OSDN Git Service
Pengfei Qu [Thu, 22 Oct 2015 08:55:04 +0000 (04:55 -0400)]
VP9 HWDec:change the name vp8_xxx to vpx_xxx and share it with VP9
Signed-off-by: Pengfei Qu <Pengfei.Qu@intel.com>
Signed-off-by: Sirisha Muppavarapu <sirisha.muppavarapu@intel.com>
Signed-off-by: peng.chen <peng.c.chen@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
Pengfei Qu [Thu, 22 Oct 2015 08:51:38 +0000 (04:51 -0400)]
VP9 HWDec: add VP9 HW decoding pipeline
v2: fix sengment feature in inter frame
v3: update the prob buffer logic
v4: fix intra-only feature issue
Signed-off-by: Pengfei Qu <Pengfei.Qu@intel.com>
Signed-off-by: Sirisha Muppavarapu <sirisha.muppavarapu@intel.com>
Signed-off-by: peng.chen <peng.c.chen@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
Pengfei Qu [Thu, 22 Oct 2015 08:40:51 +0000 (04:40 -0400)]
VP9 HWDec:add vp9 parameter check function before decoding
Signed-off-by: Pengfei Qu <Pengfei.Qu@intel.com>
Signed-off-by: Sirisha Muppavarapu <sirisha.muppavarapu@intel.com>
Signed-off-by: peng.chen <peng.c.chen@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
Pengfei Qu [Thu, 22 Oct 2015 08:39:07 +0000 (04:39 -0400)]
VP9 HWDec:add vp9 surface structure and related parameter
Signed-off-by: Pengfei Qu <Pengfei.Qu@intel.com>
Signed-off-by: Sirisha Muppavarapu <sirisha.muppavarapu@intel.com>
Signed-off-by: peng.chen <peng.c.chen@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
Sirisha Muppavarapu [Thu, 22 Oct 2015 08:31:40 +0000 (04:31 -0400)]
VP9 HWDEC:add vp9 table file
Signed-off-by: Pengfei Qu <Pengfei.Qu@intel.com>
Signed-off-by: Sirisha Muppavarapu <sirisha.muppavarapu@intel.com>
Signed-off-by: peng.chen <peng.c.chen@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
Sirisha Muppavarapu [Thu, 28 May 2015 15:08:52 +0000 (08:08 -0700)]
VP9 HWDec: Added profile and entrypoint for VAProfileVP9Profile0 in the driver
v2: don't export VP9 decoding on BXT because VP9 decoding isn't enabled (Peng)
Signed-off-by: Sirisha Muppavarapu <sirisha.muppavarapu@intel.com>
Signed-off-by: Pengfei Qu <Pengfei.Qu@intel.com>
Signed-off-by: peng.chen <peng.c.chen@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
Xiang, Haihao [Tue, 7 Jul 2015 08:32:14 +0000 (16:32 +0800)]
Add support for P010 in vaPutSurface() in a X window system
Currently it converts P010 to RGBA32 because the drawable to vaPutSurface() is RGBA color
v2: P010 has interleaved UV planar (Peng)
v3: change commit log
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Signed-off-by: peng.chen <peng.c.chen@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
Cc: Emil Velikov <emil.l.velikov@gmail.com>
Peng Chen [Mon, 15 Jun 2015 14:28:22 +0000 (22:28 +0800)]
support HEVC 10bits decoding
v2: code cleanup
v3: store shift in an int to make it more readable (Emil)
Signed-off-by: Peng Chen <peng.c.chen@intel.com>
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
Cc: Emil Velikov <emil.l.velikov@gmail.com>
Sirisha Muppavarapu [Thu, 29 Jan 2015 14:41:04 +0000 (06:41 -0800)]
Initial support for Broxton in the intel-driver
Added PCIIDs, Device info and the relevant code.
Signed-off-by: Sirisha Muppavarapu <sirisha.muppavarapu@intel.com>
[Peng: disable MPEG-2 encoding on BXT]
Signed-off-by: peng.chen <peng.c.chen@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
Lim Siew Hoon [Wed, 18 Nov 2015 08:56:22 +0000 (16:56 +0800)]
Fix for ring hung issue in SNB platform.
