CreateSurfaces2: forward function return to caller

we did not check following function's return:
i965_surface_native_memory
i965_suface_external_memory
i965_check_alloc_surface_bo
It will be in trouble if underlay buffer allocation is failed.

1.6.1.pre1 for development

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

Merge branch 'v1.6-branch' into master

vpp/avs: fix license terms (MIT).

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

H264: Use macroblock pair to calculate H264 decoding parameter under MBAFF flag

Based on the H264 spec the macroblock pair should be used to calculate
the corresponding parameters under MBAFF.(mb-adaptive frame-field).
Otherwise the wrong parameter is sent to GPU HW.

Fix the GPU hang issue in https://bugs.freedesktop.org/show_bug.cgi?id=91207

Tested-by: Lim, Siew Hoon <siew.hoon.lim@intel.com>
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>

libva-intel-driver 1.6.0

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

dec/hevc: Use a different equation for metadata tile column buffer on BSW

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>

dec/hevc: Use the same equation for each metadata buffer no matter slice type

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>

Gen75_vpp: Don't assert on BGRA input and output

version 1.6.0.pre2

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

FIX:HEVC enc cause GPU hang when lcu_size=32 with 720x480 and 176x144

Signed-off-by: Qu,Pengfei <Pengfei.Qu@intel.com>
(cherry picked from commit 0f6143e2e4bd70217887e1ad86a76200ad8b38a8)

VP8 HWEnc: Modify qp threshold value for mode cost calculatation

The patch is helpful to improve quality when qp is lower than the
threshold value.

Signed-off-by: Zhong Li <zhong.li@intel.com>
(cherry picked from commit e797089446c1f5b71b239b9046d76e054dfcba59)

VPP: Configure SURFACE_STATE based on HW requirement to fix out-of-bound access issue

Otherwise some pixles will be regarded as out-of-bound and the
access will be dropped.

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Acked-by: Xiang haihao <haihao.xiang@intel.com>
(cherry picked from commit d5cd551887c8f4e5408327aba993e9927dd7d297)

BDW/CHV/SKL: Follow the HW alignment requirement to access GPU buffer correctly

Otherwise the buffer access is incorrect.

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Acked-by: Xiang haihao <haihao.xiang@intel.com>
(cherry picked from commit fbab204929326567c876a48c0552430dd67a9852)

VPP: Configure SURFACE_STATE based on HW requirement to fix out-of-bound access issue

Otherwise some pixles will be regarded as out-of-bound and the
access will be dropped.

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Acked-by: Xiang haihao <haihao.xiang@intel.com>

BDW/CHV/SKL: Follow the HW alignment requirement to access GPU buffer correctly

Otherwise the buffer access is incorrect.

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Acked-by: Xiang haihao <haihao.xiang@intel.com>

FIX:HEVC enc cause GPU hang when lcu_size=32 with 720x480 and 176x144

Signed-off-by: Qu,Pengfei <Pengfei.Qu@intel.com>

VP8 HWEnc: Modify qp threshold value for mode cost calculatation

The patch is helpful to improve quality when qp is lower than the
threshold value.

Signed-off-by: Zhong Li <zhong.li@intel.com>

Update NEWS

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

Update the dependency on VA API

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

Add vp8_inter_frame_gen8.asm and vp8_intra_frame_gen8.asm to the core file list

Otherwise the packakge created by 'make dist' isn't buildable.

Remove vp8_inter_frame_gen9.asm and vp8_intra_frame_gen9.asm as they are the same
of vp8_inter_frame_gen8.asm and vp8_intra_frame_gen8.asm.

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

Add Android.mk to the corresponding EXTRA_DIST

'make dist' can include all Android.mk files in the release package

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=90541
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

Fix 'make dist'

Otherwise the tar package created by 'make dist' is not buildable

(cherry picked from commit 2d0ce24c66c14c3f77bef54bdc4605241e0527bd)

Update NEWS

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit 3c66aa58ccb0fcb393472eeaa6a9188216683a79)

dec/vp8: fix segmentation update

Segment-based adjustments only occur if segmentation is enabled ($9.3).

