configure: fix GLX and EGL tests.

Migrate checks for GLX and EGL to pkg-config checks or use default
include and libraries for testing.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

Revert "declare vaGetDisplay in va.h instead of va_xxx.h"

vaGetDisplay() belongs to winsys headers.

This reverts commit 52c87f898b210cddde70231021097c5a0bb6fdc3.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

API: h264 decode: fix the definition of slice_data_bit_offset.

Align with the existing practice in various media stacks available
in Linux today. i.e. codec layer submits slice data that represents
the original bitstream, thus including any emulation prevent bytes.
However, the slice_data_bit_offset represents an offset relative
to the buffer that contains the NAL unit byte and the rest of the
slice with any emulation prevention bytes removed.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

API: h264 decode: clarify slice_data_bit_offset field.

slice_data_bit_offset represents a bit offset from the NAL Header Unit
byte to the begining of the slice_data(), thus including any emulation
prevention bytes. Baiscally, this represents the number of bits parsed
for slice_header().

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

Allow 0.32.x drivers.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

API: bump version to 0.33.0 (libva 1.1.0).

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

test/egl: Fix compiling error if GLESv1_CM isn't installed in the default location

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Signed-off-by: Hai Lan <hai.lan@intel.com>

Demo code v4l/h264 encode/decode using Libva

See README.txt for details

Signed-off-by: Sven Putteneers <svenx.putteneers@intel.com>
Signed-off-by: Lecluse, Philippe <Philippe.Lecluse@intel.com>

update vainfo

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

A simple JPEG example to use VAAPI

It is based on Luc Saillard <luc@saillard.org>'s JPEG program

Usage: loadjpeg <input_filename.jpeg>

Signed-off-by: Zhao Halley <halley.zhao@intel.com>
Singed-off-by: Wind Yuan <feng.yuan@intel.com>
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

Fix comment for JPEG quantization table

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

Remove the macros for JPEG components

The range of component id is 0-255 (see B.2.2)

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

Directly use component id instead of component index in VASliceParameterBufferJPEG

The user application doesn't need to calculate the index anymore.

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

Modify the interface of Huffman table for JPEG decoding

The user application will send up to 4 huffman table to driver,
and driver will select the right huffman table for each component
via dc_selector and ac_selector.

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

Support 16-bits quantization element for JPEG decoding

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

Modify the interface of Quantization table for JPEG decoding

The user application will send up to 4 quantization table to driver,
and driver will select the right quantization table for each component
via quantiser_table_selector. The quantization elements are specified
in zig-zag scan order

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

refine the JPEG decoding interface

va: add data structure for JPEG decoding

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

include va_egl.h instead of va.h

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

fix the check for EGL

It is possible that EGL/egl.h doesn't exist however EGL library is
available. We don't want to build EGL related stuff in this case

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

Don't build egl case without EGL

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

Remove dependency on X11/Xlib.h

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

Remove legacy DRI support

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

Don't build the EGL test case without GLESv1_CM

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

egl/va_egl_x11: fix the number of the allocated elements

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

Array bound check

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

Update VA/EGL test case as well

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

Update the VA/EGL implementation in VAAPI

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

Update VA/EGL interfaces

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

Move va_egl.h to va/ to avoid compile error

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

a test case to use VA/EGL interfaces

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

VA/EGL interfaces

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

New hooks to create/free native pixmap

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

declare vaGetDisplay in va.h instead of va_xxx.h

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

rename NativeDisplay to VANativeDisplay

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

va_tpi.h: corret a typo (Kbuf->KBuf)

Signed-off-by: Austin Yuan <shengquan.yuan@intel.com>

configure: add check for EGL.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

configure: simplify summary.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

configure: cosmetics (clean up configure options).

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

configure: cosmetics (sort AC_OUTPUT Makefiles).

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

doc: add Doxygen templates.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

libva: bump version for development.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

1.0.15.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

debian: add package for vainfo.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

debian: don't use simple-patchsys (upstream has no patches).

