Commit graph

3295 commits

Author SHA1 Message Date
Matthew Waters
16cba63d43 gl/utils: fixup matrix math again for column major
e4bf9ed8f0 was not quite right and changed
the wrong thing. Intead we needed to change the multiplication order
and should have kept the previous to/from matrices as is done in this
patch.
2018-07-10 20:13:22 +10:00
Sebastian Dröge
3dd95b1311 oggdemux: Make sure that events are writable before changing their seqnum 2018-06-18 13:28:15 +03:00
Matthew Waters
6f04c3ac87 glalpha: fix compilation with msvc
It doesn't define M_PI
2018-06-07 14:23:55 +10:00
Matthew Waters
4bd93dc4d8 gldownloadelement: fix build with msvc
msvc doesn't like #ifdef inside macro expansion
2018-06-06 23:36:08 +10:00
Matthew Waters
e4bf9ed8f0 gl/utils: Fix NDC conversion matrices for column-majorness
The matrices were converting the wrong values with non-diagonal-only matrices.
e.g. a typical yflip matrix in [-1,1]^3 such as
 1  0  0  0
 0 -1  0  0
 0  0  1  0
 0  0  0  1

Would have actually required a matrix like this in [0,1]^3
 1  0  0  0
 0 -1  0  0
 0  0  1  0
 0 -2  0  1

Which is
1. not consistent with our multiplication convention and would require
   transposing matrices or changing our multiplication order (from what is
   generally used on opengl matrix guides/tutorials).
2. Produces incorrect values when input with actual vertices accounting for
   the difference in multiplication order.  e.g. some vertices multiplied by
   the yflip matrix using vertex * yflip(== transpose(yflip) * vertex):

     vertex:       -> result:           expected:
     vec4(1,0,1,1) -> vec4(1,-2,1,1)    vec4(1,1,1,1)
     vec4(1,1,1,1) -> vec4(1,-3,1,1)    vec4(1,0,1,1)

With the updated values, we now get the expected values.

Includes a test for this behaviour and the example above
2018-06-06 23:36:08 +10:00
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
Jan Schmidt
567bbe3c07 glcolorbalance: Support OES textures for input/passthrough
glcolorbalance is in the default GL path inside glimagesink,
so has been causing an possibly-unnecessary extra texture copy
on Android for a while now. If we're just doing passthrough,
we can support OES directly. If not, they'll be transformed
to 2D textures and colourbalanced.
2018-05-07 01:16:27 +10:00
Matthew Waters
c150928756 glbasefilter: expose finding the local GL context
And use it to attempt to find the GL context sooner for upload and color
conversion caps transformations.
2018-05-05 21:24:25 +10:00
Matthew Waters
af5b0ee7d0 gl: add glalpha element that is similar to the alpha element
It performs similarly to the existing alpha element however performs
calculations in floating point rather than with small (guint8) integers
so some differences are to be expected.

https://bugzilla.gnome.org/show_bug.cgi?id=794070
2018-05-05 21:24:25 +10:00
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
Tim-Philipp Müller
762733c3d8 meson: gl: fix 'invalid keyword argument' meson warnings
Required is not a valid kwarg for cc.has_header()
2018-04-20 11:44:55 +01:00
Mathieu Duponchelle
32ca1701b7 alsamidisrc: unref buffer_list before early return 2018-04-13 01:07:21 +02:00
Mathieu Duponchelle
44ea6f4c06 vorbisenc: do not map input buffer in WRITE mode 2018-04-12 15:47:27 +02:00
Sebastian Dröge
43145292d4 gl: Disable glmixerbin for the time being too
Otherwise we have one copy in gst-plugins-bad and one (unused) here,
which makes static linking unhappy.
2018-03-25 12:48:43 +03:00
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
Tim-Philipp Müller
ae9911617e Revert "alsasrc: lock calls to snd_pcm_delay() with mutex as in alsasink"
This reverts commit 4f88125b3d.

