Bump version for development

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

Intel driver 1.2.2

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

1.2.2.pre2

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

VPP: Use the right top/bottom field flag used for DI

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=72518
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

H.264: Support Constrained Baseline profile instead of Baseline profile

GENx doesn't support FMO/ASO, so remove the support
of Baseline profile for conformance testing. In addition, add the support
for Constrained Baseline profile.

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

Upload the constant buffer on ILK/CTG so that subpicture can work

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit 2a61b0d17072bdc5b58608e3dfa2c4f9f80dcec4)

Support the smpte240m color standard for conversion from YUV to RGB

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit 34627c96f331f7a344270c3d51b634f5f166073e)

Support the BT709 color standard for conversion from YUV to RGB

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit 4c43ff9234a0a6f18744078d2e743cfa0cf8f34c)

Constant buffer passes YUV2RGB CSC matrix instead of hardcoded matrix

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit 0aa6ccc405726e5521452ee437be6b3cc6fabdee)

Increase the size of constant buffer for PS thread to pass more info

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit a6372850042e2a9201f4893193f77c9a55a6598e)

Fix the error in render shader on Ivy/Haswell

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit 7af528d17924cf2ec855c0d0b1550b6c3d095682)

Enlarge the size of array misc_param

And check the type before storing misc parameters

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

dec/mpeg2: ignore slices which aren't in raster scan order on SNB

Sometimes codec layer incorrectly fills slice parameters due to
the corrupted video

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=71276
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

SNB doesn't support MPEG-2 encoding

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=72016
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

Workaround for SNB

Backporting from d0184b5 in xf86-video-intel

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit e622ecedf169bccddc8910b45d92dbec7675441e)

VPP: remove some assert()

Instead check the input parameters and return corresponding error status if failed

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit 40fa7d9ede00e804f15df4b7b805c7345a925e17)

Correct the usage of width/height in struct object_surface

Add comments for width/height, orig_width/orig_height as well

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit f886f24eaaacba9544fa5f6405b7382c686f3a1f)

VPP: use the target widht/height to calculate the horizontal/vertical step on IVB

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit ba020ed2c7730980649d9131e41e9677f603c52a)

render: add support for brightness/contrast/hue/saturation

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit 04ecb6e79f4382d96eb5d4b51733049d420f592a)

VPP: change the default values for Saturation and Contrast

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit ce0c8f0019e8545d0db529b0f28338be4b8adc15)

Remove the unused variable to avoid the warning

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit 0d37a309bd99f6bded4df922d0ece22bf3bb1757)

Use GPU to construct MFX command buffer for H264 encoding on Haswell

This is to optimze the performance of h264 encoding. The GPU can
accelerate the construction of MFX command buffer for H264 encoding.

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit 87bc38d4300212dea51b5635f184aa1ae37fa71c)

Encoding reuses aux_batchbuffer instead of allocating another new buffer

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit 052ce2930cd4661b7ce62902e6553eec0e2db9f1)

Calculate the required space of batch buffer to avoid buffer overflow in encoding

The required size is based on the number of macroblocks and slice parameter.
Then it can avoid that too large buffer is allocated or possible overflow.

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit 8acdfd023e50af37a5642e2517683c34accd78b0)

Return the error instead of assert in vaEndPicture

This is to fix the crash issue caused by the incorrect parameter.

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit 1cee858036a87837deddc87586701ed869f96261)

Add more strict check to fix crash issue caused by invalid parameter

This is to fix the crash issue in https://bugs.freedesktop.org/show_bug.cgi?id=70397

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit 80d665eb670fd700d03f9a2486e452947177a058)

Fix one error of VME shader for MPEG2 encoding

Otherwise the MPEG2 encoding will use the incorrect prediction result for the
macroblocks in the first row if the MVP is used.

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit 44889dc0f3054cce226d5c09d431022fdffe3aac)

Fix one error of VME shader for H264 encoding

Otherwise the h264 encoding will use the incorrect prediction result for the
macroblocks in the first row if the MVP is used.

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>

Fix an incorrect makefile rule for VME shader on Ivybridge

Otherwise when the corresponding source file is modified, the binary shader
is not updated.

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>

VPP: add vebox motion compensation support on HSW

Signed-off-by: Zhong Li <zhong.li@intel.com>
(cherry picked from commit 2c7c7c4d20014342538a80bfd1525f9bef5ea971)

Conflicts:

src/i965_drv_video.c

Fix the Reference Index Mapping Table L0/L1

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit 3ffbe0297e98a88db18ae90ba5c1f8c429183baf)

Follow the input Picture/Slice parameters to generate slice header/data

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit f5a694e64d0163178c28dc25d9a3e7b9b1b5d162)

Pass the reference frame index in List0/1 into the PAK command

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit 68380a7f141bedcc0f6fbbbcee2f5e42b6ade0e0)

Select a reference frame from the reference list0/1

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit 3a51e5271773a637ef63ca397285ebdf326daba2)

