Víctor Manuel Jáquez Leal [Mon, 21 Nov 2016 17:42:02 +0000 (18:42 +0100)]
vaapidecode: negotiate after destroying allocator
This is related with bug 758907 when no vaapipostproc is used (no
vaapidecodebin). In order to negotiate downstream we need to destroy
the source pad allocator, otherwise the same allocated buffers are
used, failing the mapping.
Víctor Manuel Jáquez Leal [Mon, 21 Nov 2016 15:35:34 +0000 (16:35 +0100)]
vaapivideomemory: remove GST_VAAPI_TYPE_VIDEO_INFO
Remove redundant GST_VAAPI_TYPE_VIDEO_INFO, since it is a duplicate of
GST_TYPE_VIDEO_INFO created before gstreamer 1.6, where the boxed type
was created.
https://bugzilla.gnome.org/show_bug.cgi?id=774782
Víctor Manuel Jáquez Leal [Mon, 21 Nov 2016 11:51:25 +0000 (12:51 +0100)]
plugins: update the src pad allocator video info
Update the size, stride and offset of the source pad allocator video
info, so the pool could set the correct GstVideoMeta
https://bugzilla.gnome.org/show_bug.cgi?id=774782
Víctor Manuel Jáquez Leal [Mon, 21 Nov 2016 11:36:27 +0000 (12:36 +0100)]
vaapivideomemory: add missing documentation
https://bugzilla.gnome.org/show_bug.cgi?id=774782
Víctor Manuel Jáquez Leal [Mon, 21 Nov 2016 11:29:26 +0000 (12:29 +0100)]
plugins: use early return without goto
https://bugzilla.gnome.org/show_bug.cgi?id=774782
Víctor Manuel Jáquez Leal [Mon, 21 Nov 2016 10:25:21 +0000 (11:25 +0100)]
plugins: add allocator to allocation query
This patch adds the created allocator to the allocation query either
in decide_allocation() and propose_allocation() vmehtods.
With it, there's no need to set the modified allocator's size in the
pool configuration.
https://bugzilla.gnome.org/show_bug.cgi?id=774782
Julien Isorce [Wed, 19 Oct 2016 14:15:01 +0000 (15:15 +0100)]
vaapivideobufferpool: set correct buffer size
We should set the correct buffer size when we are configuring the pool,
otherwise the buffer will be discarded when it returns to the pool.
Indeed when the ref-count of a buffer reaches zero, its pool will queue
it back (and ref it) if, and only if, the buffer size matches the
configured buffer size on the pool.
This issue can be debugged with GST_DEBUG=*PERF*:6, see gstbufferpool.c
https://bugzilla.gnome.org/show_bug.cgi?id=774782
Víctor Manuel Jáquez Leal [Thu, 10 Nov 2016 12:26:31 +0000 (13:26 +0100)]
vaapivideomemory: check for memory allocator
When calling gst_vaapi_video_memory_copy() the allocator of the memory
to copy should be allocated by the vaapi allocator.
This patch does this verification.
Víctor Manuel Jáquez Leal [Thu, 10 Nov 2016 12:25:30 +0000 (13:25 +0100)]
vaapivideomemory: code style fixes
A cosmetic commit for enhance readability of the casts and method
preconditions.
Víctor Manuel Jáquez Leal [Wed, 9 Nov 2016 18:49:22 +0000 (19:49 +0100)]
vaapivideomemory: unroll gst_vaapi_video_allocator_free()
Instead of having a gst_vaapi_video_memory_free() that is only going to
be called by gst_vaapi_video_allocator_free(), let's just remove the first
and merged into the second.
Víctor Manuel Jáquez Leal [Wed, 9 Nov 2016 18:29:12 +0000 (19:29 +0100)]
vaapivideomemory: avoid virtual methods casting
Use the expected virtual method signatures for readability.
Víctor Manuel Jáquez Leal [Wed, 9 Nov 2016 17:58:20 +0000 (18:58 +0100)]
vaapivideomemory: remove unimplemented methods
Remove unimplemented method for allocator mem_share() and mem_is_span().
