OSDN Git Service
Wang, Ce [Thu, 27 Mar 2014 04:25:36 +0000 (12:25 +0800)]
Added HEVC decode API
Signed-off-by: Wang, Ce <ce.wang@intel.com>
Signed-off-by: Jonathan Bian <jonathan.bian@intel.com>
Signed-off-by: Penne <penne.y.lee@intel.com>
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Lionel Landwerlin [Thu, 27 Nov 2014 10:39:50 +0000 (10:39 +0000)]
x11: check x pixmap is still alive before DRI2DestroyDrawable
Calling vaTerminate() after the pixmaps, given to vaPutSurface(), have
been destroyed results in errors coming from the X server.
These errors happens because LibVA tries to destroy the DRI2Drawables
related to the XPixmaps passed to vaPutSurface(). VA_DRI2Error() has
some code to catch these errors and ignore them. The problem is that
LibX11 doesn't trigger the extension error handlers unless you're
waiting for a response using _XReply(), and there is no defined answer
to a DRI2DestroyDrawable message in the DRI2 protocol.
The attached patch tries to work around that issue by checking whether
the XPixmap is still alive before proceeding with DRI2DestroyDrawable.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Xiang, Haihao [Wed, 26 Nov 2014 08:32:55 +0000 (16:32 +0800)]
Fix while loop
Otherwise i may be 32 after enter the loop and access
dri_state->drawable_hash[32].
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Zhao Yakui [Fri, 21 Nov 2014 07:32:21 +0000 (15:32 +0800)]
Encoding/avcenc: Optimize the allocated VA surfaces to reduce the memory pressure
When it needs to encode one new frame, it will try to reuse the previously
used surface for the reconstructed frame. In such case it is helpfl to
reduce the memory pressure under multi-channel 4K encoding.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Sirisha Muppavarapu [Tue, 11 Nov 2014 07:08:44 +0000 (23:08 -0800)]
JPEG Encode: Added Unit Test for testing jpeg encode feature.
(cherry picked from commit
9d01b4468f89391892e31efc5bbec774ea781242)
Austin Yuan [Sat, 8 Jun 2013 03:16:03 +0000 (11:16 +0800)]
vatrace: trace JPEG encode parameters
Signed-off-by: Austin Yuan <shengquan.yuan@intel.com>
(cherry picked from commit
e468728f9e49603f91ccbce3a232aa3bd98d1fd4)
Conflicts:
va/va_trace.c
va/va_trace.h
Jonathan Bian [Fri, 24 May 2013 22:36:58 +0000 (15:36 -0700)]
Extended existing JPEG encoding data structures and added configuration attribute for JPEG encoding.
It is the combination of
1d93d44 and
970064cf
Signed-off-by: Jonathan Bian <jonathan.bian@intel.com>
Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
[Resovle conflicts & remove irrelevant attributes]
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Xiang, Haihao [Fri, 14 Nov 2014 08:51:32 +0000 (16:51 +0800)]
libva 1.4.2.pre1 for development
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Conflicts:
configure.ac
Emil Velikov [Thu, 18 Sep 2014 16:30:32 +0000 (17:30 +0100)]
va: hide most private va_Trace* symbols
Hide all but va_TracePutSurface symbols from the user.
The latter one is used internally by va_x11 and va_android.
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Emil Velikov [Thu, 18 Sep 2014 16:30:31 +0000 (17:30 +0100)]
va_x11: hide private symbols
The following are not part of the API thus should be hidden.
- dri2Close
- VA_FGLRXGetClientDriverName
- VA_NVCTRLQueryDirectRenderingCapable
- VA_NVCTRLGetClientDriverName
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Emil Velikov [Thu, 18 Sep 2014 16:30:30 +0000 (17:30 +0100)]
glx: Hide private symbol va_glx_init_context
There is a note below the symbol ATTRIBUTE_HIDDEN,
additionally the header providing it is not installed.