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

Add NEWS file.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

va: fix compiler warnings (drop unused variables).

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

Add auto-generated Debian packaging.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

libva: bump version for development.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

libva: allow pre-releases.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

libva: fix reported version as the one from VA-API.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

vainfo: report both VA-API and libva versions.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

API: fix versioning.

We have to consider three distinct versions at this time: the VA-API
version, the library package version number, and the DSO version.

* VA-API version:
- increment major for any ABI change (which shall not occur!)
- increment minor for any interface change (e.g. new or modified function)
- increment micro for any other change (e.g. new flag, new codec definitions)
- reset micro version to zero when minor version is incremented
- reset minor version to zero when major version is incremented

* libva package version number:
- major version is automatically generated from VA-API major version
- minor version is automatically generated from VA-API minor version
- increment micro for any library release
- reset micro version to zero when VA-API major or minor version is incremented

* DSO version:
The SONAME shall remain to libva.so.1 for VA-API 0.x.y as long as the ABI
is not changed. Thus, the library name is generated as libva.<x>.<y>.0 where
<x> = VA-API major version + 1
<y> = 100 * VA-API minor version + VA-API micro version

For example:
VA-API 0.32.0 generates libva.so.1.3200.0 (libva 1.0.14 as of today)
VA-API 0.34.1 generates libva.so.1.3401.0 (e.g. libva 1.2.1)
VA-API 1.2.13 generates libva.so.2.213.0  (e.g. libva 2.2.13)

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

test: fix build warnings on Android

Change-Id: Ibc4231395af8dfcdef403d44b98c644349e2ba04
Signed-off-by: Austin Yuan <shengquan.yuan@gmail.com>

android: fix <va/va_version.h> dependencies

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

android: fix generation of <va/va_version.h>

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

va: generate __vaDriverInit_*() function name at run-time.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

va: don't explicitly include "config.h"

sysdeps.h is the internal header file to include that will eventually
include config.h, if necessary. e.g. we don't use it on Android.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

build: drop generated files.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

vainfo: use system or user-defined DISPLAY to create the X display.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

dummy_drv_video: fix last commit.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

dummy_drv_video: only build if requested.

pkgconfig: move files to a specific directory.

configure: drop check for intel-gen4asm (moved to intel-driver).

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

va_tpi: correct the first argument of vaCreateSurfaceFromKBuf

"VADisplay dpy" --> "VADriverContextP ctx"

Signed-off-by: Austin Yuan <shengquan.yuan@gmail.com>

va_fglrx: fix driver name detection.

VA_FGLRXGetClientDriverName() used to return success even if no
valid driver name was found, thus returning a NULL driver name.

Also fixed the X display name matching routine to filter out the
screen number, which may not be in the string, and make sure to
compare screen numbers from what ADL returns too.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

vatrace/vafool: refine it

remove va_fool_getframe.c which looks strange
correct some coding style for file va.c
simplify va_fool.c for both decode and encode
refine the surface dump of va_trace.c

Signed-off-by: Austin Yuan <shengquan.yuan@gmail.com>

va_tpi: vaCreateSurfaceFromKBuf for streaming the video playback to remote monitor

For such use case, the pipeline is: decode->encode->send to network.
The local decode/playback and the encode process are seperated, thus it
needs the buffer sharing b/w the decode and encode in seperate process.
This API creates/wraps the decoded surface into an encode surface. Its
input parameter is the kernel buffer handle of decoded surface which is
only understandable by driver implementation.

Change-Id: I7953ecf3b7333c1cdb6af5a0f45b054c8840f8e2
Signed-off-by: jiguoliang <guoliangx.ji@intel.com>
Signed-off-by: Austin Yuan <shengquan.yuan@gmail.com>

va.h: TopFieldOrderCnt/BottomFieldOrderCnt should be signed instead of unsigned

POC should be integer.

