Commit graph

20917 commits

Author SHA1 Message Date
Matthew Waters
5115eadc46 glupload: keep the parent buffer around when converting EGLImage to 2D textures
e.g. receiving and releasing a buffer from OMX too early will potentially
cause textures to be overwritten while/before they are displayed.
2016-04-12 09:41:23 +10:00
Matthew Waters
417bae7f23 glbasemixer: chain up to the parent implementation 2016-04-12 09:39:32 +10:00
Matthew Waters
ab33030250 vulkan: prettify the various versions in the debug output
output human readable numbers.
2016-04-12 09:39:12 +10:00
Tim-Philipp Müller
cc2b604b3b openh264dec: minor clean-ups
Remove unused property getters, setters; outdated comment.
2016-04-11 20:23:45 +01:00
Tim-Philipp Müller
9617a371ca openh264: remove unnecessary instance private structures
Element instance structures are not public API, so no need
for the extra indirection to hide the data.
2016-04-11 20:11:53 +01:00
Tim-Philipp Müller
41d8a49c52 openh264: fix indentation 2016-04-11 20:04:00 +01:00
Philipp Zabel
2f51985191 kmssink: enable UYVY, YUY2, and YVYU (interleaved YUV 4:2:2) formats
https://bugzilla.gnome.org/show_bug.cgi?id=761059
2016-04-11 19:57:48 +02:00
Philipp Zabel
360e9343f9 kmssink: enable NV16 (chroma-interleaved YUV 4:2:2) format
https://bugzilla.gnome.org/show_bug.cgi?id=761059
2016-04-11 19:57:48 +02:00
Philipp Zabel
bdb62b2658 kmssink: enable Y42B (planar YUV 4:2:2) format
https://bugzilla.gnome.org/show_bug.cgi?id=761059
2016-04-11 19:57:48 +02:00
Víctor Manuel Jáquez Leal
7d06cf3297 kmssink: keep last rendered buffer in memory
https://bugzilla.gnome.org/show_bug.cgi?id=761059
2016-04-11 19:57:48 +02:00
Víctor Manuel Jáquez Leal
b29f7d048c kmssink: wait for page flip or vblank
This patch requests for drmModePageFlip() for the used CRTC, if the kernel
module suppports async page flip. If it does not, the element requests for a
vblank event. A GstPoll waits for the event to happen.

https://bugzilla.gnome.org/show_bug.cgi?id=761059
2016-04-11 19:57:48 +02:00
Víctor Manuel Jáquez Leal
c419d17dbf kmssink: add dmabuf support
This patch will enable the import of dmabufs into a KMS buffer using
the PRIME kernel interface.

If the driver does not support prime import, the method is skipped.

It has been tested with a Freescale I.MX6 board.

https://bugzilla.gnome.org/show_bug.cgi?id=761059
2016-04-11 19:57:48 +02:00
Víctor Manuel Jáquez Leal
1aee6cdc25 kmssink: calculate display ratio
Get the aspect ratio given the information provided by libdrm, and with it
calculate the display ratio.

https://bugzilla.gnome.org/show_bug.cgi?id=761059
2016-04-11 19:57:48 +02:00
Víctor Manuel Jáquez Leal
620e1d2fcd kmssink: add plugin and sink element
This is simple video sink that use libdrm/libkms API to render frames.

The element uses planes to render through drmModeSetPlane().

It has been tested in an Exynos4412 board and in a Freescale I.MX6 board.