Cc: Zhao Yakui <yakui.zhao@intel.com>
Tested-by: Charles, Daniel <daniel.charles@intel.com>
Acked-by: Zhao, Yakui <yakui.zhao@intel.com>
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Signed-off-by: Sean V Kelley <seanvk@posteo.de>

HEVC/enc: Fix Multi Slice Encoding.

Fix the calculation of horizontal and vertical CTU position for next slice.

Signed-off-by: Sreerenj Balachandran <sreerenj.balachandran@intel.com>

HEVC: encoder_utils: Fixes for packed slice header generation.

-- Write no_output_of_prior_pics_flag to bitstream header for IRAP pictures
-- The slice_temporal_mvp_enabled_flag should only presnt in non-IDR frames.
-- Write collocated_from_l0_flag to bitstream header for B slice.
-- The max_num_merge_cand should be 5 - MaxNumMergeCand

Signed-off-by: Sreerenj Balachandran <sreerenj.balachandran@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>

HEVC: Fix the nal unit type generation in packed slice header.

Generate the Nal unit type for packed slice headers based
on idr_pic_flag and reference_pic_flag.

Signed-off-by: Sreerenj Balachandran <sreerenj.balachandran@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>

HEVC: Fix the Wrong slice type usage

Don't mix the slice_type values of h264 and h265.

The SLICE_TYPE values of H264 and HEVC are different.
H264: Bslice_type = 1, Pslice_type = 0
HEVC: Bslice_type = 0, Pslice_type = 1

Signed-off-by: Sreerenj Balachandran <sreerenj.balachandran@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>

VPP: Enable field rate madi/mcdi. Only second hunk is tested

Signed-off-by: Andy Furniss <adf.lists@gmail.com>

VPP: Make sure the store buffer is allocated

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

VPP: Adjust the parameter settings to DI on GEN6/GEN7

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

vpp: add support for Motion-Compensated deinterlacing on Ivybridge.

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

vpp: fix advanced deinterlacing on Sandybridge and Ivybridge.

This fixes support for Motion Adaptive deinterlacing mode on both
Sandybridge and Ivybridge platforms. In particular, correct field
ordering is now supported, and STMM ping-pong buffering is added.

v2: changed STMM surface format to Y800 for a single plane.

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

vpp: fix memory leak in DNDI code path.

The original current_out_surface was never released on exit. Main
reason for that is the legacy VPP framework that did not allow the
VADriverContextP handle to be passed down to the desired .finalize()
hook. Improved that to bring it on par with the VEBOX code path.

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

Bump version to 1.6.0.pre1

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

VP8 HWEnc: Init CBR QP at begining of gen8_mfc_vp8_init

Initialize CBR QP at begining of gen8_mfc_vp8_init(), because this
funtion will use qp to init prob_skip_false.

Signed-off-by: Zhong Li <zhong.li@intel.com>

VP8 HWEnc: Fix CBR qp setting for vme cost

1. The q_index range of vp8 is different from h264, it's need to be
mapped to avc qp range.
2. For CBR case, qp should be set to CBR calculated value.

Signed-off-by: Zhong Li <zhong.li@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>

gen8_mfd: free surface used by JPEG decode

This is a follow up to previous commit:

commit 8dda0650b55c467200c34ada127b5d9b54edea5c
Author: Zhao Yakui <yakui.zhao@intel.com>
Date:   Tue May 5 17:56:38 2015 +0800

    Decode: Free the allocated internal surface used by JPEG to avoid memory leak

For GEN8, this code path is normally disabled by default.
However, one can enable this code path by defining JPEG_WA
during compilation.  In that case, we would enable the
same memory leak.  Thus, apply the same fix in gen8_mfd to
fix it.