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Emil Velikov [Thu, 18 Sep 2014 16:30:29 +0000 (17:30 +0100)]
debian: distclean the *install files
MAINTCLEAN wraps around DISTCLEAN, so we can drop the files from the
former.
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Emil Velikov [Thu, 18 Sep 2014 16:30:28 +0000 (17:30 +0100)]
configure: use top_builddir for wayland_scanner_rules
Autotools already copies the files appropriately, and using
top_srcdir causes the files to be generated in the source
tree rather than the build one.
Spotted by 'make dist'
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Emil Velikov [Thu, 18 Sep 2014 16:30:27 +0000 (17:30 +0100)]
Fix out of tree builds
Prefix the libva.sym file or it will not be found when doing out of
tree builds.
Spotted by 'make distcheck'.
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Emil Velikov [Thu, 18 Sep 2014 16:30:26 +0000 (17:30 +0100)]
.gitignore: add mpeg2vaenc
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Xiang, Haihao [Tue, 30 Sep 2014 02:32:09 +0000 (10:32 +0800)]
Bump libva to libva 1.4.1.pre1
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Xiang, Haihao [Tue, 30 Sep 2014 01:42:06 +0000 (09:42 +0800)]
libva 1.4.0
Don't support DRM Render-Nodes in this version of libva
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit
8c48c3b44a56d8a5af5c5fa7cb8bee131f933569)
Gwenole Beauchesne [Fri, 26 Sep 2014 11:59:23 +0000 (13:59 +0200)]
NEWS: updates.
Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
Gwenole Beauchesne [Fri, 26 Sep 2014 09:53:07 +0000 (11:53 +0200)]
doc: fix scan order specification for quantization tables.
Fix the VA-API specification to define all quantization tables, resp.
matrices, to be supplied in zig-zag scan order, except for the H.264
codec where scaling lists shall be provided in raster scan order.
This clarifies the existing usage inherited from the primary VA driver
implementation (Intel Poulsbo), and subsequently adopted by the major
multimedia frameworks.
Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
Gwenole Beauchesne [Fri, 4 Nov 2011 16:33:48 +0000 (17:33 +0100)]
doc: add Encoding, VPP and some of Core APIs.
Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
Gwenole Beauchesne [Thu, 24 Nov 2011 14:16:37 +0000 (14:16 +0000)]
doc: improve Doxyfile configuration.
Gwenole Beauchesne [Fri, 18 Nov 2011 12:17:50 +0000 (12:17 +0000)]
doc: don't add timestamps in generated documentation.
Gwenole Beauchesne [Wed, 24 Sep 2014 08:31:21 +0000 (10:31 +0200)]
tests: try a render-node device first for VA/DRM displays.
Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
Reviewed-by: Sean V Kelley <sean.v.kelley@intel.com>
Andrey Larionov [Tue, 23 Sep 2014 15:42:39 +0000 (17:42 +0200)]
drm: add support for render nodes.
Allow vaGetDisplayDRM() to accept DRM Render-Nodes file descriptors,
thus skipping authentication burdens in that case. This also allows
remote users, connected through ssh for instance, to enjoy hardware
acceleration even if there is a local display opened (user logged in).
Signed-off-by: Andrey Larionov <andrey.larionov@intel.com>
[checked if fd is a render-node, fixed VA_DISPLAY_DRM_RENDERNODES]
Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
Reviewed-by: Sean V Kelley <sean.v.kelley@intel.com>
Tested-by: Rafal A Sapala <rafal.a.sapala@intel.com>
Xiang, Haihao [Mon, 8 Sep 2014 17:37:28 +0000 (01:37 +0800)]
Bump VA API to 0.36
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Zhao, Yakui [Wed, 14 May 2014 01:45:06 +0000 (09:45 +0800)]
Test/encode/avcenc: Follow the H264 spec to fix wrong SEI parameter setting for CBR HRD
V1->V2: Follow Haihao's comments to add some comments and remove some
magic number.
