1Mesa 24.3.0 Release Notes / 2024-11-21 2====================================== 3 4Mesa 24.3.0 is a new development release. People who are concerned 5with stability and reliability should stick with a previous release or 6wait for Mesa 24.3.1. 7 8Mesa 24.3.0 implements the OpenGL 4.6 API, but the version reported by 9glGetString(GL_VERSION) or glGetIntegerv(GL_MAJOR_VERSION) / 10glGetIntegerv(GL_MINOR_VERSION) depends on the particular driver being used. 11Some drivers don't support all the features required in OpenGL 4.6. OpenGL 124.6 is **only** available if requested at context creation. 13Compatibility contexts may report a lower version depending on each driver. 14 15Mesa 24.3.0 implements the Vulkan 1.3 API, but the version reported by 16the apiVersion property of the VkPhysicalDeviceProperties struct 17depends on the particular driver being used. 18 19SHA checksums 20------------- 21 22:: 23 24 SHA256: 97813fe65028ef21b4d4e54164563059e8408d8fee3489a2323468d198bf2efc mesa-24.3.0.tar.xz 25 SHA512: 20168ae4c278776a60d5febf53b3367cf08bffffb40ef2054821e68d7a8c37a07871d097ab17555f41a4fe716f0de7df95ad7d452b1ed57db6527838eb839ba4 mesa-24.3.0.tar.xz 26 27 28New features 29------------ 30 31- Expose Vulkan 1.3 on v3dv, both rpi4 and rpi5 32- VK_EXT_descriptor_buffer on nvk 33- VK_EXT_post_depth_coverage on nvk 34- VK_KHR_video_maintenance1 on radv 35- VK_EXT_legacy_vertex_attributes on nvk 36- GL_KHR_shader_subgroup on radeonsi 37- VK_KHR_maintenance7 on nvk 38- VK_KHR_dynamic_rendering_local_read on nvk 39- GL_ARB_timer_query on Panfrost 40- GL_EXT_disjoint_timer_query on Panfrost 41- VK_KHR_pipeline_binary on RADV 42- VK_KHR_compute_shader_derivatives on anv 43- VK_NV_compute_shader_derivatives on nvk 44- VK_KHR_compute_shader_derivatives on nvk 45- VK_KHR_compute_shader_derivatives on radv 46- VK_KHR_shader_relaxed_extended_instruction on anv, hasvk, hk, nvk, radv, tu, v3dv, lvp 47- GL_OVR_multiview and GL_OVR_multiview2 on zink 48- VK_KHR_shader_float_controls2 on radv 49- VK_KHR_shader_float_controls2 on nvk 50- VK_EXT_device_generated_commands on nvk, radv 51- VK_EXT_host_image_copy on nvk/Turing+ 52- VK_EXT_depth_clamp_control on anv, hasvk, nvk, radv 53- VK_KHR_shader_quad_control on nvk 54- GL_EXT_draw_buffers2 on etnaviv/HALTI5+ 55- GL_ARB_draw_buffers_blend on etnaviv/HALTI5+ 56- VK_KHR_fragment_shading_rate on NVK 57- GL_ARB_draw_indirect on etnaviv/HALTI5+ 58- VK_EXT_depth_clamp_zero_one on NVK 59- GL_ARB_framebuffer_no_attachments on etnaviv 60 61 62Bug fixes 63--------- 64 65- anv: Use-after-free detected by AddressSanitizer while running dEQP-VK 66- Confidential issue #11770 67- radv: DCC causes glitches in Red Dead Redemption 2 68- A5xx rendering issues with firefox 69- [ANV][Regression] Broken rendering in Flycast + Per-Pixel Alpha Sorting 70- vtn: OpTypeStruct in kernel parameters trigger assertion in glsl_types.h 71- anv: Assertion failure in \`dEQP-VK.image.extended_usage_bit_compatibility.image_format_list.s8_uint_optimal_transfer_src_bit` 72- radv: Resident Evil 6 Benchmark Tool has artifacts on 7900 XTX when DCC is enabled, game launched on 4K monitor without scaling and with FullHD settings 73- [AMD RX 6700 XT] Artifacts while upscaling games in fullscreen mode 74- Distorted pixelated graphics with Radeon RX 7900 XT with some games 75- Total War Warhammer 2 Graphical Glitch 76- Glitching artifacts in tile shaped patterns on 6700 XT, when using upscaled fullscreen game on labwc 77- anv: Page fault when using MTL simulator in dEQP-VK.ray_tracing_pipeline.data_spill.report_intersection.float32 78- mesa_cache_db.c:316:33: error: call to undeclared function 'mremap' 79- [trunk] shaders fail hard in openmw after cbfc225e2bda2c8627a4580fa3a9b63bfb7133e0 80- u_perfetto.h:33:9: error: unknown type name 'clockid_t'; did you mean 'clock_t'? 81- brw_fs_opt_copy_propagation incorrectly handles size changes of uniforms 82- nouveau paraview msaa corruption 23.1 bisected regression 83- mesa fails to build due to missing SPV_ENV_UNIVERSAL_1_6 symbol 84- Missing xshmfence dependency for X11 WSI on macOS 85- [r300g] Unigine Tropics: black stripes between center of the screen and horizon 86- [AMDGPU RDNA3] Black square artifacts in viewport renders in Blender 87- radv: test_sm67_sample_cmp_level test in vkd3d-proton fails 88- [ANV] LNL tiled corruption in background traci Blackops3-trace-dx11-1080p-high 89- [anv] Possible regression in e3814dee1ac0f90771b921a4f6f5aed10f06e8d4 90- Confidential issue #12092 91- radv: significantly less performance than amdgpu-pro in the AMD FSR2 sample 92- radv: wrong index value in radv_skip_graphics_pipeline_compile 93- Android: IMapper4 / HIDL is being deprecated 94- d3d12_video_screen.cpp undefined reference to ConvertHEVCSupportFromProfile(D3D12_VIDEO_ENCODER_PROFILE_HEVC, D3D12_VIDEO_ENCODER_CODEC_CONFIGURATION_SUPPORT_HEVC1*) 95- anv: Reduce the number of partial resolves for resource transitions with CCS_E 96- Silent Hill 2 Remake performance regression with f5805bcb 97- Disk Cache DB file descriptors are not closed on exec 98- Shader cache takes too many fds 99- Shader cache takes too many fds 100- Disk Cache DB file descriptors are not closed on exec 101- radv: \`test_instruction_msad_dxil` from vkd3d-proton fails on hawaii 102- [anv] functional regression with 02294961eeeec003fd6113c0406a80ea1a2fec72 on DG2 103- [vulkan-intel][regression] War Thunder causes GPU hang 104- RFE: rusticl: out-of-order queues 105- [nak][nouveau] - error[E0425]: cannot find value \`DRM_RS_IOCTL_NOUVEAU_EXEC` in this scope --> ../src/nouveau/compiler/nak/hw_runner.rs:193:17 106- [radeonsi] glLinkProgram terminates the application when Shader is passing Bindless Texture into function 107- gallium: crash when using images obtained from a texture with a bindless handle 108- Using DXT1 textures fails in GL ES 1, even with EXT_texture_compression_dxt1 supported 109- zink: discrepancy between reported max LOD bias and the actual allowed range 110- gallium: Crash when writing to writeonly image3D passed in via parameter 111- vulkan/queue: heap-use-after-free when enabling address sanitizer 112- r300 regression : Upside down and broken redering for Source games 113- vk_texcompress_astc_finish() may crash because astc is null 114- intel A770 dg2 silent hill 2 doesn't run 115- radv: vkGetPipelineKeyKHR global key is bugged 116- Mesa's \`intel_hang_replay` tool fails to work 117- Mesa's \`intel_hang_replay` tool fails to work 118- radv: fix dEQP-VK.texture.mipmap.3d.image_view_min_lod.base_level.* regressions on GFX6-8 119- ci: Uprev Piglit failed 120- PanVK-v10: Missing support for vkCmdDispatchIndirect 121- radv: slow packed math in nested loops due to copies 122- nvk,nak: Implement VK_KHR_shader_quad_control 123- [Turnip] Detroit Become Human gets stuck perpetually compiling a shader 124- Faaast Penguin requires force_vk_vendor set 125- [clover] linking of libMesaOpenCL.so.1.0.0 fails with LLVM/Clang 20: undefined reference to \`clang::api_notes::APINotesManager::loadCurrentModuleAPINotes(...)` 126- zink missing dep on spirv_info_h? 127- ANV: Deadlock(game) bugged invisible player/NPC models and missing textures 128- anv: dEQP-VK.mesh_shader.ext.misc.clip_geom_provoking_last and dEQP-VK.mesh_shader.ext.misc.clip_geom_and_task_shader_provoking_last fail on MTL 129- anv: dEQP-VK.mesh_shader.ext.misc.clip_geom_provoking_last and dEQP-VK.mesh_shader.ext.misc.clip_geom_and_task_shader_provoking_last fail on MTL 130- anv: dEQP-VK.subgroups.clustered.compute.subgroupclusteredmul_i64vec4_requiredsubgroupsize and friends time out due to register spilling 131- [clc] [FTBFS] [llvm20+] no matching function for DataLayout 132- building lavapipe with LLVM-19 and running any vulkan app on a Zen4 CPU shows text on terminal.. 133- PanVK-v10: Missing support for vkCmdDraw[Indexed]Indirect 134- Adreno 740v3 (Quest 3) Image Corruption 135- [regression] "anv: Change params of anv_can_fast_clear_color_view" breaks a bunch of Sparse tests on LNL 136- Update defaults for launch crash with XESS in Satisfactory 1.0 137- [bisected][zink] dEQP-GLES3.functional.fbo.multiview.samples_1 failing after feature enabled on zink 138- ACO Error: Unsupported opcode: v4 -- AMD Radeon RX 590 Series (Black Myth Wukong Benchmark Tool) 139- !31075 cause graphical artifact with Warhammer 40,000: Space Marine 2 140- PanVK-v10: VulkanCTS 1.3.9.2 - 3D Copy-and-blit tests hit GPU fault on Rock5B (RK3588) 141- [regression] paraview regressed on T860 142- PanVK-v10: Assert in cs_to_reg_tuple for panvk_v10_CmdWaitEvents2 143- PanVK-v10: VulkanCTS 1.3.9.2 - Segfault in panvk_v10_link_shaders for nocolor-tests on Rock5B (RK3588) 144- ANV: scalar FS validation failure is always hit when playing the steam game "Black Myth: Wukong Benchmark Tool" 145- PanVK-v10: VulkanCTS 1.3.9.2 - Multiple tests hit "Unknown sampler dim type"-assert on Rock5B (RK3588) 146- [REG 24.1.7->24.2.1, bisected] RADV: uVkCompute's mad_throughput benchmark hangs 147- [panfrost] rk3588 Ion Fury game (from gog via box64) no image 148- [panfrost] rk3588 Ion Fury game (from gog via box64) no image 149- [panfrost] rk3588 Ion Fury game (from gog via box64) no image 150- [panfrost] rk3588 Ion Fury game (from gog via box64) no image 151- [v3d] performance regression in GTK's NGL renderer from 24.1 to 24.2 152- Color attribute not restored properly when using glPushAttrib/glPopAttrib calls 153- 69c772e4ea8f (egl: link with libgallium directly) breaks VAAPI - OpenGL interop on radeonsi 154- Select VA-API device under Wayland in Multi-GPU setup 155- nine: intel-whl: ERROR: src0 is null 156- [clc] [FTBFS] [llvm20+] no matching function for GetResourcesPath 157- [AMDGPU RDNA3] Viewport compositing generates severe graphical artifacts in Blender 158- v3d gbm_bo_unmap() crashes when used by TigerVNC 159- !30610 broke piglit program\@execute\@bswap with rusticl on radeonsi 160- [Regression] [BISECTED] Genshin Impact: GPU HANG: ecode 12:1:85def5fb 161- Arma Reforger causes amdgpu crash on 24.2 162- [Regression][Vulkan][TGL]Failed to render to depth image at non-zero mip level on Mesa 24.1 and later 163- [ANV] Breaking Limit benchmark does not run 164- [anv][dg2,mtl][bisected][regression] dEQP-VK.ray_tracing_pipeline.acceleration_structures.device_compability_khr.gpu_built.bottom failure 165- Artifacts on Black Myth Wukong 166- DCS world radeon rx 7900 flicker 167- radv/aco: Scheduler bug in ACO somewhere in parallel-gs 168- Once human black artifacts consume the screen 169- [Regression] Wayland-only MESA fails to build without libx11-xcb-dev, libxcb-dri3-dev, libxcb-present-dev, etc. 170- Possible race condition when generating enums.h 171- Rusticl: unsigned int division results in wrong value 172- a6xx spec\@arb_texture_view\@rendering-formats and spec\@arb_shader_image_load_store\@bitcast regression 173- [Bisected] Chromium GPU Process Crashes with Vulkan Backend 174- radv: weird dEQP-GLES31.functional.blend_equation_advanced.msaa.* failures if DRLR is enabled with RADV 175- Anv/Zink ICL hang 176- [anv] possible regression found on 982106e6768e6f69d71710e314325c813591452d 177- xe2: Non-compiler assertion failure replaying cp2077 fossil 178- iris: GfxBench5's Car Chase slow-clears LOD0 when using HIZ_CCS_WT 179- Mesa doesn't build with the following config options 180- Intel: Enable Compression on depth buffers (Xe2) 181- mpv fails with VA-API for Radeon RX 6600, Mesa 24.3.0-devel 182- Failed to cross-compile Mesa-24.2 & Rusticl with bindgen 0.65.1 and rustc 1.80.0 183- Confidential issue #11711 184- Spectacle crashes recording h264 video with 24.2-rc3 185- [nak][nouveau][nvk] FAILED: src/nouveau/compiler/nak 186- u_debug_stack_test.capture_not_overwritten fails on i386 187- [bdw,bsw][bisected] \*.spec.nv_compute_shader_derivatives.execution.derivatives* 188- a6xx spec\@arb_texture_view\@rendering-formats and spec\@arb_shader_image_load_store\@bitcast regression 189- turnip: vulkaninfo crashed 190- turnip-freereno Build error (/usr/local/include/xf86drm.h:40:10: fatal error: drm.h: No such file or directory) 191- tu: compilation failing when compiling turnip with only kgsl and X11 enabled 192- samplerCube constructor in fragment shader no longer converting uvec2 bindless texture handles ( segmentation fault , core dumped ) 193- ci/llvm18: AMD unitialized value 194- anv: gen9.5 flakiness in dEQP-VK.multiview.dynamic_rendering.depth.* 195- Turnip: shaderInt8 support 196- intel/brw: q2rtx RT pipeline fails EU validation 197- vaapi decoding corruption with green blocks 198- interpolateAt precision lowering unhandled by glsl_to_nir() 199- [anv] CS2 crashes on LNL 200- [anv] Dota2 does not start on LNL 201- Raspberry Pi 5 Vulkan 1.3 missing API checklist 202- [radeonsi][bisected][regression] glClientWaitSync() quickly times out with INT64_MAX timeout 203- d3d10umd: Build regression in 24.2.0-devel 204- zink/tu: glcts flake on a750 205- nouveau: advertises GL_EXT_memory_object without implementing \`*UID` callbacks 206- LIBGL_DRIVERS_PATH gone 207- radv: GPU hang with Tauray 208- [Regression][Vulkan][TGL][Bisected]vkCmdCopyQueryPoolResults failed to write buffer with compute pipeline on Mesa 24.1 209- Worms Revolution: not rendering explosion effects 210- crash on pushbuf_validate nvc0_blit do_blit_framebuffer 211- ci_run_n_monitor.py won't continue enabling the jobs 212- ci_run_n_monitor.py doesn't monitor manual jobs 213- piglit: cl-api-build-program crashes 214- nir: regression with bindless textures (bisected) 215- i915g: glGenerateMipmap() fails with 2048×2048 textures 216- gallium-va: VA_RC_QVBR support 217- nir/glsl: vectorize_io unpredictably vectorizes mismatched types 218- [radeonsi] Asterix & Obelix XXLRomastered: river misrendered (completely black) 219- Build fails without Vulkan 220- No dependency check for PyYAML in meson.build 221- GPU Hang in Metal Gear Rising Revengeance 222- VK_ERROR_DEVICE_LOST A770 DXVK Fallout 3 223 224 225Changes 226------- 227 228Aaron Germuth (1): 229 230- Add VK_EXT_host_query_reset to guest's allowed device extensions 231 232Aaron Ruby (15): 233 234- gfxstream: Advertise VK_KHR_external_semaphore_fd always on Linux/Android guests. 235- gfxstream: logging on Linux guests 236- gfxstream: Fix createResource call for Linux-guest VirtGpu layer 237- gfxstream: translate physicalDevices list in VkPhysicalDeviceGroupProperties to their gfxstream objects 238- gfxstream: Use vk_find_struct for other struct searches in gfxstream_vk_CreateDevice 239- Revert "Reland "Partial revert of aosp/2858589 to avoid Mesa layer for Android"" 240- gfxstream: Add "transformList" method for VkFences, in addition to VkSemaphores 241- gfxstream: Clean up VK_KHR_external_memory_fd support for Linux 242- gfxstream: Revert all unnecessary gfxstream_vk_* object types and corresponding translations in functable. 243- gfxstream: Remove VkImage from functable handle translation 244- gfxstream: Change createResource to use stride instead of bpp 245- gfxstream: Switch to using scanout path with native DRM images for Linux WSI 246- gfxstream: Add linearPeerImage for Linux guests, to use for imageLayout queries 247- gfxstream: Add vkCmdPipelineBarrier2 implementation in VkDecoderGlobalState 248- gfxstream: ResourceTracker: Fix bug in return-code check for blob wait 249 250Adam Jackson (8): 251 252- meson: Introduce the legacy-x11 option 253- glx: Move DRI2 client support behind legacy-x11 254- egl/x11: Move DRI2 client support behind legacy-x11 255- gallium/video: Move DRI2 support behind legacy-x11 256- rusticl: Add clCreateSubDevices stub 257- vtn: Handle SPV_INTEL_optnone 258- rusticl: Enable out-of-order execution 259- glx: Fix the GLX_EXT_swap_control_tear drawable attributes 260 261Aditya Swarup (3): 262 263- iris: Disable fast clear when surface height is 16k 264- anv: Disable fast clear when surface height is 16k 265- anv: Use RCS engine for copying stencil resource for gfx125 266 267Agate, Jesse (2): 268 269- amd/vpelib: Refactor structs for API change 270- amd/vpelib: Refactor frontend and backend config callback 271 272Alan Liu (2): 273 274- amd/vpelib: Amend log for tone map support check 275- amd/vpelib: Remove unused structs 276 277Alejandro Piñeiro (5): 278 279- v3dv/device: set value for maxInlineUniformTotalSize 280- v3dv: update conformanceVersion to 1.3.8.3 281- v3dv: expose Vulkan 1.3 282- panfrost: add support for DRM_PANTHOR_DEV_QUERY_TIMESTAMP_INFO on drm-shim 283- panfrost: make DRM version panfrost/panthor conditional on drm-shim 284 285Aleksi Sapon (16): 286 287- gallivm: Fix LLVMPipe codegen issues discovered on Apple Silicon 288- util: const hash_table for some readonly operations 289- llvmpipe: much faster sample function cache using RCU pointers 290- llvmpipe: half_to_float use LLVM fpext when fp16 is available 291- llvmpipe: improvements to JIT assembly dump 292- llvmpipe: FlameGraph profiling support 293- wsi: Metal support 294- wsi: Fix Metal WSI CAMetalDrawable leak 295- llvmpipe: fix quad group helper invocation masking 296- llvmpipe: correctly implement output variables loads 297- llvmpipe: implement alpha-to-coverage dithering 298- lavapipe: enable alpha-to-coverage dithering 299- llvmpipe: Fix anisotropic sampling aliasing bug 300- llvmpipe: Fix pmin calculation 301- softpipe: Fix anisotropic sampling aliasing bug 302- llvmpipe: update traces for aniso filtering fix 303 304Alessandro Astone (2): 305 306- egl/gbm: Walk device list to initialize DRM platform 307- panvk: Add missing headers to android platform 308 309Ali Homafar (1): 310 311- zink: Optimize descriptor buffers struct filling 312 313Alvin Wong (1): 314 315- hasvk: Conditionally expose VK_KHR_present_wait 316 317Alyssa Rosenzweig (403): 318 319- vk/meta: add buffer view tracking 320- vk/meta: add VK_IMAGE_VIEW_CREATE_INTERNAL_MESA flag 321- asahi: drop dead code 322- asahi: drop unused patch index buffer lowering 323- asahi: refactor prolog lowering 324- asahi: drop pointless conversion 325- asahi: drop #includes 326- libagx: fix prefix sum kernel 327- asahi: make GS inputs explicit 328- asahi: add TES->compute lowering 329- asahi: split up agx_nir_lower_vs_before_gs 330- libagx: hoist code out of loop 331- asahi: offset buffer images in software 332- ail: constify more arguments 333- ail: add ail_get_level_size_B helper 334- libagx: add statistic increment kernel 335- libagx: mask counters to 32-bit 336- libagx: rework tess i/o 337- hk: add Vulkan driver for Apple GPUs 338- agx: fix ballot extend packing 339- asahi: don't count helper invs in pipeline stats query 340- nir/opt_uniform_atomics: add fs atomics predicated? flag 341- nir/divergence_analysis: handle AGX stuff 342- agx: add some SRs 343- agx: use opt_uniform_atomics 344- nir/opt_peephole_select: allow speculatable load constant 345- nir: add derivative intrinsics 346- nir/opt_peephole_select: allow derivatives 347- nir/gather_info: handle derivative intrinsics 348- nir/lower_wpos_ytransform: handle intrinsic ddx 349- nir/opt_constant_folding: handle derivative intrinsics 350- nir/divergence_analysis: handle derivative intrinsics 351- nir/instr_set: allow derivative intrinsics 352- nir/schedule: handle derivative intrinsics 353- treewide: use new-style derivative builders 354- vtn: use new derivative builders 355- radv: use glsl function name for dFdxfine 356- amd: switch to derivative intrinsics 357- agx: convert to ddx intrinsics 358- agx: fix clang-format 359- r600: remove 64-bit ddx handling 360- r600: switch to derivative intrinsics 361- dxil: switch to derivative intrinsics 362- broadcom: switch to derivative intrinsics 363- vc4: switch to derivative intrinsics 364- brw: switch to derivative intrinsics 365- elk: switch to derivative intrinsics 366- ir3: switch to derivative intrinsics 367- nir: add nir_shader_alu_pass 368- treewide: use nir_shader_alu_pass 369- nir/lower_alu: use alu_pass 370- nir/scale_fdiv: use alu_pass 371- nir/opt_idiv_const: use alu_pass 372- nir/lower_packing: use alu_pass 373- nir/opt_reassociate_bfi: use alu_pass 374- asahi: fix warning 375- agx: factor out agx_device_key 376- agx: exploit soft fault info 377- hk,asahi: pass through soft fault to VBO lower 378- asahi: optimize out vbo clamp with soft fault 379- hk: optimize bounded loads with soft fault 380- asahi: remove agx_bo::dev 381- asahi: drop segment list parsing 382- asahi: drop mprotect in decoding 383- asahi: drop agx_bo::type 384- asahi: stop fixing up USC addresses 385- asahi: drop iThing logic 386- asahi: always pass align parameter for BOs 387- asahi: introduce agx_va data structure 388- hk: fix BDA alignment 389- asahi,hk: generalize bo_bind hook 390- hk: fix restart index overemit 391- nir,agx: add "active threads in subgroup" intrinsic 392- libagx: add drawIndirectCount clone&patch kernel 393- hk: implement drawIndirectCount 394- hk: fix vkd3d-proton descriptor-performance 395- agx: optimize popcount(ballot(true)) 396- agx: handle min_lod sources 397- hk: advertise shaderResourceMinLod 398- hk: advertise more eds features 399- agx: tweak jmp_exec_none heuristic 400- agx: fuse not into if 401- agx: let if-fusing opts interact 402- hk: denouveau 403- hk: advertise KHR_sampler_ycbcr_conversion 404- agx: fix block_image_store indirect ts 405- nir: model AGX explicit coordinate intrinsics 406- nir: add image_block_size_per_thread_agx info 407- agx: plumb imageblock stride 408- agx: allow st_tile in compute shaders 409- agx: handle explicit coordinate tib instructions 410- agx: add a comment about an unknown bit 411- asahi: pack imageblock words for compute 412- asahi: switch to block based compute blitter 413- agx: lower exact frcp 414- elk: fix compute shader derivatives 415- ntt: switch to derivative intrinsics 416- r300: switch to derivative intrinsics 417- pan/bi: switch to derivative intrinsics 418- pan/mdg: convert to derivative intrinsics 419- nak: switch to derivative intrinsics 420- vtn: infer SpvCapabilityLinkage 421- asahi/clc: drop caps list 422- nir/divergence_analysis: handle load_agx 423- nir,agx: make block image store an