This seems to have undesirable side-effects and needs more
investigation first.

https://bugzilla.gnome.org/show_bug.cgi?id=746015
2018-03-07 18:50:25 +00:00
Mathieu Duponchelle
22981e8a42 Port to latest GstAggregator segment API
The aggregator segment is now exposed on the src pad

https://bugzilla.gnome.org/show_bug.cgi?id=793944
2018-03-01 15:33:25 +01:00
Jan Schmidt
09584edc9a glimagesink: Correct PAR in output caps when transforming
When doing a 3D/multiview transformation and rescaling to
match the output window size, the resulting PAR may
not match the input any more and needs recalculating,
or else the GstSample reported to client-draw has the
wrong PAR.
2018-02-23 19:39:37 +11:00
Sebastian Dröge
1cfe070de7 gl: Link to libgstallocators for the dmabuf allocator 2018-02-17 18:34:58 +02:00
Tim-Philipp Müller
466b868459 vorbisparse: error out when headers are missing
https://bugzilla.gnome.org/show_bug.cgi?id=791606
2018-02-14 00:22:38 +00:00
Edward Hervey
aab5cccc34 vorbisdec: Improve "new headers while initialized" handling
If new headers arrive after we are initialized, we need to make
sure that they are indeed valid.

A vorbis bitstream always begins with three header packets and must
be in order.

Also some streams have unframed (invalid?) headers that might
confuse and disrupt the decoding process.

Therefore if ever we see new headers, we accumulate them and once
we get a non-header packet we check them to make sure that:
* We have at least 3 headers
* They are the expected ones (identification, comments and setup)
* They are in order
* Any other "header" is ignored

If those conditions are met, we reset and reconfigure the decoder

https://bugzilla.gnome.org/show_bug.cgi?id=784530
2018-02-13 08:41:29 +01:00
Jan Schmidt
46260654a4 glimagesink: Always display with requested stereo display mode
Even if the input is monoscopic, the app might want to display
it in a different layout, to do side-by-side for VR for example,
so if the app changes the output-multiview-mode always use that.
2018-02-09 18:16:04 +11:00
Matthew Waters
fbef9220d3 gl*bin: fix transfer semantics for the create-element signal
We can either receive an element that is floating or not and need to
accomodate that in the signal return values.  Do so by removing the
floating flag.

https://bugzilla.gnome.org/show_bug.cgi?id=792597
2018-02-09 17:15:30 +11:00
Matthew Waters
0d69b9d3cc gldownload: remove texture-target field from dmabuf/sysmem caps
https://bugzilla.gnome.org/show_bug.cgi?id=792342
2018-02-09 12:33:28 +11:00
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
Nicolas Dufresne
17b118c120 glimagesink: Allow resetting render rectangle
As documented, passing -1 to x and/or y should reset the render
rectangle to the window/display size.

https://bugzilla.gnome.org/show_bug.cgi?id=792798
2018-01-30 09:58:14 +00:00
Nicolas Dufresne
3b317ea6c2 glimagesink: Add render-rectangle property
This allow controlling the render rectangle from gst-launch-1.0.

https://bugzilla.gnome.org/show_bug.cgi?id=792798
2018-01-30 09:58:14 +00:00
Branislav Katreniak
4f88125b3d alsasrc: lock calls to snd_pcm_delay() with mutex as in alsasink
Alsasrc introduced delay_lock in commit 519f85a43e73efb8f3fb2c7be45226e
because alsa-lib is not thread safe for the same handle.
Alsasrc uses the same threading pattern, it should be locked too.

https://bugzilla.gnome.org/show_bug.cgi?id=746015
2018-01-22 20:24:00 +00:00
Edward Hervey
d36a22553b theoradec: Check for valid width/height
If width or height are zero ... there's no video :)
2018-01-17 14:35:51 +01:00
Tim-Philipp Müller
2ee4ab1ebb meson: gl: define HAVE_PNG etc. in config.h instead of c_args
So that the tests have access to them as well (gl-launch-lines.c)
2017-12-22 22:28:20 +01:00
Tim-Philipp Müller
0dd66648c3 gl: fix up a few more #if HAVE_FOO
error: "HAVE_PNG" is not defined, evaluates to 0 [-Werror=undef]
2017-12-19 16:51:24 +00:00
Tim-Philipp Müller
8227135f3b gl: hook up to build system
Tests and documentation will follow separately.