Signed-off-by: Zhao, Yakui <yakui.zhao@intel.com>
(cherry picked from commit
51901e340cc467a2d562491121e0de38f5a066fe)
Zhao Yakui [Wed, 23 Apr 2014 07:41:25 +0000 (15:41 +0800)]
test/encode/avcenc: Follow the idea in h264encode to generalize avcenc test case
The following is generalized:
a. the switch between encoding order and displaying order.
b. Reference list0/1
c. It can insert more IDR frames intead of one IDR frame.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit
615c2a0f3210ec567fde5b0bac60c3a17cff3aed)
Conflicts:
test/encode/avcenc.c
Zhao Yakui [Wed, 23 Apr 2014 07:41:24 +0000 (15:41 +0800)]
test/encode/avcenc: update the SPS parameter for encoding
It will simplify the calculation of POC and frame_num.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit
d794d040b8f7311d7b19d62826fbd3725d3bc7b9)
Zhao Yakui [Wed, 16 Apr 2014 04:17:24 +0000 (12:17 +0800)]
test/avcenc: Fix the incorrect parameter setting for SPS under CBR mode
Currently the parameter setting is incorrect for CBR, which causes that
it will complain the underflow/overflow during encoding.
V1->V2: Follow Haihao's comment to update some incorrect usage about the
frame_bit_rate.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit
5e10a246287fa2dd3fae2bf756a5b33087cc097c)
Zhao Yakui [Wed, 16 Apr 2014 04:17:23 +0000 (12:17 +0800)]
test/avcenc: Remove the dynamic shrink of coded buffer
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit
296e5a548b260fcca5713bbb3ed6477f63ac977f)
Austin Yuan [Tue, 25 Mar 2014 23:44:03 +0000 (07:44 +0800)]
vatrace: fix segfault of VASurfaceAttribExternalBufferDescriptor dumpping
Signed-off-by: Austin Yuan <shengquan.yuan@intel.com>
(cherry picked from commit
e286c3264ebf7582792ff56537dd14c9f49903cf)
Austin Yuan [Thu, 27 Feb 2014 07:36:17 +0000 (15:36 +0800)]
vafool: fix the vaMapBuffer for the buffer from vaDeriveImage
It is not desired to run into FOOL path for the vaMapBuffer
from vaDeriveImage
Signed-off-by: Austin Yuan <shengquan.yuan@intel.com>
(cherry picked from commit
2896166fefa36772ea6946ebc5df85024928dd6c)
Austin Yuan [Wed, 26 Feb 2014 08:47:09 +0000 (16:47 +0800)]
vatrace: fix a overwrite
Signed-off-by: Austin Yuan <shengquan.yuan@intel.com>
(cherry picked from commit
7f1b4be6623f5fe65063581252d2b68bd77fd076)
Austin Yuan [Wed, 26 Feb 2014 08:40:41 +0000 (16:40 +0800)]
vafool: simplify VAFOOL
Signed-off-by: Austin Yuan <shengquan.yuan@intel.com>
(cherry picked from commit
ce00603a8d58f2e25c7a6bb312f60fda0833e042)
Austin Yuan [Thu, 17 Oct 2013 02:57:23 +0000 (10:57 +0800)]
vatrace: make LIBVA_TRACE_SURFACE work even without LIBVA_TRACE
Signed-off-by: Austin Yuan <shengquan.yuan@intel.com>
(cherry picked from commit
e39ebdd95062995e4bdf082d50bd7d9005a3be6c)
Austin Yuan [Tue, 25 Feb 2014 00:50:50 +0000 (08:50 +0800)]
vatrace: make standalone LIBVA_TRACE_CODEDBUF work w/o LIBVA_TRACE
Signed-off-by: Austin Yuan <shengquan.yuan@intel.com>
(cherry picked from commit
22925d9b7d1730468fc7435bee96d9d9c39f5987)
Qiming Shi [Mon, 18 Aug 2014 04:38:14 +0000 (12:38 +0800)]
Enable 64bit build for Android, compatible with 32bit.