image() intrinsic 424- agx: support bindless block image store 425- agx: fix wait_pix scheduling with eMRT 426- ail: swallow up formats 427- ail: consider format for compressability 428- ail: clarify metadata buffer calculation 429- ail: add metadata width/height in tile helpers 430- ail: add enums/queries for AGX compression modes 431- asahi/clc: add bindless image intrinsics 432- libagx: add in-place decompression kernel 433- asahi,ail: pull texture view logic into ail 434- asahi,hk: deduplicate txf sampler settings 435- asahi: skip unmapped bo in fault dbg 436- asahi: optimize aligned blits 437- asahi: allow rgb9e5 rendering internally 438- asahi: factor out agx_usc_shared_non_fragment 439- asahi: allow compute blitter to handle rgb9e5 blits 440- asahi: fix fbfetch with eMRT 441- asahi: add agx_fill_decompress_push helper 442- asahi: handle cross-process eMRT 443- asahi: add missing upcast for eMRT MSAA 444- hk: be stricter about formats in copies 445- hk: rework meta for compression 446- hk: prep for block based blitter 447- hk: remove texel buffers from meta 448- hk: fix bg key with eMRT 449- hk: handle compressed eMRT 450- hk: lock device-wide scratch access 451- hk: drop irrelevant comments 452- hk: shrink cmd bo 453- hk: disable submission merging 454- hk: wire up compression 455- lima: set no_integers 456- etnaviv: set no_integers 457- freedreno/a2xx: set no_integers 458- svga: set no_integers 459- virgl: set no_integers 460- nir: use Horner's method for atan 461- nir: correct comment for atan range reduction 462- nir: simplify atan range reduction 463- nir: push down fabs for atan 464- nir: extend copysign for no-integer hw 465- nir: use copysign for atan 466- nir: simplify atan range reduction fixup 467- nir: negate an expression in atan 468- nir: fuse ffma for atan range fixup 469- nir: push up abs in atan2 calculation 470- nir: drop NaN fixup for atan 471- nir: CSE comparisons in atan2 472- ci: update traces 473- virtio/vdrm: don't spam 474- tu/virtio: silence startup spam on asahi 475- zink: fix buffer view clamp overflow 476- vulkan: merge driver internal bits 477- vulkan: export vk_robustness_disabled 478- hk: add missing dep 479- ail: pull in DRM modifier helper 480- hk: handle compression modifiers 481- hk: disable mods on funny float formats 482- hk: advertise modifiers 483- hk: switch to vk_meta for buffer copies 484- hk: drop some dead code 485- hk: drop bad comment 486- hk: quiet gcc warning 487- asahi: add tib sample setter helper 488- hk: fix attachmentless multisampling tile sizes 489- hk: ban compression with feedback loops 490- hk: lower memory model 491- agx: pull out the big hammers for device-barriers 492- agx: plumb COHERENT 493- hk: free large_bos 494- hk: don't leak NIR with VS/GS 495- hk: don't leak GS rast with rast disc 496- hk: don't leak GS subparts in an error path 497- libagx: cleanup loop vars 498- libagx: simplify tcs out calc 499- libagx: drop unused arg 500- asahi: fix UB affecting x86 builds 501- asahi: identify factor buffer size 502- asahi/decode: fix ptr casts 503- asahi/virtio: fix ptr cats 504- asahi/virtio: fix allocate-with-alignment 505- agx: fix spilling in i686 builds 506- agx: fix =spill 507- agx: clarify scratch size units 508- agx: allocate scratch in spilling unit test 509- agx: clarify spill lowering math 510- hk: fix weird formatting 511- hk: quiesce gcc uninitialized var warning 512- hk: fix debug cache key 513- hk: fix ptr cast 514- hk: shush gcc warning 515- hk: fix loop submit with NONE 516- hk: fix vertex clustering 517- hk: fix oq availability set 518- hk: fix phi ordering 519- hk: drop dead 520- hk: clarify spilling 521- hk: mutex shader compiles when debugging 522- hk: add notess perftest 523- hk: add noborder perf test 524- hk: soup up =perf 525- hk: optimize empty tiles 526- hk: optimize out empty VDM batches 527- hk: assert batches are nonempty 528- hk: parallelize after-graphics available sets 529- hk: run opt_access 530- nir: add bounds_agx opcode 531- asahi: carveout lower VAs 532- hk: use 0 pointer for 0-sized buffers 533- agx: handle bounds_agx 534- hk: clean up before lowering bounds checks 535- hk: smarten bounds check lowering 536- agx: optimize robust access 537- agx: clean up packs from bounds check 538- asahi/lower_vbo: optimize robustness2 539- asahi: fix speculation of rgb32 loads 540- asahi,hk: disable rgb32 buffer textures 541- asahi: optimize trivial load_invocation_id 542- agx: handle hw tess eval shaders 543- asahi: add tess assert 544- asahi: add nosoft debug 545- agx: increase lower vars to scratch thresh 546- hk: fix =spill with tess 547- agx: add spilling debug helpers 548- agx: fix spilling around exports 549- agx: add helper to pad 16-bit to 32-bit 550- agx: use padding helper 551- agx: factor out reserved_region 552- agx: fix RA shuffling algorithm 553- agx: add helper to visualize reg file 554- asahi/genxml: add nicer error checking 555- asahi: fix AGX_MAX_OCCLUSION_QUERIES 556- asahi,hk: fix st_tile mask size 557- agx: validate phi sources are defined 558- agx: validate sizes are consistent in the IR 559- agx: add pass manager 560- agx: split up optimizer 561- agx: use reserved_size 562- asahi: #define the ABI 563- asahi: move sample mask to r1l 564- agx: quiesche assert with r1l preload 565- agx: fix corner case scratch allocation 566- agx: don't reserve regs if we won't use them 567- agx: fix shuffling with large reserved region 568- agx: print pack assertions to stderr 569- agx: don't propagate constants from trivial phis 570- hk: bump max input/output components 571- agx: expand macro 572- agx: fix backwards copyprop with phis 573- agx: add unit test for tricky backwards phi case 574- hk: fix tess with notess 575- hk: drop dead 576- hk: flush with events 577- agx: gimme c23 578- agx: fix corner with uniform source lowering 579- agx: use reserved_size more 580- agx: commonize reg_to_ssa update 581- agx: factor out insert_copy 582- agx: add another RA torture mode 583- agx: fix shuffling with partial clobbered vector 584- agx: add reg to agx_index 585- agx: reset kill bits in liveness 586- agx: strengthen assert 587- agx: assert a subtle invariant 588- agx: produce regful SSA in RA 589- agx: lower swaps late 590- agx: commonize ra_class 591- agx: fix shuffles of vectors 592- agx: validate RA 593- hk: fix shadowing 594- hk: fix partial EOT program 595- hk: fix ZLS across split CRs 596- hk: add helper to dump draw 597- agx: fix load_first_vertex 598- hk: split out HK_PERFTEST 599- hk: add nobarrier test 600- hk: add =batch perftest 601- hk: add norobust test 602- hk: parametrize sysmem heap fraction 603- hk: reduce heap size 604- asahi: optimize !robustness with soft fault 605- asahi: optimize !robust VBO 606- hk: optimize !robust VBO 607- hk: use vk_meta for compressed MSAA copies 608- hk: fix eMRT decompress of msaa 609- hk: end pregfx even if there's no gfx 610- hk: fix emrt compressed msaa arrays 611- hk: enable compressed MSAA 612- hk: reformat driconf 613- util,hk: add border colour driconf 614- hk: use real opaque black with integers 615- hk: drop #include 616- hk: perf debug resolves 617- hk: perf debug some meta 618- hk: Reset rendering attachment locations in BeginCommandBuffer/Rendering 619- asahi,hk: implement remapping in epilog 620- hk: advertise KHR_dynamic_rendering_local_read 621- asahi: rm include 622- agx: inline texture states into EOT 623- hk: fix store_op_none 624- hk: support geom/tess subgroups 625- hk: drop FSR todo 626- hk: check arch in pipeline cache 627- hk: fail probing silently 628- hk: drop gunk 629- nir: extract round component helper 630- nir/opt_algebraic: optimize patterns hit with OpenCL 631- util/ralloc: add total_size helper 632- nir: correct comment 633- nir/serialize: fix name 634- asahi: split out genxml header 635- asahi: fix spacing 636- asahi: assume compiler is not stupid 637- asahi: normalize enum formatting 638- asahi: migrate f-strings 639- asahi: fix indentation 640- asahi: cleanup validation 641- asahi: cleanup generate valid 642- asahi: drop extra new line 643- asahi: guard some apple code 644- asahi: guard darwin stuff 645- asahi: fix newlines 646- asahi: add exact fields to genxml 647- asahi: convert to exact 648- asahi: drop trailing 649- asahi: be explicit about LOD clamps 650- asahi: drop zero defaults 651- asahi: drop macOS defaults 652- asahi: drop PBE default 653- asahi: drop fragment face 2 defaults 654- asahi: set exact on some fixed? unknown fields 655- asahi: drop comment 656- asahi: drop comment about blob weirdness 657- asahi: drop dated comments 658- asahi: idetnify Clamp W 659- asahi: set clamp W? explicitly 660- asahi: drop unused default 661- asahi: be explicit about usc reg unk 662- asahi: be explicit about frag properties 663- asahi: automatically set default to encode to zero 664- asahi: drop implied defaults 665- asahi: get rid of default from the XML 666- asahi: make wrapping less silly 667- asahi: don't do extra runtime validation for exact 668- asahi: defer fixed->float to tess 669- asahi: rm tessellator branch 670- asahi,agx: clang-format 671- agx: split up f16 ops 672- agx: special case mov_imm 673- agx: drop encoding_16 674- agx: encoding_32 -> encoding 675- agx: consistent ffma name 676- asahi: presumably fix indirect indexed tess 677- libagx: extend indirect tess to handle indexed 678- hk: plumb indirect_local 679- hk: add mechanism to test indirects 680- hk: wire up indirect tess 681- hk: drop store_op_dontcare w/a 682- hk: switch to 64-bit queries 683- hk: use push size macro 684- hk: bump max push size 685- asahi: fix no16 flag 686- agx: fix uniform packing with local_load 687- agx: pack agx_cf_binding 688- agx: move binary_size into info 689- hk: fix alpha-to-coverage with sample shading 690- hk: allow tess modes in either stage 691- hk: make tess partitioning dynamic 692- libagx: fix missing statics 693- libagx: don't rely on loop unroll in txs 694- libagx: drop generated VDM tess path (for now?) 695- libagx: make points mode dynamic 696- libagx: always tessellate clockwise 697- asahi/clc: strip nir 698- asahi: drop unused 699- asahi: pad blit key 700- asahi: use indirect ptr helper more 701- asahi: add the BO even for internal dispatches 702- asahi: fix indirect buffer UAF 703- asahi: fix extremely subtle UAF 704- asahi: assert guard previously-subtle code 705- asahi: don't leak blit shaders 706- asahi: don't leak binaries 707- asahi: don't leak linked shaders 708- asahi: don't leak drm version 709- nir,agx: lower fmin/fmax in NIR 710- agx: optimize wait_pix a bit 711- agx: negate iadd/imsub constants 712- agx: test immediate packing opt 713- agx: run algebraic later 714- asahi,agx: rework constant promotion upload 715- agx: don't upload constant padding at the start 716- agx: fold more inots 717- hk: enable constant promotion 718- nir/divergence_analysis: add AGX atomics 719- clc: add mesa_clc tool 720- asahi/clc: ingest spir-v 721- asahi: fix a2c with sample shading, harder 722 723Amber (4): 724 725- ir3: add encoding support for 64-bit atomics introduced in a7xx. 726- ir3: add support for 64 bit atomics 727- nir: add pass to lower atomic arithmetic to a loop with cmpxchg. 728- tu, ir3: Implement VK_KHR_shader_atomic_int64 for a7xx. 729 730Andrew Woloszyn (2): 731 732- Added some fixes for zircon tests in the Android CTS. 733- Fix for dynamic rendering. 734 735Anil Hiranniah (1): 736 737- panfrost: Fix a memory leak in the CSF backend 738 739Antonio Ospite (2): 740 741- android: simplify building libgallium_dri on Android 742- android: fix build by removing references to dri-search-path 743 744Asahi Lina (9): 745 746- asahi: Add 1queue debug option 747- asahi: Print queue ID with batch debug 748- asahi: Add more batch debug 749- asahi: Fix non-async flush 750- asahi: Relax cross-context flush 751- asahi: Update to UAPI 10011 752- asahi: Do not sync against our own queue 753- agx: Fix implicit sync for virtgpu 754- agx: Fix queue destroy op for virtgpu 755 756Assadian, Navid (8): 757 758- amd/vpelib: Change Max DS support to 4:1 759- amd/vpelib: Add API function to get taps 760- amd/vpelib: Fix CS translation for geometric downscaling 761- amd/vpelib: Fix whitepoint for geometric downscaling 762- amd/vpelib: Add 420 semi-planar 12bit handling 763- amd/vpelib: Document public API structures 764- amd/vpelib: Add better rotation and mirror capability check 765- amd/vpelib: Add mirror & rotation capability check function for vpe1.1 766 767Bas Nieuwenhuizen (1): 768 769- radv: Disable EXT BDA capture and replay. 770 771Benjamin Cheng (1): 772 773- radv/video: add event support for VCN4 774 775Benjamin Herrenschmidt (1): 776 777- dril: Fixup order of pixel formats in drilConfigs 778 779Benjamin Lee (6): 780 781- nvk: implement vkCmdDispatchIndirect on pre-turing 782- panvk: don't dirty preloaded tiles for writeback 783- panvk: fix combined image/sampler descriptor arrays 784- panvk: allow resuming secondary cmdbufs with dynamic rendering 785- panvk: flush sync point before executing secondary cmdbufs 786- panvk: hack to improve depth clipping with small viewport depth range 787 788Benjamin Otte (2): 789 790- nvk: Don't emit critical messages during init 791- pvr: Don't emit critical warning all the time 792 793Bernhard Rosenkränzer (1): 794 795- dril: Fix return types of peglInitialize and peglTerminate 796 797Bo Hu (2): 798 799- goldfish: handle null handle 800- gfxstream: global_state_wrapped_decoding of vkCreateComputePipelines 801 802Boris Brezillon (157): 803 804- panvk: Adjust RGB component order for fixed-function blending 805- util/hexdump: Squeeze repeated patterns instead of zero-ed chunks 806- panvk: Make sure we emit blend descriptors even if no FS shader is provided 807- pan/genxml: Fix Resource descriptor alignment 808- pan/decode: Select the right registers when dumping local storage descriptors 809- pan/decode: Interpret LOAD_MULTIPLE opcode when decoding a command stream 810- pan/cs: Allow lazy root chunk allocation 811- pan/cs: Provide cs_root_chunk_{gpu_addr,size}() and cs_is_empty() helpers 812- pan/shader: Fix FS varyings.input_count initialization on v9+ 813- pan/bi: Support native layer_id store/load on Valhall 814- pan/tex: Allow format reinterpretation on v9+ 815- vk/sync: Use the proper type in vk_filter_{src,dst}_access_flags2() 816- nir/format_convert: Promote input to 32-bit before packing integers 817- vk/format: Add missing R10G10B10A2_SINT entry in the pipe -> vk table 818- vk/image: Add a vk_image_buffer_range() helper 819- vk/image: Fix the view extent of uncompressed views of compressed images 820- vk/meta: Let meta users pass the color attachment write mask explicitly 821- panvk: Don't spam the error output if the device is not for us 822- panvk: Rename the error labels in create_device() 823- panvk: Add missing vk_device_finish() calls 824- pan/texture: Allow uncompressed view of compressed image 825- panvk: Rework the NULL test in panvk_buffer_{gpu_ptr,range}() 826- panvk: Constify the shader pointers in the graphics state 827- panvk: Add helpers to save/restore the compute/graphics state for meta operations 828- panvk: Rework the modifier selection 829- panvk: Switch to vk_meta 830- panvk: Lower load_multisampled_pan 831- panvk: Add support for render-pass resolve operations 832- panvk: Fix depth-stencil AFBC support 833- panvk: Allow compressed formats 834- panvk: Allow ZS formats on 1D/3D images 835- panvk: Lower var copies before lowering IOs 836- panvk/ci: Enable clearing tests 837- panvk: Prepare the blend logic for Valhall 838- panvk: Collect allocated push sets at the command level 839- panvk: Prepare panvk_cmd_pool for CSF 840- panvk: Make panvk_queue_finish() a per-arch function 841- panvk: Allow more descriptor sets on Valhall 842- panvk: Move panvk_debug_adjust_bo_flags() to panvk_device.h 843- panvk: Conditionally register an host address when tracking user memory 844- panvk: Prepare the cmd_desc_state logic for Valhall 845- panvk: Prepare cmd_meta for Valhall 846- pan/cs: Make sure we're not passed cs_now() to asynchronous instructions 847- pan/cs: Accept val=undef when cs_branch_label() condition is ALWAYS 848- pan/cs: Add helpers to emit match (AKA switch-case) blocks 849- pan/cs: Allow sparse register set passed to loads/stores 850- pan/cs: Add an optional load/store tracker to make sure we don't omit waits 851- pan/cs: Add inline specifiers to a bunch of cs\_ helpers 852- pan/cs: Add a facility to restrict register access 853- pan/kmod: Don't cap VM bind operations to one 854- pan/desc: Add a struct for valhall/bifrost to the union in pan_tiler_context 855- pan/desc: Make vertex_count a Midgard field 856- panvk: Fix the varying pool label 857- panvk: Kill the set_dirty() macro 858- panvk: Drop the NO_MMAP flag when PANVK_DEBUG=trace 859- panvk: Fix valhall_pack_buf_idx() 860- panvk: Don't open-code vk_index_type_to_bytes() 861- panvk: Don't make panvk_device size gen dependent 862- panvk: Make sure push sets are reset when picked from the CommandPool 863- panvk: Prepare the panvk_device logic for CSF 864- panvk: Set the submit mode to THREADED_ON_DEMAND 865- panvk: Move Cmd{Bind,Push}DescriptorSets implementation to cmd_desc_state.c 866- panvk: Move cmd_bind_shaders() to panvk_vX_shader.c 867- panvk: Move CmdPushConstants2KHR to panvk_vX_cmd_push_constant.c 868- panvk: Be robust against allocation failures in a command buffer context 869- panvk: Don't pretend we can compile for v9 870- panvk: Transition to explicit VA assignment on v10+ 871- panvk: Add a read-write non-cached memory pool for CSF events 872- panvk: Add a CSF backend for panvk_queue/cmd_buffer 873- panvk: Advertize v10 support 874- panvk/csf: Fix missing TSD pointer initialization on RUN_IDVS 875- panvk/csf: Fix inverted address/destination register in CmdSetEvent2() 876- pan/va: Accept DIM_SUBPASS[_MS] in valhall_tex_dimension() 877- panvk: Temporarily disable VK_FORMAT_D32_SFLOAT_S8_UINT on Valhall 878- panvk: Fix valhall_pack_buf_idx() for SSBO store/atomic operations 879- panfrost: Emit an SSBO table on v9+ 880- nir/lower_ssbo: Add an option to conditionally lower loads 881- nir/lower_ssbo: Extend the load_ssbo_address intrinsic to pass an offset 882- panfrost: Pass the architecture to panfrost_nir_lower_sysvals() 883- pan/va: Optimize SSBO lowering on Valhall 884- panvk/csf: Unconditionally clean L2 and LS caches when closing a CS 885- pan/cs: Fix buffer overflow in cs_block_end() 886- pan/cs: Make sure we don't overflow the chunk capacity 887- pan/cs: Make sure cs_alloc_ins() is never passed num_instrs=0 888- pan/cs: Make the CS loop helpers more robust 889- pan/cs: Pass the block to end to cs_block_end() 890- pan/cs: Drop the cs_builder::blocks::cur field 891- pan/cs: Don't use a list for our block stack 892- pan/cs: Add helpers for if/else blocks 893- panvk/csf: Replace a cs_while() by a cs_if() 894- panvk/csf: Pass a non-zero flush-id to benefit from cache flush reduction 895- panvk/csf: Fix TLS pointer copy 896- panvk: Use nir_metadata_control_flow 897- panvk: Lower get_ssbo_size() on Valhall 898- panvk: Pass the robustness state to nir_lower_descriptors() 899- panvk: Hook up descriptor array bounds checking 900- vk/meta: Make sure texel is 32-bit in build_buffer_to_image_cs() 901- panfrost: Don't turn 3D/cube images into 2D arrays 902- pan/decode: Interpret CS_BRANCH instructions 903- pan/desc: Add layer_offset field to pan_tiler_context::valhall 904- pan/genxml: Fix layer_offset definition on v9+ 905- panvk/csf: Fix multi-layer rendering 906- panvk: Protect access to the virtual address heap 907- pan/va: Fix nir_op_pack_uvec4_to_uint 908- pan/genxml: Remove surface descriptors on v6/v7 909- pan/genxml: Fix surface descriptor alignment on v6 910- pan/genxml: Fix resource table alignment constraint on v9 911- pan/kmod: Try to use local storage in panthor_kmod_vm_bind() 912- panvk: Don't reset errno in panvk_kmod_zalloc() 913- panvk: Stop including pan_blend.h from panvk_device.h 914- panvk: Use the proper alignment for texture payloads 915- panvk: Check allocation failures in panvk_shader_upload() 916- panvk: Report allocation failures in the CreateImageView path 917- panvk: Fix FB initialization when using non-compressed views of compressed images 918- panvk: Fix stencil attachment resolve on JM 919- panvk/v10: Add missing cleanup_queue() in queue_finish() 920- panvk: Don't call queue_finish() on non-initialized queues 921- panvk: Don't leak vertex shader program descriptors 922- panvk: Let panvk_priv_bo_create() return a VkResult 923- panvk: Catch indirect allocation failures 924- panvk: Make sure RUN_IDVS has a valid tiler descriptor 925- panvk: Fix bounds checking in upload_shader_desc_info() 926- panvk/csf: Fix prepare_vp() when viewport->x is negative 927- panvk: Advertise KHR_create_renderpass2 support 928- vk/meta: Fix buffer -> image copy using a graphics pipeline 929- zink: Adjust the surface size of non-compressed views 930- vk/image: Fix the extent adjustment of non-compressed views 931- vk/meta: Support VkShaderExt objects to allow tracking internal shaders 932- vk/meta: Allow tracking of driver-specific objects in the meta list 933- vk/meta: Make some helpers public 934- panvk: Add a debug flag to force image copies through the gfx pipeline 935- panvk: Add an helper to create internal shaders 936- panvk: Keep our blend shaders in vk_meta_device 937- panvk: Keep our copy_desc shader in vk_meta_device 938- panvk: Emit textures needed for FB preload at image view creation time 939- panvk: Store attachment image views in the graphics state 940- panvk: Implement a custom FB preload logic 941- panfrost: Move pan_blitter.