Víctor Manuel Jáquez Leal [Wed, 9 Nov 2016 17:54:47 +0000 (18:54 +0100)]
vaapivideomemory: fail if frame map can't get plane
If map() vmethod in GstVideMeta cannot get the plane data, return false,
thus the caller will not try to read invalid memory.
https://bugzilla.gnome.org/show_bug.cgi?id=774213
Víctor Manuel Jáquez Leal [Wed, 9 Nov 2016 17:39:06 +0000 (18:39 +0100)]
vaapivideomemory: lock map and unmap operations
In order to avoid race condition when two threads call map/unmap the same
VA surface, this patch mutex these operations.
https://bugzilla.gnome.org/show_bug.cgi?id=774213
Víctor Manuel Jáquez Leal [Wed, 9 Nov 2016 16:37:06 +0000 (17:37 +0100)]
vaapivideomemory: refactor vaapi memory unmapping
There were duplicated code in gst_video_meta_unmap_vaapi_memory() and
gst_vaapi_video_memory_unmap() when unmapping.
This patch refactors both methods adding the common function
unmap_vaapi_memory(). This also ensures, if direct rendering is enabled, it
is correctly reset.
Additionally, only when mapping flag has the WRITE bit, it set the image as
current, which was done in gst_video_meta_map_vaapi_memory() but no in
gst_vaapi_video_memory_map().
In order to make this, the mapping flags were required, so instead of
overloading mem_unmap() virtual function, mem_unmap_full() is overloaded.
https://bugzilla.gnome.org/show_bug.cgi?id=774213
Víctor Manuel Jáquez Leal [Wed, 9 Nov 2016 12:54:23 +0000 (13:54 +0100)]
vaapivideomemory: refactor vaapi memory mapping
There were duplicated code in gst_video_meta_map_vaapi_memory() and
gst_vaapi_video_memory_map() when doing the READ and WRITE mapping.
This patch refactors both methods adding the common function
map_vaapi_memory().
Additionally, only when flag has the READ bit it calls
ensure_images_is_current(), which was done in
gst_video_meta_map_vaapi_memory() but no in
gst_vaapi_video_memory_map().
https://bugzilla.gnome.org/show_bug.cgi?id=772151
Víctor Manuel Jáquez Leal [Thu, 27 Oct 2016 16:22:18 +0000 (18:22 +0200)]
vaapidecode: unref allowed_sinkpad_caps at close()
The variable member allowed_sinkpad_caps is constructed querying the
current VA display. Bearing that in mind, the variable shall be freed
when the VA display changes or is removed.
This patch moves the freeing of allowed_sinkpad_caps to close(), when
the VA display is freed.
Hyunjun Ko [Fri, 11 Nov 2016 02:40:09 +0000 (11:40 +0900)]
vaapisink: finish event thread at stop()
The thread that handles window's events should be finished during
pipeline's shutdown, otherwise it will remain alive during pipeline
re-activation, leading to unexpected problems.
This patch fixes failures of intensive_state_change scenario of
gst-validate
https://bugzilla.gnome.org/show_bug.cgi?id=774241
Víctor Manuel Jáquez Leal [Tue, 8 Nov 2016 08:35:00 +0000 (09:35 +0100)]
vaapipostproc: enhance debug message
"gst_pad_push" is not a good description of the event.
Hyunjun Ko [Tue, 8 Nov 2016 01:05:32 +0000 (10:05 +0900)]
postproc: honor gst_pad_push() return value
Returning GST_FLOW_ERROR always when gst_pad_push fails might lead to
deadlock during seek.
This patch returns the same error of gst_pad_push() and log out the
return value.
https://bugzilla.gnome.org/show_bug.cgi?id=774030
Víctor Manuel Jáquez Leal [Fri, 4 Nov 2016 15:26:18 +0000 (16:26 +0100)]
vaapidecode: guard GST_VAAPI_MAKE_GLTEXUPLOAD_CAPS
In commit 6d11a00 were introduced a regression when gstreamer-vaapi is
compiled with out EGL/GLX support: it shall not support
GST_VAAPI_MAKE_GLTEXUPLOAD_CAPS.
This patch guards the inclusion of GST_VAAPI_MAKE_GLTEXUPLOAD_CAPS in the
allowed src caps for vaapedecode if EGL/GLX.
Víctor Manuel Jáquez Leal [Fri, 4 Nov 2016 11:55:23 +0000 (12:55 +0100)]
vaapivideomemory: increment map counter only if succeeded
Previously the frame map counter increased independently if the map succeeded
or not. This leaded to critical messages and crashes if the frame was unable
to be mapped, but the counter increased.