This issue is trigger by the fixed from
commit id:
7deaf55d3f927e32e0b2280601dae106c7b9e3d8
use to fix segmentation fault issue by checking dst_object->bo
is NULL and VA_STATUS_ERROR_UNIMPLEMENTED in i965_proc_picture_fast
function will causing GPU hung in ring.
By disable the PPGTT, no more GPU hang issue.
If return VA_STATUS_ERROR_INVALID_SURFACE for checking only
dst_obj->bo is NULL, it is working fine without purposely to
disable PPGTT.
End up decided to use VA_STATUS_ERROR_INVALID_SURFACE instead
VA_STATUS_ERROR_UNIMPLEMENTED for dst_obj->bo checking in
i965_proc_picture_fast.
Signed-off-by: Lim Siew Hoon <siew.hoon.lim@intel.com>
Reviewed-by: Kelley, Sean V" <sean.v.kelley@intel.com>
Pengfei Qu [Mon, 9 Nov 2015 04:04:27 +0000 (12:04 +0800)]
HEVC ENC:add CBR feature.
Signed-off-by: Pengfei Qu <pengfei.qu@intel.com>
[Haihao: Fix build warning]
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Xiang, Haihao [Mon, 2 Nov 2015 01:52:58 +0000 (09:52 +0800)]
Update README
It adds platform definitions and the supported codecs, fixes the required
libva version.
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Lim Siew Hoon [Mon, 16 Nov 2015 06:16:35 +0000 (14:16 +0800)]
Fix correct coordinate dst_rect->x with 4 alignment
Signed-off-by: Lim Siew Hoon <siew.hoon.lim@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
Lim Siew Hoon [Fri, 13 Nov 2015 13:05:33 +0000 (21:05 +0800)]
Fix the segmentation fault causing by NULL buffer object
Add if checking for dst_obj_surface->bo is NULL in
i965_proc_picture_fast function.
https://bugs.freedesktop.org/show_bug.cgi?id=92811
Signed-off-by: Lim Siew Hoon <siew.hoon.lim@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
Lionel Landwerlin [Tue, 3 Nov 2015 14:50:21 +0000 (14:50 +0000)]
shaders: vpp: gen8/9: use yuv to rgb matrix to compute rgb values
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
Reviewed-by: Sean V Kelley <seanvk@posteo.de>
Lionel Landwerlin [Thu, 22 Oct 2015 15:27:15 +0000 (16:27 +0100)]
shaders: vpp: gen7: use yuv to rgb matrix to compute rgb values
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
Reviewed-by: Sean V Kelley <seanvk@posteo.de>
Lionel Landwerlin [Tue, 3 Nov 2015 15:05:26 +0000 (15:05 +0000)]
vpp: gen8/9: send yuv to rgb coefs
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
Reviewed-by: Sean V Kelley <seanvk@posteo.de>
Lionel Landwerlin [Thu, 22 Oct 2015 15:13:36 +0000 (16:13 +0100)]
vpp: gen7: send yuv to rgb coefs
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
Reviewed-by: Sean V Kelley <seanvk@posteo.de>
Lionel Landwerlin [Tue, 3 Nov 2015 15:10:08 +0000 (15:10 +0000)]
vpp: gen7/8/9: add yuv to rgb conversion matrix to registers send to EUs
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
Reviewed-by: Sean V Kelley <seanvk@posteo.de>
Lionel Landwerlin [Thu, 22 Oct 2015 14:50:49 +0000 (15:50 +0100)]
render: extract out yuv->rgb coefs
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
Reviewed-by: Sean V Kelley <seanvk@posteo.de>
Julien Isorce [Tue, 13 Oct 2015 08:10:46 +0000 (09:10 +0100)]
vaDeriveImage: properly set VAImage format
This is required to allow gstvaapipostproc
outputing RGBx va surfaces.
https://bugs.freedesktop.org/show_bug.cgi?id=92088
Signed-off-by: Julien Isorce <j.isorce@samsung.com>
Xiang, Haihao [Tue, 13 Oct 2015 05:45:39 +0000 (13:45 +0800)]
Fix YUY2/UYVY info
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Sreerenj Balachandran [Mon, 12 Oct 2015 08:38:00 +0000 (11:38 +0300)]
vpp/gen9: Use 48 bit address relocation for STATE_BASE_ADDRESS
Use 48 bit address relocation for Surface state address, Dynamic state address
and Instruction base address.