{c,h} to the gallium driver 942- panvk: Make panvk_pool_free_mem() error proof 943- panvk/csf: Fix no-fragment IDVS 944- panvk/csf: Fix dirty checking in prepare_ds() 945- panvk/csf: Fix cmd_emit_dcd() in the FB preload logic 946- panvk/csf: Fix a buffer/stack-overflow when PANVK_DEBUG=sync 947- panvk: Advertise dynamic rendering support 948- panvk/csf: Fix the clear-only RUN_FRAGMENT case 949- pan/util: Discard depth/stencil writes when early fragment test is forced 950- panvk: Set .is_blit=true when compiling preload shaders 951- panvk: Fix the ZStencil descriptor initialization in the preload logic 952- panvk: Fix the hierarchy_mask selection 953- panvk: Fix the maximum sampler LOD bias 954- pan/va: Don't set bit 26:27 on texture instructions 955- pan/va: Add support for nir_texop_lod 956- pan/bi: Add support for nir_texop_lod 957- pan/desc: Fix the CRC validity update when clean tiles are not written 958- pan/cs: Fix lazy allocation support 959- pan/bi: Expose the packed TextureOperationDescriptor in bifrost_texture_operation 960- panvk/csf: Fix a wait-LS operation in finish_cs() 961 962Boyuan Zhang (3): 963 964- frontends/va: get cu_qp_delta from vaapi 965- radeonsi/vcn/enc:enable hevc cu_qp_delta for vcn2+ 966- radeonsi/vcn: fix loop filter across slices 967 968Brendan (5): 969 970- amd/vpelib: Create virtual stream concept 971- amd/vpelib: Multiple instance support in caching framework 972- amd/vpelib: Update configs to be per-pipe specific 973- amd/vpelib: Visual confirm on input 974- amd/vpelib: Fix duplicate config writer bug 975 976Caio Oliveira (91): 977 978- intel/brw: Simplify \@file annotations 979- intel/brw: Move remaining compile stages to their own files 980- intel/brw: Move and reduce scope of run_*() functions 981- intel/brw: Reduce scope of some VS specific functions 982- intel/brw: Reduce scope of some FS specific functions 983- intel/brw: Reduce scope of some GS specific functions 984- intel/brw: Reduce scope of some TCS specific functions 985- intel/brw: Reduce scope of some TES specific functions 986- intel/brw: Reduce scope of some MESH specific functions 987- intel/brw: Move printing functions to its own file 988- intel/brw: Move out of fs_visitor and rename print instructions 989- intel/brw: Remove workgroup_size() helper from fs_visitor 990- intel/brw: Move regalloc and scheduling functions out of fs_visitor 991- intel/brw: Move calculate_cfg out of fs_visitor 992- intel/brw: Move VARYING_PULL_CONSTANT_LOAD from fs_visitor to fs_builder 993- intel/brw: Move interp_reg and per_primitive_reg out of fs_visitor 994- intel: Let compiler set indirect_ubos_use_sampler 995- intel/brw: Don't retype load_subgroup_invocation result to signed 996- intel/brw: Use CSE for LOAD_SUBGROUP_INVOCATION 997- meson: Update Lua dependency version to 5.3 998- intel/brw: Also return the size of the assembled shader 999- intel: Add executor tool 1000- intel/executor: Use PRIx64 to fix building in 32-bit 1001- intel/brw: Use %td format for pointer difference 1002- meson: Only use fallback for Lua if building freedreno tools 1003- intel/brw: Avoid modifying the shader in assign_curb_setup if not needed 1004- intel/brw: Remove unused variable 1005- intel/brw: Use a helper for common VEC pattern 1006- intel/brw: Remove more uses of variable length arrays 1007- intel/brw: Move encoding of Gfx11 3-src inside the inst helpers 1008- intel/brw: Explicitly map brw_reg_file into hardware values 1009- intel/brw: Make BAD_FILE the zero value for brw_reg_file 1010- intel/brw: Remove unused ARF values and helpers 1011- intel/brw: Remove long register file names 1012- intel/brw: Pack brw_reg struct 1013- intel/brw: Simplify fs_inst annotation 1014- intel/brw: Remove special treatment for 2-src in emit() helper 1015- intel/brw: Use linear walk when shader requires DERIVATIVE_GROUP_LINEAR 1016- intel/executor: Fix SWSB for sync.nop 1017- compiler: Allow derivative_group to be used for all stages in shader_info 1018- vk/runtime: Allow "require full subgroups" for Mesh/Task 1019- intel/executor: Small fixes to the help message 1020- spirv: Rename vtn_type::deref to vtn_type::pointed 1021- spirv: Remove (pointed) type from vtn_pointer 1022- spirv: Rename vtn_pointer::ptr_type to vtn_pointer::type 1023- vulkan: Update XML and headers to 1.3.295 1024- spirv: Update headers and metadata from latest Khronos commit 1025- spirv: Prefer symbols from Khronos extension for compute shader derivatives 1026- nir: Allow Mesh/Task to use implicit LOD when DERIVATIVE_GROUP is set 1027- spirv: Allow Mesh/Task to use derivative execution modes 1028- anv: Advertise VK_KHR_compute_shader_derivatives 1029- intel/executor: Fix a couple of memory leaks in the tool 1030- intel/brw: Only validate GRF boundary crossing restriction for GRFs 1031- intel/brw: Consider pipe when comparing SWSB in tests 1032- intel/brw: Add SWSB MATH pipe to assembler 1033- intel/brw: Add DUMP flag to brw_assemble 1034- intel/executor: Dump both pre-processed source and assembly 1035- intel/brw: Skip per-primitive inputs when computing flat input mask 1036- intel/brw: Remove Gfx8- from test_eu_validate.c 1037- intel/brw: Update DPAS validation tests for Xe2 1038- intel/brw: Enable EU validation and compaction tests for Xe2 1039- intel/brw: Enable mi_builder test for Xe2 1040- spirv: Update SPIR-V grammar to use aliases 1041- intel/brw: Add ERROR helper variant that returns to EU validation 1042- intel/brw: Create a struct to hold a decoded brw_inst in eu_validation 1043- intel/brw: Add opcode to brw_hw_decoded_inst 1044- intel/brw: Add num_sources to brw_hw_decoded_inst 1045- intel/brw: Add has_dst to brw_hw_decoded_inst 1046- intel/brw: Add exec_size to brw_hw_decoded_inst 1047- intel/brw: Add access_mode to brw_hw_decoded_inst 1048- intel/brw: Add dst/srcs reg file to brw_hw_decoded_inst 1049- intel/brw: Add dst/srcs type to brw_hw_decoded_inst 1050- intel/brw: Add pred/cmod/sat to brw_hw_decoded_inst 1051- intel/brw: Add dst/src0 address_mode to brw_hw_decoded_inst 1052- intel/brw: Add abs/negate to brw_hw_decoded_inst 1053- intel/brw: Add dst/srcs register numbers to brw_hw_decoded_inst 1054- intel/brw: Add vstride/width/hstride to brw_hw_decoded_inst 1055- intel/brw: Remove the remaining DO_SRC macro from EU validation 1056- intel/compiler: Rename shader_stats structs 1057- intel: Add statistic for Non SSA registers after NIR to BRW 1058- intel/brw: Print Non-SSA regs after NIR in debug output 1059- intel/brw: Fix Gfx9 3-src validation to handle FIXED_GRF 1060- intel/brw: Add phases to backend 1061- intel/brw: Validate some instructions exists only up until some phases 1062- intel/brw: Reduce scope of has_source_and_destination_hazard 1063- intel/brw: Add SHADER_OPCODE_REDUCE 1064- intel/brw: Add SHADER_OPCODE_*_SCAN 1065- intel/brw: Move emit_scan/emit_scan_step near its usage 1066- intel/brw: Add fs_builder::BROADCAST() helper 1067- intel/brw: Ensure BROADCAST() value respect register alignment 1068- intel/brw: Add SHADER_OPCODE_VOTE_* 1069 1070Carlos Santa (2): 1071 1072- intel/hang_replay: fix the typo in the ioctl call 1073- intel/hang_replay: remove EXEC_OBJECT_WRITE 1074 1075Casey Bowman (5): 1076 1077- vulkan/screenshot-layer: Add Vulkan screenshot layer 1078- vulkan/screenshot-layer: Fix image index selection 1079- vulkan/screenshot-layer: Increase buffer sizes 1080- vulkan/screenshot-layer: Fix memory leaks 1081- vulkan/screenshot-layer: Remove extraneous malloc operations 1082 1083Chan, Roy (2): 1084 1085- amd/vpelib: Fix compilation warnings 1086- amd/vpelib: Revise the config sharing handling 1087 1088Chang, Tomson (2): 1089 1090- amd/vpelib: Add and fix collaborate sync data 1091- amd/vpelib: Fix color fill performance issue on VPE1.1 (#419) 1092 1093Charmaine Lee (1): 1094 1095- svga: sync up with the latest svga include files 1096 1097Chenyu Chen (1): 1098 1099- amd/vpelib: Remove unused define macro 1100 1101Chia-I Wu (32): 1102 1103- panvk: remove unused panvk_physical_device::master_fd 1104- panvk: refactor panthor_kmod_dev creation 1105- panvk: query device ids of drm device nodes 1106- panvk: advertise VK_EXT_physical_device_drm 1107- panvk: fill in VkExternalBufferProperties 1108- panvk: fill in VkExternalImageFormatProperties 1109- panvk: remove panvk_GetPhysicalDeviceImageFormatProperties 1110- panvk: advertise VK_KHR_external_memory_fd 1111- panvk: advertise VK_EXT_external_memory_dma_buf 1112- panvk: advertise VK_KHR_external_{fence,semaphore}_fd 1113- panvk: remove panvk_GetImageDrmFormatModifierPropertiesEXT 1114- panvk: fix VkDrmFormatModifierPropertiesListEXT query 1115- panvk: check VkPhysicalDeviceImageDrmFormatModifierInfoEXT 1116- panvk: advertise VK_KHR_image_format_list 1117- panvk: advertise VK_EXT_image_drm_format_modifier 1118- panvk: remove unused is_sampler and is_texture 1119- panvk: fix descriptor set layout hash 1120- panvk: fix texture size/levels/samples query 1121- panvk: fix gl_VertexIndex 1122- panvk: fix scissor box 1123- panvk: disable depth write when depth test is disabled 1124- panvk: remove an incorrect assert in collect_cs_deps 1125- Revert "panvk: Set the submit mode to THREADED_ON_DEMAND" 1126- panvk/jm: assert that the submit mode is not threaded 1127- panvk: add support for VK_KHR_timeline_semaphore 1128- panvk: fix truncated access flags in collect_cs_deps 1129- panvk: expand meta stage and access flags 1130- panvk: use force_fb_preload for unaligned preload 1131- panvk: force_fb_preload should insert a barrier 1132- panvk: fix missing same-subqueue wait for CmdWaitEvents2 1133- panvk: fix dummy sampler handle for vs 1134- panvk: ensure res table is restored after meta 1135 1136Christian Gmeiner (70): 1137 1138- etnaviv: isa: left shift is 3 bit long 1139- dri: fix driver names 1140- etnaviv: Rework uniform handling for UBO addresses 1141- meson: Add missing inc's to idep_nir_headers 1142- nir_lower_mem_access_bit_sizes: Support load_kernel_input 1143- ci: austriancoder farm maintenance 1144- nak: Move bitset to compiler crate 1145- nak: Move cfg to compiler crate 1146- etnaviv: Move halti determination to drm 1147- etnaviv: Switch etna_compiler_create(..) to etna_core_info 1148- etnaviv: Extend shader structs with etna_core_info 1149- etnaviv: Drop halti from etna_specs 1150- etnaviv: isa: Add conv instruction 1151- etnaviv: Drop has_sin_cos_sqrt and has_sign_floor_ceil 1152- etnaviv: Switch to shader_core_count from etna_core_info 1153- etnaviv: Switch to vertex_cache_size from etna_core_info 1154- etnaviv: Switch to vertex_output_buffer_size from etna_core_info 1155- etnaviv: npu: Switch to use etna_core_info 1156- etnaviv: npu: Drop not used spec values 1157- etnaviv: Switch to num_constants from etna_core_info 1158- etnaviv: Switch to max_registers from etna_core_info 1159- etnaviv: Switch to stream_count from etna_core_info 1160- etnaviv: Switch to etna_core_has_feature(..) for npot_tex_any_wrap 1161- etnaviv: Switch to etna_core_has_feature(..) for has_halti2_instructions 1162- etnaviv: Drop assignment of get_timestamp 1163- broadcom/simulator: Add support for asahi based systems 1164- compiler/rust: Copy MappedInstrs from NAK 1165- etnaviv: Update headers from rnndb 1166- etnaviv: nir: Make use of nir_lower_fragcolor 1167- etnaviv: Report number of render targets 1168- etnaviv: nir: Make use of reported number of render targets 1169- etnaviv: Call etna_resource_level_mark_changed(..) for all rts 1170- etnaviv: Extend frag_rb_swap for MRTs 1171- etnaviv: Extend compiled_shader_state for MRTs 1172- etnaviv: Extend compiler for MRTs 1173- etnaviv: Extend etna_record_flush_resources(..) for MRTs 1174- etnaviv: Extend etna_update_blend_color(..) for MRTs 1175- etnaviv: Extend etna_update_zsa(..) for MRTs 1176- etnaviv: Extend etna_set_framebuffer_state(..) for MRTs 1177- etnaviv: Extend etna_update_ts_config(..) for MRTs 1178- etnaviv: Emit MRT states 1179- etnaviv: Emit PS_OUTPUT_REG2 if num_rt is 8 1180- etnaviv: blend: Store information per render target 1181- etnaviv: blend: Move logicop_enable assignment 1182- etnaviv: blend: Add support for MRTs 1183- etnaviv: blt/rs: Just pass the color to the callee 1184- etnaviv: rs: Extend to support MRTs 1185- etnaviv: blt: Extend to support MRTs 1186- etnaviv: shader: Apply output remapping 1187- etnaviv: shader: Apply output saturation 1188- etnaviv: Relink shader when framebuffer state changed 1189- etnaviv: TS usage for MRT needs HALTI2 1190- etnaviv: Limit max supported render targets 1191- etnaviv: Report correct PIPE_CAP_MAX_RENDER_TARGETS value 1192- etnaviv: Enable EXT_draw_buffers2 for HALTI5+ 1193- etnaviv: Enable ARB_draw_buffers_blend for HALTI5+ 1194- etnaviv: Improve split sampler check 1195- etnaviv: blt: Convert assert to real if 1196- etnaviv: Implement ARB_draw_indirect 1197- v3d: Add a few function traces 1198- vc4: Add a few function traces 1199- docs: Update the list of drivers with CPU tracepoints 1200- etnaviv: nir: Enforce stricter swizzle for virtual scalar x register 1201- etnaviv: nir: Switch to derivative intrinsics 1202- compiler/rust: Copy NirInstrPrinter from NAK 1203- etnaviv: isa: Add clamp0_max instruction 1204- etnaviv: isa: Add img_load instruction 1205- etnaviv: Enable ARB_framebuffer_no_attachments 1206- etnaviv: isa: Add img_store instruction 1207- etnaviv: Fix incorrect pipe_nn creation 1208 1209Colin Marc (2): 1210 1211- vulkan/video: set HEVC nuh_temporal_id_plus1 and nal_unit_type correctly 1212- radv/video: set TemporalId correctly 1213 1214Collabora's Gfx CI Team (10): 1215 1216- Uprev Piglit to 4a62c26721a47552a96416a134b789a813dd51a6 1217- Uprev Piglit to 0453436872b6e4d502c2e87817addb95e0d77e3b 1218- Uprev Piglit to f11abb664bfcad09586f32f411b90331e23be2e5 1219- Uprev Piglit to 1d8fab9b49cc487ebcc949119eeb166fb3fa7591 1220- Uprev Piglit to 6533b66878c8cb9361c554b7ff940518ab12c816 1221- Uprev Piglit to a3826de3c26a279599d15b018a9a3e75ca46f4f8 1222- Uprev Piglit to 93b4bd2e0aaab1c22ae3e1a23f9e057a8f7451b2 1223- Uprev Piglit to 7ce69da1199d12ed0ddaa251ed489750523798fb 1224- Uprev Piglit to 791e420b2628c1e35eea81b3bafdb1c904a141e8 1225- Uprev Piglit to c2b31333926a6171c3c02d182b756efad7770410 1226 1227Connor Abbott (60): 1228 1229- egl/x11: pass render fd to screen create for zink 1230- freedreno/ci: Combine and document failures due to test bug 1231- ci: Move two failing loader-related tests to all-skips.txt 1232- ci: Uprev VK-CTS to 1.3.9.0 1233- tu: Fix off-by-one in UBO CP_LOAD_STATE size 1234- ir3, tu: Use a UBO for VS primitive params on a750+ 1235- Revert "tu/a750: Disable HW binning when there is GS" 1236- ir3: Fix nir_op_vec8/vec16 1237- freedreno/computerator: Fix enabling wave128 1238- freedreno/a7xx: Document compute dispatch tiling registers 1239- freedreno, tu, ir3: Enable tiled workgroup item dispatch on a7xx 1240- ir3: Move reconvergence backedge handling 1241- tu: Fix passthrough D24S8 attachments 1242- tu: Treat partially-bound depth/stencil attachments as passthrough 1243- tu: Add a750 flush workaround and re-enable UBWC for storage images 1244- freedreno: Rename CP_SET_CTXSWITCH_IB to CP_SET_AMBLE 1245- ir3: Increase compute const size on a7xx 1246- tu: Initial support for VK_KHR_calibrated_timestamps on a750 1247- freedreno/a6xx+: Add CP_CONTEXT_SWITCH_CNTL bitfields 1248- ir3: Ban conversions with mismatching sizes 1249- Reapply "ci/lima: Temporarily disable" 1250- ir3, turnip: Support VK_*_compute_shader_derivatives on a7xx 1251- tu: Don't WFI after initializing RB_CCU_CNTL on a7xx 1252- tu: Split out register setting from tu6_init_hw() 1253- tu: Write bin size for each bin 1254- tu: Re-emit LRZ state before each bin 1255- tu: Simplify device startup CS creation 1256- tu: Emit CP_SET_AMBLE packets 1257- freedreno: Bump kernel uapi 1258- tu/virtio: Make virtio_simple_ioctl take a vdrm_device 1259- tu: Allow being preempted on a7xx 1260- tu: Add TU_DEBUG=hiprio 1261- vulkan/state: Handle NULL in DS input attachment mapping correctly 1262- vulkan/state: Track the input attachment count 1263- v3dv: Don't misuse nir_variable::data.index 1264- spirv: Make the default input attachment index ~0 1265- nir/lower_input_attachments: Handle unscaled input attachments with no index 1266- ir3: Fix non-bindless s2en texture/sampler order 1267- tu: Support color attachment remapping 1268- tu: Support dynamic input attachments 1269- tu: Make input attachments always contain a real descriptor 1270- tu: Fix flushes for feedback_invalidate case 1271- tu: Track possible feedback loops for dynamic renderpasses 1272- tu: Expose VK_KHR_dynamic_rendering_local_read 1273- ir3, tu: Plumb through support for per-shader robustness 1274- tu: Implement VK_EXT_pipeline_robustness 1275- freedreno/fdl: Extend 2bpp UBWC special case to 1bpp 1276- freedreno/a6xx: Remove dead fd6_get_ubwc_blockwidth() call 1277- freedreno/fdl: Add UBWC config struct 1278- freedreno: Add default UBWC config values 1279- virtio/drm: Update header 1280- tu: Gather UBWC config 1281- freedreno/fdl: Add tiling/untiling implementation for a6xx/a7xx 1282- tu: Add a flag for cached non-coherent BOs 1283- tu: Implement VK_EXT_host_image_copy 1284- freedreno: Add compute constlen quirk for X1-85 1285- tu: Implement VK_PIPELINE_CREATE_2_VIEW_INDEX_FROM_DEVICE_INDEX_BIT_KHR 1286- tu: Don't invalidate CS state for 3D blits 1287- ir3: Fix detection of nontrivial continues 1288- tu: Re-emit visibility stream before each render pass 1289 1290Constantine Shablia (4): 1291 1292- vk/meta: Add copy/fill/update helpers 1293- pan/va: Define the TEX_GRADIENT instruction 1294- pan/va: Lower nir_texop_txd to TEX_GRADIENT with derivs followed by TEX_SINGLE 1295- pan/bi: Lower nir_texop_txd to TEXC in GRDESC_DER mode followed by sampling TEXC 1296 1297Corentin Noël (2): 1298 1299- ci: Allow to pass the PIGLIT_RUNNER_OPTIONS variable 1300- virgl: Avoid a race condition on handle removal 1301 1302Craig Stout (18): 1303 1304- util: fix fossilize_db compile error 1305- util: fix 32bit compiler error in fossilize_db 1306- [guest] Check that the singleton VirtGpuDevice matches desired capset 1307- Reapply "[guest] Check that the singleton VirtGpuDevice matches desired capset" 1308- Fix dispatchable object layout on Fuchsia 1309- Exclude Android entry points on Fuchsia 1310- [guest] Fix Fuchsia build 1311- [guest] Add skeleton Fuchsia VirtGpu implementation 1312- [guest] Build fixes for Fuchsia 1313- [guest] Fuchsia: include missing Vulkan entry points 1314- [guest] Remove Fuchsia syslog_static dependency 1315- [guest] Fix Fuchsia build, add missing header 1316- [guest] Fuchsia: open magma device 1317- [guest] Fuchsia: initialize queries 1318- [guest] Fix Fuchsia build 1319- tu: add OS guards to drm_format_mod 1320- u_gralloc: include dep_android_mapper4 as needed 1321- meson: add log dependency to dep_android 1322 1323Daniel Almeida (6): 1324 1325- nouveau/headers: Run rustfmt on generated files 1326- nouveau/headers: Add a Rust struct for each method 1327- nouveau/headers: Add an nv_push crate in Rust 1328- util: memstream: add fflush support 1329- nak: memstream: move into common code 1330- util: u_memstream: add tests 1331 1332Daniel Schürmann (55): 1333 1334- aco/reindex_ssa: free memory of previous live variable sets 1335- aco: implement aco::validate_live_vars() 1336- aco/scheduler: enable live variables validation when ACO_DEBUG=validate-livevars is set 1337- aco: don't attempt to spill dead phis 1338- aco/cssa: fix kill flags during lowering to CSSA 1339- aco: change signature of get_live_changes() and get_temp_registers() 1340- aco/cssa: update RegisterDemand and validate live variable information 1341- aco/cssa: short-cut some trivial case 1342- aco: call nir_copy_prop() and nir_opt_dce() before instruction selection 1343- nir: add nir_block \*nir_src_get_block(src) helper 1344- nir/opt_if: optimize phis between similar IFs 1345- nir/opt_if: merge IFs which have phis between them 1346- aco/ra: add add_rename() helper 1347- aco/ra: avoid rename map lookups in most cases 1348- aco/ra: use auto for std container iterators 1349- aco/ra: use arena allocator for hash maps 1350- aco/live_var_analysis: Don't attempt to re-insert the same temporary twice into live set 1351- aco: introduce new Operand flag 'Clobbered' 1352- aco/live_var_analysis: use Clobbered flag to calculate additional operand demand 1353- aco: introduce Operand flag 'CopyKill' 1354- radv: promote VK_NV_compute_shader_derivatives -> VK_KHR_compute_shader_derivatives 1355- aco/live_var_analysis: don't set lateKill for p_interp_gfx11 m0 operand 1356- aco: simplify Definition constructors 1357- aco/lower_to_hw: don't allocate new temporaries 1358- aco/reindex_ssa: remove update_live_out parameter 1359- aco: remove Program::allocationId 1360- aco/live_var_analysis: don't test whether phis are assigned to VCC 1361- aco/ra: add RegisterFile::fill_killed_operands(Instruction*) helper 1362- aco/ra: manually fill killed operands when required 1363- aco/ra: remove unnecessary check for duplicate precolored operands 1364- aco: introduce 'isPrecolored' flag for Operand and Definition 1365- aco: set Precolored flag before register allocation 1366- aco/live_var_analysis: check for isPrecolored flag rather than isFixed 1367- aco/ra: make use of Precolored flag 1368- aco/ra: always block register file for precolored operands 1369- aco/live_var_analysis: check isFixed() for definitions in order to set needs_vcc 1370- aco/spill: fix faulty assertions 1371- aco/spill: use float division for score() calculation rather than integers 1372- nir/divergence: separately indicate whether loops have divergent continues or breaks 1373- nir/opt_uniform_atomics: don't preserve divergence information 1374- nir: remove nir_update_instr_divergence() 1375- nir: change signature of nir_src_is_divergent() 1376- treewide: use nir_src_is_divergent() rather than checking the divergence of the SSA 1377- nir/divergence: require nir_metadata_block_index 1378- nir/divergence: Introduce and set nir_def::loop_invariant 1379- nir: consider loop invariance in nir_src_is_divergent() 1380- nir/divergence: calculate divergence without requiring LCSSA form 1381- nir/divergence: skip expensive nir_src_is_divergent() check in most cases 1382- nir/divergence: Don't assume that LCSSA phis are not loop-invariant 1383- treewide: don't lower to LCSSA before calling nir_divergence_analysis() 1384- aco: Respect addressible SGPR limit in VS