Commit graph

22187 commits

Author SHA1 Message Date
Alessandro Decina
3d8d60baa8 vtdec: make vtdec_hw fallback to software on renegotiation
When renegotiating mid stream - for example with variable bitrate
streams - and therefore destroying and recreating VTSessions, the
hw decoder might become temporarily unavailable.

To deal with this and avoid erroring out on bitrate changes,
vtdec_hw now falls back to using the software decoder if the hw
one was available at some point but isn't anymore. At
renegotiation/bitrate change time, it will still retry to open
the hardware one.
2016-06-02 16:30:02 +10:00
Alessandro Decina
4a83686a57 vtdec: fix switching from GLMemory to Sysmem
When renegotiating from GLMemory to Sysmem do teardown the texture_cache.

Fixes: https://bugzilla.gnome.org/show_bug.cgi?id=766190
2016-06-02 13:15:05 +10:00
Alessandro Decina
bf444301f9 vtdec: optimize renegotiation
::negotiate can be called several times before the CAPS event is sent downstream
so use the currently configured output state caps instead of the pad current
caps when deciding whether to recreate the VTSession or not.

This leads to creating/destroying less VTSessions which makes renegotiation more
reliable especially when using hw decoding.
2016-06-02 13:15:05 +10:00
Reynaldo H. Verdejo Pinochet
9e442d5cc0 dvbsrc: remove comment on self-explanatory code 2016-06-01 13:52:48 -07:00
Reynaldo H. Verdejo Pinochet
46de31d0f8 dvbsrc: avoid out-bound write on PID filter array
There's no need for an end-of-list marker in the filter
PIDs array if full, as the absolute maximum number of
elements (MAX_FILTERS) is known.

CID #1362441
2016-06-01 13:23:53 -07:00
Tim-Philipp Müller
889a1a9b90 androidmedia: fix error debug message when camera doesn't exist
Makes no sense to include the system error here since errno
will likely not be set and then it says 'system error: success'
which is confusing.

https://bugzilla.gnome.org/show_bug.cgi?id=767087
2016-05-31 20:41:14 +01:00
Justin Kim
fe62233b83 ahcsrc: release resources in 'finalize' function
In general, 'dispose' function is used for dropping all references
and 'finalize' is called for releasing instances.

https://bugzilla.gnome.org/show_bug.cgi?id=763309
2016-05-31 12:58:06 +01:00
Guillaume Desmottes
0467923415 player: inhibit signals after gst_player_stop() has been called
Also wait for the state change to STOP to have been announced before
destroying the player so it won't appear as leaked by leak detector
tools.

https://bugzilla.gnome.org/show_bug.cgi?id=766607
2016-05-30 12:43:22 +03:00
Guillaume Desmottes
b00f0d0180 player: handle uri-loaded in test
Had to adapt the existing tests because of this new callback.

https://bugzilla.gnome.org/show_bug.cgi?id=766607
2016-05-30 12:43:22 +03:00
Scott D Phillips
3edd641667 h265parse: Don't assume contiguous id's in make_codec_data
vps/sps/pps id's are not required to be used contiguously.

https://bugzilla.gnome.org/show_bug.cgi?id=766891
2016-05-30 12:42:24 +03:00
Reynaldo H. Verdejo Pinochet
1d8673af62 dvbsrc: add sample ATSC launch line 2016-05-29 23:31:05 -07:00
Havard Graff
63a05aff6a gl: glquery: cast to silence compiler warning
https://bugzilla.gnome.org/show_bug.cgi?id=766973
2016-05-28 22:28:06 +01:00
Havard Graff
7ecfd7e24d gltestsrc: gltestsrc.h already defines GstGLTestSrc
And redefinition is not allowed.

https://bugzilla.gnome.org/show_bug.cgi?id=766973
2016-05-28 22:20:51 +01:00
Havard Graff
45a832686b player: use correct _NONE enum
https://bugzilla.gnome.org/show_bug.cgi?id=766973
2016-05-28 21:16:27 +01:00
Tim-Philipp Müller
1b58bbf3ad h264parser: maintain minimal ABI compat
Because we can.

https://bugzilla.gnome.org/show_bug.cgi?id=723352
2016-05-28 10:54:13 +01:00
Sebastian Dröge
b5f8b556e9 h264parser: Remove unused fps_num/fps_den fields
Instead the newly added function should be used to calculate
the framerate properly.

https://bugzilla.gnome.org/show_bug.cgi?id=723352
2016-05-28 10:54:01 +01:00
Tim-Philipp Müller
7d46d67c59 smoothstreaming: update fps calculation for h264 codec parser API changes
Use new gst_h264_video_calculate_framerate() API instead of fps_n/fps_d
fields in SPS struct which are to be removed.