Refined the Android.mk to compatible with both 32 and 64 bit android
build.
Signed-off-by: Tinghong Li <tinghong.li@intel.com>
Signed-off-by: Qiming Shi <qiming.shi@intel.com>
(cherry picked from commit
03de01dce6f2b88b9c5ee54705aeebe17c6c5919)
Gwenole Beauchesne [Tue, 3 Dec 2013 23:00:57 +0000 (00:00 +0100)]
API: add support for buffer exports.
Add interfaces for low-level buffer exports to suport interop with
external APIs like EGL or OpenCL (OCL). Theory of operations:
- vaAcquireBufferHandle(): locks buffer for external API usage. This
is a synchronization point. Upon successful return, useful buffer
details are returned in a somewhat implementation-specific way to
support interop with external APIs.
- vaReleaseBufferHandle(): unlocks buffer after usage from external
API, and deallocates any resources that were needed.
Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
(cherry picked from commit
544d31d2fc096a838c21f4684cec05746cfdd7b6)
Conflicts:
va/va.h
Xiang, Haihao [Wed, 30 Jul 2014 02:13:03 +0000 (10:13 +0800)]
Update comment for quality_level
Allow user set VAEncMiscParameterBufferQualityLevel::quality_level
to 0 to use default quality level
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit
b05e38f64c5aa2d6edfc2bd1d6e1f4339a54095a)
Conflicts:
va/va.h
Jonathan Bian [Fri, 14 Jun 2013 05:04:30 +0000 (22:04 -0700)]
Added VAConfigEncQualityRange attribute and VAEncMiscParameterBufferQualityLevel to support encoder quality settings.
(cherry picked from commit
24a14f9d13e6fe0c25d4094c4884aa4076ed05bc)
Conflicts:
va/va.h
Sameer Kibey [Wed, 27 Aug 2014 16:45:49 +0000 (00:45 +0800)]
Added OTC hybrid UMD driver to the list of supported drivers
This will be used for hybrid codecs like VP8 h-encode. No change
to existing KMD interface.
Signed-off-by: Sameer Kibey <sameer.kibey@intel.com>
(cherry picked from commit
98669760e62f50ca3e38e99835692aee5f89a4bf)
Xiang, Haihao [Mon, 4 Aug 2014 02:41:59 +0000 (10:41 +0800)]
avcenc: only support chroma_format_idc with 1
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Xiang, Haihao [Mon, 4 Aug 2014 02:00:13 +0000 (10:00 +0800)]
h264encode: check driver supports the RateControl control mode or not
https://bugs.freedesktop.org/show_bug.cgi?id=76720
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Xiang, Haihao [Thu, 31 Jul 2014 07:20:05 +0000 (15:20 +0800)]
h264encode: Make it runnable with the latest libva-intel-driver (1.3.2)
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
pylee [Wed, 30 Oct 2013 15:41:33 +0000 (08:41 -0700)]
Add vaSurfaceAttribUsageHint to give the driver a hint of intended usage to optimize allocation (e.g. tiling).
(cherry picked from commit
6cdc2e54d40afeb1550b920a6a6221bafeec95ec)
Austin Yuan [Sat, 7 Sep 2013 10:22:04 +0000 (18:22 +0800)]
Add frame_type to VP8 encoder
From penne.y.lee at intel.com
Signed-off-by: Austin Yuan <shengquan.yuan@intel.com>
(cherry picked from commit
c61d986851140a6d1e3f9a6d788bcb7a9407dffc)
Austin Yuan [Fri, 2 Aug 2013 03:28:33 +0000 (11:28 +0800)]
va_enc_vp8: update the vp8 interface again
previous update missed some changes
Signed-off-by: Austin Yuan <shengquan.yuan@intel.com>
(cherry picked from commit
34035a18a7e963da536e15e9dea32f77cde8577f)
Jonathan Bian [Thu, 18 Jul 2013 20:13:38 +0000 (13:13 -0700)]
Removed fields in VP8 structures that duplicate VAEncMiscParameter structures. Added reference frames to support non auto reference frame implementation. Changed VAQMatrixBufferVP8 to use quantization index deltas.