Signed-off-by: Sreerenj Balachandran <sreerenj.balachandran@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
Sreerenj Balachandran [Mon, 12 Oct 2015 08:37:59 +0000 (11:37 +0300)]
vpp/gen8: Use 48 bit address relocation for STATE_BASE_ADDRESS
Use 48 bit address relocation for Surface state address, Dynamic state address
and Instruction base address.
Signed-off-by: Sreerenj Balachandran <sreerenj.balachandran@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
Sreerenj Balachandran [Mon, 12 Oct 2015 08:37:58 +0000 (11:37 +0300)]
intel_batchbuffer: Add utility fuction for supporting 48-bit address relocations in Gen8+
There are Gen8+ instruction which requires 48bit address relocation
(eg: Surface State Address in STATE_BASE_ADDRESS instruction). Add the batchbuffer
utility funcation for the relocation based on Mesa's batch buffer implementation.
Signed-off-by: Sreerenj Balachandran <sreerenj.balachandran@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
Sreerenj Balachandran [Mon, 12 Oct 2015 08:37:57 +0000 (11:37 +0300)]
vpp/gen9: Fix the incorrect loading of kernel module.
The correct i965_kernel interace index for pp_pl3_load_save_pl3_gen9 is PP_PL3_LOAD_SAVE_PL3.
Signed-off-by: Sreerenj Balachandran <sreerenj.balachandran@intel.com>
Sreerenj Balachandran [Mon, 12 Oct 2015 08:37:56 +0000 (11:37 +0300)]
vpp/gen8: Fix the incorrect loading of kernel module.
The correct i965 kernel interface index for pp_pl3_load_save_pl3_gen8 is PP_PL3_LOAD_SAVE_PL3.
Signed-off-by: Sreerenj Balachandran <sreerenj.balachandran@intel.com>
Lim Siew Hoon [Wed, 30 Sep 2015 11:07:13 +0000 (19:07 +0800)]
Avoid NULL ptr deref in i965_decoder_wrapper_picture()
Signed-off-by: Lim Siew Hoon <siew.hoon.lim@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
Xiang, Haihao [Mon, 28 Sep 2015 08:45:38 +0000 (16:45 +0800)]
AVS: avs has more phases for sharp filter on SKL
It fixes the broken VA_FILTER_SCALING_HQ on SKL
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Reviewed-by: Zhao, Yakui <yakui.zhao@intel.com>
Qu,Pengfei [Tue, 29 Sep 2015 03:29:30 +0000 (11:29 +0800)]
FIX: Follow user's setting on filter flag
Signed-off-by: Qu,Pengfei <Pengfei.Qu@intel.com>
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Xiang, Haihao [Fri, 25 Sep 2015 01:06:19 +0000 (09:06 +0800)]
Use the right function to release allocated resources
Otherwise it might result in memory leak
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Tested-by: Liu, YangbinX <yangbinx.liu@intel.com>
Xiang, Haihao [Wed, 23 Sep 2015 05:40:41 +0000 (13:40 +0800)]
Encode: Update all used flag register channels in VME shader on GEN8+
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Reviewed-by: Zhao, Yakui <yakui.zhao@intel.com>
Xiang, Haihao [Mon, 21 Sep 2015 00:54:11 +0000 (08:54 +0800)]
Make sure the pointer is valid before dereferencing it
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Reviewed-by: Zhao, Yakui <yakui.zhao@intel.com>
Xiang, Haihao [Fri, 18 Sep 2015 04:12:03 +0000 (12:12 +0800)]
Update the NEWS
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Xiang, Haihao [Tue, 8 Sep 2015 01:18:14 +0000 (09:18 +0800)]
Silence compiler warning
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Xiang, Haihao [Mon, 7 Sep 2015 07:21:49 +0000 (15:21 +0800)]
Check pointer returned from calloc()
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Xiang, Haihao [Mon, 7 Sep 2015 01:29:38 +0000 (09:29 +0800)]
Check pointer against NULL first before dereferencing it
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Reviewed-by: Zhao, Yakui <yakui.zhao@intel.com>
Sean V Kelley [Fri, 11 Sep 2015 01:21:22 +0000 (18:21 -0700)]
i965: Remove duplicate check on platform in hybrid codec proxy
Check is made in hybrid driver for respectve platforms.