Signed-off-by: U. Artie Eoff <ullysses.a.eoff@intel.com>
Signed-off-by: Sean V Kelley <seanvk@posteo.de>
Tested-by: Sean V Kelley <seanvk@posteo.de>

Decode: Free the allocated internal surface used by JPEG to avoid memory leak

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Sean V Kelley <seanvk@posteo.de>
Tested-by: Sean V Kelley <seanvk@posteo.de>

VP8 HWEnc: Fix vp8 mv incorrectly stored issue

As vp8 spec, all luma motion vectors are doubled stored.
This patch has much benefit to P frame qualtity.

Signed-off-by: Zhong Li <zhong.li@intel.com>

Fix the build on Android

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

VP8 HWEnc: Alloc larger memory for internal buffer

It is to prevent the internal bitstream buffer overflow when qp is low.

Signed-off-by: Zhong Li <zhong.li@intel.com>

VP8 HWEnc: Add CBR Support

Add bit rate control (CBR) for vp8 hw encoding

v1->v2: follow haihao's comment to update this patch

Signed-off-by: Zhong Li <zhong.li@intel.com>

HEVC/dec: correct the frame store index

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

HEVC/dec: Fix collocated_ref_idx and collocated_from_l0_flag

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

HEVC/dec: Make sure the derived value for ChromaLog2WeightDenom is in the the range of 0 to 7

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

HEVC/dec: Fix weight/offset in HCP_WEIGHTOFFSET

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

Fix segmentation fault

The segmentation fault occurs when CBR is selected without HRD parameter

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

HEVC enc:Added 4K&2K support; Added Profile&Level 5.1 above

Signed-off-by: Qu,Pengfei <Pengfei.Qu@intel.com>

Allow va{Get,Put}Image() to use derived image

The surface can't be the same surface from which the image has
been derived.

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Reviewed-by: Sean V Kelley <seanvk@posteo.de>
Tested-by: Sean V Kelley <seanvk@posteo.de>

Fix the quality factor.

The following are added in this commit:
    1) Added lower bound to the quality factor.
    2) Added extensive comments on scaling and normalization of the quality factor.

check that intel-gen4asm tool is actually present

intel-gen4asm tool has to be present on the host tools so that
it can actually be used. Cross compiling shows problems when the
gpu-tools version is met but the tool is not available on the PATH

So, first check that the tool is available and later check for version
before defining HAVE_GEN4ASM.

Signed-off-by: Daniel Charles <daniel.charles@intel.com>
(cherry picked from commit 038097c00ad7b4f2e0fd572ee20f255dd378b5cb)

Conflicts:

configure.ac

vpp: add support for "low-power" mode.

Add support for "low-power" mode expressed with VA_PROC_PIPELINE_FAST
flag set to VAProcPipelineParameterBuffer.pipeline_flags. The purpose
is to discard any complex operation that would consume too many HW
resources.

While doing so, also try to optimize for cases where we only want to
perform format conversion, scaling and basic bob-deinterlacing.

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

gen7_mfd: Optimise start code search

Code inspired from GStreamer

See https://bugs.freedesktop.org/show_bug.cgi?id=89507

Signed-off-by: Olivier Crete <olivier.crete@collabora.com>

Add support for HEVC decoding on CHV

We can re-use SKL HEVC decode pipeline

Signed-off-by: Sean V Kelley <seanvk@posteo.de>
Cc: haihao.xiang@intel.com
Cc: focus.luo@intel.com
(cherry picked from commit 78171ec3b8e73071405d9ff6ecbddc05a6787001)

MPEG2 ENC: Remove gen9 pak pipeline code

SKL pak pipeline of mpeg2 encoding is same as gen8,
so remove the redundant code

Signed-off-by: Zhong Li <zhong.li@intel.com>
(cherry picked from commit ac46de469dd5eeaed34be0510631228a07748ec3)

VP8 DEC: Change error concealment method

Use intra prediciton as error concealment method, instead of inter P
copy.