This patch increases the map counter only if the map operation occurred.
https://bugzilla.gnome.org/show_bug.cgi?id=773939
Víctor Manuel Jáquez Leal [Thu, 3 Nov 2016 16:30:46 +0000 (17:30 +0100)]
plugins: set negotiation caps in src allocator
When the allocator is created, it stores the allocation caps. But sometimes
the "allocation caps" may be different from the "negotiation caps".
In this case, the allocator should store the negotiation caps since they
are the ones used for frame mapping with GstVideoMeta.
When vaapispostproc is used, this is not a problem since the element is assume
to resize. But when using a vaapi decoder only, with a software renderer, it
fails in this case.
https://bugzilla.gnome.org/show_bug.cgi?id=773323
Víctor Manuel Jáquez Leal [Tue, 19 Jul 2016 14:40:10 +0000 (16:40 +0200)]
vaapidecode: generate source pad caps
Just as vaapipostproc, VA decoder's context can be queried to get the possible
raw formats, so, the src caps can negotiate the exact caps that the context
supports.
Víctor Manuel Jáquez Leal [Thu, 25 Feb 2016 17:57:30 +0000 (18:57 +0100)]
libs: decoder: add _get_surface_formats()
This function exposes the available formats of the surfaces in the the current
context to the plugins.
Víctor Manuel Jáquez Leal [Thu, 18 Feb 2016 18:32:58 +0000 (19:32 +0100)]
libs: context: ensure context formats
This patch ensures to get the formats, as filter does, available in the
decoder / encoder context.
The context fills up the array as soon it is created, otherwise the pipeline
could get stalled (perhaps this is a bug in my HSW backend).
https://bugzilla.gnome.org/show_bug.cgi?id=752958
Víctor Manuel Jáquez Leal [Thu, 18 Feb 2016 18:20:10 +0000 (19:20 +0100)]
libs: move get_surface_formats to utils_core
The query of all the supported formats for a VA config were only used by the
postprocessor (vaapifilter). But, in order to enable the vaapidecoder to
negotiate a suitable raw format with downstream, we need to query these
formats against the decoder's config.
This patch is the first step: moves the code in filter's ensure_image() to a
generic gst_vaapi_get_surface_formats() in vaapiutils_core, so it can be
shared later by the decoder.
https://bugzilla.gnome.org/show_bug.cgi?id=752958
Víctor Manuel Jáquez Leal [Tue, 19 Jul 2016 16:56:15 +0000 (18:56 +0200)]
plugins: remove set_sinkpad_dmabuf_allocator()
Since when the sink pad allocator is created, it is decided if the required
one is vaapi allocator or dmabuf allocator, there is no need to force its set
again.
Víctor Manuel Jáquez Leal [Thu, 27 Oct 2016 09:25:55 +0000 (11:25 +0200)]
plugins: ensure display when getting raw caps
When running gst-discoverer-1.0, in certain media, vaapipostroc is stopped
meanwhile it is transforming caps. The problem is that stop() calls
gst_vaapi_plugin_base_close(), which nullifies the element's va display, but
the va display is used in tranform_caps() when it is extracting the possible
format conversions. This display disappearing generates warning messages.
This patch holds a local reference of va display at ensure_allowed_raw_caps()
hence it doesn't go away meanwhile it is used, even if the
gst_vaapi_plugin_base_close() is called in other thread.
https://bugzilla.gnome.org/show_bug.cgi?id=773593
Víctor Manuel Jáquez Leal [Thu, 3 Nov 2016 11:54:23 +0000 (12:54 +0100)]
plugins: fix code style for errors
Víctor Manuel Jáquez Leal [Wed, 19 Oct 2016 17:04:20 +0000 (19:04 +0200)]
libs: fix code style for errors
Víctor Manuel Jáquez Leal [Thu, 3 Nov 2016 08:31:17 +0000 (09:31 +0100)]
plugins: update GstGL deprecated symbol
GST_GL_TYPE_CONTEXT was deprecated. Now it is GST_TYPE_GL_CONTEXT.
Víctor Manuel Jáquez Leal [Fri, 21 Oct 2016 09:48:54 +0000 (11:48 +0200)]
plugins: direct render when raw video
Enable the direct rendering with linear surfaces if the negotiated src caps
are video/x-raw without features.
Pass also the caps, since they are needed to know the requested caps features.