Signed-off-by: Sean V Kelley <seanvk@posteo.de>
Xiang, Haihao [Thu, 10 Sep 2015 02:57:32 +0000 (10:57 +0800)]
1.6.2.pre1 for development
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Xiang, Haihao [Thu, 10 Sep 2015 02:56:28 +0000 (10:56 +0800)]
Merge branch 'v1.6-branch' into fdo--master
Conflicts:
configure.ac
Xiang, Haihao [Wed, 9 Sep 2015 08:51:36 +0000 (16:51 +0800)]
libva-intel-driver 1.6.1
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Lim Siew Hoon [Wed, 9 Sep 2015 02:10:58 +0000 (10:10 +0800)]
Fix if issue of SLICE_TYPE_I statement in vme_pipeline_programing
Signed-off-by: Lim Siew Hoon <siew.hoon.lim@intel.com>
Reviewed-by: Zhao, Yakui <yakui.zhao@intel.com>
Xiang, Haihao [Wed, 9 Sep 2015 04:19:21 +0000 (12:19 +0800)]
Replace --enable-wrapper with --enable-hybrid-codec
hybrid-codec is more meaningful.
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit
2d42512bd6b7382c5effe21a5f9999742d98db88)
Xiang, Haihao [Wed, 9 Sep 2015 04:19:21 +0000 (12:19 +0800)]
Replace --enable-wrapper with --enable-hybrid-codec
hybrid-codec is more meaningful.
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
Xiang, Haihao [Sun, 6 Sep 2015 07:49:46 +0000 (15:49 +0800)]
Update NEWS
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Sameer Kibey [Sun, 6 Sep 2015 01:39:16 +0000 (09:39 +0800)]
Remove file access check in loading wrapper
Remove the check for file access in the wrapper.
Otherwise this call will fail on Chrome OS with sandboxing
enabled.
Signed-off-by: Sameer Kibey <sameer.kibey@intel.com>
Signed-off-by: Sean V Kelley <seanvk@posteo.de>
(cherry picked from commit
b63c28904e2568c0c50b5047b5a90d32d02b9fca)
Zhao Yakui [Sun, 6 Sep 2015 01:39:15 +0000 (09:39 +0800)]
Export the profile/entrypoints for the wrapped backend drivers
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Sean V Kelley <seanvk@posteo.de>
(cherry picked from commit
ad7b5a0233d8ead9a2ca141f3fa75f4bda1045c1)
Zhao Yakui [Sun, 6 Sep 2015 01:39:14 +0000 (09:39 +0800)]
Call the wrapped backend driver for the given decoding
The wrapped vaBeginPicture/vaRenderPicture/vaEndPicture are called.