(cherry picked from commit
fbd3de9b22491689c6c3e5f1c305d7af76444e45)
Jonathan Bian [Wed, 10 Apr 2013 23:26:37 +0000 (16:26 -0700)]
Added support for VP8 encoding.
(cherry picked from commit
d46a07340a641770487e7624bebc49f3258ea2ec)
Conflicts:
va/Android.mk
Xiang, Haihao [Thu, 18 Jul 2013 05:13:31 +0000 (13:13 +0800)]
va: release VADisplayContext at the end of vaTerminate()
Otherwise it will result in invalid reading in va_TraceEnd()
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit
d4988142a3f2256e38c5c5cdcdfc1b4f5f3c1ea9)
Tested-by: Daniel Charles <daniel.charles@intel.com> on ChromiumOS
Gwenole Beauchesne [Tue, 3 Jun 2014 14:57:13 +0000 (16:57 +0200)]
Bump VA-API to 0.35.2.
Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
Gwenole Beauchesne [Tue, 3 Jun 2014 14:50:58 +0000 (16:50 +0200)]
Merge H.264 MVC profiles from staging branch.
Cherry-pick'ed from:
64f0223: vainfo: add the support for the new VA profiles
b6a6e24: reorder VAProfile{H264MultiviewHigh,StereoHighVP8Version0_3}
7a5a81b: test/basic: update to new API
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
Xiang, Haihao [Fri, 9 May 2014 04:53:26 +0000 (12:53 +0800)]
bva: bump version to 1.3.2.pre1 for development
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Xiang, Haihao [Fri, 9 May 2014 04:38:02 +0000 (12:38 +0800)]
libva 1.3.1
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Xiang, Haihao [Mon, 5 May 2014 04:17:46 +0000 (12:17 +0800)]
Update NEWS
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Xiang, Haihao [Mon, 5 May 2014 04:07:33 +0000 (12:07 +0800)]
Bump VA-API to 0.35.1
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Gwenole Beauchesne [Fri, 25 Apr 2014 09:20:32 +0000 (11:20 +0200)]
Fix VA_ENC_PACKED_HEADER_RAW_DATA value.
Attribute values for VAConfigAttribEncPackedHeaders config shall
represent a valid maskable set of bits. So, make sure "raw" packed
header mask is 0x10, not 0x0c.
Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
Xiang, Haihao [Wed, 6 Feb 2013 04:35:57 +0000 (12:35 +0800)]
Fix the broken install
install -d /usr/share/doc/libva/html
install -m 0644 html-out/* /usr/share/doc/libva/html
install: omitting directory `html-out/search'
make[2]: *** [install-html-local] Error 1
make[2]: Leaving directory `/home/xhh/graphics/ex/vaapi/libva/doc'
make[1]: *** [install-am] Error 2
make[1]: Leaving directory `/home/xhh/graphics/ex/vaapi/libva/doc'
make: *** [install-recursive] Error 1
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit
ee948e105b97a03b60168b4e5aed3c64a14c37e6)
Jonathan Bian [Mon, 8 Apr 2013 04:00:13 +0000 (21:00 -0700)]
Added STE filter to the VPP pipeline.
Rinat Ibragimov [Tue, 15 Apr 2014 03:01:43 +0000 (11:01 +0800)]
Add a workaround for a bug in xorg-server/glx
See details at https://bugs.freedesktop.org/show_bug.cgi?id=76755
Signed-off-by: Rinat Ibragimov <ibragimovrinat@mail.ru>
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Sameer Kibey [Mon, 24 Mar 2014 22:57:51 +0000 (15:57 -0700)]
Doxygen tags in libva.
Updated the doxygen tags in libva for generating
doxygen based documentation.