Víctor Manuel Jáquez Leal [Mon, 24 Oct 2016 18:09:59 +0000 (20:09 +0200)]
plugins: move src allocator error to instantiator
Just as we did in ensure_sinkpad_allocator(), let's move the error message
into the ensure_srcpad_allocator() from the caller,
gst_vaapi_plugin_base_decide_allocation()
Víctor Manuel Jáquez Leal [Thu, 20 Oct 2016 17:37:01 +0000 (19:37 +0200)]
plugins: enable direct upload if raw video
Enable the direct upload with linear surfaces if the negotiated sink caps are
video/x-raw without features.
Víctor Manuel Jáquez Leal [Fri, 21 Oct 2016 09:21:04 +0000 (11:21 +0200)]
pluginutil: add gst_caps_is_video_raw()
Víctor Manuel Jáquez Leal [Mon, 24 Oct 2016 17:25:27 +0000 (19:25 +0200)]
plugins: receive caps in ensure_sinkpad_allocator()
Instead of receiving the GstVideoInfo structure as parameter, get the original
GstCaps from ensure_sinkpad_buffer_pool(), in this way we could decide better
which allocator instantiate.
Víctor Manuel Jáquez Leal [Thu, 20 Oct 2016 17:31:58 +0000 (19:31 +0200)]
vaapivideomemory: destroy derived image at unmap
If the allocator was configured to use direct upload or rendering, the
generated derived image created at mapping needs to be destroyed after
unmapping, because, in order to process the surface, it should not be marked
as "busy" by the driver.
Víctor Manuel Jáquez Leal [Fri, 21 Oct 2016 09:57:55 +0000 (11:57 +0200)]
vaapivideomemory: enhance logs for direct modes
Print, conditionally, only the enabled direct mode.
Víctor Manuel Jáquez Leal [Thu, 20 Oct 2016 15:02:49 +0000 (17:02 +0200)]
vaapivideomemory: add direct upload flag
Adds the direct-upload flag in the GstVaapiVideoAllocator and
GstVaapiVideoMemory.
It still doesn't apply any functional change.
Víctor Manuel Jáquez Leal [Thu, 20 Oct 2016 14:49:22 +0000 (16:49 +0200)]
vaapivideomemory: set direct rendering at run-time
The way to experiment with the direct rendering is through and internal
compiler pre-processor flag.
The current change set enables a way to specified at run-time, as a flag
passed to the allocator at instanciation time.
Víctor Manuel Jáquez Leal [Thu, 20 Oct 2016 16:09:59 +0000 (18:09 +0200)]
vaapivideomemory: log in perf category when copy
Log in performance category when the derive image handling fails, falling back
to memory copy.
Víctor Manuel Jáquez Leal [Thu, 20 Oct 2016 14:31:21 +0000 (16:31 +0200)]
vaapivideomemory: error log is derive image fails
Instead of a silently failure of the derive image, this patch log an error
message according to the failure.
Víctor Manuel Jáquez Leal [Thu, 20 Oct 2016 10:52:18 +0000 (12:52 +0200)]
vaapivideomemory: store surface alloc flags in qdata
For sake of consistency, we should add the requested surface allocation flags
to the object's qdata structure.
Víctor Manuel Jáquez Leal [Thu, 20 Oct 2016 10:22:06 +0000 (12:22 +0200)]
vaapivideomemory: category init when object define
Move the Gstreamer debug category initialize to the GObject definition.
Víctor Manuel Jáquez Leal [Thu, 3 Nov 2016 07:31:16 +0000 (08:31 +0100)]
libs: vaapitexturemap: trivial code-style fix
Víctor Manuel Jáquez Leal [Wed, 2 Nov 2016 19:01:09 +0000 (20:01 +0100)]
libs: display: egl: avoid recreate native display
Instead of passing the native descriptor of the display, just pass the received
GstVaapiDisplay and reuse it.
Víctor Manuel Jáquez Leal [Wed, 2 Nov 2016 14:38:52 +0000 (15:38 +0100)]
plugins: log the GstVaapiDisplay name
Now that GstVaapiDisplay is descendant of GstObject, it has a human-friendly
name. Log it instead of the memory address.
Hyunjun Ko [Wed, 2 Nov 2016 09:37:00 +0000 (18:37 +0900)]
libs: window: egl: pass native va display
When creating a GstVaapiWindowEGL, it also creates native window by its own
native display. It should pass the native display, either X11 or Wayland.
https://bugzilla.gnome.org/show_bug.cgi?id=768266
Hyunjun Ko [Thu, 13 Oct 2016 03:53:17 +0000 (12:53 +0900)]
libs: display: GstVaapiDisplay as GstObject descendant
This patch is to change the inheritance of GstVaapiDisplay to GstObject,
instead of GstVaapiMiniObject. In this way we can use all the available
infrastructure for GObject/GstObject such as GstTracer, GIR, etc.