Currently the wrapper is for the VP9 decoding.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Sean V Kelley <seanvk@posteo.de>
(cherry picked from commit
12a09a2efe213af350f29f342da6842d66dab63e)
Zhao Yakui [Sun, 6 Sep 2015 01:39:13 +0000 (09:39 +0800)]
Add the support of wrapper the VA Buffer to the backend driver
The obj_context is passed when calling vaCreateBuffer, which can be
used to check whether it belongs to the wrapped context. If one VA buffer
belongs to wrapped buffer, the callback function of backend driver will
be called for the corresponding VA buffer.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Sean V Kelley <seanvk@posteo.de>
(cherry picked from commit
f75ddc58b16cb93e85e737792c431415288cafb1)
Zhao Yakui [Sun, 6 Sep 2015 01:39:12 +0000 (09:39 +0800)]
Add one function to wrapper VA surface into backend driver
Currently only NV12 surface is supported as the decoding uses the NV12 format.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Sameer Kibey <sameer.kibey@intel.com>
Signed-off-by: Sean V Kelley <seanvk@posteo.de>
(cherry picked from commit
44e3ec2a98e99ccd839bd05cba45f0b8c9afe437)
Zhao Yakui [Sun, 6 Sep 2015 01:39:11 +0000 (09:39 +0800)]
Call the wrapped function to free wrapped surface/buffer
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Sean V Kelley <seanvk@posteo.de>
(cherry picked from commit
464ebff117df40b04c4f8da560c311e470d5a112)
Zhao Yakui [Sun, 6 Sep 2015 01:39:10 +0000 (09:39 +0800)]
Call wrapped driver to handle config/context related with VP9 decoding
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Sean V Kelley <seanvk@posteo.de>
(cherry picked from commit
815d17f944b13ebc1eca62572dbec0c49ea01999)
Zhao Yakui [Sun, 6 Sep 2015 01:39:09 +0000 (09:39 +0800)]
Wrapper the DriverContextP of backend driver
The default directory of backend driver is used.(LIBVA_DRIVERS_PATH).
Only when the backend driver exists, the wrapper is initialized and added.
Otherwise it won't be initialized.
And the option of "enable-wrapper" is added, which is used to determine whether
the wrapper of backend driver is supported.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Sean V Kelley <seanvk@posteo.de>
(cherry picked from commit
3bbaff8b4180d2510c8c77d9e4d19bf84853f31a)
Zhao Yakui [Sun, 6 Sep 2015 01:39:08 +0000 (09:39 +0800)]
Add one field for Config/Context/Surface/Buffer to store the wrapper content
This is to do the preparation that the libva-intel-driver can be the user of
other backend driver.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Sean V Kelley <seanvk@posteo.de>
(cherry picked from commit
2f254ecd154d59553a1e3949f66161bade6ca2b7)
Lim Siew Hoon [Wed, 26 Aug 2015 11:43:34 +0000 (19:43 +0800)]
Clean up unused local variable 'i965'
Signed-off-by: Lim Siew Hoon <siew.hoon.lim@intel.com>
(cherry picked from commit
020b341cf959c7ae10a233972fcbe29c47d86e68)
Lim Siew Hoon [Wed, 26 Aug 2015 09:20:34 +0000 (17:20 +0800)]
Clean up the duplicate extra two line of code
Signed-off-by: Lim Siew Hoon <siew.hoon.lim@intel.com>
(cherry picked from commit
cec8f05596a4bf11070fffeec1b4c7e5b7d44323)
Lim Siew Hoon [Wed, 26 Aug 2015 08:27:59 +0000 (16:27 +0800)]
Fix if statement checking sampling factor for YUV422V_4Y
Signed-off-by: Lim Siew Hoon <siew.hoon.lim@intel.com>
(cherry picked from commit
d7c2ba36a1bce5d4bd2e40fd161eae160946bd76)
Lim Siew Hoon [Wed, 26 Aug 2015 04:31:21 +0000 (12:31 +0800)]
Fix suspiciously placed semicolon at if statement.
Signed-off-by: Lim Siew Hoon <siew.hoon.lim@intel.com>
(cherry picked from commit
eb06298ecadf53a752abbb99208a37bda3e90c30)
Lim Siew Hoon [Wed, 26 Aug 2015 03:13:44 +0000 (11:13 +0800)]
Fix memset initialize for huff_co_table and huff_si_table array
Signed-off-by: Lim Siew Hoon <siew.hoon.lim@intel.com>
(cherry picked from commit
e12ddd354454a6491b24be202f0f56d401ab0a1b)
Lim Siew Hoon [Wed, 26 Aug 2015 02:55:18 +0000 (10:55 +0800)]
Fix memset initialize value for huff_size_table and huff_code_table
Signed-off-by: Lim Siew Hoon <siew.hoon.lim@intel.com>
(cherry picked from commit
d69346b33e31efac0e7338bc24718eef76f329da)
Lim Siew Hoon [Thu, 20 Aug 2015 10:29:09 +0000 (18:29 +0800)]
Fix klockwork critical message hit on calloc function usage
The calloc function maybe return NULL, it will causing
memory access violation if continue using NULL C structure.