Signed-off-by: Sameer Kibey <sameer.kibey@intel.com>
Xiang, Haihao [Tue, 25 Mar 2014 00:53:14 +0000 (08:53 +0800)]
libva: bump version for development
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Xiang, Haihao [Mon, 24 Mar 2014 08:10:07 +0000 (16:10 +0800)]
libva 1.3.0
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Xiang, Haihao [Tue, 18 Mar 2014 03:04:47 +0000 (11:04 +0800)]
va: keep compatibility with the backend driver built against 0.34.x
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Xiang, Haihao [Mon, 17 Mar 2014 04:26:04 +0000 (12:26 +0800)]
va: Bump VA-API to 0.35 and libva to 1.3.0.pre1
Update NEWS as well
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Austin Yuan [Tue, 25 Feb 2014 01:01:42 +0000 (09:01 +0800)]
vatrace: fix a missing field of va_TraceMsg
Signed-off-by: Austin Yuan <shengquan.yuan@intel.com>
(cherry picked from commit
2bf493a4838b419a61fa5cc6925fab4d4786dbcd)
Austin Yuan [Fri, 21 Feb 2014 05:57:30 +0000 (13:57 +0800)]
vatrace: simplify coded buffer trace
Signed-off-by: Austin Yuan <shengquan.yuan@intel.com>
(cherry picked from commit
d46a3d927ce66e0f7b61c8c184cf80b6d926327e)
Conflicts:
va/va_trace.c
Austin Yuan [Fri, 21 Feb 2014 05:22:36 +0000 (13:22 +0800)]
vatrace: Remove VA_TRACE_LOGSIZE because it is useless
Signed-off-by: Austin Yuan <shengquan.yuan@intel.com>
(cherry picked from commit
733bcaeef5ef7a5a0f2f3b12e9565e4c178d58fd)
Conflicts:
va/va_trace.c
Austin Yuan [Tue, 18 Feb 2014 02:39:00 +0000 (10:39 +0800)]
vatrace: print VASurfaceAttribExternalBufferDescriptor
Signed-off-by: Austin Yuan <shengquan.yuan@intel.com>
(cherry picked from commit
fae9c44816a4c3cfc480d2879d1b4a0c1c3a1527)
liubolun [Wed, 22 Jan 2014 07:10:23 +0000 (15:10 +0800)]
Dump max frame size in libva trace.
Change-Id: I4b38bc5930d653c96b45e4ce91856bfaf0cb277a
Signed-off-by: liubolun <bolun.liu@intel.com>
(cherry picked from commit
753421b8b3e9bd4529986dc0418e39e05dc2f394)
wfeng6 [Thu, 16 Jan 2014 08:58:06 +0000 (16:58 +0800)]
Add more VA Buffer print to ease debug
Add more VA Buffer print to ease debug:
1) correct the RefList1 print
2) the luma_weight, luma_offset, chroma_weight and
chroma_offset should always be printed
Change-Id: Id122e5fa186d24623d1c541c24c761ab2d807aac
Signed-off-by: wfeng6 <wei.feng@intel.com>
(cherry picked from commit
8d1eec87d3bf1e4ce13e4302c944569a22129c5d)
Zhao Halley [Thu, 24 Oct 2013 07:22:21 +0000 (15:22 +0800)]
vatrace: add trace for VP8 decoding
(cherry picked from commit
303ab444365e7de29d471b28331e1c214877b63a)
Conflicts:
va/va_trace.c
Signed-off-by: Zhao Halley <halley.zhao@intel.com>
Signed-off-by: Liu Bolun <bolun.liu@intel.com>
[Haihao: fix conflicts with some modifications come from
bc49c42bcbfb24286c06b3e7cdb171b2a01c6f80]
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Austin Yuan [Thu, 17 Oct 2013 02:42:54 +0000 (10:42 +0800)]
vatrace: fix UV data offset issue of LIBVA_TRACE_SURFACE_GEOMETRY
Signed-off-by: Austin Yuan <shengquan.yuan@intel.com>
(cherry picked from commit
c954ff868057e02f620470d64c1f2071e2cfbb4e)
Austin Yuan [Thu, 17 Oct 2013 02:40:23 +0000 (10:40 +0800)]
vatrace: refine the log format
Signed-off-by: Austin Yuan <shengquan.yuan@intel.com>
(cherry picked from commit
f78e4a6a84178c4561a5b1c8dbb7773e5a575d0f)
Xiang, Haihao [Mon, 16 Sep 2013 05:08:04 +0000 (13:08 +0800)]
vatrace: Partial dumping of VAProcPipelineParameterBuffer
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit
f5c913765e6af38d835cacf339054ccc60bddefb)
Conflicts:
va/va_trace.c
Xiang, Haihao [Mon, 12 Aug 2013 06:52:14 +0000 (14:52 +0800)]
vatrace: check the pointer against NULL before using it
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit
7ae3f1f3666e3571774a9b98ee6571433dcc73d5)
Xiang, Haihao [Mon, 12 Aug 2013 06:28:01 +0000 (14:28 +0800)]
vatrace: Initialize the local variable 'i'
Otherwise the variable is used without being initialized.