Signed-off-by: Zhong Li <zhong.li@intel.com>
(cherry picked from commit 8d8b4a577fbaa8d278fbc5fe213023cfeea5d57a)

Disable the context check to avoid JPEG ENC failed

https://bugs.freedesktop.org/show_bug.cgi?id=88728
Signed-off-by: Qu,Pengfei <Pengfei.Qu@intel.com>
(cherry picked from commit a8ca5f911c7c19b5ede3f96ac6113606b9dd4d95)

Just Style alignment

Signed-off-by: Qu,Pengfei <Pengfei.Qu@intel.com>
(cherry picked from commit 22439f1fcebbd2c1d563503720b1123a4bb97160)

HEVC ENC:Added HEVC support in API function

Signed-off-by: Qu,Pengfei <Pengfei.Qu@intel.com>
(cherry picked from commit d4f1087a0606c7c8d2d5fd65d9f46990f41ce93a)

HEVC ENC:Added pipeline init

Signed-off-by: Qu,Pengfei <Pengfei.Qu@intel.com>
(cherry picked from commit 2759cabfa57c6e78ab2033cad4e7c2cbf2b9d4e2)

HEVC ENC:Added VME pipeline

Signed-off-by: Qu,Pengfei <Pengfei.Qu@intel.com>
(cherry picked from commit 2a85633e34a4e66b18634e7b53cc91aa44ad2c2b)

Conflicts:
src/gen6_mfc_common.c

HEVC ENC:Added PAK pipeline

Signed-off-by: Qu,Pengfei <Pengfei.Qu@intel.com>
[Fix PAK-BSE data start offset in HCP_SLICE_STATE]
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit 3f47222004c6c2e575b170c7c7c3be00f61df302)

HEVC ENC:Added slice header generated internally

Signed-off-by: Qu,Pengfei <Pengfei.Qu@intel.com>
(cherry picked from commit 2c54b96fe89aa0839f8c34935a84b79c41fde55d)

Conflicts:
src/i965_encoder_utils.h

HEVC ENC:Added PAK context for HEVC

Signed-off-by: Qu,Pengfei <Pengfei.Qu@intel.com>
(cherry picked from commit 41f23f3dcfe24ca9493b36cb00bab5317584b099)

VP8 HWEnc: Add BSW VP8 HWEnc support

Add BSW vp8 encoding support, and let SKL and BDW use the same PAK pipeline.

Signed-off-by: Zhong Li <zhong.li@intel.com>
(cherry picked from commit c2be56ae6f3628ea246a1dd02e5cac18da84df56)

VP8 HWEnc: declare copyright and license of vp8_probs.h

Signed-off-by: Zhong Li <zhong.li@intel.com>
(cherry picked from commit a52a5ffe1dc62ff12989e329860ae81ced247f00)

VP8 HWEnc: vp8 encode quality optimization

1. Add mv and mode cost calc
2. Support 16x16 and 4x4 intra prediction
3. Support multi-mode intar prediction
4. Support 16x16 New_MV mode inter prediction

Signed-off-by: Zhong Li <zhong.li@intel.com>
(cherry picked from commit a25bda914d4041a96252bd24a597558af3237b15)

VP8 HWEnc: Calc vp8 coded size by internal buffer

Signed-off-by: Zhong Li <zhong.li@intel.com>
(cherry picked from commit 379282cbc79b65c994c22009412472bdeddbb4cc)

VP8 HWEnc: Add P frame encoding support

Signed-off-by: Zhong Li <zhong.li@intel.com>
(cherry picked from commit e4737bb3b29c5ac9fb01f4efccd743d260673e27)

Conflicts:
src/shaders/vme/Makefile.am

VP8 HWEnc: Build VP8 PAK pipeline and enabling I frame

Signed-off-by: Zhong Li <zhong.li@intel.com>
(cherry picked from commit a18ce4664113d5b1a9b29ed45bf137df6b7a7898)

Conflicts:
src/i965_encoder_utils.c

VP8 HWEnc: Build vp8 gen9 encoding vme pipeline

Signed-off-by: Zhong Li <zhong.li@intel.com>
(cherry picked from commit d8588862d85414ef412fd5e7ae1fb9ca78b69e84)

Conflicts:
src/i965_device_info.c

VP8 HWEnc: Add vp8 gen9 intra frame encoding shader

Signed-off-by: Zhong Li <zhong.li@intel.com>
(cherry picked from commit 722fe89eae811743d7d46f9d11162aa3a5c3ba53)

Conflicts:
src/shaders/vme/Makefile.am

1.5.2.pre1 for development

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

Add new SKL PCI ids

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>

jpeg_enc: Fix the column raster conversion of quatization matrix.