Add assert function to do checking on its.
bugzilla:
https://bugs.freedesktop.org/show_bug.cgi?id=91699
Signed-off-by: Lim Siew Hoon <siew.hoon.lim@intel.com>
(cherry picked from commit
d117b9d30a00c811eb7e636c772979de846d30c8)
Xiang, Haihao [Mon, 3 Aug 2015 06:32:48 +0000 (14:32 +0800)]
encode/hevc: Check slice type first in case user passes wrong parameters to driver
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit
9170fed321780f394e8eae4a789c31120c4e957c)
Xiang, Haihao [Mon, 3 Aug 2015 06:32:34 +0000 (14:32 +0800)]
encode/hevc: Check the correct BO before unreferencing it
v2: Change the log message per Emil's suggestion
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit
1c5a8431e581eb26240c765d28f89e06138ca879)
Joel Holdsworth [Wed, 26 Aug 2015 23:21:12 +0000 (00:21 +0100)]
Added LIBVA_DRIVERS_PATH configure argument
(cherry picked from commit
c1f6411d3e0337226e157a51f8aab6a9c4e8b434)
Zhao Yakui [Tue, 25 Aug 2015 05:35:21 +0000 (13:35 +0800)]
Decoding: Export the attribute of slice mode for decoding
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Reviewed-by: Xiang Haihao <haihao.xiang@intel.com>
(cherry picked from commit
29f4234504fd99299997a3fc2f01393fb77030b7)
Zhao Yakui [Tue, 25 Aug 2015 05:35:16 +0000 (13:35 +0800)]
Remove the redundant check in vaBeginPicture call
The corresponding check is already executed when calling vaCreateConfig.
So the redundant check is removed.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Reviewed-by: Xiang Haihao <haihao.xiang@intel.com>
(cherry picked from commit
e751897e64d1d37d6c592b71a8cc7ea8bc8a0b7f)
Zhao Yakui [Tue, 25 Aug 2015 05:35:08 +0000 (13:35 +0800)]
Add one field to store the expected VA_RT_FORMAT for surface
The VA_RT_FORMAT_XXX is also used to derive the corresponding layout of
VA SURFACE.For example: VA_RT_FORMAT_YUV420/422, RGB32 and so on.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Reviewed-by: Xiang Haihao <haihao.xiang@intel.com>
(cherry picked from commit
02a2d4cb26e9a651b33a6e35c665b287155bbb83)
Zhao Yakui [Fri, 14 Aug 2015 04:13:13 +0000 (12:13 +0800)]
VPP: Initialize the uninitialed surface to avoid NULL GPU buffer
Sometimes the surface allocation is deferred. In such case it should be
checked correctly before it is using. Otherwise the corresponding
GPU buffer is NULL.
This is to fix the regression issue reported by XBMC.
Tested-by: Zaverel <zaverel@free.fr>
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit
7c47a8f74878904e0d2fbb5da966552f783d74ab)
Zhao Yakui [Fri, 14 Aug 2015 04:13:08 +0000 (12:13 +0800)]
VPP: Flush batch buffer to assure that previous GPU operation is submitted
Tested-by: Zaverel <zaverel@free.fr>
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit
67e122a4b100365d5d8bb6377c73e39f6c44abf9)
Xiang, Haihao [Thu, 23 Jul 2015 02:05:37 +0000 (10:05 +0800)]
encode/hevc: fix HEVC encode on SKL GT3
GT3 has 2 BSD rings, but HEVC commands can be only dispatched to BSD ring 0.
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit
611d8ea9d75dc026c203e3ebe53b434769d4587c)
Zhao Yakui [Wed, 15 Jul 2015 21:16:41 +0000 (17:16 -0400)]
VPP: Combine the CSC/Scaling together to optimize the performance
Currently it is divided into two steps when the video format is invovled.
In such case it is not efficient.
V1->V2: Follow Haihao's comment to call back the missing operation of
handle background color
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
(cherry picked from commit
4bda4667fd635df5bd27ce01353a3b6d6bf85508)
Xu Guangxin [Wed, 1 Jul 2015 01:49:11 +0000 (09:49 +0800)]
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.
