- update to 23.1.1:
* Add VA HEVC block sizes and features caps
* Remove VP dependency for codec softlet lib on win/Linux
- update to 22.6.6:
* Revert "vaGetImage optimization on D-card."
- updated Supplements
- update to intel-media-22.5.4:
* Official quarterly release including some enhancements and key fixes.
* Alchemist(DG2)/ATS-M New Features and Enhancement
* Enhanced HEVC & AVC decode error handling robustness.
* Fixed VP9 dynamic resolution change decode random corruption issue.
* Enabled Memory Decompression for ADLS and ADLN.
* Fixed MPEG2 decode crash issue.
* Fixed AV1 decode film grain hang issue.
* Fixed color fill corruption issue.
* Fixed first VPP operation color artifacts.
* Enhanced I420 and UYVY format support in creating surface and derive
image.
* Fixed aux table l2 page fault
- update to 22.6.4:
* Alchemist(DG2)/ATS-M New Features and Enhancement
* Enhanced HEVC & AVC decode error handling robustness.
* Fixed VP9 dynamic resolution change decode random corruption issue.
* Enabled Memory Decompression for ADLS and ADLN.
* Fixed MPEG2 decode crash issue.
* Fixed AV1 decode film grain hang issue.
* Fixed color fill corruption issue.
* Fixed first VPP operation color artifacts.
* Enhanced I420 and UYVY format support in creating surface and derive
image.
* Fixed aux table l2 page fault
- needed for jira#PED-1174 (Video decoding/encoding support
(VA-API, ...) for Intel GPUs is outside of Mesa)
- Update to version 2.6.1
* Revert "[Decode] Legacy MI interface removal"
- specfile cleanup
- updated Supplements
- Update to version 2.6.0:
* Revert "[Decode] Virtual Node Assign Policy Optimization"
- Code changes from version 2.5.4:
* Enabled Memory Decompression for ADLS and ADLN.
* Fixed MPEG2 decode crash issue.
* Fixed AV1 decode film grain hang issue.
* Fixed color fill corruption issue.
* Fixed first VPP operation color artifacts.
* Enhanced I420 and UYVY format support in creating surface and derive image.
* Fixed aux table l2 page fault
- Remove u_libva-2.16.0.patch
- adding _constraints in the hope to reserve enough disk space;
trying with 7GB for now ...
- u_libva-2.16.0.patch
* fixes build against libva 2.16.0
* culprit: https://github.com/intel/libva/commit/8682f9e30f2fabf2ccc6f7609db035ed1af44703
- No code changes
- Update to version 22.4.4 was part of Intel oneVPL GPU Runtime
2022Q2 Release 22.4.4
- updated supplements.inc
- Update to version 22.4.4:
* Enabled HDR10 and HVS support
* Added RPL-P platform enabling
* Added HDR10 capability report
- disabling Werror from build no longer needed; therefore commented
out this sed line for now ...
- Update to version 22.4.2:
* [Encode] AVC RC mode - Implement abs QP map (MBQP) and CQP QP
- Enable abs QP map mode caps
- Implement programming for abs QP map mode
- removed no longer needed Werror-initialize-in-right-order.patch
- Update to version 22.1.1:
* New Features and Enhancement:
- Enabled Alchemist/ATS-M platform decoding and video
processing features
- Added ADL-N platform support
- Enhanced AV1 decoding robustness for error clips handling
- Added vaCopy caps reporting
- Enabled GPU copy for small resolution in vaMap/unMap
- Optimized GetImage perf for NV12 format
- Added HEVC sub-features caps reporting
- Improved compatibility by disabling compression when creating
surface
- Improved debuggability by enabling OCA support
* Bugs fixed:
- Fixed multiple layer composition corruption issue
- Fixed OCA stability issue in multi-thread scenario
- Fixed render copy mem leak
- Update to version 21.3.5:
* Enabled vaCopy by GPU HW
* Added 0YUV decode output format support
- mention baselibs.conf as source in specfile
- Fix 32bit build (dependency for Steam in-home streaming hardware decode)
- fixed JIRA number in a previous changelog
- Update to 21.3.4
* Fix regression:libva conformance failures since github issue#1229
- covers jira#SLE/SLE-19361
- Update to 21.3.3:
* Supported platforms:
BDW/SKL/APL(BXT)/KBL/CFL/WHL/CML/ICL/EHL/JSL/TGL/RKL/DG1/SG1/ADL-S/ADL-P
* Extensive new features and enhancements, see README.md
- Move generated supplements list to supplements.inc
- update to 20.3.0:
* Enabled AV1 4:2:0 8bit and 10bit decoding for both driver and MSDK, up to 8K video decoding and 16K still picture decoding.
* Added Y212 and Y412 color format support for 12 bit decoding.
* Added several RePAK related changes for VP9 performance improvements including Pak-only RePAK and adaptive-RePAK
* Reconstructed Surface Variant formats configuration is now handled by the media-driver
* Added HuC Error Status Report improvements for VP9 VDEnc encoding
* Added scaling interpolation method(nearest, bilinear, advanced)
* Added resource usage type to specify cache policy usage
* Enabled SFC bilinear scaling.