Apparently H264 content in MSS is always non-interlaced/progressive,
so we can just pass 0 for field_pic_flag and don't need to parse any
slice headers first if there's no external signalling. But even if
that's not the case the new code is not worse than the existing code.

https://msdn.microsoft.com/en-us/library/cc189080%28VS.95%29.aspx

https://bugzilla.gnome.org/show_bug.cgi?id=723352
2016-05-28 10:29:20 +01:00
Reynaldo H. Verdejo Pinochet
e35bc2c2b4 dvbsrc: use single marker at end of filtering PID list
Avoids at least ~100 unneeded assignment operations at runtime
2016-05-26 16:18:56 -07:00
Reynaldo H. Verdejo Pinochet
faadd0f12a dvbsrc: simplify reporting of set polarity 2016-05-26 16:18:56 -07:00
Reynaldo H. Verdejo Pinochet
6a4fa2de75 dvbsrc: fix bandwidth-hz property description
Bandwidth Hz is no longer a DVBT-only property
2016-05-26 16:18:56 -07:00
Reynaldo H. Verdejo Pinochet
31a52b6c7c dvbsrc: fix usage of PES & DVR acronyms everywhere
Additionally, improve message on gst_poll_new() failure
2016-05-26 16:18:14 -07:00
Matthew Waters
e597e45008 gldisplay: always unref the context on a collision
Otherwise we will leak GstGLContext's when adding the same context more than
once.

Fixes a regression caused by 5f9d10f603 in the
gstglcontext unit test that failed with:

Assertion 'tmp == NULL' failed
2016-05-26 20:19:13 +10:00
Nicolas Dufresne
ff3b60a969 Revert "aggregator: Start the task when linked"
This reverts commit 302580c381.
2016-05-25 13:38:47 -04:00
Nicolas Dufresne
203e893e10 caopengllayersink: Don't cache buffer pool
Pools cannot be used by the two elements at the same time.

https://bugzilla.gnome.org/show_bug.cgi?id=766611
2016-05-25 13:35:59 -04:00
Nicolas Dufresne
302580c381 aggregator: Start the task when linked
Until now we would start the task when the pad is activated. Part of the
activiation concist of testing if the pipeline is live or not.
Unfortunatly, this is often too soon, as it's likely that the pad get
activated before it is fully linked in dynamic pipeline.

Instead, start the task when the first serialized event arrive. This is
a safe moment as we know that the upstream chain is complete and just
like the pad activation, the pads are locked, hence cannot change.

https://bugzilla.gnome.org/show_bug.cgi?id=757548
2016-05-25 13:35:00 -04:00
Xavier Claessens
b2cfc56e83 gl: win32: Add debug category in gstglcontext_wgl.c
https://bugzilla.gnome.org/show_bug.cgi?id=766867
2016-05-25 23:10:44 +10:00
Haihua Hu
a5cb746983 glimagesink: support video rotation using transform matrix
Add "rotate-method" to glimagesink and apply transform matrix
to vertex coordinate to control rotation.

https://bugzilla.gnome.org/show_bug.cgi?id=765795
2016-05-25 18:28:20 +10:00
Matthew Waters
061a157ef5 glvideomixer: fix race retrieving the GL context from the display
_get_gl_context() can be called concurrently from either propose_allocation() or
decide_allocation().  If it so happens that this happens at the same time,
the check for whether we already had a GL context was outside the lock.  Inside
the lock and loop, the first thing that happens is that we unref the current GL
context (if valid) as if there was a conflict adding it to the display.  If the
timing was unlucky, subsequent use of the GL context would be referencing an
already unreffed GL context object resulting in a critical:

g_object_ref: assertion 'object->ref_count > 0' failed

https://bugzilla.gnome.org/show_bug.cgi?id=766703
2016-05-25 18:28:20 +10:00
Matthew Waters
5f9d10f603 gldisplay: add some debugging about retrieving GL contexts 2016-05-25 18:28:20 +10:00
Sebastian Dröge
6f9b9e153a eglimage: Re-add accidentially removed GST_DEFINE_MINI_OBJECT_TYPE() 2016-05-25 10:01:38 +03:00
Guillaume Desmottes
586509d486 player: fix suburi getter debug message
The property is a string, not a boolean.

https://bugzilla.gnome.org/show_bug.cgi?id=766607
2016-05-25 09:55:28 +03:00
Sebastian Dröge
3af9fd7322 player: Don't set subtitle URI directly but dispatch to the player thread
Otherwise we do state changes and everything in the application thread, which
might block and more important can cause reentrant signals.

