* Promote OCL false sharing usages L3 to WB together with 2-way
on xe2 (#249)
* Mark GO:L3 as Uncachebale on MTL (#248)
* Increment minor number for WCL support (#247)
* Enable Multi-engine compression 64KB WA (#175)
* Update XAdapater to be L3 Cacheable (#246)
* Add new BMG device IDs (#244)
* Introduce WCL support (#240)
* Correct "Use GO:L3 on OCL Usages (#242)
- fixed source line in specfile
- Update to version 22.7.2
* Add BMG Device ID (#232)
* Add BMG Device ID (#233)
* Add new BMG Device ID (#234)
* Disable compression on GMM_FORMAT_I420 format (#236)
* Handle Coherent and compressible resources (#235)
- Update to version 22.7.1:
* Use GO:L3 on OCL Usages (#231)
* Add Size Check When checking NoOptimizationPadding (#230)
* Add MOCS variable for Xe2 (#228)
* Force Tile4 instead of Linear for flipchain resources (#229)
* Fix for type incompatibilities (#227)
* Fix to resolve Media and NEO sharing issue with Xe2compression
on BMG (#226)
- Update to version 22.6.0
* Introduce PTL Support (#225)
- Update to version 22.5.5:
* Add new media compression mode to fix assertion (#224)
* Adjust BaseWidth for RGB24 format (#64)
* Enable IsCpuCacheable in Linux (#214)
* PAT MOCS Programming for test (#223)
* Make R10G10B10_XR_BIAS_A2_UNORM as presentable (#222)
* Add new DG2 Device ID (#221)
* Update reserved PATIndex (#220)
- update to 22.5.2 - changes since 22.3.20 (jira#PED-10205)
* Introduce LunarLake Support (#188)
* Remove media compressed formats (#202)
* Introduce Battlemage basic enabling support (#203)
* Introduce Battlemage full support (#205)
* Update README with new platform support info (#208)
* Fix Debug build type(#210)
* Check if BaseWidth and BaseHeight is zero (#213)
* Xe2 Caching demotion for L3XD|L4WT|NC request (#212)
- update to 22.3.20:
* fail if pTextureCalc is NULL
* add RGB format perftag
* add more device IDs
* Revert "Failing the resource allocation if the U plane offset
exceeds the HW"
- Update to version 22.3.18:
* Introduce ARL-H support (#172)
* Added new ADL-N Device IDs (#189)
* Add new Xe_HP Device IDs (#190)
* Handle out of memory case (#193)
* Correct DG2 Depth/Stencil MSAA 16x 8x Texture Allocation (#192)
* Failing the resource allocation if the U plane offset exceeds
the HW limitations (#191)
- update to 22.3.17:
* Fix possible integer overflow
* Assert and return on divide by zero (#187)
* Update DG2 device ID macro (#183)
- update to 22.3.16:
* Fix to align HAlign to 128bytes for Linear Packed YUV format
* Add PVC device ID
* Guard FormatTable to be within bound
- This includes latest version of one of the components needed for
Video (processing) hardware support on Intel GPUs (bsc#1217770)
- update to 22.3.14:
* Add more DG2 Device IDs
- update to 22.3.13:
* Fix Multiadapter ULT issue reproducibilty
* Add new PVC Device ID
- update to 22.3.11:
* Add PVC Device ID
* Add more MTL Device IDs
* Replace WaDefaultTile4 with a feature flag
* Multi-threaded fixes for GMM Context info
* Use GmmResGetAuxSurfaceOffset64
* Fix UnifiedAuxSurvface checks to allow HiZ w/o CCS
* add ARL-S support
* Add device IDs for RPL
- update to 22.3.5:
* Change Scratch L1 Caching to WB
* Update Cachepolicy for MTL
* Add forceful fallback to support Tile4 using an Ftr flag
* Enable L3 for VP_INPUT_PICTURE_RENDER usage
- update to 22.3.4:
* remove unused tilewidth, mip4height and mip5height variables
- update to 22.3.3:
* Initialize Filedescriptor
* Disable MultiAdapter ULT testcases
- update to 22.3.2:
* Fix tiling selection
* Align SURFTYPE_3D surfaces
* Correct max surface size
- enable building for x86_64 micro-level architectures
- update to 22.3.1:
* Fix memory leak Destroy allocated resources for ULT
- needed for jira#PED-1174 (Video decoding/encoding support
(VA-API, ...) for Intel GPUs is outside of Mesa)
- Update to version 22.3.0:
* Support for default build types
* Add ATS-M Device Ids
* Adding more dg2 device Ids
* Fixing XE_HPC macro usage for Cache Policy settings
* Fix QPtich calculations for CCS
* PVC PAT table implementations
* Initialize NumPATRegisters
* Add PVC Device IDs
* Fix GetPrivatePATEntry API
* Introducing MTL Support
- No code changes
- Update to version 22.1.4 was part of Intel oneVPL GPU Runtime
2022Q2 Release 22.4.4
- Update to version 22.1.4:
* No upstream changelog available
- Update to version 22.1.2:
* No upstream changelog available
- Update to version 22.0.1:
* No upstream changelog available.