prologs 1385- aco/ssa_elimination: perform jump threading after parallelcopy insertion 1386- aco/ssa_elimination: remove some redundant checks during jump threading 1387- aco: move jump threading optimization into separate pass 1388- aco/README: add descriptions of recently added passes 1389 1390Daniel Stone (146): 1391 1392- build: Check for PyYAML in Meson build 1393- dri: Allow INVALID for modifier-less drivers 1394- gbm/dri: Remove erroneous assert 1395- ci: Create new 'performance' stage 1396- ci/amd: Move manual/nightly jobs to postmerge stage 1397- ci/intel: Move manual/nightly jobs to postmerge stage 1398- ci/broadcom: Move manual/nightly jobs to postmerge stage 1399- ci/zink: Move manual/nightly jobs to postmerge stage 1400- ci/llvmpipe: Move manual/nightly jobs to postmerge stage 1401- ci/nouveau: Move manual/nightly jobs to postmerge stage 1402- ci/etnaviv: Move manual/nightly jobs to postmerge stage 1403- ci/panfrost: Move manual/nightly jobs to postmerge stage 1404- ci/run_n_monitor: Add --include-stage argument 1405- ci/run_n_monitor: Add --exclude-stage filtering 1406- ci/run_n_monitor: Remove --force-manual 1407- ci/run_n_monitor: Add --dry-run argument 1408- ci: Make per-build dependencies optional 1409- doc/llvmpipe: Update URL to fix linkcheck 1410- ci/fdno: Add some flaky tests for a618 1411- ci/dozen: Add a couple more skips 1412- ci/windows: Don't upload shader cache to artifacts 1413- ci/{lvp,venus}: Skip massive-alloc tests 1414- intel/tests: Reduce load from anv_tests 1415- util/tests: Reduce load from sparse-array test 1416- ci/angle: Clean up stray open-coded path 1417- ci/lava: Fix path to structured logger 1418- ci/lava: Make it easier to run pytest locally 1419- ci/lava: Don't dump venv into .venv 1420- ci/lava: Allow passing more args to pytest 1421- ci/lava: Fix pytest not passing farm value 1422- ci/lava: Fix fastboot tests 1423- ci/alpine: Fix shellcheck errors 1424- ci: Fix shellcheck warnings in firmware download 1425- ci: Bring back yamllint and shellcheck 1426- ci: Run pytest once for both LAVA and bin/ci 1427- ci: Bring back shader-db and drop LTO 1428- ci/intel-gpu-freq: Fix shellcheck warning 1429- ci/vulkan: Backport dEQP commit to make it less OOMy 1430- ci/lima: Temporarily disable 1431- ci: Add Python requirements include 1432- ci: Add pyutils container 1433- ci/python: Allow empty PYTHONPATH 1434- ci/lava: Use new pyutils container 1435- doc/llvmpipe: Fix broken link syntax 1436- docs/gallium: Fix broken GL extension link 1437- doc/linkcheck: Exclude cgit.fd.o 1438- doc/vc4: Remove reference to old shader-db 1439- ci/windows: Increase dozen-deqp fraction 1440- ci/svga: Add two Piglit flakes seen in the wild 1441- ci/svga: Only run half the Piglit tests 1442- ci/virgl: Special-case llvmpipe parallelisation 1443- ci: Avoid subshell for executing HWCI_TEST_SCRIPT 1444- ci/bare-metal: Don't move structured log file 1445- ci: Always create results dir from init 1446- ci/android: Use common $RESULTS_DIR for cuttlefish 1447- ci/deqp: Use common $RESULTS_DIR 1448- ci/dxvk: Use common results dir 1449- ci/piglit: Use common $RESULTS_DIR 1450- ci/trace: Move trace cleanup to Piglit runner 1451- ci/gtest: Use common $RESULTS_DIR 1452- ci/kdl: Use common $RESULTS_DIR 1453- ci/devcoredump: Use common $RESULTS_DIR 1454- ci/xorg: Capture Xorg log in results artifacts 1455- ci/kdl: Fix KDL install location 1456- ci: Upgrade ci-kdl 1457- ci: Use new arguments to ci-kdl to avoid child management 1458- ci: Make token parsing more robust 1459- ci: Make common rules common 1460- ci: Make setup-test-env safe for set -u 1461- ci/piglit: Make build-piglit safe for set -u 1462- ci/deqp: Make build-deqp safe for set -u 1463- ci/apitrace: Make build-apitrace safe for set -u 1464- ci/va: Make build-va-tools safe for set -u 1465- ci/crosvm: Make build-crosvm safe for set -u 1466- ci/wayland: Make build-wayland safe for set -u 1467- ci/libdrm: Make build-libdrm safe for set -u 1468- ci/windows: Make build-directx-headers safe for set -u 1469- ci/wine: Make build-wine safe for set -u 1470- ci/libclc: Make build-libclc safe for set -u 1471- ci/vulkan: Make build-vulkan-validation safe for set -u 1472- ci/vkd3d: Make Proton build safe for set -u 1473- ci/angle: Make build-angle safe for set -u 1474- ci/deqp-runner: Make build-deqp-runner safe for set -u 1475- ci/skqp: Make build-skqp safe for set -u 1476- ci/skqp: Reorder some things 1477- ci/skqp: Don't hardcode LLVM version 1478- ci/kernel: Make download-prebuilt-kernel safe for set -u 1479- ci/swrast: Fix library paths for llvmpipe/softpipe 1480- ci/swrast: Archive Xvfb logs 1481- ci/zink: Fix library path 1482- ci/virgl: Fix library path 1483- ci/{b2c,swrast,layered}: Strip unnecessary things from runtime containers 1484- egl/x11: Downgrade warning to info 1485- ci: Disable trace jobs using old Wine 1486- ci: Remove non-Proton Wine 1487- ci/vk: Strip and optimise validation layers 1488- ci: Strip yet more unnecessary things from the rootfs 1489- ci/deqp: Flatten fraction/shard sed into a single pass 1490- ci/deqp: Compress caselists with zstd 1491- ci/angle: Update ANGLE, reduce build times 1492- ci: Don't carry static libraries around for tests 1493- ci/shellcheck: Don't overwrite SCRIPTS_DIR with relative path 1494- ci/shellcheck: Don't exit on first failure 1495- ci/python: Fix section emission 1496- ci: Don't move results directory 1497- ci: Don't print structured log data URL 1498- ci/lava: Remove pointless messages 1499- ci/lava: Quieten timestamp warnings from tar 1500- ci/lava: Break section-header print into separate function 1501- ci/lava: Add section for device wait 1502- ci/lava: Rename lava_boot section 1503- ci/lava: Hide more boot details into sections 1504- ci/lava: Explicitly pass UTC timezone 1505- ci/lava: Truncate printed times 1506- ci/lava: Fix colour definitions 1507- ci/lava: Be a little less enthusiastic with bold 1508- ci/lava: Change default section colour to cyan 1509- ci/hw: Wrap pre-test setup in collapsed section 1510- ci/lava: Print relative timestamps in sections 1511- ci/lava: Coalesce post-processed job information 1512- ci/bare-metal: Move preparation steps into section 1513- ci/bare-metal: Use section for boot messages from cros-servo 1514- ci/bare-metal: Truncate printed times and prefixes 1515- ci/bare-metal: Use section for boot messages from poe-powered 1516- ci/crosvm: Make crosvm-runner safe for set +u 1517- ci/crosvm: Replace SCRIPT_DIR with SCRIPTS_DIR 1518- ci/traces: Use sections to quiet execution 1519- ci: Quieten post-test cleanup 1520- ci: Make section emission really quiet 1521- ci: Quieten test execution 1522- ci/lava: Source setup-test-env 1523- ci/mold: Use appropriate parallelism and strip binaries 1524- ci: Really remove Mesa drivers from rootfs 1525- ci/firmware: Allow calling firmware as sourced script 1526- ci/angle: Use relative paths 1527- ci/skqp: Fix relative paths 1528- ci: Use sections for test container builds 1529- ci/skqp: Set $SKQP_ARCH from build-skqp.sh 1530- ci/rpi: Add flakes seen in the wild 1531- ci/fdno: Add flake seen in the wild 1532- ci/lava: Remove duplicate build download 1533- ci/lava: Use LAVA rootfs overlays for build/per-job 1534- ci/lava: Provide LAVA rootfs URL directly 1535- ci/lava: Provide list of overlays to submitter 1536- ci/lava: Refactor is-rootfs-uploaded function 1537- build: Don't run wayland-protocols tests 1538 1539Daniel Svensson (1): 1540 1541- zink: add spirv_info_h dep to libzink. 1542 1543Danylo Piliaiev (20): 1544 1545- tu/a750: Allow mutable images to have UBWC with all compatible formats 1546- tu: Enable UBWC for D24S8 with USAGE_SAMPLED and formatless border color 1547- freedreno: Add enum for fd_lrzfc_layout::dir_track values 1548- tu/u_trace: Add support for u_trace inidirect params 1549- tu/u_trace: dispatch indirect dims and LRZ status as indirect params 1550- freedreno: Clarify RB_BLIT_INFO::TYPE field 1551- tu/a7xx: Use generic clear for LOAD_OP_CLEAR 1552- tu/a7xx: Use generic clear for CmdClearAttachments 1553- tu/a7xx: Use generic clear for image clears 1554- tu/a7xx: Use BLIT_EVENT_STORE_AND_CLEAR when appropriate 1555- tu: Fix ZPASS_DONE having stale value with generic clear+early frag test 1556- ir3: Respect aligment of required consts when allocating ubo,preamble 1557- u_trace: Fix trace_payload_as_extra_func desync between drivers 1558- tu/a750: Workaround GPU fault when fast-clearing R8G8 formats 1559- freedreno/rnndec: Consider array length when finding by reg name 1560- freedreno/cffdec: Add option to dump bindless descriptors 1561- nir/opt_16b_tex_image: Sign extension should matter for texel buffer txf 1562- util/vma: Fix util_vma_heap_get_max_free_continuous_size calculation 1563- ir3: Fix cat5 parsing with a1.x src present 1564- nir/nir_opt_offsets: Do not fold load/store with const offset > max 1565 1566Dave Airlie (27): 1567 1568- draw/orcjit: supply stub function for tcs coro 1569- llvmpipe/cs/orcjit: add stub function name for coro 1570- gallivm/sample: fix sampling indirect from vertex shaders 1571- nvc0: fix null ptr deref on fermi due to debug changes. 1572- anv/video: add mode costs for h264 encoding 1573- vulkan/video: handle KHR_video_maintenance1 lack of profile list. 1574- radv/video: handle inline queries for vulkan video encode. 1575- radv/video: add KHR_video_maintenance1 support 1576- radv/video: fix reporting video format props for encode. 1577- radv: move video decode enable test into a flag 1578- radv: turn video decode/encode on for VCN4 with latest fw 1579- llvmpipe: handle stride properly on lvp udmabuf imports 1580- llvmpipe: make sure to duplicate the fd handle before giving out 1581- vulkan/video: fix vui encoding 1582- vl/bitstream: use an int32_t for se encoding. 1583- radv/anv/video: handling encoding both sps and pps in same buffer 1584- radv: Fix radeon_enc_code_ue with values over 2^16 1585- radv/video: fix encode reference slot counting 1586- radv/video/enc: report pps overrides in feedback for h265 1587- radv/video: use the h264 defines for macroblock w/h 1588- radv/video/enc: report pictureAccessGranularity of CTB size. 1589- radv/video: add encode field for vcn4 1590- radv/video: handle missing h265 feedback struct. 1591- zink: free modifiers 1592- zink: handle mutable check without crashing if format list not found 1593- zink: whitespace fixups 1594- va: add some h264 fields 1595 1596David Heidelberg (66): 1597 1598- freedreno: Enable Adreno 306A 1599- freedreno: support lua pkgconfig variants 1600- freedreno: drop hack for old dEQP 1601- freedreno: Introduce emit_blit_fini 1602- freedreno: introduce color format parameter to emit_blit_buffer_dst 1603- freedreno: Implement fast clear_buffer for Adreno 6xx and 7xx 1604- freedreno: drop dEQP hacks 1605- intel/genxml: fix length of HCP_FQM_STATE for gen20 and 125 1606- llvmpipe: Silence "possibly uninitialized value" warning for ssbo_limit (cont) 1607- meson/addrlib: allow unintialized callbacks 1608- ci/alpine: use llvm variables 1609- ci/alpine: build test with LLVM 18 1610- ci/alpine: use packaged spirv-llvm-translator-dev 1611- ci/alpine: enable ORCJIT 1612- ci/alpine: libclc on Alpine offering necessary targets spirv-mesa3d-;spirv64-mesa3d- 1613- freedreno: Implement get_compute_state_info for Adreno 6xx/7xx 1614- freedreno: update night expectations 1615- etnaviv: rename enums_h appropriately 1616- etnaviv: build dependency for the etnaviv tests 1617- freedreno: Convert to SPDX-License-Identifier instead of pasting whole license 1618- freedreno: use unicode © instead of DOS-like (C) copyright sign 1619- bin/gen_release_notes: adjust checksums section 1620- ci/freedreno: add Adreno 306/307 occasional timeout 1621- ci/iris: implement nightly CL testing using piglit on ADL 1622- panfrost: drop leftover definition after pan_nir_lower_64bit_intrin removal 1623- nir_lower_mem_access_bit_sizes: Assert when 0 components or bits are requested 1624- freedreno/ir3: Use nir_lower_mem_access_bit_sizes instead custom lowering 1625- ci/intel: Officially switch intel-adl-cl to pre-merge 1626- ci/freedreno: move disabled a530 entries back to main gitlab-ci.yml 1627- ci/freedreno: mark jobs to be retested with patched 6.11 kernel 1628- ci/freedreno: re-enable piglit tests for Adreno 530 in nightly runs 1629- ci/freedreno: mark a530_piglit as postmerge job 1630- panfrost/midgard: Implement nir_lower_mem_access_bit_sizes pass 1631- panfrost/midgard: Support 64bit pack/unpack 1632- panfrost/midgard: Lower fisnormal 1633- panfrost/midgard: Lower \*hadd instructions 1634- ci/panfrost: Implement basic nightly OpenCL testing on T860 1635- ci/panfrost: switch to 6.11 with timestamp patches for traces 1636- ci/panfrost: update furmark checksum for T860 traces 1637- ci/panfrost: update rock-5b from rc5 to rc7 1638- freedreno/ir3: mad.x24 is not safe to lower 1639- freedreno/ir3: Do not allow 16-bit mad.x24 1640- pvr: unify code into one condition 1641- ci/piglit: keep the include_test.h used for OpenCL testing 1642- ci/fedora: when doing non-LTO build, we need C_ARGS too 1643- ci: enable IPO/LTO only for nightly builds 1644- ci: restrict number of maximum threads 1645- util: add aarch64 fast half-float to float conversion 1646- amd: Pass addrlib cpp args to the tests 1647- osmesa: Fix OSMesaGetDepthBuffer() test without depth 1648- ci/etnaviv: unify job naming with the rest of the CI 1649- ci/freedreno: switch to dash instead of underscore, same as rest of the CI 1650- ci/alpine: stay trendy, Alpine flies on LLVM 19 now 1651- build: Fix LTO with Android NDK 1652- ci/build: debian-build-testing should follow same limit as other jobs 1653- ci/fedora: add workaround for buggy GCC warning triggered by r600 1654- ci/build: Do not do LTO on any pre-merge job 1655- freedreno/ir3: Lower fisnormal op 1656- ci/freedreno: Introduce OpenCL testing for Adreno 618, 660, and 750 1657- freedreno: python fixes 1658- util: Drop 3Dnow optimisation leftovers 1659- util: Remove MMX/MMXext detection code 1660- util: Drop ancient Intel CPU detection 1661- util: drop XOP detection code 1662- llvmpipe: align with u_cpu_detect struct changes 1663- compiler/rust: drop duplicated bindgen check 1664 1665David Rosca (147): 1666 1667- radeonsi/vcn: Only enable filler data for CBR 1668- radeonsi/vcn: Bump enc FW interface version for VCN3 and VCN4 1669- radeonsi/vcn: Add support for QVBR rate control mode 1670- radeonsi/vcn: Fix MB/CTB calculation for slice encoding 1671- dri: Use --whole-archive when linking libdri 1672- radeonsi/vcn: Allow dynamic change of QVBR quality level 1673- radeonsi/vcn: Add decode DPB buffers as CS dependency 1674- gallium: Add header_flags to pipe_h2645_enc_picture_desc 1675- frontends/va: Don't check header type for packed header buffers 1676- frontends/va: Keep track if VPS/SPS/PPS/AUD was sent 1677- frontends/omx: Request SPS PPS for IDR pictures 1678- radeonsi/vcn: Only insert headers when requested for H264/5 1679- frontends/va: Fix use after free with EFC 1680- radeonsi: Don't allow DCC for encode in is_video_target_buffer_supported 1681- vulkan/format: Add VK_FORMAT_G12X4_B12X4R12X4_2PLANE_420_UNORM_3PACK16 1682- radv/video: Add support for 12-bit AV1 decode 1683- gallium: Change pipe_h265_enc_rate_control to array 1684- frontends/va: Assert maximum number of temporal layers 1685- frontends/va: Get per temporal layer params for HEVC 1686- frontends/va: Parse packed header slice for HEVC TemporalId 1687- frontends/va: Fix leaks with multiple coded buffer segments 1688- util/rbsp: Fill bits again if reading more than 16 leading zeros 1689- frontends/va: Implement DPB management for H264/5 encode 1690- frontends/va: Only use parameters for first slice in H264/5 encode 1691- frontends/va: Parse H264 slice packed header 1692- frontends/va: Parse HEVC slice packed header 1693- frontends/omx: Adjust to encoding changes 1694- radeonsi: Add GPU copy path to si_video_resize_buffer 1695- radeonsi/vcn: Switch to app DPB management for H264 and HEVC encode 1696- radeonsi/vcn: Deduplicate header encoding functions 1697- radeonsi/vcn: Fix radeon_enc_code_ue with values over 2^16 1698- radeonsi/vcn: Use pipe header params in HEVC header encoder 1699- radeonsi/vcn: Use pipe header params in H264 header encoder 1700- radeonsi: Enable transform_skip HEVC feature support 1701- radeonsi: Implement buf_ofst_info in si_video_resize_buffer GPU copy path 1702- radeonsi/vcn: Support H264 constrained intra prediction 1703- d3d12: Remove get_feedback_fence implementation 1704- frontends/va: Implement sync buffer/surface timeout for encode feedback 1705- pipe: Add int return value to video end_frame 1706- frontends/va: Check for errors from end_frame in vlVaEndPicture 1707- radeonsi/vcn: Implement get_feedback_fence 1708- pipe: Use correct struct for h264/h265 dpb entry 1709- frontends/va: Allocate surface buffers on demand 1710- frontends/va: Allow drivers to allocate and use encode DPB surface buffers 1711- egl/wayland: Fix flush after blit in swap buffers multi-GPU case 1712- frontends/va: Fix locking in vlVaDeriveImage 1713- frontends/va: Fix locking in vlVaQueryVideoProcPipelineCaps 1714- vulkan/wsi: Fix stack-use-after-scope in x11_surface_create_swapchain 1715- frontends/va: Parse more H264 PPS for transform_8x8_mode_flag 1716- radeonsi/vcn: Stop hardcoding values in VCN version overrides 1717- radeonsi/vcn: Add support for H264 8x8 transform on VCN5 1718- radv/video: Don't use op_close in encode 1719- radv/video: Don't use multiple tasks for one encode operation 1720- ci: Disable building OMX driver 1721- Remove OMX 1722- docs: Remove OMX frontend 1723- glx: Destroy base screen after deinitScreen 1724- pipe: Add pipe_h264_nal_unit_type and pipe_h265_nal_unit_type enums 1725- frontends/va: Support raw packed headers for H264 and HEVC 1726- frontends/va: Allow multiple NALUs in coded buffer segments 1727- radeonsi/vcn: Support raw packed headers for H264 and HEVC 1728- radeonsi: Remove unused radeon_temporal.h 1729- radeonsi/vcn: Don't copy the pipe enc structs 1730- pipe: Remove video enc header_flags 1731- meson/megadriver: Add megadriver_libdir argument 1732- targets/va: Build va driver into libgallium when building with dri 1733- targets/vdpau: Build vdpau driver into libgallium when building with dri 1734- radeonsi/vce,uvd_enc: Stop using obsolete rate control params 1735- radeonsi/vce: Stop using base.level and base.max_references 1736- radeonsi/uvd_enc: Stop using base.level 1737- radeonsi/vcn: Stop using base.level for H264 level_idc 1738- d3d12: Stop using base.level for H264 level_idc 1739- frontends/va: Set default encoding parameters at context creation 1740- Revert "frontends/va: Process VAEncSequenceParameterBufferType first in vaRenderPicture" 1741- frontends/va: Create encoder at context creation 1742- pipe: Remove unused fields in video rate control 1743- frontends/va: Reset intra refresh in beginPicture 1744- radeonsi: Fix reported minimum width for video encode 1745- frontends/va: Remove fallback if app doesn't enable packed headers 1746- radeonsi: Disable EFC on VCN 2.2 1747- radeonsi/vcn: Add rc_per_pic_ex encode command 1748- ac: Fix typo RENCDOE -> RENCODE 1749- ac: Add ac_vcn_init_enc_cmds 1750- ac: Add remaining VCN encode defines 1751- radeonsi/vcn: Use ac_vcn_enc_init_cmds and AV1 defines from ac 1752- ac: Add VCN IB parser 1753- radeonsi/vcn: Enable IB parsing with AMD_DEBUG=ib 1754- radeonsi/vcn: Don't reuse context with multiple VCN instances 1755- frontends/va: Fix AV1 packed header parsing 1756- pipe: Add video fence_wait 1757- radeonsi/uvd: Implement fence_wait 1758- radeonsi/vcn: Implement fence_wait 1759- radeonsi/vpe: Implement fence_wait 1760- d3d12: Implement fence_wait 1761- r600/uvd: Implement fence_wait 1762- frontends/va: Use fence_wait instead of get_*_fence 1763- pipe: Remove video get_*_fence 1764- radeonsi/vcn: Allow per-frame QP and max frame size changes 1765- radeonsi/vcn: Use correct initial DPB size for HEVC encode 1766- radeonsi/vcn: Don't hardcode nal_ref_idc 1767- radeonsi/vcn: Fix some small coverity issues 1768- radeonsi/vcn: Fix out of bounds write when invalidating QP map regions 1769- radeonsi/vcn: Fix out of bounds read in H264 decode 1770- frontends/va: Implement DPB management for AV1 1771- frontends/va: Parse more AV1 frame and sequence header params 1772- frontends/va: Support raw packed headers for AV1 1773- radeonsi/vcn: Switch to app DPB management for AV1 1774- frontends/va: Remove now unused ref fields for AV1 encode 1775- radeonsi/vcn: Support raw packed headers for AV1 1776- radeonsi/vcn: Fix coded size in sequence header on VCN5 1777- radeonsi/vcn: Add support for VCN5 AV1 compound 1778- radeonsi/vcn: Add support for VCN5 dpb tier2 1779- frontends/va: Fix parsing HEVC st_ref_pic_set 1780- frontends/va: Parse H264 prefix for temporal_id 1781- frontends/va: Calculate VBV params for other layers relative to base layer 1782- radeonsi/vcn: Implement per-layer rate control for HEVC 1783- radeonsi/vcn: Select layer for rate control 1784- frontends/va: Fix some small coverity issues 1785- frontends/va: Fix NULL check in intra refresh params 1786- frontends/va: Fix parsing leb128 when using more than 4 bytes 1787- frontends/va: Don't relock mutex in vlVaEndPicture 1788- pipe: Fix video enc max DPB size for H264 1789- frontends/va: Evict unused surfaces from encode DPB 1790- radeonsi/vcn: Only enable skip mode with matching references 1791- frontends/va: Add support for AV1 still picture encode 1792- radeonsi/vcn: Add support for AV1 still picture encode 1793- radeonsi/uvd_enc: Implement fence_wait 1794- radeonsi/vce: Implement fence_wait 1795- radeonsi: Enable VIDEO_CAP_ENC_SUPPORTS_ASYNC_OPERATION on VCE/UVD 1796- radeonsi/uvd_enc: Allocate session buffer in VRAM 1797- radeonsi: Support HEVC features and block sizes for UVD 1798- radeonsi/vcn: Enable VCN4 AV1 encode WA 1799- radeonsi/vce: Remove dual instance support 1800- frontends/va: Remove H264 encode delayed flush 1801- frontends/va: Stop using PIPE_VIDEO_CAP_ENC_SUPPORTS_ASYNC_OPERATION 1802- pipe: Remove PIPE_VIDEO_CAP_ENC_SUPPORTS_ASYNC_OPERATION 1803- frontends/va: Move encode fence to coded buffer 1804- radeonsi/vcn: Don't try to override HEVC SPS conformance window 1805- radeonsi/vcn: Gracefully handle encode errors and report to frontend 1806- radeonsi/vcn: Rework encode padding and session init params 1807- radv/video: Fix H264 slice control 1808- radv/video: Fix HEVC slice control 1809- radv/video: Report correct encodeInputPictureGranularity 1810- radv/video: Avoid selecting rc layer over maximum 1811- radeonsi/vcn: Use correct frame context buffer for preencode on VCN5 1812- frontends/va: Create surfaces with correct fourcc for RT format 1813- frontends/va: Use compositor blit with different number of planes 1814 1815Deborah Brouwer (15): 1816 1817- ci/lava: Detect a6xx gpu recovery failures 1818- ci: add firmware files to rootfs 1819- ci/intel: add i915/MTL firmware to rootfs 1820- ci: remove xfail program\@build\@include-directories 1821- ci: remove obsolete build-hang-detection script 1822- ci/deqp-runner: fix option to build from a commit 1823- ci/deqp-runner: print version info 1824- ci/deqp-runner: update instructions for uprevving 1825- ci/deqp-runner: uprev from 0.20.0 to 0.20.2 1826- ci/deqp-runner: implement max-fails 1827- freedreno/ci: remove redundant skip files 1828- ci: simplify .baremetal-arm32-asan-test 1829- ci: simplify .baremetal-arm64-asan-test 1830- freedreno/ci: convert a630-gles-asan to toml suite 1831- ci/b2c: update RESULTS_DIR for .b2c-test jobs 1832 1833Denis (1): 1834 1835- freedreno: Enable A505 1836 1837Derek Foreman (5): 1838 1839- meson: Update wayland-protocols to 1.38 1840- vulkan/wsi/wayland: Use presentation timing v2 when available 1841- vulkan/wsi/wayland: Use fifo protocol for FIFO 1842- vulkan/wsi/wayland: Pace frames with commit-timing-v1 1843- adv+zink/ci: Add a recent flake 1844 1845Dmitry Baryshkov (3): 1846 1847- freedreno/ir3: Treat MESA_SHADER_KERNEL in the same way as compute 1848- freedreno/ir3: Lower the hadd operations 1849- rusticl: Enable on freedreno 1850 1851Dmitry Osipenko (6): 1852 1853- util/mesa-db: Fix missing O_CLOEXEC 1854- util/mesa-db-multipart: Open one cache part at a time 1855- util/mesa-db: Open DB files during access time 1856- util/mesa-db: Fix crash on compacting empty DB 1857- util/cache_test: Add mesa-db test for adding cache entry bigger than empty cache 1858- ci/zink: Mark glx-multithread-clearbuffer flake on ADL 1859 1860Donald Robson (2): 1861 1862- pvr: Stop creating waits when there are no waits 1863- pvr: Fixed creation of waits in queue submission 1864 1865Dylan Baker (35): 1866 1867- compilers/clc: Add missing break statements. 