Track the used reference surface

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit 97e1b531d85bd7b7d3bc1d3e5a7c3355af87a204)

Clean up for setting up reference surface state

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit 939b6bef6d8f8ecfee589cf70fde51f7a34175a1)

Fix the reference for list1

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit a45edbef143808ee925ef7708c516e6df21fa36b)

Check the reference surface id against VA_INVALID_SURFACE

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit 0c2def319f52bdb222e5480d81feea486cbf3e11)

Indent the code of encoding

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit 6ad68f55e9ba49af541a5e4d86a305bbd0f22d63)

fix the broken package generated by make dist

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

fix make dist

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

1.2.2.pre1 for development

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

Intel driver 1.2.1

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

VPP: fix the check for DI filter flags on SNB/IVB

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit ea3c725a7f9fa66e7ef05e1c08af30c28f8c1aa0)

support 422H/422V/411P/444P format for wayland video output
(cherry picked from commit 7a02190eda3abc245ba918f6078a121373f50703)

VPP: also remove an assert on forward temporal reference on HSW

Do the same thing of commit 8a0a61c

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=69669
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit f6685c309d94fb7679c9772703c8790cb71cdd73)

VPP: Remove an assert on forward temporal reference on IVB

Add a warning and return VA_STATUS_ERROR_INVALID_PARAMETER instead

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit 8a0a61ce47e086ac156602c3daaf70f20e4ca222)

VPP: MADI on IVB

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit 219452451ede9d52d8940be55cd38feda50baea7)

Conflicts:

src/i965_drv_video.c

VPP: track the frame sequence for DI on IVB

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit c008b4222d593594330fbb087204ff53722f9765)

VPP: move the BO for STMM into sub-contexts

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit 6ce827df0a9d6cb841308813ec54a019e3c63d37)

VPP: Create separate sub-context for each processing

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit 4faf6bf47f8e4e2fe587e3bb6a004340edd59c4c)

Conflicts:

src/i965_post_processing.c
src/i965_post_processing.h

VPP: the similar fix to Bob DI on SNB

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit c55cc476b1f75ddf0504a8b85ac80c0168585c5c)

VPP: Update DEINTERLACE_SAMPLER_STATE on IVB

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit b05729d0feea5ceb4217997f26442d5a8d94fa48)

VPP: VAProcFilterParameterBufferDeinterlacing::flags overrides VAProcPipelineParameterBuffer::filter_flags

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit 5334ca07c5ad1a2edeb279784bde3d03cbb10c49)

Update NEWS

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

allow vaDeriveImage() to work for UYVY formats

Hi,

I would push the following as obvious around next week.

Add support for UYVY format to vaDeriveImage(). Also remove dead code
along the way, i.e. packed YUV 4:2:2 formats have a single plane.

Regards,
Gwenole.

vpp: fix output filter count from QueryVideoProcFilters().

When vaQueryVideoProcFilters() returns VA_STATUS_ERROR_MAX_NUM_EXCEEDED,
i.e. when the caller allocated too few entries for the filters argument,
then the num_filters argument shall be adjusted to correct number of
entries that need to be re-allocated.

https://bugs.freedesktop.org/show_bug.cgi?id=67292

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

vpp: add basic processing support for packed YUV to packed YUV.

Add support for video processing from packed YUV 4:2:2 formats (YUY2, UYVY)
to packed YUV 4:2:2 formats. In particular, add support for HW accelerated
vaPutImage() for packed YUV formats.

https://bugs.freedesktop.org/show_bug.cgi?id=67338

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
[Haihao: the shader for IVB+ has been added in 0eb2288]
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

VPP: Packed 4:2:2 to packed 4:2:2 on IVB+

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

VPP: cleanup code

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

Return the status of image processing

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

More conversions between two images with different pixel formsts

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

Convert RGBx/RGBA into non-NV12 format

It is the combined conversions of RGBx/RGBA->NV12 and NV12->non-NV12.
It would be better to implement RGBx/RGBA->YUV444 and YUV444->non-444 later

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

Support B frame for reference frame

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

Remove the unnecessary shader binary for MPEG2 encoding on Haswell/Ivb

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit 42bb613e72d235bcbe141c906dec9431e4c29661)

Optimize quantization rounding precision for MPEG2 encoding on Ivy

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit 05ea96da3e0b09648bfdeb35967f6ab9bb3b23e4)

Optimize the VME shader for MPEG2 encoding on Ivb

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit 15db142076321e9523db0c8be8e6bae5e1b64c8a)

Optimize MPEG2 encoding on Ivb

This is backported from Haswell. The scoreboard/walker/cost-center
is applied.