In addition, a new debug category for GstVaapiDisplay is created to make it
easier to trace debug messages. It is named "vaapidisplay" and it transverse
all the VA display backends (DRM, GLX, EGL, Wayland, ...)
This patch is a step forward to expose GstVaapiDisplay for users in a future
library.
https://bugzilla.gnome.org/show_bug.cgi?id=768266
Signed-off-by: Víctor Manuel Jáquez Leal <victorx.jaquez@intel.com>
Sebastian Dröge [Tue, 1 Nov 2016 16:54:54 +0000 (18:54 +0200)]
Back to development
Sebastian Dröge [Tue, 1 Nov 2016 16:19:32 +0000 (18:19 +0200)]
Release 1.10.0
Víctor Manuel Jáquez Leal [Thu, 27 Oct 2016 15:13:48 +0000 (17:13 +0200)]
vaapidecodebin: resurrect disable-vpp property
https://bugzilla.gnome.org/show_bug.cgi?id=773589
Víctor Manuel Jáquez Leal [Thu, 27 Oct 2016 14:32:23 +0000 (16:32 +0200)]
vaapidecodebin: name the internal queue
https://bugzilla.gnome.org/show_bug.cgi?id=773589
Víctor Manuel Jáquez Leal [Thu, 27 Oct 2016 14:27:45 +0000 (16:27 +0200)]
vaapidecodebin: remove unused variables
Since vaapipostproc is only registered if the driver supports it, all the
support for dynamic loading were removed. Though some leftovers remained.
https://bugzilla.gnome.org/show_bug.cgi?id=773589
Víctor Manuel Jáquez Leal [Thu, 27 Oct 2016 10:53:54 +0000 (12:53 +0200)]
docs: document environment variables
https://bugzilla.gnome.org/show_bug.cgi?id=773544
Víctor Manuel Jáquez Leal [Thu, 27 Oct 2016 10:31:49 +0000 (12:31 +0200)]
docs: replace vaapidecode with each codec
In the spirit of the codec split, this patch removes the documentation of
vaapidecode and adds a page per each possible decoder.
Nonetheless, only those available in the compilation system are going to be
instrospected, because the rest are not registered.
Víctor Manuel Jáquez Leal [Thu, 27 Oct 2016 09:06:06 +0000 (11:06 +0200)]
docs: add missing long descriptions
Víctor Manuel Jáquez Leal [Tue, 25 Oct 2016 12:32:44 +0000 (14:32 +0200)]
vaapipostproc: use GST_*_OBJECT when possible
Since we can have several vaapipostproc operating in a pipeline, it is useful
to know which one is generating the logging message.
https://bugzilla.gnome.org/show_bug.cgi?id=773497
Víctor Manuel Jáquez Leal [Tue, 19 Jul 2016 15:00:23 +0000 (17:00 +0200)]
vaapidecode: rename member to allowed_sinkpad_caps
vaapidecode has a member named allowed_caps, but this name is not enough
explicit. This patch renames allowed_caps to allowed_sinkpad_caps.
No functional changes were included.
https://bugzilla.gnome.org/show_bug.cgi?id=773497
Víctor Manuel Jáquez Leal [Thu, 20 Oct 2016 16:12:04 +0000 (18:12 +0200)]
plugins: fix code style for errors
https://bugzilla.gnome.org/show_bug.cgi?id=773497
Víctor Manuel Jáquez Leal [Thu, 20 Oct 2016 15:01:57 +0000 (17:01 +0200)]
vaapivideomemory: comment style
https://bugzilla.gnome.org/show_bug.cgi?id=773497
Víctor Manuel Jáquez Leal [Thu, 20 Oct 2016 09:19:50 +0000 (11:19 +0200)]
vaapivideomemory: rename input parameter
In order to clarify the use of flag as input parameter, it is renamed to
surface_alloc_flag, since it is used when creating a VA surface with certain
properties.
https://bugzilla.gnome.org/show_bug.cgi?id=773497
Víctor Manuel Jáquez Leal [Tue, 25 Oct 2016 17:22:03 +0000 (19:22 +0200)]
vaapidecode: rename element description
So encoders and decoders have similar descriptions.