https://bugzilla.gnome.org/show_bug.cgi?id=766607
2016-05-25 09:53:15 +03:00
Sebastian Dröge
b730f7e89d eglimage: Ensure that the debug category is always initalized
Before the initializer was only run if dmabuf support was used.

https://bugzilla.gnome.org/show_bug.cgi?id=766794
2016-05-25 09:46:37 +03:00
Arjen Veenhuizen
113d5c143c gltransformation: make the pivot-z property READWRITE
Instead of just being READABLE.

https://bugzilla.gnome.org/show_bug.cgi?id=766818
2016-05-24 23:30:09 +10:00
Guillaume Desmottes
dff46e3239 gltestsrc: fix src_impl leak
https://bugzilla.gnome.org/show_bug.cgi?id=766661
2016-05-24 21:30:19 +10:00
Tim-Philipp Müller
db90f4686f g-i: pass compiler env to g-ir-scanner
It's what introspection.mak does as well. Should
fix spurious build failures on gnome-continuous.
2016-05-24 00:55:11 +01:00
Jan Schmidt
d0d22a9b2c pnmdec: Fix ASCII parsing
Parse gray16 properly in ascii mode, and fix
some bugs around reading data in chunks when
ascii values cross chunk boundaries
2016-05-24 07:21:44 +10:00
Jan Schmidt
1079e1c48c pnm: Add a basic unit test
Test that we can run an encode / decode cycle in various
formats without errors
2016-05-24 07:21:44 +10:00
Jan Schmidt
a2fde272f5 pnmenc: Make output caps match the srcpad template
Don't output 'image/pnm', which isn't even in the template
caps. Instead, output a mime-type consistent with what we
are encoding.
2016-05-24 07:21:44 +10:00
Jan Schmidt
fa2d98afe2 pnmenc: Tidy up code, support BE/LE GRAY16 properly
Remove code that dealt with odd strides separately - there's
not really any overhead to just using 1 codepath for both matched
and unmatched stride output.

Add separate codepaths for BE vs LE GRAY16 input so they're
handled properly
2016-05-24 07:21:44 +10:00
Jan Schmidt
30503f07ae pnmdec: Actually output LE or BE GRAY16 as negotiated
Add codepaths to output GRAY16 in little or big endian
as negotiated.

Move all output format negotiation into a central
function and clean it up
2016-05-24 07:21:44 +10:00
Dimitrios Katsaros
88a9b151c1 pnmdec: Implementation of GRAY16 handling
https://bugzilla.gnome.org/show_bug.cgi?id=757022
2016-05-24 07:21:44 +10:00
Dimitrios Katsaros
61440e58c1 pnmenc: Implementation of GRAY16 handling
https://bugzilla.gnome.org/show_bug.cgi?id=757022
2016-05-24 07:21:44 +10:00
Víctor Manuel Jáquez Leal
f778a3f19a kmsbufferpool: error only if no allocator
Do not expect an allocator from the configuration, but expect to have already
one set.
2016-05-23 22:06:29 +02:00
Víctor Manuel Jáquez Leal
3b872650aa kmssink: frame copy log in performance category
Log the message when the frame is going to be copy (worse case) under the
performance log category.

https://bugzilla.gnome.org/show_bug.cgi?id=766466
2016-05-23 22:06:29 +02:00
Víctor Manuel Jáquez Leal
e87772dd4c kmssink: use trace level for buffer render log
Instead of using debug level for logging each buffer rendered, use trace
level.

https://bugzilla.gnome.org/show_bug.cgi?id=766466
2016-05-23 22:06:29 +02:00
Arun Raghavan
18f8895f33 player: Fix documentation for gst_player_video_info_get_framerate()
Document out parameters and fix function name.
2016-05-23 17:17:55 +05:30
Guillaume Desmottes
0f19a48c98 player: use gst_check_init() in test
Calling GST_DEBUG() in test rely on the default category to be defined,
which is done in gst_check_init().

https://bugzilla.gnome.org/show_bug.cgi?id=766607
2016-05-21 09:57:54 +03:00
Aaron Boxer
e76705dbde openjpegenc: enable MCT for RGB video
https://bugzilla.gnome.org/show_bug.cgi?id=766732
2016-05-21 09:44:22 +03:00
Reynaldo H. Verdejo Pinochet
5a47f4a8bc dvbsrc: group DVB-T2 substream-id check with its corresponding set_prop() 2016-05-20 14:08:03 -07:00