The mixer elements in the opengl plugin need to stay
in -bad for now since they use GstVideoAggregator.

https://bugzilla.gnome.org/show_bug.cgi?id=754094
2017-12-19 12:01:48 +00:00
Tim-Philipp Müller
26b3955aba gl: use #ifdef HAVE_* instead of #if HAVE_FOO 2017-12-19 12:01:41 +00:00
Tim-Philipp Müller
59fe9c9122 gl: remove GST_USE_UNSTABLE_API bits 2017-12-19 12:01:37 +00:00
Tim-Philipp Müller
729b500208 opus: remove unused build variable and define 2017-12-19 12:01:32 +00:00
Tim-Philipp Müller
cdbb261d9c Move OpenGL library and plugin from -bad
Merge branch 'opengl-move'

https://bugzilla.gnome.org/show_bug.cgi?id=754094
2017-12-19 12:00:43 +00:00
Víctor Manuel Jáquez Leal
8f0bf85cfe gldownload: fix wrong enum
When compiling with clang, an enum conversion error is triggered
since GstVideoFrameFlags are not GstVideoFlags.

This patch sets GST_VIDEO_FRAME_FLAG_NONE to the added video meta.

https://bugzilla.gnome.org/show_bug.cgi?id=791251
2017-12-09 19:32:31 +00:00
Nicolas Dufresne
dc5dd864e6 gldownload: Add missing ifdef for dmabuf and egl
This fixes the build for platforms that don't support one or the other.
2017-12-09 19:32:31 +00:00
Matt Fischer
c98f92ee03 gldownload: Add dmabuf exporting
This patch adds code to gldownload to export the image as a
dmabuf if requested.  The element now exposes memory:DMABuf as
a cap feature, and if it is selected, the element exports the
texture to an EGL image and then a dmabuf. It also implements a
fallback to system memory download in case the exportation failed.

https://bugzilla.gnome.org/show_bug.cgi?id=776927
2017-12-09 19:32:31 +00:00
Tim-Philipp Müller
b5dee0fe5b Remove GstAggregator from -bad, moved to core
https://bugzilla.gnome.org/show_bug.cgi?id=739010
2017-12-09 19:32:31 +00:00
Matthew Waters
1abd0804f4 gl/caopengllayer: use public GstGLContext instead of Cocoa-specific one
Allows keeping the GstGLCAOpenGLLayer public but not the winsys-specific
context/display/window.
2017-12-09 19:32:31 +00:00
Mathieu Duponchelle
d5addcb352 aggregator: Remove klass->sinkpads_type
This posed problems for the python bindings (and possibly others).

Instead, subclasses now use add_pad_template_with_gtype.

https://bugzilla.gnome.org/show_bug.cgi?id=789986
2017-12-09 19:32:31 +00:00
Matthew Waters
0063e63fe3 gltestsrc: guard stop in gl thread
So we don't result in a critical when we've never created the GL context:

gst_gl_context_thread_add: assertion 'GST_IS_GL_CONTEXT (context)' failed
2017-12-09 19:32:31 +00:00
Matthew Waters
5793dfe881 gltestsrc: free in the GL thread on stop
Calling GL functions not on the GL thread may crash in some GL
implementations.

https://bugzilla.gnome.org/show_bug.cgi?id=789665
2017-12-09 19:32:30 +00:00
Tim-Philipp Müller
77277c46f9 gl: use new gst_element_foreach_sink_pad()
Instead of gst_aggregator_iterate_sinkpads() which will
soon be removed.

https://bugzilla.gnome.org/show_bug.cgi?id=785679
2017-12-09 19:32:30 +00:00
Ponnam Srinivas
601e274c95 glmixer: Unmap video frame in error case
https://bugzilla.gnome.org/show_bug.cgi?id=788127
2017-12-09 19:32:30 +00:00
Haihua Hu
29973b7d6c glvidemixer: need reconfigure output gemotry after caps renegotiated 2017-12-09 19:32:30 +00:00
Matthew Waters
9fbd1930b7 Revert "glvideomixer: need update output geometry after src caps reconfigure"
This reverts commit d6e538dc56.
2017-12-09 19:32:30 +00:00
Haihua Hu
60eeef12a0 glvideomixer: need update output geometry after src caps reconfigure
Need update output geometry when sink caps changed and use
gst_structure_set to update caps if structure is fixed

https://bugzilla.gnome.org/show_bug.cgi?id=787820
2017-12-09 19:32:30 +00:00