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit
f5a6a2867b88b36936bf721682b674e2979065d3)
Austin Yuan [Sun, 14 Jul 2013 04:04:09 +0000 (12:04 +0800)]
vatrace: fix compile warnings and refine some format
Signed-off-by: Austin Yuan <shengquan.yuan@intel.com>
(cherry picked from commit
d68c9fb7eebe195d96e113abbf654f782a1450d8)
Conflicts:
va/va_trace.c
Austin Yuan [Thu, 4 Jul 2013 04:59:25 +0000 (12:59 +0800)]
vatrace/vafool: save the context data into display context
The static allocation of context data has issue in
multiple threads environment. Put the allocation into
display context strcuture
Signed-off-by: Austin Yuan <shengquan.yuan@intel.com>
(cherry picked from commit
e1ab961b20583c6c54ad05e2fe6bab619462106f)
Conflicts:
va/va_trace.c
Austin Yuan [Fri, 28 Jun 2013 07:24:18 +0000 (15:24 +0800)]
vatrace: always printf RefPicList
Signed-off-by: Austin Yuan <shengquan.yuan@intel.com>
(cherry picked from commit
e9e685fe752b9865ba9e28cb63e18ce3f8aed2a0)
Xiang, Haihao [Thu, 20 Jun 2013 06:31:57 +0000 (14:31 +0800)]
va_trace: avoid to dereference a NULL pointer
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit
e635bf060b68694174cbf24c6b06c573b629ddcc)
Austin Yuan [Sat, 8 Jun 2013 03:16:03 +0000 (11:16 +0800)]
vatrace: trace vaQuerySurfaceAttributes
Signed-off-by: Austin Yuan <shengquan.yuan@intel.com>
(cherry picked from commit
e468728f9e49603f91ccbce3a232aa3bd98d1fd4)
[Haihao: Remove JPEG encode from this commit]
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Xiang, Haihao [Fri, 14 Feb 2014 04:36:16 +0000 (12:36 +0800)]
vainfo: Query the supported profile list before listing profile/entrypoint
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Xiang, Haihao [Fri, 14 Feb 2014 04:15:25 +0000 (12:15 +0800)]
vainfo: clean up resource in the error handling path
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Xiang, Haihao [Wed, 27 Feb 2013 05:12:40 +0000 (13:12 +0800)]
vainfo: Add the support for the new VA profiles
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Gwenole Beauchesne [Mon, 3 Mar 2014 09:15:23 +0000 (10:15 +0100)]
tests: basic: add missing string representation for VP8 profiles.
Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
Gwenole Beauchesne [Mon, 3 Mar 2014 09:11:58 +0000 (10:11 +0100)]
va: drop VAEntrypointMax enumeration.
This is an oversight, and that was supposed to be removed earlier.