1868- mesa: fix memory leak when using shader cache 1869- util/glsl2spirv: fixup the generated depfile when copying sources 1870- util/glsl2spirv: only copy the input file if we're going to change it 1871- tgsi_to_nir: free disk cache value if the size is wrong 1872- crocus: properly free resources on BO allocation failure 1873- crocus: check for depth+stencil before creating resource 1874- mesa/program: remove memcpy where src == dst 1875- docs: add release notes for 24.2.2 1876- docs: update sha sums for 24.2.2 1877- docs: update sha sums for 24.2.2 1878- iris: Run checks that do not require resources before creating them 1879- anv: if queue is NULL in vm_bind return early 1880- intel: replace \`(uint64_t - uint64_t) > 0` with \`uint64_t > uint64_t` 1881- anv: assert we don't write past the end of an array 1882- anv: enforce state->cmd_buffer is never null in emit_Simpler_shader_init_fragment 1883- intel/elk: delete copy constructor and copy-assignment-operator 1884- compiler/glsl: explicitly delete copy constructor and copy-assign 1885- util/fossilize_db: don't use check_files_opened_successfully for foz_prepare 1886- anv: remove useless uint >= 0 check 1887- hasvk: remove useless uint >= 0 check 1888- intel/perf: delete dead code. 1889- meson: use \`add_project_arguments` instead of \`global` 1890- anv: enable VK_EXT_ycbcr_2plane_444_formats 1891- release: push 24.3 out two weeks 1892- anv: assert that we don't read off the end color_att array 1893- anv: when handling descriptor sets, layout cannot be NULL if dynamic_offets are used 1894- VERSION: bump for 24.3.0-rc1 release 1895- .pick_status.json: Update to ced2404cb433deaa84cf6cf9edce212733117c0b 1896- .pick_status.json: Update to 5e0b81413de588803c9a0736acd8decd40d19ab6 1897- .pick_status.json: Update to b0c9789bc1ed808d29f642e9445599dc85896444 1898- VERSION: bump for 24.3.0-rc2 release 1899- .pick_status.json: Update to b8d253b30c0a8c1878f2429f0a4df0a068bbf0aa 1900- .pick_status.json: Update to 8845255881f03df9a695fae164a64201c4211d35 1901- .pick_status.json: Update to 9b779068c3a7a8ded00f1d2519fd01dd1d05f3d8 1902 1903Echo J (3): 1904 1905- d3d10umd: Use pipe_resource_usage enum in translate_resource_usage() 1906- util: Fix the integer addition in os_time_get_absolute_timeout() 1907- compiler/rust: Force native build for the compiler_proc library 1908 1909Eric Engestrom (249): 1910 1911- VERSION: bump to 24.3 1912- docs: reset new_features.txt 1913- loader: gc loader_get_extensions_name() and __DRI_DRIVER_{GET\_,}EXTENSIONS defines 1914- radeonsi/ci: skip timing out test 1915- freedreno/ci: double job timeout for a306 1916- freedreno/ci: document extra variants of failing tests on a618 and a630 1917- anv+zink/ci: mark some tests as fixed 1918- anv+zink/ci: document two tests, one failing and one crashing 1919- anv+zink/ci: mark a couple of tests as flaky 1920- venus/ci: skip timing out test 1921- nak: fix meson typo 1922- venus: initialize bitset in CreateDescriptorPool() 1923- v3d/ci: mark spec\@amd_performance_monitor\@vc4 tests as flaky 1924- meson: xcb & xcb-randr are needed by the loader whenever x11 is built 1925- docs: add stub header for u_format_gen.h 1926- v3d,v3dv: reuse dep_v3d_hw from simulator/meson.build 1927- v3d,v3dv: figure out whether we're using the simulator only once 1928- v3d,v3dv: simplify USE_V3D_SIMULATOR/using_v3d_simulator logic 1929- vc4,v3d,v3dv: avoid compiling in unused \`sim_file` field in the {vc4,v3d}_screen/v3dv_physical_device struct 1930- ci/baremetal: fix logic for retrying boot when it failed 1931- v3d/ci: mark spec\@amd_performance_monitor\@api as flaky on rpi4 1932- v3d/ci: mark glx\@glx-visuals-depth as flaky on rpi4, just like the -stencil variant 1933- meson: don't select the deprecated \`swrast` option ourselves 1934- meson: improve wording of "incompatible llvm options" error 1935- ci: remove llvmpipe in the job that disables llvm 1936- rpi4/ci: mark fs variant of arb_texture_buffer_object as flaky too 1937- docs/meson: add missing newline at the begining of code-block 1938- bin/ci: update python-gitlab to 4.x 1939- ci/prepare-artifacts: use find's -exec instead of iterating over its output with xargs 1940- ci/prepare-artifacts: check if we're about to overwrite ci config/expectations files 1941- llvmpipe/ci: bump timeout for nightly job to 1h 1942- ci: drop dead VKD3D_CONFIG 1943- ci/vkd3d: rename vkd3d runner script to vkd3d-runner.sh to match the other ones 1944- ci/vkd3d: store vkd3d "errors" in the results file instead of spamming the job log 1945- ci/vkd3d: use GPU_VERSION to identify the list of failures, unifying with deqp-runner.sh 1946- ci/vkd3d: move test expectations file to the results folder 1947- ci/vkd3d: move expectation file creation logic out of the test results handling 1948- ci/vkd3d: be more specific for what to grep 1949- ci/vkd3d: rename vkd3d expected failures list to be more explicit 1950- ci/vkd3d: add support for \*-skips.txt list files 1951- ci/vkd3d: add support for \*-flakes.txt list files 1952- pick-ui: move NominationType values around so that None is 0 1953- ci/poe-powered: rename --test-timeout to --test-timeout-minutes to be more explicit 1954- ci/bare-metal: rename fastboot & cros-servo TEST_PHASE_TIMEOUT to TEST_PHASE_TIMEOUT_MINUTES to be coherent 1955- ci/poe-powered: allow setting a different boot timeout than 5 minutes 1956- rpi/ci: shorten the boot timeout to 90 seconds 1957- docs: add release notes for 24.1.5 1958- docs: update calendar for 24.1.5 1959- docs: add sha256sum for 24.1.5 1960- nvk/ci: bump vkcts job timeout 1961- meson,ci: remove dead \`kmsro` option in \`gallium-drivers` 1962- ci: remove dead start-x.sh script 1963- rpi3/ci: remove {version,render}_check from [piglit.env] 1964- ci/deqp-runner: fix --fraction-start in non-parallel jobs 1965- ci: pass MESA_SPIRV_LOG_LEVEL from job to the test 1966- android: fix build in multiple ways 1967- docs: update calendar for 24.2 1968- docs: add release notes for 24.2.0 1969- docs: add sha256sum for 24.2.0 1970- docs: update calendar for 24.1.6 1971- docs: add release notes for 24.1.6 1972- docs: add sha256sum for 24.1.6 1973- ci/container: define LLVM_VERSION in the alpine container job 1974- ci/build: reuse alpine llvm version to make sure it stays coherent 1975- ci/build: add comment to explain why s390x is useful 1976- ci/build: document \@ajax as a stakeholder for s390x 1977- ci/build: drop softpipe from s390x job 1978- nvk/ci: add vkd3d job on the ga106 1979- llvmpipe/ci: mark now-skipping test as no longer failing 1980- nvk+zink/ci: mark \`spec\@arb_sample_shading\@samplemask .*\@0\\..* partition` as fixed 1981- nvk/ci: document regressions 1982- turnip/ci: document regressions 1983- turnip/ci: skip timing out tests 1984- turnip/ci: document flakes 1985- nvk/ci: mark a bunch of tests as fixed 1986- turnip/ci: add a750_vk flakes 1987- docs/envvars: fix reference 1988- broadcom/ci: set default TEST_PHASE_TIMEOUT_MINUTES so that jobs never timeout without uploading results 1989- broadcom/ci: increase the timeout of the nightly job v3d-rpi4-gl:arm32 by 10min 1990- ci/build: add honeykrisp (asahi vk) to several builds 1991- intel/ci: don't trigger anv-jsl-full & anv-tgl-full on GL changes 1992- nvk/ci: document CTS bug 1993- ci: add check for misleading indentation in ci toml files 1994- ci: don't run hardware tests if the toml files don't pass validation 1995- ci: fix toml-lint rules to not block nightly pipelines 1996- r300/ci: move postmerge jobs to -postmerge stage 1997- lavapipe/ci: move postmerge jobs to -postmerge stage 1998- venus/ci: move postmerge jobs to -postmerge stage 1999- ci/rules: make every job exist as manual in fork pipelines 2000- kopper: only include loader_dri3_helper.h when xcb is enabled 2001- zink+nvk/ci: document new variant of test failing 2002- zink+nvk/ci: document all the flakes seen in the last week 2003- turnip/ci: document all the a750 flakes seen in the last week 2004- vc4: Add missing libvc4_neon build dependencies 2005- vc4/meson: simplify neon build now that the android build system doesn't exist anymore 2006- ci/deqp-runner: restore CC after temporarily overriding it 2007- ci/deqp: simplify command to list local deqp patches 2008- ci/deqp-runner: be less verbose in the loop printing the deqp builds info 2009- ci/deqp-runner: build from git checkout even on linux 2010- ci/deqp-runner: add infra to apply patches 2011- ci/image-tags: re-generate all the images building deqp-runner 2012- ci/build: fix ppc64le and s390x jobs rules 2013- ci: take igalia farm offline 2014- docs: add release notes for 24.2.1 2015- docs: add sha sum for 24.2.1 2016- docs: update calendar for 24.2.1 2017- nvk/ci: drop skip of fixed vkd3d test 2018- ci/vkd3d: make the job output less spammy, focus on what is useful to print 2019- ci/vkd3d: use upstream test-runner.sh for process isolation 2020- ci: run only one vkd3d test at a time 2021- ci/vkd3d: check for tests listed in fails but not actually failing 2022- v3dv/ci: fix test timeout for v3dv-rpi5-vk-full:arm64 2023- etnaviv/ci: fix gc2000_piglit test timeout 2024- freedreno/ci: fix a630_*_full tests timeout 2025- nouveau/ci: fix gk20a-gles test timeout 2026- freedreno/ci: drop unused TEST_PHASE_TIMEOUT_MINUTES in LAVA jobs 2027- ci/bare-metal/poe: change the default TEST_PHASE_TIMEOUT_MINUTES to \`CI_JOB_TIMEOUT - 5min` 2028- broadcom/ci: drop TEST_PHASE_TIMEOUT_MINUTES that match the default value 2029- broadcom/ci: use TEST_SETUP_AND_UPLOAD_MARGIN_MINUTES 2030- etnaviv/ci: drop TEST_PHASE_TIMEOUT_MINUTES that match the default value 2031- freedreno/ci: drop TEST_PHASE_TIMEOUT_MINUTES that match the default value 2032- nouveau/ci: drop TEST_PHASE_TIMEOUT_MINUTES that match the default value 2033- ci/build: untangle debian-s390x from debian-ppc64el 2034- freedreno/ci: add a630 flake 2035- docs: add release notes for 24.1.7 2036- docs: update calendar for 24.1.7 2037- docs: add sha sum for 24.1.7 2038- ci: move build jobs that are used by tests to an earlier stage 2039- ci: merge build-x86_64 & build-misc 2040- ci: do not start build-only jobs until the critical build-for-tests jobs are done 2041- nvk/ci: mark -dEQP-VK.drm_format_modifiers.export_import* as fixed 2042- radeonsi/ci: bump timeout for nightly job glcts-vangogh-valve 2043- Revert "freedreno/ci: drop TEST_PHASE_TIMEOUT_MINUTES that match the default value" 2044- freedreno/ci: fix test timeout for a306_piglit 2045- radeonsi/ci: mark a bunch of subgroups tests as failing 2046- venus/ci: drop redundant flakes definitions 2047- venus/ci: add flake and skip timing out test 2048- zink+nvk/ci: bump zink-nvk-ga106-valve timeout as more tests are being run 2049- zink+nvk/ci: mark a ton of tests as fixed 2050- zink+turnip/ci: fix .zink-turnip-valve-manual-rules 2051- turnip/ci: add vkd3d job on the a750 2052- ci: add virgl & venus to the s390 build 2053- docs: Fix linkcheck failures 2054- zink+nvk/ci: mark test as fixed 2055- zink+nvk/ci: add flakes seen recently 2056- docs: add release notes for 24.2.3 2057- docs: update calendar for 24.2.3 2058- docs: add sha sum for 24.2.3 2059- amd/ci: track changes to the global driver \`*-skips.txt` files 2060- broadcom/ci: track changes to the global driver \`*-skips.txt` files 2061- llvmpipe/ci: track changes to the global driver \`*-skips.txt` files 2062- lavapipe/ci: track changes to the global driver \`*-skips.txt` files 2063- intel/ci: track changes to the global driver \`*-skips.txt` files 2064- v3dv/ci: drop skip of fixed test 2065- docs/release-calendar: add 24.3 branchpoint and rcs 2066- zink+nvk/ci: document spec\@egl_ext_surface_compression\@create as crashing 2067- zink+nvk/ci: add flakes seen recently 2068- lavapipe/ci: document regression 2069- llvmpipe/ci: document regression 2070- radeonsi/ci: document spec\@egl_ext_surface_compression\@create as crashing 2071- radeonsi/ci: mark KHR-GL46.shader_image_load_store.basic-allTargets-atomic as fixed 2072- ci/build: move debian-clang-release to priority build-for-tests stage 2073- lavapipe/ci: document ray tracing regression 2074- mr-label-maker: add gfxstream 2075- ci/deqp: make sure the git commit hash of deqp is stable across rebuilds 2076- ci: fix windows container jobs being missing on forks 2077- ci/build: don't wait on the windows build anymore 2078- turnip/ci: mark vkd3d test_planar_video_formats as fixed 2079- zink+nvk/ci: mark spec\@egl 1.4\@egl-ext_egl_image_storage as fixed 2080- zink+nvk/ci: document more flakes seen lately 2081- egl: fix dri2_from_names() call 2082- docs: add release notes for 24.2.4 2083- docs: update calendar for 24.2.4 2084- docs: add sha sum for 24.2.4 2085- Revert "ci/build: move debian-clang-release to priority build-for-tests stage" 2086- ci/build: move debian-build-testing to priority build-for-tests stage 2087- ci/build: mention the .build-for-tests-jobs exception 2088- docs: add release notes for 24.2.5 2089- docs: update calendar for 24.2.5 2090- docs: add sha sum for 24.2.5 2091- ci: drop spec\@egl_ext_surface_compression\@create from the rest of the expected fails 2092- zink+nvk/ci: document regression 2093- zink+nvk/ci: document more flakes seen 2094- zink+radv/ci: document more flakes seen 2095- nvk/ci: document more flakes seen 2096- ci: run shader-db only when one of the drivers it tests is modified 2097- zink+nvk/ci: mark a few tests as fixed 2098- zink+nvk/ci: document more flakes seen 2099- ci/{deqp,piglit}-runner: make zstd respect FDO_CI_CONCURRENT when set, and fallback to the current "all the threads" 2100- ci/prepare-artifacts: use more than one thread when compressing the install tarball 2101- freedreno/ci: add {freedreno,turnip}{,-manual}-rules to simplify the rest of the code 2102- freedreno/ci: use {freedreno,turnip}{,-manual}-rules 2103- ci: rename "freedreno" farm to "google-freedreno" 2104- {freedreno,intel}/ci: rename "premerge-skips.txt" to "merge-skips.txt" to accurately reflect what they are 2105- {freedreno,intel}/ci: add missing tracking of \`merge-skips.txt` files 2106- ci: simplify ci-tron priority variable 2107- ci: simplify setting lava job priority 2108- ci: fix shader-db rules 2109- ci/b2c: allow B2C_DTB_URL to be set to an empty string to unset it 2110- ci/b2c: make initramfs & kernel urls optional 2111- ci: move yaml-toml-shell-test out of build-for-tests stage 2112- ci: move comment into the appropriate section 2113- zink+turnip/ci: update a750 results 2114- docs: update link to \`perf` wiki 2115- nvk/ci: add back a crash that was mistakenly removed from the expectations 2116- zink+nvk/ci: document ext_egl_image_storage regression in c06a55fd...2fb4aed9 (likely !31585) 2117- all-drivers/ci: drop duplicate flakes lines 2118- zink+nvk/ci: drop duplicate flakes lines 2119- zink+nvk/ci: add flakes seen recently 2120- turnip/ci: add more dEQP-VK.renderpass2.fragment_density_map.* flakes seen recently 2121- nvk/ci: add flakes seen recently 2122- lavapipe/ci: group & sort fails 2123- lavapipe/ci: group & sort skips 2124- lavapipe/ci: skip builtin ray query tests that take too long and time out 2125- lavapipe/ci: document regression in the commit range 765d1c47...366f63fd 2126- ci: move shellcheck options to .shellcheckrc 2127- ci: rename "merge-skips" to "slow-skips" as they're about to be used outside of merge piplines 2128- ci: skip slow tests on all non-"full" jobs 2129- ci: drop dead "load jwt in the environment just before exiting the job" code 2130- egl: error out during setup if the configuration is invalid 2131- freedreno/ci: abort a750 testing when a hang is detected 2132- freedreno/ci: add nightly freedreno gl testing on a750 2133- ci/build: deduplicate meson build command 2134- ci/build: handle build timeout ourselves to give more time to nightly LTO builds 2135- ci/build: lower most job's timeout from 30 to 15 min 2136- docs: add release notes for 24.2.6 2137- docs: update calendar for 24.2.6 2138- docs: add sha sum for 24.2.6 2139- ci: only end current section if there is a current section 2140- ci: initialize state_x (tracking \`set -x`) 2141- ci: consistently restore \`-x` after temporarily disabling it 2142- ci: rename state_x to previous_state_x to be clear about what it contains 2143- ci/b2c: properly escape double-quoted container command 2144- ci/b2c: make dut boot quieter on ci-tron jobs 2145- ci: properly quote ccache dir and path 2146- ci: don't replace ld with ld.gold if ld.gold doesn't exist 2147- ci: bump ci-templates 2148- amd/ci: document flakes seen lately 2149- nvk/ci: document flakes seen lately 2150- freedreno/ci: document flakes seen lately 2151- broadcom/ci: document flakes seen lately 2152- meson: drop variable initialized twice 2153- meson: add dependencies needed by wsi_common_x11.c even on non-drm platforms 2154- ci: drop unused extra args in build-vkd3d-proton.sh 2155- ci: replace plain \`meson` with explicit \`meson setup` 2156- meson: bump spirv-tools version needed to v2022.1 2157- ci: move error handling functions at the end 2158- ci: use quiet alias for commands 2159- ci: raise priority of release manager pipelines 2160 2161Eric R. Smith (19): 2162 2163- panfrost: make PAN_MESA_DEBUG=dump really show all GPU memory 2164- panfrost: fix texture.border_clamp regression for valhall 2165- panfrost: use RGB1 component ordering for R5G6B5 pixel formats 2166- panfrost: Make the unit to use for instructions explicit 2167- panfrost: Update bifrost_isa.py to handle some Valhall constructs 2168- panfrost: refactor valhall ISA parser 2169- panfrost: split pseudo instructions from Bifrost and Valhall 2170- panfrost: fix crash when disassembling shader for trace 2171- panfrost: Add b2i16 conversion to table 2172- panfrost: move pan_arch definition out of genxml to pan_props.h 2173- panfrost: move lower_image_ms to after lower_atomics 2174- panfrost: implement image2dms support for valhall 2175- panfrost: add support for image2DMSArray on bifrost 2176- panfrost: improve swizzle inversion 2177- panfrost: print human readable versions of some swizzle fields 2178- panfrost: Add back A8_UNORM format for valhall 2179- panfrost: fix SAMPLE_ALPHA_TO_ONE 2180- panvk: Support blend shaders for alpha_to_one_enable 2181- panfrost: fix earlyzs settings for alpha_to_coverage 2182 2183Erico Nunes (13): 2184 2185- dri: fix sun4i-drm driver name 2186- lima/ppir: add folding lowering for src modifiers 2187- lima/ppir: add folding lowering for dest modifiers 2188- lima/ppir: remove nir_legacy 2189- lima/ppir: add support and folding for fclamp_pos 2190- lima/ppir: add support and folding trunc to a modifier 2191- lima/ppir: switch to derivative intrinsics 2192- docs/ci: pass -D platforms=[] to avoid dependencies 2193- v3dv: unconditionally expose VK_EXT_queue_family_foreign 2194- v3dv: match render and display device for wsi present 2195- lima/ci: define FDO_HTTP_CACHE_URI for caching-proxy 2196- lima/ci: update piglit ci expectations 2197- Revert "ci/lima: Temporarily disable" 2198 2199Erik Faye-Lund (54): 2200 2201- docs/panfrost: mention that Mali-G610 is conformant 2202- mesa/main: use extension-helper 2203- mesa/main: validate GL_UNSIGNED_INT_5_9_9_9_REV 2204- mesa/main: validate GL_UNSIGNED_INT_2_10_10_10_REV 2205- mesa/main: validate GL_UNSIGNED_INT_10_10_10_2 2206- mesa/main: check depth/stencil formats 2207- mesa/main: validate