https://bugzilla.gnome.org/show_bug.cgi?id=761059
2016-04-11 19:57:48 +02:00
Matthew Waters
65770c2af9 vulkan: add a wayland winsys implementation 2016-04-11 14:18:47 +10:00
Matthew Waters
e46bd98fe6 vulkan/xcb: getProcAddress winsys functions
There's no guarantee that they will always be exposed by the vulkan loader.
2016-04-11 14:18:47 +10:00
Matthew Waters
b76fcfd3f7 vkupload: copy necessary buffer metadata
timestamps, flags, etc
2016-04-11 14:18:47 +10:00
Alessandro Decina
b73d2586ce libgstgl: cocoa, eagl: use libdispatch to schedule GL calls
Use libdispatch instead of GMainLoop to dispatch GL calls. libdispatch is more
optimized and cuts a lot of poll()/pthread_* overhead.
2016-04-11 08:40:03 +10:00
Alessandro Decina
38cb69279b glwindow: reduce the number of GMutexes and GConds in send_message()
Don't create many short lived locks/conds in gst_gl_window_send_message. This is
a micro optimization to save a bunch of pthread_* calls which are expensive on
OSX/iOS and possibly other platforms.
2016-04-11 08:40:03 +10:00
Tim-Philipp Müller
519c942eb3 win32: update .def file for new gl api 2016-04-10 22:18:38 +01:00
Matthew Waters
09808f9780 vkswapper: add necessary wait stages to the submission info 2016-04-08 21:46:50 +10:00
Matthew Waters
e4377e5c8f vkswapper: add srgb formats to our list of formats
intel returns them for their surface formats.
2016-04-08 21:46:50 +10:00
Matthew Waters
61d1b31136 vulkan: silence a validation warning filling out the image barrier 2016-04-08 21:46:50 +10:00
Matthew Waters
cecbcd704b vulkan: use the correct validation layer name
The threading validation layer has changed names.
2016-04-08 21:46:50 +10:00
Sebastian Dröge
da999d7af8 decklink: Initialize capture_time to NONE so it can be used if no video frame is provided but audio is
CID 1358390
2016-04-08 14:08:31 +03:00
Matthew Waters
761171b6d2 vulkan: only warn on not found layers
don't error out completely

https://bugzilla.gnome.org/show_bug.cgi?id=764545
2016-04-08 17:56:50 +10:00
Matthew Waters
b32ca4b550 vulkan: don't g_error when failing to find validation layers
Aborting the application is never a good thing to do for recoverable errors.

Just warn and error out instead.

https://bugzilla.gnome.org/show_bug.cgi?id=764545
2016-04-08 17:42:46 +10:00
Matthew Waters
dc1a326090 nvenc: add some rate control modes/properties
Supported modes are constant quantizer, cbr, vbr and vbr with a minimum quantizer.
2016-04-08 15:24:20 +10:00
Matthew Waters
065f4dceb1 nvenc: move codec config initialization from the implementation to the base class
Supports a better separation of configuration parameters
2016-04-08 15:08:18 +10:00
Matthew Waters
c45fc2ef53 nvenc: add preset selection
Some presets are not always supported on all devices and will cause an error if
used.  Specifically, the LOSSLESS presets are known to not work everywhere.
2016-04-08 15:08:18 +10:00
Marcin Kolny
ac8a14d1c8 chromaprint: emit notify::fingerprint signal when fingerprint is ready
In addition to adding the fingerprint to the tags.

https://bugzilla.gnome.org/show_bug.cgi?id=733233
2016-04-07 20:44:09 +01:00
Matthew Waters
aa2b23fe39 videoaggregator: repect the result of find_best_format in the default update_caps
We weren't using the result of find_best_format at all.

Also, move the find_best_format usage to the default update_caps() to make
sure that it is also overridable.

https://bugzilla.gnome.org/show_bug.cgi?id=764363
2016-04-07 20:30:25 +10:00
Nicolas Dufresne
81960ea853 rfbsrc: Fix fallback to GstVideoBufferPool
The replacement pool need to be added to the query, otherwise the
baseclass won't see it. We also need to properly remove the rejected
pools.

https://bugzilla.gnome.org/show_bug.cgi?id=763441
2016-04-05 21:40:58 -04:00
Nicolas Dufresne
7e293f15dd rfbsrc: Implement decide_allocation virtual
This way we can use the base class for buffer allocation, hence use
fill() instead of create() virtual. This also adds a strict check on the
select pool buffer size as we don't support strides and padding.

This is based on initial patch proposed by Sebastien Dröge, from which I
also fixed a buffer pool leak.

https://bugzilla.gnome.org/show_bug.cgi?id=763441
2016-04-05 21:21:29 -04:00
Nicolas Dufresne
d3d34b5a8c rfbsrc: Fix caps negotiation
As we currently only use the server reported "natural" format, caps
negotiation should simply be limited to telling the base class which
format to use. Fix the negotiation by moving the associated code
into negotiate() virtual function. Also, use gst_base_src_set_caps()
rather then setting it on the pad directly. Also protect against this
method being called multiple time (we can't renegotiate for now).