User applications cannot rely on this value to statically allocate
an array of possible VA entrypoints. The correct interface is to
use vaMaxNumEntrypoints() and dynamically allocate that array.
Rationale: VA entrypoints values are maintained sparsely, so this
can yield to a large number overall.
Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
Gwenole Beauchesne [Mon, 3 Mar 2014 09:05:00 +0000 (10:05 +0100)]
va: include <stdint.h> header.
Always include <stdint.h> header by default as we are using standard
fixed-size integral types from now on.
Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
Zhong Li [Mon, 20 Jan 2014 03:07:16 +0000 (11:07 +0800)]
vp8 dec: add clear notation for qindex
Signed-off-by: Zhong Li <zhong.li@intel.com>
Zhao, Halley [Tue, 3 Dec 2013 05:43:56 +0000 (13:43 +0800)]
vp8 dec: refine va_dec_vp8.h
add clear notation for: key_frame/VABoolCoderContextVPX.count/num_of_partitions/partition_size[0]
remove mb_skip_coeff since it can be deduced from mb_no_coeff_skip if it is required
Signed-off-by: Zhao, Halley <halley.zhao@intel.com>
Tianmi Chen [Mon, 1 Jul 2013 09:38:35 +0000 (17:38 +0800)]
libva: change the types of all fields in VABoolCoderContextVPX
BZ: 120268
Change the types of all fields in VABoolCoderContextVPX to
"unsigned char" to avoid potential bugs.
Change-Id: I483f115080cefe45ae49503ff5737c7dfff305a8
Signed-off-by: Tianmi Chen <tianmi.chen@intel.com>
Jonathan Bian [Thu, 14 Mar 2013 23:36:38 +0000 (16:36 -0700)]
Changes to VASliceParameterBufferVP8 to make it consistent with slice parameter buffer structure for other formats.
Jonathan Bian [Mon, 11 Mar 2013 05:34:51 +0000 (22:34 -0700)]
Added VASliceParameterBufferVP8 to describe data partition related fields.
Jonathan Bian [Thu, 26 Apr 2012 17:21:49 +0000 (10:21 -0700)]
Add VP8 decoding API
Added new profile, buffer types and data structures for VP8 decoding
Signed-off-by: Jonathan Bian <jonathan.bian@intel.com>
[Fixed conflicts and remove unused profiles]
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Rinat Ibragimov [Sun, 12 Jan 2014 15:45:17 +0000 (19:45 +0400)]
do not request any GLX_DEPTH_SIZE in create_tfp_surface
Window depth was confused with OpenGL depth buffer.
Signed-off-by: Rinat <ibragimovrinat@mail.ru>
Zhao Yakui [Fri, 5 Jul 2013 02:46:25 +0000 (10:46 +0800)]
MPEG2ENC: Initialize the f_code based on supported level instead of hacked setting
The f_code of MPEG2 picture parameter indicates the max search range of the
motion vector prediction. The small value will cause the unoptimized motion
vector prediction. So remove the hack setting.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit
4fa9b3ba9d45a021e282f761d061e1151158c3ea)
Victor Ying [Tue, 23 Jul 2013 06:25:00 +0000 (08:25 +0200)]
va: add new ABGR/XBGR fourcc codes.
Add VA_FOURCC_ABGR and its companion VA_FOURCC_XBGR definition. This
represents a packed RGB format with 32 bpp, i.e. 8-bit per component.
The components order in memory is A B G R.
Signed-off-by: Victor Ying <victor.ying@intel.com>
Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
Gwenole Beauchesne [Mon, 15 Jul 2013 08:45:16 +0000 (10:45 +0200)]
drm: improve check for authentication.
On Linux systems, the drmGetClient() function would return the thread ID
instead of the actual process ID.
Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
Gwenole Beauchesne [Thu, 27 Jun 2013 09:12:54 +0000 (11:12 +0200)]
build: fix make dist.
Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
Gwenole Beauchesne [Thu, 27 Jun 2013 09:12:00 +0000 (11:12 +0200)]
libva: bump version for development.
Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>