desktop gl format/types 2208- mesa/main: validate bgr formats 2209- mesa/main: validate abgr format 2210- mesa/main: validate integer-formats 2211- mesa/main: validate half-float types 2212- mesa/main: switch from deny-list to allow-list 2213- mesa/main: add gles-compatible check helpers 2214- mesa/main: tighten glsl-version checks 2215- mesa/main: tighten gl-version checks 2216- panvk: remove out-of-date detail 2217- panvk: drop unused define 2218- panvk: error-check get_fb_descs calls 2219- mesa: make compressed-format emulation optional 2220- panfrost: unify compressed formats 2221- panfrost: store texfeat_bit in panfrost_format 2222- panfrost: check fmt.bitfeat_bit for compressed-support 2223- panfrost: do not re-fetch compressed formats 2224- panvk: check against texfeat_bit 2225- panvk: implement KHR_get_memory_requirements2 2226- panvk: drop needless function implementation 2227- panvk: expose already-implemented KHR_maintenance1 2228- i915: do not calculate native instructions 2229- mesa: remove superfluous "native" limits and stats 2230- mesa: clean up setup of MaxAddressRegs constant 2231- panvk: implement and expose KHR_maintenance2 2232- panvk: expose KHR_relaxed_block_layout support 2233- panvk: expose KHR_bind_memory2 support 2234- docs/features: fix bad sorting of drivers 2235- panvk: expose KHR_external_semaphore and KHR_external_semaphore_capabilities 2236- panvk: expose KHR_external_memory and KHR_external_memory_capabilities 2237- panvk: expose KHR_external_fence and KHR_external_fence_capabilities 2238- panvk: fix broken wsi 2239- panvk: avoid signed integer underflow 2240- panvk: drop needless assert 2241- panvk: put conditional outside of define 2242- panvk: assert on missing vs 2243- panvk/csf: only look at fs if it's required 2244- panfrost: drop needless assign 2245- panfrost: add an assert in render-target setup 2246- panvk: drop duplicate dirty-test 2247- panvk/jm: respect depth-clip state 2248- panvk/csf: respect depth-clip state 2249- panvk: expose EXT_depth_clip_enable 2250- docs/features: claim vk 1.0 for panvk 2251- docs/features: mark off more panvk extensions 2252- panvk: enable KHR_16bit_storage 2253- docs/features: mark off missing panvk feature 2254- glx: avoid null-deref 2255 2256Erwin Jansen (1): 2257 2258- Enable GFXStream bazel build 2259 2260Evan (3): 2261 2262- amd/vpelib: Color Keyer Implementation 2263- amd/vpelib: Luma AND Color Keyer Full Support 2264- amd/vpelib: Input Format Adjustment 2265 2266Faith Ekstrand (309): 2267 2268- nvk: Fix indirect cbuf binds pre-Turing 2269- nvk: Don't advertise sparse residency on Maxwell A 2270- nak/sm50: Split IAdd2 into IAdd2 and IAdd2X 2271- nak: Implement ineg on sm50 2272- nak: Add an AtomCmpSrc to AtomOp::CmpExch 2273- nak/sm50: Support AtomOp::CmpExch 2274- nak/sm50: Improve encoding of OpFFma 2275- nak/sm50: Various encoding fixes 2276- nak/sm50: Implement OpPixLd 2277- nak: Lower fsqrt in NIR on Maxwell A and earlier 2278- nak/sm50: Implement OpCCtl 2279- nak/nir: Split 64-bit conversions pre-Volta 2280- nak: Drop the nvfuzz binary 2281- nak: Plumb tessellation parameters through ShaderStageInfo 2282- nak: Rework fragment shader stage info 2283- nak: Move a few more things to VtgIoInfo 2284- nak: Move creation of nak_shader_info to ShaderBin::new() 2285- nvk: Reject sparse images on Maxwell A and earlier 2286- nil: Add a helper for populating the null descriptor 2287- nvk: Create the zero page earlier in device creation 2288- nvk: Call into NIL to populate the null descriptor 2289- nouveau/winsys: Uninline nouveau_ws_bo_unmap() 2290- nak: Pull the QMD code out into its own crate 2291- nak: Make ShaderBin a public struct 2292- nak/meson: Rename libnak_bindings_gen to _libnak_bindings_rs 2293- nak: Add a bare HW shader runner 2294- nak: Add an implementation of the ACORN random number generator 2295- nak: Add dst_type decorations 2296- nak: Add a Foldable trait 2297- nak: Implement Foldable for OpIAdd3 and OpIAdd3X 2298- nak: Initialize DEBUG on-demand 2299- nak: Add a framework for running hardware tests 2300- nak: Add some opcode tests 2301- nak: Implement Foldable for OpPrmt 2302- nak: Implement Foldable for OpShf 2303- nak: Fix shf.l.high on Maxwell 2304- nak: Add 64-bit shift helpers 2305- nak: Add tests for 64-bit shifts 2306- nak: Use .wrap for 64-bit shifts 2307- nak: Fix shl64 for pre-Volta 2308- meson/megadriver: Don't invoke the megadriver script with no drivers 2309- nak/spill_values: Don't assume no trivial phis 2310- nak: Drop the old encode_sm*.rs files 2311- nak/sm50: Add control-flow sync ops 2312- nak: Add more NIR wrappers for walking the NIR CFG 2313- nak/sm50: Emit sync instructions for control-flow 2314- nak: Disable opt_jump_thread pre-Volta 2315- nak: Don't emit barrier ops for shared memory pre-Volta 2316- nak/sm70: Implement OpAtom with Dst::None as RED 2317- nak/sm50: Implement OpAtom with Dst::None as RED 2318- nak: Use the RED path for atomics with unused destinations 2319- nak: IMul and IMad are variable latency pre-Volta 2320- nak/calc_instr_deps: Set a delay of 13 for flow ops pre-Volta 2321- nak: Align shader data to the min CBuf alignment 2322- nak: Add tests for nak_builder::isetp64 2323- nak: Add tests for nak_builder::iadd64 2324- nak: Add tests for OpISetP 2325- nak/hw_tests: Add debug printing 2326- nak/sm50: Set the .x bit for isetp 2327- nak/builder: Fix isetp64 for pre-Volta 2328- nvk: Squash txq results to 0 for null descriptors 2329- nak: Fix OpCCtl encoding on Maxwell 2330- nak/nir: Don't emit CCtl for release barriers on Maxwell 2331- nvk: Don't push bound UBOs on Volta and earlier 2332- nak: Implement From<bool> for Pred 2333- nak/sm50: OpIAdd2 only supports one negate modifier 2334- nak/sm50: OpLop2 does not have .X 2335- nak: Add tests for OpLop2 and OpLop3 2336- nak/sm50: Fix the encoding of ffma with cbuf in src2 2337- nak: Implement depth clip control pre-Volta 2338- nvk: Use nvk_buffer_addr_range() for index buffers 2339- nvk: Fix zero-size vertex/index buffer bindings pre-Turing 2340- nak/sm50: Add support for OpAL2P 2341- nak: Move the runner into the main nak crate 2342- nak: Plumb through the call/return stack size 2343- nvk: Take CRS into account when allocating the SLM area 2344- nak/nir: Fix the reverse execlist iterator 2345- nak/nir: Add a few more NIR helpers 2346- nak/nir: Rework CRS handling 2347- nak: Add target labels to CRS pop instructions 2348- nak: Add an optimization for CRS ops 2349- nak: OpKill also needs a delay on Maxwell 2350- nak/sm50: Add execution delays for more ops 2351- nak: SrcMod::INeg cannot be folded 2352- nak: Add tests for nak_builder::ineg64 2353- nak/builder: Allow source modifiers in ineg64() 2354- nak: Lower iadd64(x, -y) to isub64(x, y) in NIR 2355- nak/hw_tests: Explicitly test equal cases in test_isetp64 2356- nak: Add tests for OpFlo 2357- nak: Assert no overflow when legalizing iadd with too many ineg modifiers 2358- nak: Add source and destination types for carry values 2359- nak/hw_tests: Respect src_mod and src_swizzle in op auto-tests 2360- nak/hw_tests: Use better test values for iadd tests 2361- nak: Add tests for OpIAdd2 and OpIAdd2X 2362- nak/sm50: Set the not bit for OpLop2::srcs[0] 2363- nak: Use OpFoldData::get_u32_bnot_src() more places 2364- nak: Add tests for OpPopC and enable SrcMod::BNot 2365- nak/nir: Preserve rounding modes in split_64bit_conversions() 2366- nak: Rename OpFSOut to OpRegOut 2367- nak: pre-Volta uses r0 to output the final GS handle 2368- nak: Add a PredSetOp::eval() helper 2369- nak: Add tests for OpPSetP 2370- nak: Non-constant offsets are allowed on patch loads 2371- nak: Rework isetp64 on sm50 again 2372- nak: Disallow isetp.x pre-Volta 2373- nak: Manually flush denorms for nir_op_fquantize2f16 pre-Volta 2374- nak/sm50: Set f2f.high 2375- nvk: Default to NAK on Maxwell+ 2376- nvk: Disable sparse D32S8 cubes pre-Turing 2377- nak: Move the QMD code back into the main nak crate 2378- nak/qmd: Return the complete QMD dispatch size layout 2379- mme/fermi: Don't try to access zero-size std::vector 2380- nouveau/mme: Add support for MUL on Fermi 2381- nouveau/mme: Add support for multiplication on Fermi 2382- nak/hw_runner: Use byte_offset() 2383- nak/hw_tests: Use ptr::from_ref() and ptr::cast() 2384- nak: Use nak_builder::fsetp() in op_fquantize2f16 2385- nvk: Fix indirect dispatch on Turing+ 2386- vulkan: Update XML and headers to 1.3.292 2387- nvk: Move nvk_sample_location to NAK 2388- nak: Sample locations are byte-aligned 2389- nak,nir: Rework gl_SampleMaskIn[] handling 2390- nvk: Don't force sample shading when gl_SampleMaskIn[] is read 2391- nvk: Set POST_PS_INITIAL_COVERAGE 2392- compiler/rust: Copy the Rust NIR wrappers from NAK 2393- compiler/rust/nir: Implement NIR types directly 2394- compiler/rust/nir: Add explicit lifetimes 2395- nak: Switch to the common NIR wrappers 2396- nak: Move DstsAsSlice::is_uniform() to its own trait 2397- nak: Replace the guts of Srcs/DstsAsSlice with a new AsSlice trait 2398- nak,compiler: Move AsSlice to common code 2399- nvk: Require color or depth/stencil attachment support for input attachments 2400- nvk: Support STORAGE_READ_WITHOUT_FORMAT on buffers 2401- nvk: Move descrptor structs into a separate header 2402- nvk: s/device/dev/ in nvk_buffer_view.c 2403- nvk/nvkmd: Advertise the usable VA range 2404- nvk: Add a VK_EXT_descriptor_buffer buffer view cache 2405- nvk: Refactor image intrinsic lowering a bit 2406- nvk: Lower descriptors for VK_EXT_descriptor_buffer buffer views 2407- nvk: Use nvk_edb_buffer_view_descriptor for EDB descriptor set layouts 2408- nvk: Use the EDB buffer view path with NVK_DEBUG=edb_bview 2409- nvk: Implement GetDescriptorLayoutSize/BindingOffsetEXT() 2410- nvk: Refactor some descriptor set helpers 2411- nvk: Implement GetDescriptorEXT 2412- nvk/descriptor_table: Add support for requesting a specific index 2413- nvk: Implement descriptor capture/replay 2414- nvk: Properly indent a comment 2415- nvk: Rework descriptor set bindings 2416- nvk: Implement descriptor buffer binding 2417- nvk: Add support for embedded immutable samplers 2418- zink: Align descriptor buffers to descriptorBufferOffsetAlignment 2419- nvk: Advertise VK_EXT_descriptor_buffer 2420- nvk: Invalidate MME data for descriptor buffers 2421- nak: Pass -C default-linker-libraries when building tests 2422- ci: Add NVK to debian-arm64-build-test 2423- vulkan: Add null descriptor bits to vk_pipeline_robustness_state 2424- nvk: Plumb the whole vk_pipeline_robustness_state through to nvk_ubo/ssbo_addr_format 2425- nvk: Enable shader bounds checking when nullDescriptor is enabled 2426- nouveau/headers: Add more nv_push asserts 2427- nouveau/winsys: Fix an undefined shift 2428- nouveau/mme: Fix add64 of immediates on Fermi 2429- nouveau/mme: Better handle the too many immedaties case on Turing+ 2430- nouveau/mme: Don't allow PRMT and MTHD in the same op on Turing+ 2431- nouveau/mme: Disassemble MERGE more like the mme_merge() 2432- nouveau/mme: Add NOT and AND_NOT ops 2433- nouveau/mme: Allow shifting immediates on Fermi 2434- nouveau/mme: Use a designated initializer in mme_fermi_decode() 2435- nouveau/mme/tu104: Break the state simulator into a pluggable component 2436- nouveau/mme/fermi: Break the state simulator into a pluggable component 2437- nouveau/mme: Add a pluggable state interface 2438- nouveau/mme/tu104: Properly handle mthd increment 2439- nouveau/mme: Add a generic simulator function 2440- nvk: Reserve MME scratch area for communicating with FALCON 2441- nvk/mme: Add a unit test framework for driver macros 2442- nvk: Emit NVK_MME_SCRATCH_DRAW_BEGIN as part of flush_ia_state() 2443- nvk: Increase the context init size 2444- nvk: Move the guts of BindIndexBuffer2KHR to a macro 2445- nvk: Move the guts of BindVertexBuffers2 into a macro 2446- nvk: Move setting VB enables and strides into macros 2447- nvk/mme: Add helpers for masked scratch updates 2448- nvk: Add macros to sort out tess parameters 2449- nvk: Move the ANTI_ALIAS_CONTROL logic to the MME 2450- nvk: Stop leaking the descriptor table in-use set 2451- nvk: Disable conditional rendering around CopyQueryPoolResults 2452- nvk: Expose a new nvk_compile_nir_shader() helper 2453- nvk: Refactor compute shader dispatch 2454- nvk: Add a helper for dispatching compute shaders 2455- nvk: Use the new dispatch helper for queries 2456- nvk: Treat NULL vertex/index buffer pointers as null descriptors 2457- nil,nvk: Disable modifiers for B10G11R11_UFLOAT and E5B9G9R9_UFLOAT 2458- nir: Preserve fp_fast_math in nir_opt_vectorize() 2459- v3dv: Update a debug message 2460- vtn: Move initialization of mesa_spirv_debug to a helper 2461- vtn: Allow SPIR-V debug options in release builds 2462- vtn: Add a debug flag to dump SPIR-V assembly 2463- v3dv: Drop the SPIR-V dumper 2464- radv: Use the SPIR-V printer in spirv_to_nir 2465- vulkan/pipeline: Handle VIEW_INDEX_FROM_DEVICE_INDEX_BIT 2466- lvp: Stash the PipelineCreateFlags2KHR in the pipeline 2467- dzn: Use vk_pipeline_shader_stage_to_nir() 2468- vulkan: Take a VkPipelineCreateFlags2KHR in vk_pipeline_*shader_stage*() 2469- vulkan: Handle VIEW_INDEX_FROM_DEVICE_INDEX_BIT in the runtime 2470- nvk: Fix a typo in a comment 2471- nvk: Use 4 bits per value for the anti alias values 2472- nvk: Hash minSampleShading in nvk_hash_graphics_state() 2473- nvk: Advertise VK_KHR_maintenance7 2474- nvk/nvkmd: Add an nvkmd_mem_init() helper 2475- nvk/nvkmd: Plumb map_out through to the back-end nvkmd_mem_ops::map 2476- nvk/nvkmd: Plumb map pointers through to back-end unmap funcs 2477- nvk/nvkmd: Plumb map_flags through to unmap/overmap 2478- nvk/nvkmd: Don't inline nvkmd_mem_[un]map() 2479- nvk/nvkmd: Split client and internal mappings 2480- nvk/nvkmd: Reference count internal mappings 2481- nvk: Don't do linear<->tiled copies for rendering suspend/resume 2482- nvk: Take depth image layer counts from the VkRenderingInfo 2483- nvk: Add all_linear to nvk_rendering_state 2484- vulkan: Add a helper for getting VkRenderingAttachmentLocaiontInfoKHR 2485- vulkan: Allow pColorAttachmentLocations == NULL in CmdSetRenderingAttachmentLocationsKHR() 2486- vulkan: Expose a vk_cmd_set_rendering_attachment_locations() helper 2487- nvk: Reset rendering attachment locations in BeginCommandBuffer/Rendering 2488- nvk: Emit SET_CT_SELECT based on the dynamic color location map 2489- nvk: Adverties VK_KHR_dynamic_rendering_local_read 2490- nak: Add a NAK_SUBGROUP_SIZE #define 2491- nak: Move subgroup_id and num_subgroups to lower_system_values 2492- nak: Shuffle gl_LocalInvocationID for quad derivatives 2493- nvk: Advertise VK_NV/KHR_compute_shader_derivatives 2494- nvk: Only advertise compute_shader_derivatives when NAK is used 2495- vulkan: Handle variable-length property arrays more generically 2496- nil: Rework calculations of the sizes of miplevels 2497- nil: Return a more accurate size in level_size_B 2498- nil: Assert array_len == 0 in Extent4D<Bytes>::size_B() 2499- nvk: Advertise VK_KHR_shader_float_controls2 2500- nvk: Silence a maybe-uninitialized warning 2501- nvk: Add an NVK_DEBUG=gart flag 2502- vulkan: Add a stages field to vk_pipeline 2503- vulkan: Add a vk_pipeline_ops::get_shader method 2504- vulkan/pipeline: Patch through INDIRECT_BINDABLE_BIT 2505- nvk: Fill out shader push data at compile time 2506- nvk: Use VkShaderStageFlags for shaders_dirty 2507- nvk: Disable CBuf pushing when INDIRECT_BINDABLE_BIT is set 2508- nvk: Add an implementation of VkIndirectExecutionSetEXT 2509- nvk/nvkmd: Always map read/write for internal maps 2510- nvk/nvkmd: Add a mechanism to track and look up nvkmd_mems 2511- nvk: Track and dump device-generated commands if possible 2512- nvk: Move flush_push_descriptors to the front of flush_state() 2513- nvk: Export graphics state flushing helpers 2514- nvk: Expose an nvk_cmd_flush_cs_qmd() helper 2515- nvk: Implement CmdProcess/ExecuteGeneratedCommandsEXT 2516- nvk: Handle indirect dispatch in pipeline barriers 2517- nvk: Advertise VK_EXT_device_generated_commands 2518- nvk: Handle aspects in D32_S8_UINT copies 2519- nvk: Assume a GOB height of 8 for copies 2520- nvk,nil: Replace gob_height_is_8 and is_tiled with a new GOBType enum 2521- vulkan: Use SEVERITY_WARNING_BIT for vk_error*() 2522- nil: Add a couple Extent4D and Offset4D methods 2523- nil: Add tiled memcpy helpers 2524- vulkan: Add a comment explainint SEVERITY_WARNING_BIT for vk_error*() 2525- nvk/nvkmd: Stop leaking the primary FD 2526- nir: Add an option to lower quad vote 2527- nak,nvk: Implement VK_KHR_shader_quad_control 2528- nvk: Fix a comment in SET_VIEWPORT_CLIP_CONTROL 2529- vulkan/queue: Don't use vk_semaphore in threaded payload stealing 2530- vulkan/queue: Move has_binary_permanent_semaphore_wait into the sumbit struct 2531- vulkan/queue: Use _mem_signal_temp instead of signal_mem_sync 2532- vulkan: Add a vk_queue_submit_has_bind() helper 2533- vulkan/queue: Use a builder pattern for vk_queue_submit 2534- vulkan/queue: Split vk_queue_submit into create and submit 2535- vulkan/queue: Merge submits when possible 2536- nouveau/mme: Test ifs with immediate parameters 2537- nouveau/mme: Evaluate methods immediately in the Turing sim 2538- nouveau/mme: Turing can't have emit and state in the same op 2539- nil: Add a helper to get samples from a SampleLayout 2540- nvk: Add an nvk_cmd_emit_sample_layout() helper 2541- nvk: Rework setup of sample masks 2542- nouveau/class_parser: Add more method arrays 2543- nil: Add D3D sample layouts 2544- nvk: Support D3D sample modes 2545- nil: Use D3D sample modes by default 2546- nvk: Implement GetPhysicalDeviceFragmentShadingRatesKHR() 2547- nvk: Initialize VARIABLE_PIXEL_RATE_SAMPLE_ORDER 2548- vulkan: Add a vk_fragment_shading_rate_is_disabled() helper 2549- nvk: Rework setting of min_sample_shading 2550- nvk: Emit FSR state 2551- nvk: Add support for binding fragment shading rate images 2552- nak: Get rid of some dead code warnings 2553- nak: Rename SV_VERTEX_COUNT to PRIM_TYPE 2554- nak: Add support for gl_ShadingRateEXT 2555- nak: Pass a nak_compiler into sysval and attrib helpers 2556- nak: Add support for gl_PrimitiveShadingRateEXT 2557- nvk: Set VARIABLE_PIXEL_RATE_SHADING_TABLE_SELECT 2558- nvk: Advertise VK_KHR_fragment_shading_rate 2559- vulkan/queue: Check for _mem_signal_temp before we submit 2560- nvk: Add an nvkmd_engines bitfield to nvk_queue 2561- nvk: Only set texture/sampler tables and SLM for enabled engines 2562- nvk: Re-sort the features table 2563- nvk: Advertise 64-bit atomics on buffer views 2564- nak: Remove annotations before calc_instr_deps() 2565- nak: Handle annotations in legalization 2566- compiler/rust: Enable unit tests 2567- compiler/rust: Add a unit test for the memstream abstraction 2568- compiler/rust: Fix a bad cast in the memstream abstraction 2569- nvk: Only wait on the upload queue if there are commands 2570- nvk: Only submit the state pushbuf if there are commands 2571- nvk: move nvk_queue_state_update to nvk_queue_submit_exec 2572- nvk: Use nvk_queue_submit_simple() for nvk_queue_state_update() 2573- nak/bindings: Use an enum for IOCTL numbers 2574- nvk: Fix host copies for array images 2575- nil: Add a nil_sample_offset() helper 2576- nvk: Put a sample map in the descriptor for MSAA storage images 2577 2578Felix DeGrood (5): 2579 2580- anv: fix src_hash dumping for compute shaders 2581- anv/measure: ignore events from reused command buffers 2582- intel/measure: add nogl feature 2583- intel/measure: defer file open until first write 2584- intel/measure: increase size of filename malloc to account for \\0 2585 2586Francisco Jerez (8): 2587 2588- intel/dev: Add devinfo flag for TBIMR push constant workaround. 2589- intel/brw: Implement null push constant workaround. 2590- anv/gfx12.5: Pass non-empty push constant data to PS stage for TBIMR workaround. 2591- iris/gfx12.5: Pass non-empty push constant data to PS stage for TBIMR workaround. 2592- iris: Pin pixel hashing table BO from iris_batch submission instead of from iris_state. 2593- intel/brw/gfx12.5+: Fix IR of sub-dword atomic LSC operations. 2594- intel/brw/xe2+: Adjust performance analysis divergence weight due to EU fusion removal. 