This change also moves some network code that was being run during the
application state change call, to be run on the streaming thread.

https://bugzilla.gnome.org/show_bug.cgi?id=739598
2016-04-05 15:36:31 -04:00
Nicolas Dufresne
52b50d0952 rfbsrc: Implement unlock() method
This prevent stalls when we try to stop the source while waiting
for the server to send more updates.
2016-04-05 13:35:38 -04:00
Nicolas Dufresne
89f3f162a5 rfbsrc: Check for connection being closed
Although it's not very well documented, g_input_stream_read_all() will
set the number of bytes read to 0 if the connection is closed rather
then returning an error.
2016-04-05 13:35:25 -04:00
Nicolas Dufresne
0eb5722af6 rfbsrc: Check for read/write error
Check for read/write error. This prevent undefined behaviour that rely
on unitialized buffer.
2016-04-05 13:35:18 -04:00
Nicolas Dufresne
ba93b86011 rfbsrc: Add a internal method to disconnect
This also removes the disconnected boolean hack.
2016-04-05 13:35:11 -04:00
Nicolas Dufresne
ff5dc55c8a rfbsrc: Don't ignore errors
This prevents recursion on error. This used to happen as we
don't change the state when something fails. We end up running
and failing in the same state forever.
2016-04-05 13:34:57 -04:00
Nicolas Dufresne
459bdaae3f rfbsrc: Serialize write operations
Currently we send key events from seperate threads. IOStream does not
allow concurrent write operations, so protect this operation using a
mutex.
2016-04-05 13:34:38 -04:00
Nicolas Dufresne
c8d2e234c4 rfbsrc: Add 3.7/3.8 security type negotiation 2016-04-05 13:34:34 -04:00
Nicolas Dufresne
21b322c8f3 rfbsrc: Port to GSocketClient
Using GSocketClient we can simplify a lot the read/write operation.
This also provide an GSocketConnection (a GIOStream) which can then
be used with the GTlsClientConnection for secure connections. Note
that we use _write_all() to ensure all bytes have been read. This is
to follow the fact the none of the _send() calls check the return
value.
2016-04-05 13:34:07 -04:00
Nicolas Dufresne
4e0a5c101f rfbsrc: Remove useless library wrapping
We use a sinlge .c file, simplify everything by removing that
static library.
2016-04-05 13:34:01 -04:00
Nicolas Dufresne
29f1ca8602 rfbsrc: Cleanup dead code
Remove all the unused stuff, the include header, definition and C code
that don't contribute anything to this element.
2016-04-05 13:33:50 -04:00
Nicolas Dufresne
7b4b8a95b4 rfbsrc: Properly fallback to 3.3
If the major version is not 3, then the minor version provided
should be ignored, as it is meaningless for our fallback.
2016-04-05 13:33:38 -04:00
Nicolas Dufresne
b83a6967c9 rfbsrc: Cleanly handle security negotiation failure
When the security cannot be negotiated, the server returns
security type of 0 (failure). In that case, the next step is
to read the error reason string.
2016-04-05 13:33:24 -04:00
Vincent Penquerc'h
8ac261841c mpdparser: clamp segment durations to each other and period end
5.3.2.1 in the spec.

https://bugzilla.gnome.org/show_bug.cgi?id=751792
2016-04-05 17:01:01 +01:00
Florin Apostol
7d8cc26c91 dashdemux: unit testing reproducing segment duration overflow
unit test reproducing https://bugzilla.gnome.org/show_bug.cgi?id=751792

With minor changes by Vincent Penquerc'h.
2016-04-05 17:01:00 +01:00
Sebastian Dröge
f06647fd0c decklinkaudiosrc: Don't accept packets without timestamps after a discont
We have no idea which timestamps they are supposed to have so the only thing
we can do at this point is to drop them. Packets without timestamps happen if
audio was captured but no corresponding video, which shouldn't happen under
normal circumstances.

https://bugzilla.gnome.org/show_bug.cgi?id=747633
2016-04-05 14:49:51 +03:00