE conversion by adding a dedicated fast path instead of
using the generic converter. Implement transform_ip function in order to do the
endian swap in place.

This saves buffer allocation for the intermediate format, can be done in place
and also performs the conversion in one step instead of unpack-convert-pack.

For all bit widths the naive algorithm is implemented, which provides the best
performance when compiled with -O3. ORC was considered but eventually removed
as it requires a dedicated function for in-place conversion (due to the
"restrict" parameters).

A more complex algorithm for the 24-bit conversion with unrolled loop and
32-bit processing is implemented in the #if 0 section. It performs better if
compiled with -O2. With -O3 however the naive algorithm performs better.

https://bugzilla.gnome.org/show_bug.cgi?id=773073
2016-11-28 17:24:17 +02:00
Petr Kulhavy
640c54d8f8 audio-convert: simplify the chain free process
It is not needed to store a pointer to every single chain element to free it.
Instead walk the channel list backwards and free the chain elements one by one.

Rename GstAudioConverter->chain_pack to chain_end.

https://bugzilla.gnome.org/show_bug.cgi?id=773073
2016-11-28 17:24:13 +02:00
Garima Gaur
b2b8e77566 subtitleoverlay: Fix caps memory leak when failing to get sinkpad from subtitle renderer
https://bugzilla.gnome.org/show_bug.cgi?id=775224
2016-11-28 13:56:33 +02:00
Sebastian Dröge
4efc15dffb discoverer: Extract video information from caps manually without GstVideoInfo
The caps might not be fixated (which is required by GstVideoInfo) and we
would assert otherwise. However the caps often contain useful
information in the already-fixed parts that we can use here.
2016-11-28 10:12:49 +02:00
Sebastian Dröge
bd74b102fa discoverer: Also stop waiting for subtitles if we get EOS
We're not going to get a buffer or GAP event anymore after EOS and would
wait forever otherwise.
2016-11-28 10:04:38 +02:00
Sebastian Dröge
0846eb5b27 riff-media: Check if caps are NULL before using them for the first time, not afterwards
Otherwise we'll get a g_critical() before erroring out cleanly on
https://samples.mplayerhq.hu/A-codecs/ATRAC3/SND0.AT3
2016-11-26 13:54:41 +02:00
Tim-Philipp Müller
69c1e0d7a9 common: use https protocol for common submodule
https://bugzilla.gnome.org/show_bug.cgi?id=775110
2016-11-26 11:22:46 +00:00
Miguel Paris
036bdf5d2d rtpbuffer: Fix ensure_buffers() if whole packet is in one GstMemory
When gst_rtp_buffer_add_extension_onebyte_header() is used over a
GstRtpBuffer that only contains a memory for the whole packet,
ensure_buffers function crashes at the next point:

mem = gst_memory_copy (rtp->map[i].memory, offset, rtp->size[i]);

when i==2 because the payload is not mapped.

In addition the offset is calculated subtracting in the wrong direction.

https://bugzilla.gnome.org/show_bug.cgi?id=774959
2016-11-25 13:15:24 +02:00
Sebastian Dröge
47fdb15074 video-info: Add unit test for overflow checks
And also prevent overflows caused by allowing uint width/height in
gst_video_info_set_format() but storing them as (signed!) ints.
2016-11-24 15:40:22 +02:00
Sebastian Dröge
3d7a566640 video-info: And change the overflow check to not actually overflow itself 2016-11-24 15:12:40 +02:00
Sebastian Dröge
681d97aed7 video: Handle errors in gst_video_info_set_format() / gst_video_info_align()
https://bugzilla.gnome.org/show_bug.cgi?id=774588
2016-11-24 15:06:39 +02:00
Sebastian Dröge
17cdd369e6 video-info: Sanity check the frame size to prevent overflows
https://bugzilla.gnome.org/show_bug.cgi?id=774588
2016-11-24 15:06:39 +02:00
Ulf Olsson
a4c1dfe4cd mikey: Generate the correct SRTP policy
https://bugzilla.gnome.org/show_bug.cgi?id=774911
2016-11-23 17:44:29 +01:00
Sebastian Dröge
7e4697ec98 rtspdefs: Assert on the excepted MD5 digest string length
CID 1394494.
2016-11-23 18:26:29 +02:00
Matthew Waters
2fdccfd64f typefind: bounds check windows ico detection
Fixes out of bounds read

https://bugzilla.gnome.org/show_bug.cgi?id=774902
2016-11-23 21:37:27 +11:00