Use temp array for the new raster order calculation, otherwise
it simply generate wrong values.

Note: There is opportunity for more optimization, for eg
we can avoid the usage of mulitple for loops here and there.

Signed-off-by: Sreerenj Balachandran <sreerenj.balachandran@intel.com>
Reviewed-by: Sirisha Muppavarapu <sirisha.muppavarapu@intel.com>

jpeg_enc: Fix the quatisation matrix scaling.

The misplaced parentheses are causing wrong value assignment
to the quatization matrix.

This will allow the ecoding when quality > 50.
Otherwise it will simply generate garbage in encoded video
for any quality factor greater than 50

Signed-off-by: Sreerenj Balachandran <sreerenj.balachandran@intel.com>
Reviewed-by: Sirisha Muppavarapu <sirisha.muppavarapu@intel.com>

jpeg_enc: Avoid integer overflow while doing quality factor scaling

For eg: The uint8_t will simple overflow if submitted
quality factor is 1 (5000/1).

Signed-off-by: Sreerenj Balachandran <sreerenj.balachandran@intel.com>
Reviewed-by: Sirisha Muppavarapu <sirisha.muppavarapu@intel.com>

Do not print warnings on stdout

The driver should not use stdout as this pipe is used by
a lot of programs to exchange data and should instead print
debug messages to stderr

Signed-off-by: Michael Müller <michael@fds-team.de>

decode/VP8: HW needs 1 extra byte for each partition

Signed-off-by: Zhong Li <zhong.li@intel.com>
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Reviewed-by: Sean V Kelley <seanvk@posteo.de>
Tested-by: Sean V Kelley <seanvk@posteo.de>

libva-intel-driver: Implement max width and height in QuerySurfaceAttributes

With this, we can get hardware supported resolution for video encode and
decode. This CL is used to prepare GetSupportedResolution in chrome.

Signed-off-by: Heng-ruey Hsu <henryhsu@chromium.org>
Reviewed-by: Wu-cheng Li <wuchengli@chromium.org>
Reviewed-by: Sean V Kelley <seanvk@posteo.de>
Cc: Pawel Osciak <posciak@chromium.org>
Cc: Stéphane Marchesin <marcheu@chromium.org>

dri: Update region's width/height on every PutSurface

Currently we expect that if a surface changes size, we will be given a
new buffer with a new name. This is not always true as the ddx may try
to keep the same buffer for the client if it is large enough for the new
size. Since we always know the width/height of the buffer, we can pass
that information along every time into the intel_region struct.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
Cc: Dmitry Ermilov <dmitry.ermilov@intel.com>
Reviewed-by: Sean V Kelley <seanvk@posteo.de>
Tested-by: Sean V Kelley <seanvk@posteo.de>

The SEND with EOT message must use R112-R127 for message payload on GEN7+

EOT message on GEN9 has used R112-R127, but the same message on GEN7-GEN8 still uses
other registers. To avoid potential issues, use R112-R127 on GEN7-GEN8 too

In additition, GEN7-GEN9 use the same 'write' message, so factor out the same
code to exa_wm_write.g7i in this patch

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>

VC1: No overlap smoothing if the overlap flag is 0

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>

Disable upper bound check for decoding on BDW+

We saw pixel corruption in stress test with the wrong up-bound address.
For simplicity, disable upper bound check in the driver.

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Tested-by: Zhao Yakui <yakui.zhao@intel.com>
Reviewed-by: Sean V Kelley <sean.v.kelley@intel.com>

Use VMask instead of DMask in 3DSTATE_PS on GEN8+

Mesa uses VMask instead of DMask in 3DSTATE_PS.
  "Initialize the execution mask with VMask.  Otherwise, derivatives are
  incorrect for subspans where some of the pixels are unlit.  We believe
  the bit just didn't take effect in previous generations."