- Disable Werror from build, which avoids having to create build fixes
patches for every release
- remove u_20.1.1-build-fixes.patch, u_20.2.0-build-fixes.patch (obsolete)

- Update to version 20.2.0
* Official quarterly release including some enhancements and key
fixes. (needed for jira#SLE/SLE-12712)
- u_20.2.0-build-fixes.patch
* fixed more -Werror build errors
- updated Supplements lines in specfile
- moved profiles for setting env variable LIBVA_DRIVER_NAME to
/usr/etc/profiles.d on TW (boo#1173006)
- u_20.1.1-build-fixes.patch
* fixed remaining -Werror build errors
- Update to version 20.1.1:
* Enabled HEVC SCC Decoding.
* VP SFC driver refactor for modularity/structure improvement.
* Enhanced media buffer and 2D surface copy to improve
robustness.
* Enhanced Tile mode(Tile64, TileX, Tile4) support.
* MOS driver refactor for modularity/robustness enhancement.
* HEVC VDEnc Encoding
- Added AYUV/ARGB10/Y410 format support for increased MV
predictor range of TU1.
- Enabled DirtyROI for better performance of static region.
* VP9 VDEnc Encoding: Enabled scalability for better performance
in 5k+ and range-extension real-time encoding.
* Various bugfixes.
- For changes from previous releases, please see the included
README.md
- Drop patches fixed upstream:
* Werror-implicit-function-not-valid-for-C++.patch
* U_Encode-Add-some-missing-device-IDs-for-CML-and-ICL.patch
- Rebase Werror-initialize-in-right-order.patch
- generate-supplements.sh/specfile: fixed supplements lines
(boo#1164851)
- U_Encode-Add-some-missing-device-IDs-for-CML-and-ICL.patch
* add support for support AML-Y (bsc#1160967)
- This release fullfills requirements for jsc#SLE-8838
- Update to version 19.2.1
* Fixed known GPU hang issue when encoding 444 8/10bit in ICL
issue#677
* Fixed several compile warnings.
- Werror-implicit-function-not-valid-for-C++.patch,
Werror-initialize-in-right-order.patch
* needed as buildfixes
- Update to version 19.2.0:
* Release pVEInterf to fix the mem leak issue.
* Fix memory leak during 3D surface allocation.
* Move the lock earlier to make surface heap status aligned when
surface is destroyd.
* [VP] UMD_OCA Implement:
- Add UMD_OCA support in mos.
- Add UMD_OCA support for VP RCS case (No CM based) and
Vebox/SFC case.
- It is disabled by default.
* [VP] Remove dupliatcated defines on composition.
* Remove duplicated defines and revert part codes.
* Full support of Gen9 CML kernel integration.
* Stay the same compile option as Gen9.
* Readme: update feature list for default and open source build
* [Encode]:
- Fix segment fault issue in Jpeg encode cases on Linux.
- Fix Dynamic Scaling Media Reset Issue.
- Update modalias supplements.
- Updated to version 18.4.1:
* Supported Platforms:
- BDW/SKL/APL(BXT)/KBL/CFL/WHL/CNL/ICL
* Improvements:
- WHL: Added WHL support and enabled its VP8 encoding.
- ICL: Added ICL open source compute shaders for video processing.
Enabled ICL HDR10 Tone Mapping using close source compute shaders.
- DRM uAPI relative code is refactored to better comply to i915
Open Source Userspace Requirement.
- Added X11 build option detection.
- Enabled software swizzling support to improve surface reading
performance of vaGetImage.
- Added ARGB,Y410 image format support.
- Added versioning information to CMRT library.
- BRC enhancement:
AVC hardware encoding: added QVBR support.
AVC shader encoding: added AVBR and max frame size BRC support.
VP8 encoding: corrected fps calculation error.
* Fixed issues:
- GStreamer/gstreamer-vaapi integration:
Fixed the issue about KBL HEVC P010_10LE encoding kernel state
initialization failure for CODECHAL_HEVC_MBENC_PENC.
Fixed the issue about aspect ratio being unable to retain in
video processsing PutSurface for vaapisink display.
- Corrected RT format reporting for HEVC main10 profile.
- Corrected caps reporting of max width/height for VP9 decoding
from 2k (2048x2048) to 8k (8192x8192).
- Add sub-packages libigfxcmrt7 and libigfxcmrt-devel
- Updated supplements accordingly
- added hardware supplements in order to install the package on
appropriate hardware (Broadwell, Skylake, Broxton/Apollo Lake,
Kaby Lake, Coffee Lake, Cannonlake, Ice Lake at this time)
- added script to generate the above information easily
- Updated to version 18.3.0:
* Adding ICL support
* Adding dynamic loading support of Gmmlib
* Better support to ffmpeg and gstreamer:
+ AVC low power encoding
+ MPEG2 encoding
+ VP8 decoding
+ VC1 decoding
+ JPEG encoding and decoding
* Adding Dockerfile and docker images
- Use shared lib of gmmlib; added buildreq accordingly
- initial package: media-driver Q2'18 release