Signed-off-by: Zhao yakui <yakui.zhao@intel.com>
(cherry picked from commit cc7452d14f7faa1d5b5fa8c16db3e3cb3fd4f53d)

Rewrite the VME shader for MPEG2 encoding on Ivy

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit 6200c9a7779c1309f5a85b7c62aec1b9796793c6)

Remove the dead code in file of gen7_vme.c

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit d15582b54486847811f3feab46ddd51181561776)

Pass the constant buffer info for MPEG2 encoding correctly on Ivb

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit 7690091889eac91dcab53e0318f9810c25071e18)

Optimize quantization rounding precision for MPEG2 encoding on haswell

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit 24d8bf31c8aeb326bc8b33c1ac9700ec1d169666)

Optimize the VME shader for MPEG2 on Haswell

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit 19e93152f0e10f94ecaf3ddecf95c1dc7b97dfed)

Configure the cost-center of MPEG2 VME shader on haswell

This is derived from the neighbour macroblock based on MPEG2 spec.

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit a51860aba2d2713a9c8c817c430ffa93abd5145a)

Use the scoreboard/walker to assure MB dependency for MPEG2 encoding

If MVP is added for MPEG2 encoding, it must be assured that the left macroblock
should be already finished before processing the current macroblock.
And this needs the scoreboard/walker mechanism to assure MB dependency.

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit 5d2163d02d395fc0a794d834979a06c287bf9ba5)

configure the dynamic VME MV/mode cost for MPEG2 encoding on Haswell

Currently it uses the constant VME MV/mode cost when executing the mode/motion
vector prediction for MPEG2 encoding on Haswell, which causes that the unoptimized
mode/MV is used for MPEG2 encoding.

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit 6fec7e353704dc23c9675966467caffa95f792fe)

Enable the Intra-prediction for MPEG2 P-B frame

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit c074d4d61ad931d044fbe0836a45c49768090b4b)

Restrict the MV search range based on MPEG2 encoding LEVEL

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit c05073b1f8764271ccf4fe1aa037f881dedd3818)

Rewrite inter-frame shader for MPEG2 encoding on HSW to follow MPEG2 spec

Now the MPEG2/H264 uses the same mode/motion vector prediction shader. But
the MV search region of mpeg2 is different with that on H264, which causes
that the wrong mode/motion vector prediction is used for MPEG2.

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit 6842f08aa375b5942cee4b9d06421609c212895a)

Convert 422H/422V/411P/444P into other formats for internal using

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit 232ef48766c1f91a87a3d41f951fd2ac26dcf2ae)

A separate batch buffer for video processing

It is easy to result in multithread issue if the rendering code
and video processing code share the same batch buffer

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit ce0984814269e0923f44196e47f1c7cc2dddc55c)

Use the right wight/height to initialize the internal buffers for MPEG-2 encoding

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit 3ecbff585af918d96959ce791eec29be25360d91)

Cleanup profile tracking in encoder

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit edd25a94e92b9cec23594dc978691506a1c8cfab)

Rename the macros

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit 3ab97be8db1b8e55d0d5b95f577863416a87c6ff)

Enable the Bay Trail platform.

This patch adds PCI IDs for Bay Trail (sometimes called Valley View).
As far as the video driver is concerned, it's very similar to
Ivybridge GT1 except VP8 decoding support.
(cherry picked from commit b3afeef8092dc4eb7cb73fce672ddf7a55205f34)

VPP: remove needless functions and parameters in gpe pipeline
(cherry picked from commit ab0546e76967e5e7c465569f90e192b560678d8c)

Fixes valgrind warning

"Conditional jump or move depends on uninitialised value(s)"

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit cbd00deb6c5cad58ebd5e6ce5b89aaaded0f78a5)

Release the private driver data when call vaTerminate()

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit c735d9e58dd49c9a92ad0042b5649a9d3fe7c2c4)

VPP: check the filter when query the video filter capabilities

Return VA_STATUS_ERROR_UNSUPPORTED_FILTER if an unsupported filter
was supplied

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit 428723853f88b4d5cac436f5fd62e6cc64a9e8e9)

Insert a phantom slice for H.264 deocdong on SNB

If the first slice does't start at 0, a phantom slice is added
before the first slice.

This fixes the GPU hang issue mentioned in https://bugs.freedesktop.org/show_bug.cgi?id=63946
(not the original issue).

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Tested-by: Krzysztof Kotlenga <pocek@users.sf.net>
(cherry picked from commit 11115e3f0427d056367c1c5946585e3f7cead662)

Check the returned pointer from malloc() before using it

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit 1caf179b1425b13cacaa421c688c6df8369668c6)

Add the dependency to the ring supported by the underlying OS for VPP filters

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit a532539cbc7048f5c01b64dfe239f1570123c959)

Check whether VEBOX is supported by the underlying OS

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit c586c80d29d8860011d95e78d1609ff3683f3cc4)

Won't build the shaders for Sharpening on HSW

This fixes https://bugs.freedesktop.org/show_bug.cgi?id=66258

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

build: fix make dist for packaging.

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

Udate the micro version

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

Bump version for development

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

Intel driver 1.2.0

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

NEWS: updates

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

Update the dependency on libdrm

libdrm 2.4.45 is required to support VEBOX on HSW

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