(cherry picked from commit
12b3d9ef379e6159e30c2aafc040f1137cd4141a)
Xiang, Haihao [Wed, 8 Jul 2015 01:33:22 +0000 (09:33 +0800)]
1.6.1.pre1 for development
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
(cherry picked from commit
d06c17e9c0d4a11a3efaead7da63765148bf938e)
Gwenole Beauchesne [Tue, 7 Jul 2015 08:04:37 +0000 (10:04 +0200)]
vpp/avs: fix license terms (MIT).
Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
(cherry picked from commit
fb5e1cd86047e2683f169cd52cc27861447321cf)
Zhao Yakui [Mon, 6 Jul 2015 17:45:47 +0000 (01:45 +0800)]
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>
(cherry picked from commit
7bb3658212f49a6f8d3d61b0b5d161aca73c2a11)
Xiang, Haihao [Mon, 29 Jun 2015 05:05:51 +0000 (13:05 +0800)]
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>
(cherry picked from commit
8d11fd7dd7952551b2d57f900ce1f408a35616a1)
Xiang, Haihao [Mon, 29 Jun 2015 05:00:57 +0000 (13:00 +0800)]
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>
(cherry picked from commit
6ff572fb94fce0b941d56eb7c0f45d82c1672ea3)
fritsch [Wed, 24 Jun 2015 08:18:11 +0000 (10:18 +0200)]
Gen75_vpp: Don't assert on BGRA input and output
(cherry picked from commit
e49960198a4c53ad96e36e0952db514764af06f6)
Sameer Kibey [Sun, 6 Sep 2015 01:39:16 +0000 (09:39 +0800)]
Remove file access check in loading wrapper
Remove the check for file access in the wrapper.
Otherwise this call will fail on Chrome OS with sandboxing
enabled.
Signed-off-by: Sameer Kibey <sameer.kibey@intel.com>
Signed-off-by: Sean V Kelley <seanvk@posteo.de>
Zhao Yakui [Sun, 6 Sep 2015 01:39:15 +0000 (09:39 +0800)]
Export the profile/entrypoints for the wrapped backend drivers
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Sean V Kelley <seanvk@posteo.de>
Zhao Yakui [Sun, 6 Sep 2015 01:39:14 +0000 (09:39 +0800)]
Call the wrapped backend driver for the given decoding
The wrapped vaBeginPicture/vaRenderPicture/vaEndPicture are called.
Currently the wrapper is for the VP9 decoding.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Sean V Kelley <seanvk@posteo.de>
Zhao Yakui [Sun, 6 Sep 2015 01:39:13 +0000 (09:39 +0800)]
Add the support of wrapper the VA Buffer to the backend driver
The obj_context is passed when calling vaCreateBuffer, which can be
used to check whether it belongs to the wrapped context. If one VA buffer
belongs to wrapped buffer, the callback function of backend driver will
be called for the corresponding VA buffer.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Sean V Kelley <seanvk@posteo.de>
Zhao Yakui [Sun, 6 Sep 2015 01:39:12 +0000 (09:39 +0800)]
Add one function to wrapper VA surface into backend driver
Currently only NV12 surface is supported as the decoding uses the NV12 format.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Sameer Kibey <sameer.kibey@intel.com>
Signed-off-by: Sean V Kelley <seanvk@posteo.de>
Zhao Yakui [Sun, 6 Sep 2015 01:39:11 +0000 (09:39 +0800)]
Call the wrapped function to free wrapped surface/buffer
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Sean V Kelley <seanvk@posteo.de>
Zhao Yakui [Sun, 6 Sep 2015 01:39:10 +0000 (09:39 +0800)]
Call wrapped driver to handle config/context related with VP9 decoding
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Sean V Kelley <seanvk@posteo.de>
Zhao Yakui [Sun, 6 Sep 2015 01:39:09 +0000 (09:39 +0800)]
Wrapper the DriverContextP of backend driver
The default directory of backend driver is used.(LIBVA_DRIVERS_PATH).