2595- intel/fs/xe2: Fix up subdword integer region restriction with strided byte src and packed byte dst. 2596 2597Frank Binns (9): 2598 2599- pvr: minor vulkan driver formatting fixes 2600- pvr: drop 'Imagination' from deviceName 2601- pvr: fix image size calculation when mipLevels is 1 2602- pvr: fix bo refcounting for imported buffers 2603- pvr: fix use of uninitialised value when setting up tex info swizzle 2604- pvr: remove set but not used struct member 2605- pvr: add handling for retry error in pvr_srv_rgx_submit_transfer2() 2606- pvr: add handling for retry error in pvr_srv_rgx_kick_compute2() 2607- pvr: ensure stencil clear value fits TA_STATE_ISPA.sref field 2608 2609Friedrich Vock (5): 2610 2611- radv/rt: Only do ploc atomicCompSwap once per workgroup 2612- radv/rt: Don't atomicAdd local prefix sums 2613- aco: Fix 1D->2D dispatch conversion on <gfx9 2614- radv/rt: Skip all AABB code when no_skip_aabbs is not set 2615- vulkan/rmv: Correctly set heap size 2616 2617GKraats (8): 2618 2619- i915g: fix generation of large mipmaps 2620- i915g: fix mipmap-layout for npots 2621- i915g: fix max_lod at mipmap-sampling 2622- i915g: Screen corruption with ENOBUFS caused by fence register shortage 2623- i915g: fix count of buffers at i915_drm_batchbuffer_validate_buffers 2624- X11: fix crash of gnome-shell if mesa is compiled with legacy-x11=dri2 2625- i915g: fix texture3d npot mipmaps 2626- i915g: fix GL_TEXTURE_BASE_LEVEL MAX_LEVEL MIN_LOD and MAX_LOD 2627 2628Ganesh Belgur Ramachandra (7): 2629 2630- radeonsi: fix eptich on chips without image opcodes (e.g. gfx940) 2631- amd/common: skip lane size determination for chips without image opcodes (e.g. gfx940) 2632- amd: fix incorrect PIPE_INTERLEAVE_BYTES size for CDNA chips 2633- radeonsi: fix decode corruptions with AMD_IMAGE_OPCODES=false on non-CDNA based chips 2634- amd,radeonsi: use new pass manager to handle midend optimizations 2635- amd,radeonsi: reduce legacy::PassManager use to only run backend passes 2636- amd: remove the redundant target library info instance in LLVM compiler 2637 2638Georg Lehmann (126): 2639 2640- radv/rt: remove one VALU from traversal loop 2641- spirv: ignore more function param decorations 2642- aco/optimizer: update temp_rc when converting to uniform bool alu 2643- aco: add CompilationProgress::after_lower_to_hw 2644- aco: validate temp_rc 2645- aco/gfx11+: apply neg to vinterp 2646- aco: reorder dpp for ddx/ddy 2647- aco/gfx11.5+: use vinterp for fddx/fddy 2648- aco/gfx11+: don't use VOP3 v_swap_b16 2649- aco: test xor swap16 path 2650- nir/peephole_select: ignore masked/quad swizzle without fetch_inactive 2651- aco: optimize 64bit find_lsb/find_msb 2652- aco: optimize dd[xy]_fine if it's only used by abs 2653- nir/lower_int64: replace uadd_sat with ior for find_lsb64 and ufind_msb64 2654- aco/gfx10+: set lateKill for sgprs used by wave64 VALU writing a mask 2655- aco: only set latekill in live_var_analysis 2656- aco/tests: don't test dpp constant propagation with row shift 2657- aco/tests: do not use add to tests neg modifer 2658- aco/print_ir: use neg() for constants 2659- aco/tests: parse neg(constant) in vop3p test 2660- aco/tests: allow literals with resolved swizzles in vop3p test 2661- aco/tests: do not use mul with constant to tests neg modifier 2662- nir: optimize pack_uint_2x16 of pack_half(a, 0) 2663- zink: switch to derivative intrinsics 2664- aco/ra: unconditionally replace literal with sgpr when promoting to VOP3 2665- aco/ra: use parallelcopy to copy literal instead of s_mov 2666- aco/ra: fix copying 64bit literal to sgprs 2667- aco/ra: always reuse def register for literal copy 2668- aco: use v_cvt_pk_u8_f32 for f2u8 2669- nir/opt_algebraic: reassociate constants in ior(iand) chains 2670- radv/ci: update trace checksums 2671- aco/ra: do not reuse killed vcc 2672- nir/opt_sink: do not sink load_ubo_vec4 out of loops 2673- nir/opt_sink: do not sink inverse_ballot out of loops 2674- nir: merge out of loop decision with nir_can_move_instr logic 2675- nir/instr_set: fix fp_fast_math 2676- aco/isel: add function to create builder for alu 2677- aco/ir: add float control definition flags 2678- aco/vn: handle sz/inf/nan preserve 2679- aco/isel: set per instruction float control modes 2680- aco/tests: use new float control flags 2681- aco/optimizer: use new per definition float control flags 2682- aco: remove per block inf/nan/sz control 2683- aco/tests: fix omod signed zero tests 2684- radv: advertise VK_KHR_shader_float_controls2 2685- nir: add amd shared append/consume 2686- amd/nir: add ac_nir_opt_shared_append 2687- aco/assembler: support ds_append/ds_*_addtid 2688- aco: implement nir_shared_append/consume_amd 2689- aco: use ac_nir_opt_shared_append 2690- aco/isel: use upper bound for v_mul_u32_u24 2691- aco/gfx10+: work around non uniform ds_append wave64 result 2692- spirv: explicitly lower derivatives to zero 2693- amd: add offset to is_subgroup_invocation_lt_amd 2694- ac/lower_ngg: use is_subgroup_invocation_lt_amd offset 2695- aco/ssa_elimination: don't check for VALU limitation when optimizing branching sequence 2696- aco/ssa_elimination: optimize branching sequence with SALU that has multiple definitions 2697- aco: optimize lanecount_to_mask 2698- nir: sink is_subgroup_invocation_lt_amd 2699- nir: replace nir_opt_remove_phis_block with a single source version 2700- nir: make nir_instr_clone usable with load_const and undef 2701- nir/opt_remove_phis: rematerialize constants 2702- radv: remove nir_opt_reuse_constants call 2703- nir: delete nir_opt_reuse_constants 2704- vc4/ci: documment new failure 2705- aco: use s_pack_hl for shfr16 2706- aco: use s_pack_lh for bitfield_select(0xffff) 2707- radv: call nir_opt_remove_phis in radv_optimize_nir_algebraic 2708- nir/opt_move_discards_to_top: handle ddx/ddy intrinsics 2709- elk: remove alu fddx/fddy check 2710- ir3: remove alu fddx/fddy check 2711- spirv: remove alu fddx/fddy from comment 2712- mesa/prog_to_nir: use derivative builder 2713- nir: always emit ddx intrinsics 2714- nir: remove uses_fddx_fddy 2715- nir: remove alu ddx/ddy 2716- nir: remove nir_op_is_derivative 2717- aco/ir: add function to parse depctr waits 2718- aco: do not track ALU delay across jumps 2719- aco/insert_delay_alu: consider more implicit waits 2720- aco/insert_delay_alu: do not delay lane mask fast forwarding 2721- aco/statistics: update branch issue cycles 2722- aco/print_ir: use parse_depctr_wait 2723- meson: remove selinux option 2724- ci: remove selinux from containers 2725- aco: fix 64bit extract_i8/extract_i16 2726- aco/insert_exec: don't pretend WQMState is a bit mask 2727- aco/insert_exec: replace pair with a named struct 2728- aco/insert_exec: remove stray break_cond variable 2729- aco/insert_exec: untangle add_branch_code control flow 2730- aco/insert_exec: remove unused includes 2731- aco/ir: rework Operand equality to return true for equal fixed non-temp ops 2732- aco/insert_exec: remove get_exec_op 2733- nir/opt_algebraic: optimize -0.0 + a 2734- aco/ssa_elimination: also optimize branching sequence with s_and without saveexec 2735- aco/insert_exec: avoid phis for masks in exec 2736- aco/insert_exec: don't always reset top exec 2737- aco/insert_exec: reuse old exec temp in loop pre-header 2738- aco/insert_exec: reuse old exec temp instead using s_and_saveexec 2739- aco/ssa_elimination: don't avoid saving exec when optimizing branching sequence 2740- radv: don't use v_mqsad_u32_u8 on gfx7 2741- aco: remove heuristic that restricts VOP2/C with 2 sgprs 2742- aco/ssa_elimination: add instr_accesses helper 2743- aco/ssa_elimination: don't assume exec writes can be removed based on block kind 2744- aco: optimize conditional divergent breaks at the end of loops 2745- aco: move exec copy out of waterfall loops 2746- ac/llvm: handle shared atomic base offset 2747- nir,radv: optimize shared atomic offsets 2748- nir/opt_algebraic: create ubfe with non constant mask 2749- radv: use nir_opt_fragdepth 2750- nir: add SYSTEM_VALUE_PIXEL_COORD 2751- radv: add support for load_pixel_coord 2752- ac/llvm: implement load_pixel_coord 2753- aco: implement load_pixel_coord 2754- nir: add nir_opt_frag_coord_to_pixel_coord 2755- radv: use nir_opt_frag_coord_to_pixel_coord 2756- nir/opt_frag_coord_to_pixel_coord: optimize trunc/floor 2757- nir/lower_wpos_ytransform: remove unnecessary state variable 2758- nir/lower_wpos_ytransform: remove redundant state shader 2759- nir/lower_wpos_ytransform: clean up baryc_at_offset 2760- nir/lower_wpos_ytransform: clean up sample_pos 2761- nir/lower_wpos_ytransform: clean up wpos_adjustment 2762- nir/lower_wpos_ytransform: use more typical pass structure 2763- nir/lower_wpos_ytransform: remove reference to long removed TGSI code 2764- nir/lower_fragcoord_wtrans: use intrinsics_pass 2765- amd: lower gl_FragCoord.w rcp in NIR 2766 2767George Ouzounoudis (3): 2768 2769- d3d12: Split memory size info to device and system memory 2770- d3d12: Keep track of size and number of evictions 2771- d3d12: Add support for GL_NVX_gpu_memory_info and GL_ATI_meminfo 2772 2773Gert Wollny (2): 2774 2775- nir/opt_algebraic: Allow two-step lowering of ftrunc\@64 to use ffract\@64 2776- Revert: r600/sfn: call nir_lower_doubles explicitely" 2777 2778Guilherme Gallo (26): 2779 2780- bin/ci: crnm: Deal with \`created` status 2781- bin/ci: crnm: Fix pipeline monitoring skipping 2782- bin/ci: crnm: Reduce trace cluttering 2783- bin/ci: crnm: Don't overwrite the global sets 2784- bin/ci: crnm: Use frozen set for statuses 2785- ci/a618: Fix zink-tu-a618-full rules 2786- ci/freedreno: Rebalance limozeen jobs 2787- ci/freedreno: Replace 2 limozeen with kingoftown 2788- ci/freedreno: Rebalance jobs via fraction and parallel 2789- ci/freedreno: Add a618_gl_full 2790- ci/amd: Rebalance radeonsi-stoney-gl:x86_64 2791- ci/amd: Rebalance jobs via DEQP_FRACTION 2792- ci/intel: Rebalance jobs via parallel 2793- ci/anv: Introduce missing farm var for ADL jobs 2794- ci/anv: Update xfiles 2795- ci/lava: Make perf jobs use pyutils artifact 2796- ci/zink: Add skips for zink-anv-(adl|tgl) 2797- ci/radv: Rebalance stoney-vkcts 2798- ci: Use merge-skips files during merge pipelines 2799- ci/amd: Track merge-skips.txt changes 2800- ci/amd: Reassign radv-raven-traces from zork to dalboz 2801- ci/panfrost: Increment panfrost-g52-gl parallel no. 2802- ci/freedreno: Rebalance limozeen jobs 2803- ci/turnip: Rebalance kingoftown jobs 2804- ci/lava: Fix lava-tags parsing 2805- ci/lava: Fix LAVA job definition tests 2806 2807Gurchetan Singh (145): 2808 2809- vulkan/runtime: introduce vk_meta_object_list 2810- tu: use os_get_total_physical_memory(..) 2811- vulkan: fix glibc AOSP build 2812- vulkan/android: change vk_enum_defines.h path 2813- gfxstream: fix Linux guest build 2814- gfxstream-guest: add precise virtio-gpu capabilities 2815- gfxstream: use ASG for Vulkan streams only 2816- gfxstream: gfxstreamCapset --> vulkanCapset 2817- gfxstream: query non-Vulkan caps from the guest 2818- gfxstream/guest: add GetRingParamsFromCapset function 2819- gfxstream/guest: use GetRingParamsFromCapset 2820- gfxstream: nuke unused functions 2821- gfxstream: simplify VK_HOST_CONNECTION macro 2822- gfxstream: don't use renderControlEnc for Vulkan 2823- gfxstream: nuke ResourceTracker::Impl 2824- gfxstream: fix meson build system issues 2825- gfxstream/guest: pass through Android process names 2826- gfxstream: import Vulkan-docs from 463f8c 2827- gfxstream: vulkan-Docs: import the cerealgenerator 2828- gfxstream: vulkan-docs: track feature type 2829- gfxstream: vulkan-docs: add cereal generator 2830- gfxstream: genvk: add support for vk_gfxstream.xml 2831- gfxstream: vulkan-docs: don't auto-generate vk_android_native_buffer 2832- gfxstream: add generate-gfxstream-vulkan.sh 2833- gfxstream: add support for VK_KHR_maintenance5 + VK_EXT_host_image_copy 2834- gfxstream: use standard Vulkan headers and new auto-gen flow 2835- gfxstream: look for common/opengl headers 2836- gfxstream: cereal: [moduleType] --> [(moduleType, moduleName)] 2837- gfxstream: cereal: SUPPORTED_WRAPPERS --> SUPPORTED_MODULES 2838- gfxstream: codegen EXT_swapchain_maintenance1 + KHR_swapchain host only 2839- gfxstream: mesa: import Mesa 2840- gfxstream: mega-change to support guest Linux WSI with gfxstream 2841- gfxstream: nuke goldfish_vk_handlemap 2842- gfxstream: use absolute path for some code 2843- genvk: options.filename --> args.target 2844- gfxstream: codegen: allow generating host/guest autogen separately 2845- gfxstream: guest: cpp_args --> guest_cpp_args 2846- gfxstream: reduce use of PLATFORM_SDK_VERSION 2847- gfxstream: PLATFORM_SDK_VERSION --> ANDROID_API_LEVEL 2848- gfxstream: don't forward declare ProcessResources 2849- gfxstream: nuke VIRTIO_GPU flag 2850- gfxstream: nuke HOST_BUILD 2851- gfxstream: add GFXSTREAM_ENABLE_GUEST_GOLDFISH 2852- gfxstream: add REQUIRED_TYPES list 2853- gfxstream: always initialize rcEncoder and other things 2854- gfxstream: build the test layer via Meson 2855- gfxstream: generate development ICD for meson build 2856- gfxstream: always advertise device memory extension guest side 2857- gfxstream: convert some Vulkan to mesa_log(..) 2858- gfxstream: valgrind fix 2859- gfxstream: add VK_EXT_vertex_attribute_divisor 2860- Revert^2 "gfxstream: add VK_EXT_vertex_attribute_divisor" 2861- Revert "gfxstream: Add vkCmdPipelineBarrier2 implementation in VkDecoderGlobalState" 2862- Revert "Revert "gfxstream: Add vkCmdPipelineBarrier2 implementat..." 2863- gfxstream: ResourceTracker: fix mesa_loge(..) 2864- gfxstream: guest: add VK_EXT_color_write_enable 2865- gfxstream: guest: add XCB Surface support 2866- gfxstream: guest: support blobCmd + blobCmdSize 2867- gfxstream: don't define hasDedicatedImage + hasDedicatedBuffer multiple times 2868- Revert "Add snapshot support for vkBindImageMemory2" 2869- gfxstream: guest: enable virtio-gpu kumquat 2870- gfxstream: host + guest: bump up version of the protocol 2871- gfxstream: linux guest: support client-side blob allocations 2872- gfxstream: guest: fix useless return statement 2873- gfxstream: guest: simpler get param for kumquat 2874- gfxstream: host: introduce vkGetSemaphoreGOOGLE 2875- gfxstream: guest: add GFXSTREAM_ACQUIRE_SYNC protocol 2876- gfxstream: guest: add experimental fence passing APIs 2877- gfxstream: guest: mBlobId -> mAtomicId 2878- gfxstream: guest: start using new vkGetSemaphoreGOOGLE apis 2879- gfxstream: kumquat: unmap memory when needed 2880- gfxstream: guest: fix kumquat build 2881- gfxstream: vulkan: fix issue with GCC 2882- gfxstream: guest: nuke AndroidLock 2883- gfxstream: guest: nuke aemu/base/Optional.h 2884- gfxstream: guest: nuke unused functions 2885- gfxstream: guest: make sure signalSemaphoreValueCount is correct 2886- gfxstream: guest: vk_CmdBeginTransformFeedbackEXT fix 2887- gfxstream: host: don't abort when unbox fails 2888- gfxstream: guest: add support VK_FORMAT_A2R10G10B10_UNORM_PACK32 Wayland buffers 2889- gfxstream: guest: use MESA_SCOPE_TRACE 2890- gfxstream: guest: nuke more AEMU base stuff 2891- gfxstream: guest: kumquat: fixes in preparation for AOSP Kumquat testing 2892- gfxstream: nuke set instance for testing 2893- gfxstream: nuke libplatform stub 2894- gfxstream: guest: plumb descriptor all the way through the HostConnection 2895- gfxstream: connect to the specificied kumquat socket 2896- gfxstream: guest: add vkInitializeKumquat 2897- gfxstream: guest: nuke Android HealthMonitor 2898- gfxstream: guest: plumb the descriptor everywhere 2899- gfxstream: end2end: nuke RutabagaLayer, use kumquat for testing 2900- gfxstream: simplify GLESv1/GLESv2/EGL/Vulkan builds 2901- gfxstream: nuke ErrorLog.h 2902- gfxstream: nuke AlignedBuf.h 2903- gfxstream: nuke mWorkPool 2904- gfxstream: guest: nuke vk_enum_string_helper.h 2905- gfxstream: guest: nuke Fuchsia Goldfish Address Space 2906- gfxstream: guest: nuke unnecessary Android properties 2907- gfxstream: guest: nuke goldfish_address_space_android.impl 2908- gfxstream: guest: mesa-ify GoldfishAddressSpace 2909- gfxstream: guest: mesa-ify libplatform_virtgpu 2910- gfxstream: guest: add virtgpu suffix 2911- gfxstream: guest: remove EmulatorFeatureInfo dependency in libgfxstream_vulkan.so 2912- gfxstream: guest: simplify Vulkan sequence number logic 2913- gfxstream: sync command buffer api with Mesa upstream 2914- gfxstream: move EmulatorFeatureInfo and ExtendedRenderControl 2915- gfxstream: guest: remove dependency on EmulatorFeatureInfo 2916- gfxstream: add processPipeInit and connect to IOStream 2917- Reapply "gfxstream: nuke mWorkPool" 2918- gfxstream: guest: fix timeout issues 2919- gfxstream: guest: introduce Gralloc::getGrallocType() 2920- gfxstream: guest: remove renderControl dependency from Gralloc 2921- gfxstream: GrallocMinigbm uses platform_virtgpu 2922- gfxstream: mesa-ify libgfxstream_android 2923- gfxstream: delete duplicate virtgpu_drm.h 2924- gfxstream: guest: don't rely on HostConnection for platform helpers 2925- gfxstream: guest: introduce GfxStreamConnectionManager 2926- gfxstream: guest: vulkan: use connection manager 2927- gfxstream: guest: introduce QemuPipeStream stub 2928- gfxstream: guest: fully mesa-ify vulkan_enc 2929- gfxstream: guest: remove unnecessary includes in VkEncoder 2930- gfxstream: guest: remove unneeded dependencies in meson build 2931- gfxstream: guest: vulkan: use hardware/google/aemu 2932- gfxstream: guest: vulkan: don't assume Android 2933- gfxstream: guest: nuke android_compat directory 2934- gfxstream: move goldfish sync 2935- gfxstream: guest: Gralloc --> GfxStreamGralloc 2936- gfxstream: don't assert without host 2937- gfxstream: move include outside of ANDROID 2938- gfxstream: codegen: don't require clang-format 2939- gfxstream: delete unneccessary code 2940- gfxstream: build time guest autogeneration 2941- gfxstream: re-license files 2942- gfxstream: re-license cereal 2943- gfxstream: simplify codegen 2944- gfxstream: moar relicense 2945- mesa: gfxstream: nuke vk_icd_* 2946- mesa: add gfxstream-experimental to -Dvulkan-drivers 2947- gfxstream: add clang-format 2948- gfxstream: nuke util function 2949- gfxstream: use util/libsync 2950- util: add sync_fence_info 2951- gfxstream: use sync_fence_info 2952- gfxstream: move generate-gfxstream-vulkan.sh script 2953- gfxstream: update Kumquat API 2954 2955Hailin Zhang (1): 2956 2957- Vulkan: fix dstArrayElement index wrap issue 2958 2959Hans-Kristian Arntzen (4): 2960 2961- wsi/common: Do not update present mode with MESA_VK_WSI_PRESENT_MODE. 2962- wsi/x11: Bump maximum number of outstanding COMPLETE events. 2963- radv: Always make sure to write the pipeline binary key. 2964- vulkan/wsi/wayland: Use X11-style image count strategy when using FIFO. 2965 2966Hsieh, Mike (4): 2967 2968- amd/vpelib: Remove deprecated update_3dlut flag 2969- amd/vpelib: Refactor OPP registers 2970- amd/vpelib: Add cache mechanism for 3D Lut command 2971- amd/vpelib: Refactor 3D LUT code 2972 2973Hyunjun Ko (25): 2974 2975- intel/genxml: add a missing value for MFX_SURFACE_STATE 2976- intel/genxml: change the length of MFX_QM_STATE 2977- intel/genxml: update VDENC instructions 2978- vulkan/video: adds a bitstream writer of h264 slice header 2979- anv/query: handle VK_QUERY_TYPE_VIDEO_ENCODE_FEEDBACK_KHR 2980- anv/video: remove unnecessary macros 2981- anv/video: implemnt VkGetEncodedVideoSessionParametersKHR 2982- anv/video: initial support for h264 encoding 2983- anv: enable VK_KHR_video_encode_queue and VK_KHR_video_encode_h264 2984- intel/genxml: Add missing fields for HCP_SLICE_STATE 2985- intel/genxml: fix the length of VDENC_DS_REF_SURFACE_STATE 2986- intel/genxml: Adds more VDENC commands 2987- intel/genxml: fix some length of HCP_FQM_STATE 2988- intel/genxml: adds a value of reference pic to HCP_SURFACE_STATE 2989- intel/decoder: Handle HCP_PAK_INSERT_OBJECT 2990- vulkan/video: adds a bitstream writer of h265 slice header 2991- anv/query: consider codec when querying the encoding status. 2992- anv/video: initial support for h265 encoding 2993- anv: support h265 encoding 2994- zink: walk the chain of resources for multi-planar formats. 2995- anv/video: support VK_VIDEO_ENCODE_RATE_CONTROL_MODE_DISABLED_BIT_KHR. 2996- anv: consider VK_VIDEO_CODEC_OPERATION_ENCODE_H264_BIT_KHR when allocating mv storgae. 2997- anv: support for inline query for vulkan video 2998- anv: support VK_IMAGE_CREATE_VIDEO_PROFILE_INDEPENDENT_BIT_KHR 2999- anv: enable KHR_video_maintenance1 3000 3001Iago Toral Quiroga (22): 3002 3003- v3d: move TLB blit checks to a helper 3004- v3d: support blitting straight from tile buffer 3005- v3d: add a faster TLB blit path 3006- v3d: do not expose EXT_float_blend 3007- broadcom/compiler: emit instructions producing flags earlier 3008- broadcom: implement nir_op_uclz 3009- broadcom: drop backend implementation of nir_op_ufind_msb 3010- broadcom/compiler: skip small immediates optimization on vpm instructions 3011- broadcom/compiler: rename is_ldunif_dst to try_rf0 3012- broadcom/compiler: avoid register conflict with ldunif(a) and ldvary 3013- broadcom/compiler: don't add const offset to unifa if it is 0 3014- broadcom/compiler: don't use small immediates in geometry stages 3015- v3d: support shader precompile for compute shaders 3016- nir: make fclamp_pos_mali and fsat_signed_mali opcodes generic 3017- broadcom: add missing opcodes for fmov on mul alu for V3D 7.x 3018- broadcom: add helpers to identify availability of new unpack modifiers 3019- broadcom/compiler: implement NIR mali opcodes for clamping 3020- broadcom/compiler: generate mali opcodes for clamping on Pi5 3021- broadcom: fix pairing tmu lookup with previous ldtmu 3022- v3dv: drop device_id field 3023- v3dv: rewrite device identification 3024- v3dv: fix leak during device initialization 3025 3026Ian Forbes (1): 3027 3028- driconf: Re-enable GLX_EXT_buffer_age on vmwgfx 3029 3030Ian Romanick (48): 3031 3032- nir/range_analysis: Fix errors in fmin and fmax tables 3033- nir/algebraic: Optimize some bit operation nonsense observed in some shaders 3034- nir/algebraic: Optimize some masking of extract_u8 operations 3035- intel/brw: Fix undefined shift by 64 of uint64_t in brw_compute_first_urb_slot_required 3036- intel/brw: Fix undefined left shift of negative value in update_uip_jip 3037- intel/brw: Fix undefined left shift of large UW value in brw_imm_uw 3038- intel/brw: Fix undefined left shift of negative value in brw_texture_offset 3039- intel/elk: Fix undefined shift by 64 of uint64_t in elk_compute_first_urb_slot_required 3040- intel/elk: Fix undefined left shift of negative value in update_uip_jip 3041- intel/elk: Fix undefined left shift of large UW value in elk_imm_uw 3042- intel/elk: Fix undefined left shift of negative value in elk_texture_offset 3043- intel/brw: Eliminate dead flag writes 3044- intel/brw: Don't propagate saturate to an instruction that writes flags 3045- intel/elk: Don't propagate saturate to an instruction that writes flags 3046- intel/brw: Small code refactor in brw_fs_opt_saturate_propagation 3047- intel/brw: Use def analysis for simple cases of saturate propagation 3048- intel/brw: Move fsat instructions closer to the source 3049- intel/brw: Make 16-bit ishl, ishr, and ushr SSA friendly 3050- intel/brw: Make ifind_msb SSA friendly 3051- nir/algebraic: Improve some find_lsb and ifind_msb patterns 3052- anv: Protect against OOB access to anv_state_pool::buckets 3053- anv: Larger memory pools for huge shaders 3054- intel/brw/xe2: Allow int64 conversions 3055- nir/algebraic: Remove incorrect bfi of iand pattern 3056- nir/divergence: resource_intel is less divergent than you thought 3057- intel/brw: Copy prop from raw integer moves with mismatched types 3058- intel/brw: Run intel_nir_lower_conversions only after brw_nir_optimize 3059- intel/brw: Unconditionally run optimizations after nir_opt_uniform_subgroup 3060- intel/brw: Relax is_partial_write