- Bump somajor define to 12 and in baselibs.conf following upstream
so bump.
- Replace gcc-c++ with generic c++_compiler BuildRequires.
- Update to version 21.3.3:
* No upstream changelog available.
- Allow building on rpmbuild --target=i686.
- Drop R/TM, too hard to get right.
- mention baselibs.conf as source in specfile
- Fix 32bit build (dependency for Steam-in-home streaming)
- fixed JIRA number in previous changelog
- covers jira#SLE/SLE-19361
- Update to 21.2.2
* No upstream changelog available.
- update to 20.3.2
* baseAlignment can be greater than 64KB and needs to be aligned to 64KB
* DG1 HiZ H/V Align modification
* Set MOCS0 with WB cache value
* DG1 Placed resource alignment issues Dx12/Vulkan
* Remove '-Werror=implicit-function-declaration'
* SyncInfoLin: Initialize members
* Do not use memset for clearing an object of non-trivial type
* Added the ADL-S device ID's and phyAddr support
- update to 20.2.5
* No upstream changelog available (needed for jira#SLE/SLE-12712)
- update to 20.2.3:
* No upstream changelog available.
- Update to version 20.1.1:
* No upstream changelog available.
- Update to version 19.4.1:
* No upstream changelog available
- This release fullfills requirements for jsc#SLE-8726
- Updated to version 19.3.3
* Intel Graphics Memory Management Library 2019 Q3 Release 3
- Bump soname to 11
- Updated to version 19.2.4
* adds support for Icelake, EHL, Cometlake, Lakefield
* bug fixes
- Update to version 19.2.3:
* Memory leak fix.
* Crash fix.
* More device support.
- Bump soname to 9 following upstream changes.
- Updated to version 18.4.1:
* Cleanup, more device support
- Changed the subpackage name to follow libigdgmm.so.5
- Initial version: 18.3.0
OBS-URL: https://build.opensuse.org/package/show/X11:XOrg/gmmlib?expand=0&rev=80
* Use GO:L3 on OCL Usages (#231)
* Add Size Check When checking NoOptimizationPadding (#230)
* Add MOCS variable for Xe2 (#228)
* Force Tile4 instead of Linear for flipchain resources (#229)
* Fix for type incompatibilities (#227)
* Fix to resolve Media and NEO sharing issue with Xe2compression
on BMG (#226)
OBS-URL: https://build.opensuse.org/package/show/X11:XOrg/gmmlib?expand=0&rev=75
* Introduce LunarLake Support (#188)
* Remove media compressed formats (#202)
* Introduce Battlemage basic enabling support (#203)
* Introduce Battlemage full support (#205)
* Update README with new platform support info (#208)
* Fix Debug build type(#210)
* Check if BaseWidth and BaseHeight is zero (#213)
* Xe2 Caching demotion for L3XD|L4WT|NC request (#212)
OBS-URL: https://build.opensuse.org/package/show/X11:XOrg/gmmlib?expand=0&rev=69
- update to 22.3.11:
* Add PVC Device ID
* Add more MTL Device IDs
* Replace WaDefaultTile4 with a feature flag
* Multi-threaded fixes for GMM Context info
* Use GmmResGetAuxSurfaceOffset64
* Fix UnifiedAuxSurvface checks to allow HiZ w/o CCS
* add ARL-S support
* Add device IDs for RPL
* Enable L3 for VP_INPUT_PICTURE_RENDER usage
* Fix memory leak Destroy allocated resources for ULT
- needed for jira#PED-1174 (Video decoding/encoding support
- mention baselibs.conf as source in specfile
- fixed JIRA number in previous changelog
* No upstream changelog available.
* Intel Graphics Memory Management Library 2019 Q3 Release 3
- Updated to version 19.2.4
OBS-URL: https://build.opensuse.org/request/show/1110902
OBS-URL: https://build.opensuse.org/package/show/X11:XOrg/gmmlib?expand=0&rev=52