Change-Id: I75281aafa148c88f0de58981969e1c43ea8d2684
Signed-off-by: ywan171 <yi.a.wang@intel.com>

va: remove i965 driver from libva repository

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

va/test: Fix DSO link issue

I didn't see this issue until I upgraded my system recently.
It's evident the default behavior of the linker is changed.
Explicitly adding the missing libraries to the link step can resolve
this issue.

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

va: update the check for intel-gen4asm

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

bump to 1.0.14

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

i965_drv_video: don't export internal driver functions.

Make sure to use our internal functions. In particular, we override
some DRM functions and they have to be used. e.g. this fixes VA/GLX.

va.h: VA_DECODE_ERROR_TYPE changed to VADecodeErrorType for unified codying style (CamelCase)

VA_DECODE_SLICE_MISSING --> VADecodeSliceMissing
VA_DECODE_MB_ERROR --> VADecodeMBError

Signed-off-by: Austin Yuan <shengquan.yuan@gmail.com>

dri2: ignore BadDrawable errors.

Hi,

This patch fixes the following error.

X Error of failed request:  BadDrawable (invalid Pixmap or Window parameter)
  Major opcode of failed request:  137 (DRI2)
  Minor opcode of failed request:  4 (DRI2DestroyDrawable)
  Resource id in failed request:  0x4800001
  Serial number of failed request:  30
  Current serial number in output stream:  32

The X drawable can die before we notice it in free_drawable_hashtable(),
called through vaTerminate(). This patch is ported from Mesa (GLX/DRI2).

Regards,
Gwenole.

Fix make dist.

Fix out-of-source builds.

i965_drv_video: cosmetics (cleanup Makefile).

cosmetics: clean up Makefiles.

VA/X11: drop useless casts.

VA/GLX: drop useless casts.

vatrace: trace JPEG encode parameters

Signed-off-by: Austin Yuan <shengquan.yuan@gmail.com>

Update on vatrace/test and va.h

vatrace:print H263 decode paramters
test/putsurfac: create drawable with a geometry from commandline
va.h: add BGRA fourcc

Signed-off-by: Austin Yuan <shengquan.yuan@gmail.com>

i965_drv_video: update post processing interface

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

i965_drv_video: fixes assertion failure

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

i965_drv_video: fix next slice vertical position for field picture

This fixes https://bugs.freedesktop.org/show_bug.cgi?id=38628

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

vatrace: print MPEG4/H263 encode parameters information

Change-Id: Ib39b436865c2b665dc49c916d5f364f03e8c79b5
igned-off-by: Austin Yuan <shengquan.yuan@gmail.com>

i965_drv_video: add support for VA_SUBPICTURE_DESTINATION_IS_SCREEN_COORD.

i965_drv_video: fix subpicture scale factor for Y axis.

i965_drv_video: simplify put_surface() and put_subpicture() args.

va/test/putsurface_x11: avoid warning 'warning: ‘win’ is used uninitialized in this function'

This fixes broken putsurface on X11

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

va/test: return a value to silence compiler waring

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

va: fix warning about memset.

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

test/putsurface: more cleanup, remove #ifdef ANDROID from putsurface_common.c

Signed-off-by: Austin Yuan <shengquan.yuan@gmail.com>

test/putsurface: add UV element into the auto-generated image

Signed-off-by: Austin Yuan <shengquan.yuan@gmail.com>

Fix vaTerminate() with non DRI-based drivers.

va-android: same as X11, also always create a new VA display on android

Previously a VA display will be shared if a native display is shared by multiple threads,
it will casue some thread safety issue in a multi-threaded program.

Remove the global VADisplayContextP pDisplayContexts which is not thread safe

Signed-off-by: Austin Yuan <shengquan.yuan@gmail.com>

i965_drv_video: fix assertion failure

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

i965_drv_video: fix GPU hang issue when decoding field coded MPEG2 picture.

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>