To avoid potential issue, use VMask in libva too.

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Reviewed-by: Sean V Kelley <sean.v.kelley@intel.com>

add missing include

Without this building without HAVE_VA_X11 fails with:
[...]
In file included from i965_drv_video.c:37:0:
i965_output_wayland.h:31:26: error: unknown type name 'VADriverContextP'
i965_output_wayland.h:34:31: error: unknown type name 'VADriverContextP'
i965_drv_video.c:5243:9: error: 'i965_output_wayland_init' undeclared here (not in a function)
i965_drv_video.c:5244:9: error: 'i965_output_wayland_terminate' undeclared here (not in a function)
[....]

Fix the GetConfigAttributes() for JPEGBaseline profile

Advertise the PACKED_RAW_DATA header support for JPEGEncoding.

Update license header for some shaders

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

BDW+: disable SGVS

BDW introduces separate packets for controlling instancing and system
generated values (eg vertex id and instance id).  We don't use
instancing, but still need to disable sgvs to avoid undefined behaviour
when some other driver (mesa) uses it.

Signed-off-by: Joe Konno <joe.konno@intel.com>
Signed-off-by: Kristian Høgsberg <krh@bitplanet.net>
Reviewed-by: Daniel Charles <daniel.charles@intel.com>
(cherry picked from commit 4caa37b895d39af754e2b2307879752130731f56)

i965 render: Explicitly disable instancing for vertex elements on Skylake

We really should guarantee that vertex elements non-instanced,
especially since we don't use our own hardware context - other
programs can change this on us.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Acked-by: Sean V Kelley <seanvk@posteo.de>
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit 96278cab8cbc029d72ef2aa3b4728a722eceba1f)

i965 render: Explicitly disable instancing for vertex elements on Gen8

We really should guarantee that vertex elements non-instanced,
especially since we don't use our own hardware context - other
programs can change this on us.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Acked-by: Sean V Kelley <seanvk@posteo.de>
(cherry picked from commit e079591f2ed4dbfaad309c95aa1a1ddb00eb80e8)

Fix segmentation fault caused by 4711da5

The corresponding pointer is NULL if VAProcFilterSkinToneEnhancement
isn't enabled in the pipeline

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit 8596cb68e85a5708e05b9b49510ca52dd3d0a468)

STDE Optimization: Added code to support optimization of Skin Tone Detection and Enhancement for stde factors 0(default), 3, 6, and 9.

(cherry picked from commit 4711da55046cb9bd3fa26065869d8c3bcfc9f39a)

Fix the broken attribute setting of i965_GetConfigAttrib

Now the some attributes are incorrectly reported. It is not supported while
VA_ATTRIB_NOT_SUPPORTED flag is not returned.

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Reviewed-by: Xiang Haihao <haihao.xiang@intel.com>
(cherry picked from commit dd23d450f5b2b70c36a0053a0892c0ff1e32e8cf)

Fix scaling of NV12 surfaces when no output_region is set

The code block filling destination surface target rectangle is too early
and incorrectly fills input surface width & height as final target width
and height. This results in improper scaling of the source surface.

Move the destination code block further below the point when actual
target surface is referenced and its width and height are known and use
them to fill the destination rectangle.

Signed-off-by: Antti Seppälä <a.seppala@gmail.com>

Fix forward reference requirement for Bob deinterlacing in intel-driver

Remove stray semicolon to fix condition check for forward reference surface
count in vaQueryVideoProcPipelineCaps.

This fix allows the pipeline query to return 0 forward references for Bob
deinterlacer.

Signed-off-by: Antti Seppälä <a.seppala@gmail.com>

1.5.1.pre1 for development

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit 939ef3f0031ec404a32e926471e2961bb2079fde)

Intel driver 1.5.0

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit 83b833a19ef72b07e8a103d8e8c59fcb779f2ff6)