https://bugzilla.gnome.org/show_bug.cgi?id=773497
Víctor Manuel Jáquez Leal [Tue, 2 Aug 2016 09:32:19 +0000 (11:32 +0200)]
vaapiencode: h264, h265: rename codec name
So encoder and decoders have the same codec name.
https://bugzilla.gnome.org/show_bug.cgi?id=773497
Víctor Manuel Jáquez Leal [Fri, 29 Jul 2016 13:17:33 +0000 (15:17 +0200)]
plugins: simplify code
Merge two lines of variable declarations.
https://bugzilla.gnome.org/show_bug.cgi?id=773497
Hyunjun Ko [Fri, 7 Oct 2016 09:46:22 +0000 (18:46 +0900)]
libs: minor correction for logical consistency
GstVaapiDecode is a descendant of GstVaapiMiniObject, so, thought we should
use its methods, even though it doesn't change functionality.
GstVaapiPixmap, GstVaapiTexture and GstVaapiWindow are descendant of
GstVaapiObject, hence its methods shall be used.
https://bugzilla.gnome.org/show_bug.cgi?id=772554
Julien Isorce [Wed, 19 Oct 2016 14:39:54 +0000 (15:39 +0100)]
vaapivideomemory: add explanation about the call 'dup (dmabuf_fd)'
In short GstFdMemory is configured to call close when using
GstDmabufMemory.
https://bugzilla.gnome.org/show_bug.cgi?id=755072
Víctor Manuel Jáquez Leal [Fri, 2 Sep 2016 14:42:45 +0000 (16:42 +0200)]
build: clean up the dlopen usage
Víctor Manuel Jáquez Leal [Sat, 8 Oct 2016 12:33:59 +0000 (14:33 +0200)]
encoder: h264,h265: fix regression in offset count
In commit
dc35dafa a bug was introduced because I assumed that
GST_CLOCK_TIME_NONE is zero when is -1. This patch fixes that mistake.
https://bugzilla.gnome.org/show_bug.cgi?id=772259
Víctor Manuel Jáquez Leal [Tue, 18 Oct 2016 15:02:59 +0000 (17:02 +0200)]
docs: update README
Víctor Manuel Jáquez Leal [Tue, 27 Sep 2016 15:29:25 +0000 (17:29 +0200)]
libs: display: egl: remove unused header include
The header gmodule.h is not used since the library dynamic loading for EGL
display was removed.
https://bugzilla.gnome.org/show_bug.cgi?id=772599
Sebastian Dröge [Fri, 30 Sep 2016 10:05:20 +0000 (13:05 +0300)]
Release 1.9.90
Vineeth TM [Fri, 4 Mar 2016 07:35:11 +0000 (16:35 +0900)]
vaapi: use new gst_element_class_add_static_pad_template()
https://bugzilla.gnome.org/show_bug.cgi?id=763083
Signed-off-by: Víctor Manuel Jáquez Leal <victorx.jaquez@intel.com>
Hyunjun Ko [Thu, 22 Sep 2016 07:34:48 +0000 (16:34 +0900)]
plugins: reset textures at negotiation/shutdown
When caps reconfiguration is called, the new downstream frame size might be
different. Thus, if the downstream caps change,the display's texture map is
reset.
In addition, during pipeline shutdown, textures in texture map have to be
released, since each one have a reference to the GstVaapiDisplay object, which
is a dangerous circular reference.
https://bugzilla.gnome.org/show_bug.cgi?id=769293
Signed-off-by: Víctor Manuel Jáquez Leal <victorx.jaquez@intel.com>
Hyunjun Ko [Thu, 22 Sep 2016 07:34:38 +0000 (16:34 +0900)]
libs: display{egl,glx}: cache GstVaapiTextures
instances when created and reuse
This patch improves performance when glimagesink uploads a GL texture.
It caches the GStVaapiTexture instances in GstVaapiDisplay{GLX,EGL}, using an
instance of GstVaapiTextureMap, so our internal texture structure can be found
by matching the GL texture id for each frame upload process, avoiding the
internal texture structure creation and its following destruction.
https://bugzilla.gnome.org/show_bug.cgi?id=769293
Signed-off-by: Víctor Manuel Jáquez Leal <victorx.jaquez@intel.com>
Hyunjun Ko [Thu, 22 Sep 2016 07:33:06 +0000 (16:33 +0900)]
libs: vaapitexturemap: implement GstVaapiTextureMap
Implement GstVaapiTextureMap object, which caches VAAPI textures, so them can be
reused. Internally it is a hash table.