Only when the backend driver exists, the wrapper is initialized and added.
Otherwise it won't be initialized.
And the option of "enable-wrapper" is added, which is used to determine whether
the wrapper of backend driver is supported.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Sean V Kelley <seanvk@posteo.de>
Zhao Yakui [Sun, 6 Sep 2015 01:39:08 +0000 (09:39 +0800)]
Add one field for Config/Context/Surface/Buffer to store the wrapper content
This is to do the preparation that the libva-intel-driver can be the user of
other backend driver.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Sean V Kelley <seanvk@posteo.de>
Lim Siew Hoon [Wed, 26 Aug 2015 11:43:34 +0000 (19:43 +0800)]
Clean up unused local variable 'i965'
Signed-off-by: Lim Siew Hoon <siew.hoon.lim@intel.com>
Lim Siew Hoon [Wed, 26 Aug 2015 09:20:34 +0000 (17:20 +0800)]
Clean up the duplicate extra two line of code
Signed-off-by: Lim Siew Hoon <siew.hoon.lim@intel.com>
Lim Siew Hoon [Wed, 26 Aug 2015 08:27:59 +0000 (16:27 +0800)]
Fix if statement checking sampling factor for YUV422V_4Y
Signed-off-by: Lim Siew Hoon <siew.hoon.lim@intel.com>
Lim Siew Hoon [Wed, 26 Aug 2015 04:31:21 +0000 (12:31 +0800)]
Fix suspiciously placed semicolon at if statement.
Signed-off-by: Lim Siew Hoon <siew.hoon.lim@intel.com>
Lim Siew Hoon [Wed, 26 Aug 2015 03:13:44 +0000 (11:13 +0800)]
Fix memset initialize for huff_co_table and huff_si_table array
Signed-off-by: Lim Siew Hoon <siew.hoon.lim@intel.com>
Lim Siew Hoon [Wed, 26 Aug 2015 02:55:18 +0000 (10:55 +0800)]
Fix memset initialize value for huff_size_table and huff_code_table
Signed-off-by: Lim Siew Hoon <siew.hoon.lim@intel.com>
Lim Siew Hoon [Thu, 20 Aug 2015 10:29:09 +0000 (18:29 +0800)]
Fix klockwork critical message hit on calloc function usage
The calloc function maybe return NULL, it will causing
memory access violation if continue using NULL C structure.
Add assert function to do checking on its.
bugzilla:
https://bugs.freedesktop.org/show_bug.cgi?id=91699
Signed-off-by: Lim Siew Hoon <siew.hoon.lim@intel.com>
Xiang, Haihao [Mon, 3 Aug 2015 06:32:48 +0000 (14:32 +0800)]
encode/hevc: Check slice type first in case user passes wrong parameters to driver
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Xiang, Haihao [Mon, 3 Aug 2015 06:32:34 +0000 (14:32 +0800)]
encode/hevc: Check the correct BO before unreferencing it
v2: Change the log message per Emil's suggestion
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Joel Holdsworth [Wed, 26 Aug 2015 23:21:12 +0000 (00:21 +0100)]
Added LIBVA_DRIVERS_PATH configure argument
Zhao Yakui [Tue, 25 Aug 2015 05:35:21 +0000 (13:35 +0800)]
Decoding: Export the attribute of slice mode for decoding
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Reviewed-by: Xiang Haihao <haihao.xiang@intel.com>
Zhao Yakui [Tue, 25 Aug 2015 05:35:16 +0000 (13:35 +0800)]
Remove the redundant check in vaBeginPicture call
The corresponding check is already executed when calling vaCreateConfig.
So the redundant check is removed.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Reviewed-by: Xiang Haihao <haihao.xiang@intel.com>
Zhao Yakui [Tue, 25 Aug 2015 05:35:08 +0000 (13:35 +0800)]
Add one field to store the expected VA_RT_FORMAT for surface
The VA_RT_FORMAT_XXX is also used to derive the corresponding layout of
VA SURFACE.For example: VA_RT_FORMAT_YUV420/422, RGB32 and so on.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Reviewed-by: Xiang Haihao <haihao.xiang@intel.com>
OSDN Git Service