Note that it is GstObject based rather than GstVaapiObject, as part of the future
converstion to GstObject of most of the code.
https://bugzilla.gnome.org/show_bug.cgi?id=769293
Signed-off-by: Víctor Manuel Jáquez Leal <victorx.jaquez@intel.com>
Sreerenj Balachandran [Wed, 21 Sep 2016 06:55:53 +0000 (09:55 +0300)]
encoder: vp8: Increase the allocation size for coded buffer
We are not getting enough compression for some streams and
encoded frame end up with more size than allocated.
Assuming a compression ratio of 4, which should be good enough
for holding the frames.
https://bugzilla.gnome.org/show_bug.cgi?id=771528
Sreerenj Balachandran [Wed, 21 Sep 2016 06:52:21 +0000 (09:52 +0300)]
encoder: vp9: Fix refresh frame flag setting
While doing the mode-1 referece picture selection,
the circular buffer logic was not correctly setting the
refresh frame flags as per VP9 spec.
Make sure refresh_flag[0] get updated correclty after
each cycle of GST_VP9_REF_FRAMES.
https://bugzilla.gnome.org/show_bug.cgi?id=771507
Víctor Manuel Jáquez Leal [Wed, 14 Sep 2016 16:42:09 +0000 (18:42 +0200)]
vaapidecode: codec_data minimal size is 7
When the format of a H.264 stream is AVC3, the SPS and PPS are inside the
stream, not in the codec_data, so the size of codec_data might be 7.
This patch reduces the minimal size of the codec_data buffer from 8 to 7.
https://bugzilla.gnome.org/show_bug.cgi?id=771441
Víctor Manuel Jáquez Leal [Wed, 14 Sep 2016 14:29:01 +0000 (16:29 +0200)]
vaapidecode: reset decoder hard when set_format()
set_format() is called by upstream when the stream capabilites has changed.
Before, if the new stream is compatible with the old one the VA decoder was
not destroyed. Nonetheless, with this behavoir, the VA decoder ignores
when the upstreamer parsers gets more details of the stream, such as the
framerate. Hence, when the src caps are negotiates, the further sink caps
updates are ignored.
This patch forces the VA decoder destroying and recreation when set_format()
is called.
https://bugzilla.gnome.org/show_bug.cgi?id=770921
Sebastian Dröge [Wed, 14 Sep 2016 09:31:39 +0000 (11:31 +0200)]
configure: Depend on gstreamer 1.9.2.1
Víctor Manuel Jáquez Leal [Fri, 9 Sep 2016 10:03:37 +0000 (12:03 +0200)]
Revert "vaapivideomemory: load VA Image when mapping to write"
This reverts commit
c67edea4aba35f16d9e97c78a0b49ad1b590b112.
Jan Schmidt [Sat, 10 Sep 2016 10:52:21 +0000 (20:52 +1000)]
Automatic update of common submodule
From b18d820 to f980fd9
Jan Schmidt [Fri, 9 Sep 2016 23:58:25 +0000 (09:58 +1000)]
Automatic update of common submodule
From f49c55e to b18d820
Hyunjun Ko [Thu, 8 Sep 2016 07:16:09 +0000 (16:16 +0900)]
plugins: set allocator's image size to sinkpad bufferpool
Otherwise the buffer is always ditched by the bufferpool, losing performance.
https://bugzilla.gnome.org/show_bug.cgi?id=771035
Víctor Manuel Jáquez Leal [Wed, 7 Sep 2016 15:34:08 +0000 (17:34 +0200)]
libs: surface: ensure composite overlay is not bigger
Ensure the composition overlay rectangle (subtitles) is not bigger than
the surface where it is going to be composited and rendered.
https://bugzilla.gnome.org/show_bug.cgi?id=766978
Hyunjun Ko [Wed, 7 Sep 2016 08:51:23 +0000 (17:51 +0900)]
vaapivideomemory: load VA Image when mapping to write
When calling gst_video_frame_map() with GST_MAP_WRITE flag, it doesn't call
ensure_image_is_current(), which means it doesn't guarentee VAImage is valid
in this case.
https://bugzilla.gnome.org/show_bug.cgi?id=766978
Víctor Manuel Jáquez Leal [Tue, 6 Sep 2016 10:27:45 +0000 (12:27 +0200)]
vaapidecode: merge vc1 and wmv3 elements
This patch merges vaapivc1dec and vaapiwmv3dec into a single
vaapivc1dec. Also, removed the WMVA format, since it is not
supported by libva.
https://bugzilla.gnome.org/show_bug.cgi?id=734093
Víctor Manuel Jáquez Leal [Tue, 6 Sep 2016 09:19:05 +0000 (11:19 +0200)]
decoder: vc1: fails only on advanced profile
In commit 2eb4394 the frame coding mode was verified for progressive
regardless the profile. But the FCM is only valid in the advanced
profile. This patch checks for the advanced profile before verifying FCM for
progressive.
https://bugzilla.gnome.org/show_bug.cgi?id=769250
Víctor Manuel Jáquez Leal [Thu, 1 Sep 2016 10:39:15 +0000 (12:39 +0200)]
vaapi: DISPLAY envvar as dependency
In a multiple video cards system, a X11 environment may have different VA
capabilities. This patch tracks the DISPLAY environment variable to
invalidates the GStreamer features cache. Also tracks WAYLAND_DISPLAY.
https://bugzilla.gnome.org/show_bug.cgi?id=770357
Scott D Phillips [Fri, 26 Aug 2016 21:55:17 +0000 (14:55 -0700)]
decoder: vc1: Fail only on actual interlaced frames
In the earlier patch:
f31d9f3 decoder: vc1: Print error on interlaced content
Decoding would error out if the interlace flag was set in the
sequence bdu. This isn't quite right because a video can have this
flag set and yet not have any interlaced pictures.
Here instead we error out when either parsing a field bdu or
decoding a frame bdu which has fcm set to anything other than
progressive.
Signed-off-by: Scott D Phillips <scott.d.phillips@intel.com>
https://bugzilla.gnome.org/show_bug.cgi?id=769250
Sebastian Dröge [Thu, 1 Sep 2016 09:34:48 +0000 (12:34 +0300)]
Back to development
Sebastian Dröge [Thu, 1 Sep 2016 09:34:38 +0000 (12:34 +0300)]
Release 1.9.2
Scott D Phillips [Tue, 16 Aug 2016 08:58:38 +0000 (11:58 +0300)]
decoder: vc1: Print error on interlaced content
Interlaced video is as yet unsupported in the vc1 element. Print
an error to make that more obvious.
https://bugzilla.gnome.org/show_bug.cgi?id=769250
Víctor Manuel Jáquez Leal [Wed, 10 Aug 2016 11:29:45 +0000 (13:29 +0200)]
encoder: h264: fix C90 mixed declarations and code
Commit 4259d1a introduced this compilation error. This patch fixes it.
Víctor Manuel Jáquez Leal [Thu, 21 Jul 2016 15:38:40 +0000 (17:38 +0200)]
encoder: h264,h265,mpeg2,vp8: use gst_util_uint64_scale() for bitrate
Use gst_util_uint64_scale() to calculate bitrate instead of normal arithmetic
to avoid overflows, underflows and loss of precision.
https://bugzilla.gnome.org/show_bug.cgi?id=768458
Víctor Manuel Jáquez Leal [Tue, 5 Jul 2016 18:07:15 +0000 (20:07 +0200)]
vaapiencode: h264,h265: validate fps numerator
Validate that fps numerator is non-zero so it can be used to calculate
the duration of the B frame.
Also it gst_util_uint64_scale() is used instead of normal arithmetic in
order to aviod overflows, underflows and loss of precision.
https://bugzilla.gnome.org/show_bug.cgi?id=768458
Tim-Philipp Müller [Sat, 6 Aug 2016 11:54:17 +0000 (12:54 +0100)]
encoders: demote to RANK_NONE since not fit for autoplugging yet
Encoders claim to support a whole bunch of input formats but then
just error out if the format is not actually supported, even if
there's a converter in front. This means they're not fit for
autoplugging in encodebin or camerabin yet and therefore should
not have a rank. People can still use them in custom pipelines.
https://bugzilla.gnome.org/show_bug.cgi?id=769266
Víctor Manuel Jáquez Leal [Tue, 19 Jul 2016 17:24:08 +0000 (19:24 +0200)]
plugins: check dmabuf-import for sink pad allocator
Check earlier if upstream video source has activated the dmabuf-import
io-mode (hack to disappear soon), thus we can avoid the re-assignation of a
new allocator.