check in cmod propagation 3061- intel/brw: Enable constant propagation for a couple more logical sends 3062- intel/brw: load_offset cannot be constant on this path 3063- nir/algebraic: Optimize more comparisons with b2f 3064- nir/algebraic: Recognize open-coded bitfield_reverse in XCOM 2 3065- nir: Add a pass to generate BFI instructions from logical operations 3066- intel/brw: Use nir_opt_generate_bfi 3067- r600: Use nir_opt_generate_bfi to generate bitfield_select 3068- radeonsi: Use nir_opt_generate_bfi to generate bitfield_select 3069- radv: Use nir_opt_generate_bfi to generate bitfield_select 3070- brw/copy: Don't remove instructions w/ conditional modifier 3071- brw/builder: Add new style ALU3 builder 3072- brw/sat: Fix small typos, copy and paste, etc. 3073- brw/sat: Convert nearly all tests to use new style builders 3074- brw/algebraic: Refactor constant folding out of brw_fs_opt_algebraic 3075- brw/copy: Perform constant folding with constant propagation 3076- brw: Call brw_fs_opt_algebraic less often 3077- brw/emit: Add correct 3-source instruction assertions for each platform 3078- brw/copy: Don't copy propagate through smaller entry dest size 3079- brw/cse: Don't eliminate instructions that write flags 3080 3081Icenowy Zheng (10): 3082 3083- zink: reject Imagination proprietary driver w/o geometryShader 3084- gallivm: orcjit: keep the ownership of tm for LPJit 3085- gallivm: orcjit: use atexit to release LPJit singleton at exit 3086- util: detect LoongArch architecture 3087- gallivm: add LoongArch support to the mattrs setting code 3088- llvmpipe: add LoongArch support in ORCJIT 3089- zink: do not expose quad support when no geometry shader 3090- pvr: enlarge transfer fw_stream buffer for multicore 3091- pvr: add tpu_dm_global_registers feature 3092- pvr: emit tpu_tag_cdm_ctrl in compute stream when present 3093 3094Igor Chernyshev (4): 3095 3096- Add VK_EXT_debug_utils and VK_EXT_scalar_block_layout 3097- Enable VK_KHR_create_renderpass2 in Guest 3098- Move vkUpdateDescriptorSets to hand-written mode 3099- Do not copy invalid descriptor set image view handles 3100 3101Iliyan Dinev (2): 3102 3103- pvr: fix mipmap alignment for non-32bpp textures 3104- pvr: handle PVR_SRV_ERROR_RETRY pvr_srv_rgx_kick_render2() 3105 3106Iván Briano (13): 3107 3108- intel/rt: fix terminateOnFirstHit handling 3109- nir: add pass to convert ViewIndex to DeviceIndex 3110- anv: handle VK_PIPELINE_CREATE_VIEW_INDEX_FROM_DEVICE_INDEX_BIT 3111- anv: fix adding to wa_addr 3112- anv: be consistent about aux usage with modifiers 3113- anv: free shaders on rt pipeline compile error 3114- anv: skip rt pipeline compile if we found all shaders 3115- vulkan: use standard sample locations if there's no VkPipelineSampleLocationsStateCreateInfoEXT 3116- anv: allocate sparse descriptor buffers from the correct heap 3117- hasvk: fix non matching image/view format attachment resolve 3118- brw: fix task/mesh push constant loading 3119- intel/rt: fix ray_query stack address calculation 3120- anv: remove unused/misleading/wrong parameters from the RT trampoline 3121 3122Jami Kettunen (1): 3123 3124- nouveau/headers: Fix build without rustfmt 3125 3126Jason Macnak (49): 3127 3128- Rename imported system directory to guest 3129- Handle AHB and swapchain info in VkBindImageMemoryInfo 3130- Move platform into guest 3131- Move guest/shared/* to guest/* 3132- Add error logging to aid debugging 3133- Make VirtGpu* interfaces 3134- Wrap interactions with sync fds in an interface 3135- Update namespace in libandroidemu to gfxstream::guest 3136- Update iostream namespace to gfxstream::guest 3137- Enable building guest Gfxstream components on host 3138- Wrap AHardwareBuffer_* interactions into Gralloc interface 3139- Update VirtioGpuPipeStream to use VirtGpu abstraction 3140- Move SyncHelper to platform 3141- Revert "[guest] Check that the singleton VirtGpuDevice matches desired capset" 3142- Move info maps into ResourceTracker 3143- Promote testing framework to a proper platform backend 3144- Use libbase's property getting functions 3145- Add dispatch magic for Gfxstream linux ICD 3146- Wrap some classes in "namespace gfxstream::guest {}" 3147- Sort and dedup physical device extensions 3148- Fix "-Werror=conversion" errors for RanchuHwc 3149- Handle AHB R8 format conversions 3150- Make RutabagaLayer shared between "guest impls" 3151- Introduce Gfxstream Features to decouple Gfxstream from AEMU 3152- Ensure glProgramBinary initializes uniform and attrib info 3153- Add composition support to the end2end test framework 3154- Update emulated gralloc to use AHB format 3155- Rename VirtGpuBlob -> VirtGpuResource 3156- Explicitly specify target/bind/bpp in resource creation 3157- Partial revert of aosp/2858589 to avoid Mesa layer for Android 3158- Revert "Partial revert of aosp/2858589 to avoid Mesa layer for Android" 3159- Reland "Partial revert of aosp/2858589 to avoid Mesa layer for Android" 3160- Update vkMapMemory to not hold lock when calling into enc 3161- Remove extra semicolon in VkEncoder 3162- Support BGRA render target in CompositorVk 3163- Adds helper for tracking command buffers with device lost 3164- Adds AHB import and sampling test 3165- Add YUV AHB import and sample end2end tests 3166- Demote some logs to debug only 3167- Disable device groups 3168- Ensure mesa struct is translated to gfxstream struct on Android 3169- Handle DEPTH_STENCIL_ATTACHMENT_BIT in VK -> AHB usage 3170- Handle null handles in vkFreeCommandBuffers 3171- Fix vkAllocateCommandBuffers with failing allocation 3172- Revert "gfxstream: nuke mWorkPool" 3173- Handle external fences in vkGetFenceStatus() 3174- Add perfetto tracing to host renderer 3175- gfxstream: fix log levels in descriptor handling 3176- gfxstream: use gralloc metadata in vkGetAHBPropertiesANDROID 3177 3178Jean-Francois Thibert (4): 3179 3180- Add vkUpdateDescriptorSetWithTemplateKHR to resource tracker 3181- gfxstream: Fix usage of encoder after release in vkResetCommandBuffer 3182- gfxstream: move staging reset to end of reset command buffer 3183- Provided supported formats through gfxstream caps 3184 3185Jean-François Thibert (1): 3186 3187- Add depth formats support to import/export 3188 3189Jesse (1): 3190 3191- amd/vpelib: Config Writer hook and CDC refinement 3192 3193Jesse Natalie (18): 3194 3195- microsoft/clc: Split struct copies before vars_to_ssa in pre-inline optimizations 3196- mr-label-maker: Add d3d10umd label rules 3197- meson: Add an error message for llvmpipe without llvm draw support 3198- wgl: Add missing idep_mesautilformat 3199- d3d12: Don't use a vertex re-ordering GS for line primitives 3200- ci/windows: Use MSVC v143 build tools 3201- ci/windows: Update VK-GL-CTS to d48899f85b486a70d090af59a1453763458611d9 3202- ci/windows: Bump image tags to rebuild deps with v143 build tools 3203- compiler, vk: Support subgroup size of 4 3204- dzn: Drop custom subgroup size handling 3205- microsoft/compiler: Handle subgroup size 4 as a required wave size 3206- microsoft/compiler: Move nir_lower_undef_to_zero out of the optimization loop 3207- d3d12: Fix shader selector hash to hash array instead of pointer-to-array 3208- winsys/d3d12: Use the waitable object to decrease latency 3209- wgl: Add driconf options for controlling latency and swap interval 3210- winsys/d3d12: Implement the latency control option 3211- driconf: Disable dzn for bg3.exe 3212- dzn: Clean up dri options cache 3213 3214Jessica Clarke (3): 3215 3216- Revert "meson: Do not require libdrm for DRI2 on hurd" 3217- Revert "meson: fix with_dri2 definition for GNU Hurd" 3218- meson: egl: Build egl_dri2 driver even for plain DRI 3219 3220Jiale Zhao (1): 3221 3222- llvmpipe: add loongarch util_get_cpu_caps function 3223 3224Jiali (1): 3225 3226- amd/vpelib: Enhance output format capabilities 3227 3228Jianxun Zhang (8): 3229 3230- intel/common: Ensure SIMD16 for fast-clear kernel (xe2) 3231- intel/common: Remove blank lines in intel_set_ps_dispatch_state() (xe2) 3232- anv: Fix assertion failures on BMG (xe2) 3233- iris: Fix an assertion failure with compressed format 3234- anv: Disable compression on legacy modifiers (xe2) 3235- anv: Disable legacy CCS setup in binding (xe2) 3236- Revert "anv: Disable PAT-based compression on depth images (xe2)" 3237- Revert "iris: Disable PAT-based compression on depth surfaces (xe2)" 3238 3239Job Noorman (88): 3240 3241- ir3/postsched: improve debug state dumping 3242- ir3/postsched: fix calculation of max_delay 3243- ir3/postsched: remove unused delay field 3244- ir3/postsched: include ss/sy delay in max_delay 3245- ir3/postched: don't prioritize instructions with soft delays 3246- ir3: fix spill/reload split src/dst regs 3247- ir3: update merge set affinity in shared RA 3248- ir3: fix clearing merge sets after shared RA 3249- ir3: fix wrong dstn used in postsched 3250- ir3: fix counting of repeated registers 3251- ir3: correctly count vectorized instructions for tex prefetch 3252- ir3: add debug option to expand rpt instructions 3253- ir3: print (sat) modifier of instructions 3254- ir3: add backend support for repeated instructions 3255- ir3: add builders for repeated instructions 3256- ir3: make RA aware of repeat groups 3257- ir3: add pre-RA pass to clean up repeat groups 3258- ir3: add post-RA pass to merge repeat groups into rptN instructions 3259- ir3: lower vectorized NIR instructions 3260- nir/opt_vectorize: move rewriting of uses to a function 3261- nir/opt_vectorize: prepare for multiple try_combine functions 3262- nir/opt_vectorize: process blocks in source-code order 3263- nir/opt_vectorize: add support for phi nodes 3264- ir3: add support for vectorized NIR phi nodes 3265- ir3: replace \@load_uniform by new \@load_const_ir3 intrinsic 3266- ir3: rename \@store_uniform_ir3 to \@store_const_ir3 3267- nir/load_store_vectorize: support non-byte offset 3268- nir/load_store_vectorize: support stores without wrmask 3269- nir/load_store_vectorize: add load/store_const_ir3 3270- ir3: enable load/store_const_ir3 vectorization 3271- ir3: add support for rpt bary.f/flat.b 3272- ir3: use rpt instructions for frag coord 3273- ir3: use correct bit size for bools in emit_alu 3274- ir3/ci: remove fixed tests from a307-fails 3275- ir3: make fullsync sync after shared writes 3276- ir3/legalize: handle scalar ALU WAR hazards for a0.x 3277- zink/ci: add a618 flake 3278- freedreno: don't require binning and non-binning inputs to match 3279- ir3: remove unused outputs for binning pass in NIR 3280- ir3: don't modify const state for the binning variant in ir3_cp 3281- ir3: make ir3_const_state less error-prone to use 3282- ir3: fix calling ir3_const_state_mut in the binning VS 3283- ir3: print predicate inversion for branches 3284- ir3: print block divergence info 3285- ir3: remove unnecessary reconvergence point in scan_clusters lowering 3286- ir3: preserve block divergence info in ir3_lower_subgroups 3287- ir3: fix reconvergence points for blocks with two terminators 3288- ir3: fix adding physical edges multiple times 3289- ir3: fix reconvergence of blocks with multiple divergent predecessors 3290- ir3: recalculate reconvergence after opt_jump 3291- ir3: use physical cfg in helper_sched 3292- ir3: fix recognizing const/imm registers as a0 3293- ir3/legalize: don't add WAR dependencies for const/imm regs 3294- ir3: add assert to detect getting reg file of const/imm 3295- ir3/legalize: add needs_ss_war helper 3296- ir3/legalize: don't add (ss) for WAR hazards synced with (sy) 3297- ir3: add is_war_hazard_producer helper 3298- ir3/postsched: take WAR ss-delay into account 3299- ir3/legalize: resolve WAR hazards for stc 3300- ir3/isa: add isaspec definition for shfl 3301- ir3: add ir3_compiler::has_shfl for shfl support on a6xx+ 3302- ir3/print: add support for shfl 3303- ir3: make backend aware of shfl: 3304- ir3: add codegen for rotate 3305- tu: advertise VK_KHR_shader_subgroup_rotate 3306- ir3: disallow immediates for shfl src1 3307- nir/load_store_vectorize: fix division by zero 3308- nir/lower_int64: add nir_intrinsic_read_invocation_cond_ir3 3309- nir/lower_int64: add nir_intrinsic_rotate 3310- ir3: lower 64b intrinsics after generic 64b lowering 3311- ir3: run 64b phi lowering together with other 64b lowerings 3312- ir3: lower 64b SSBO accesses 3313- ir3: make ir3_mem_access_size_align publically available 3314- tu: lower 64b push constants before lowering IO 3315- ir3: add support for 64b reductions 3316- tu: advertise shaderInt64 3317- nir: add shuffle_{xor,up,down}_uniform_ir3 intrinsics 3318- ir3: add codegen for shuffle_{xor,up,down}_uniform_ir3 3319- ir3: add helper to get glsl type for nir_def 3320- ir3: optimize subgroup shuffles using shfl 3321- nir/lower_subgroups: move up some helper functions 3322- nir/lower_subgroups: add build_cluster_mask helper 3323- nir/lower_subgroups: scan/reduce for multiple ballot components 3324- ir3: fix physical edges of predicated branches 3325- freedreno,computerator: add support for local memory 3326- ir3: merge is_reg_gpr and reg_gpr 3327- ir3/ra: prevent moving source intervals for shared collects 3328- ir3,tu: include ir3 debug flags in shader hash key 3329 3330Jocelyn Falempe (3): 3331 3332- loader: Fix typo in __DRI_IMAGE_FORMAT_XBGR16161616 definition 3333- gbm/dri: Use PIPE_FORMAT_* instead of using __DRI_IMAGE_* 3334- gbm/dri: Fix color format for big endian. 3335 3336John Anthony (4): 3337 3338- pan/kmod: Add max_tasks_per_core to kmod props 3339- panvk: Add cmd_dispatch 3340- panvk: Pull out task axis and increment calculation for dispatch 3341- panvk: Add support for CmdDispatchIndirect 3342 3343Jonathan Marek (2): 3344 3345- freedreno: improve a6xx CP_SET_MARKER xml definition 3346- freedreno/a6xx: add missing USES_GMEM flag and BIN_RENDER_END markers 3347 3348Jordan Justen (31): 3349 3350- include/drm-uapi/README: Update README for drm uapi header files 3351- intel/brw/validate: Simplify grf span validation check by not using a mask 3352- intel/brw/validate: Update dst grf crossing check for Xe2 3353- intel/brw/validate: Convert access mask to be grf based 3354- intel/dev: Update hwconfig => max_threads_per_psd for Xe2 3355- intel/dev: Enable BMG PCI IDs (without INTEL_FORCE_PROBE) 3356- anv: Drop "not yet supported" warning for Xe2 3357- anv: Do hasvk devices check first 3358- intel/dev: Add devinfo::probe_forced based on INTEL_FORCE_PROBE 3359- anv: Don't warn about unsupported devices if INTEL_FORCE_PROBE was used 3360- intel/dev: Fix warning for max_threads_per_psd when devinfo->verx10 == 120 3361- intel/dev: Add 0xb640 ARL PCI id 3362- intel/dev: Rework DEVINFO_HWCONFIG; add DEVINFO_HWCONFIG_KV macro 3363- intel/dev: Simplify DEVINFO_HWCONFIG_KV by adding should_apply_hwconfig_item() 3364- intel/dev: Allow specifying a version when to always use hwconfig 3365- intel/dev: Use hwconfig for urb min/max entry values 3366- intel/dev: Support Xe3 device init (for intel_device_info_test) 3367- intel/dev: Add XE3_FEATURES macro 3368- intel/genxml: Start Xe3 support 3369- intel/isl: Build for Xe3 3370- intel/shaders: Build for Xe3 3371- iris: Build for Xe3 3372- intel/dev: Add Xe3 support to get_l3_list() 3373- intel/compiler: Add compiler enum for Xe3 3374- intel/dev: Add INTEL_PLATFORM_PTL platform enum 3375- intel/dev: Add PTL device info 3376- intel/dev: Add PTL PCI IDs (with FORCE_PROBE set) 3377- intel/brw: Allow Xe3 in brw_stage_has_packed_dispatch() 3378- intel/compiler: Xe2 and Xe3 use the same compaction tables 3379- anv: Build for Xe3 3380- intel/dev: Set L3 bank count for Xe2+ from Xe KMD 3381 3382Jose Maria Casanova Crespo (5): 3383 3384- v3d: v3d_resource Use LINEAR layout for importing with INVALID modifier 3385- v3d: avoid load/store of tile buffer on invalidated framebuffer 3386- v3d: initialize job local key with the 8 color buffer available in v7.1+ 3387- v3dv/ci: Add missing fails on RPi4/5 for uprev VKCTS to 1.3.10.0 3388- v3d: Enable Early-Z with discards when depth updates are disabled 3389 3390Joshua Ashton (2): 3391 3392- radv: Enable variableSampleLocations 3393- radv: Fix sample locations at 0 for X/Y 3394 3395Joshua Duong (1): 3396 3397- For goldfish pipe, compute colorBufferMemoryIndex the same as host. 3398 3399José Roberto de Souza (50): 3400 3401- intel/brw: Add a maximum scratch size restriction 3402- anv: Drop useless 'if (total_scratch > 0) {' block in cmd_buffer_ensure_cfe_state() 3403- anv: Wait for Xe exec queue to be idle before destroying it 3404- isl: Fix Xe2 protected mask 3405- anv: Propagate protected information to blorp_batch_isl_copy_usage() 3406- anv: Improve error message when pipeline creation fails during shader compilation 3407- anv: Handle internal shader compilation failure 3408- intel: Sync xe_drm.h 3409- intel/dev: Support new topology type with SIMD16 EUs 3410- docs: Move all ANV environment variables to a single section in envvars 3411- anv: Disable sparse even on Xe KMD with ANV_SPARSE 3412- docs/anv: Document ANV_SPARSE and ANV_SPARSE_USE_TRTT 3413- anv: Extend ANV_QUEUE_OVERRIDE to blit count 3414- intel: Remove INTEL_ENGINE_CLASS_COMPUTE and INTEL_ENGINE_CLASS_COPY parameters 3415- docs/anv: Remove documentation about INTEL_COMPUTE_CLASS and INTEL_COPY_CLASS 3416- anv: Add trivial_batch and query-pool to the error capture 3417- anv: Use batch_bo_pool in utrace anv_async_submit_init() calls 3418- anv: Nuke anv_utrace_submit::trace_bo 3419- intel/isl/gfx20: Alow hierarchial depth buffer write through for multi sampled surfaces 3420- anv/gfx20: Enable depth buffer write through for multi sampled images 3421- iris/gfx20: Enable depth buffer write through for multi sampled images 3422- anv: Nuke perf_metric 3423- intel/perf: Fix intel_gem.h include 3424- anv: Drop useless '>= 0' check over a unsigned 3425- anv: Add anv_device_perf_close() 3426- anv: Fix context id or exec queue used to open perf stream 3427- anv: Add warning about mismatch between query queues 3428- anv: Make sure all previous vm binds are done before execute perf query pool 3429- anv: Check if vkCreateQueryPool() is being created in a supported queue 3430- anv: Add documentation to some fields in anv_query_pool 3431- anv/query: Fix batch end value 3432- anv: Improve readbility of khr_perf_query_availability_offset() and khr_perf_query_data_offset() 3433- anv: Fix condition to clear query pool with blorp 3434- intel: Split anv_xe_wait_exec_queue_idle() and move part of it to common/ 3435- iris: Use xe_queue_get_syncobj_for_idle() 3436- anv: Optimize vkQueueWaitIdle() on Xe KMD 3437- intel/perf: Sync oa-lnl.xml 3438- intel/perf: Add BMG OA support 3439- intel/genxml: Append 'Z Async Throttle settings' to gfx125 definition of STATE_COMPUTE_MODE 3440- anv: Set all async compute registers in STATE_COMPUTE_MODE 3441- intel/genxml: Do small fixes in gfx20 definition of STATE_COMPUTE_MODE 3442- anv: Set recommended values for gfx20 async compute registers in STATE_COMPUTE_MODE 3443- iris: Program async compute registers of STATE_COMPUTE_MODE in compute engine 3444- intel/dev/xe: Set max_eus_per_subslice using topology query 3445- intel/dev: Fix max_cs_threads value on simulator 3446- intel/perf: Add OA support to ARL 3447- drm-uapi: Sync xe_drm.h 3448- intel/perf: Extend intel_perf_stream_set_metrics_id() to syncronize metrics id changes 3449- intel/perf: Add INTEL_PERF_FEATURE_METRIC_SYNC and check if KMD supports it 3450- anv: Enable perf metrics id set syncronization 3451 3452Juan A. Suarez Romero (32): 3453 3454- v3d: use operations to specify what to save in blitter 3455- v3d: add new clear blitter op 3456- v3d/ci: add failures from piglit all profile 3457- v3d/ci: add new flakes 3458- vc4: fix typo in simulator conditional declaration 3459- v3d/vc4/ci: add address sanitizer jobs 3460- v3d/ci: enable full run jobs for rpi5 3461- broadcom/ci: annotate some of the failures 3462- v3d/ci: add new flakes in rpi5 3463- v3d/ci: update number of rpi5 available for testing 3464- v3d: free dbuf 3465- broadcom/ci: use kernel8 for rpi5 full jobs 3466- v3d/ci: update expected results 3467- vc4/ci: update expected results 3468- Revert "ci: take igalia farm offline" 3469- vc4/ci: update expected results 3470- vc4: build simulator references conditionally 3471- v3d/v3dv: build simulator references conditionally 3472- v3d/ci: add new flake for rpi5 3473- Revert "v3d: never replace a mapped bo" 3474- v3d: do not rebind a sampler view already rebound 3475- v3d/ci: add new flake for rpi5 3476- v3d/ci: update expected results 3477- v3dv/ci: add new flake 3478- broadcom: do not include simulator if not available 3479- v3d/ci: add new flakes and timeouts 3480- vc4: handle nir_op_ult32 intrinsic 3481- v3d/v3dv/ci: update expected results 3482- v3d/ci: use special kernel for full jobs 3483- v3dv/ci: move asan fail to skip 3484- v3d: add new flake 3485- v3d/ci: add OpenCL failures 3486 3487Jude Shih (2): 3488 3489- amd/vpelib: Dynamic Command List Expansion 3490- amd/vpelib: Restructure CDC FE/BE 3491 3492Jules Blok (5): 3493 3494- vulkan: Add support for VK_EXT_depth_clamp_control 3495- radv: Add support for VK_EXT_depth_clamp_control 3496- anv: Add support for VK_EXT_depth_clamp_control 3497- nvk: Add support for VK_EXT_depth_clamp_control 3498- docs: Update status of VK_EXT_depth_clamp_control support 3499 3500Julia Zhang (1): 3501 3502- mesa/st: use drawable->ID as hash for drawable_ht 3503 3504Juston Li (5): 3505 3506- anv/android: remove unneeded ANB implicit import flags 3507- anv/android: handle R8G8B8X8 as R8G8B8A8 3508- anv/android: refactor out u_gralloc tiling query 3509- anv/android: resolve ANB swapchain images on bind 3510- android: look for debug/vendor prefixed options 3511 3512K900 (5): 3513 3514- gbm: split the DRI backend into a separate library and unify backend handling 3515- meson: require libdrm when building gbm 3516- meson/gbm: clean up dependencies 3517- meson: remove dri-search-path 3518- meson/gbm: fix meson devenv for GBM backends 3519 3520Kai Wasserbäch (2): 3521 3522- fix(FTBFS): clover: use \`.getDataLayout()` with LLVM >= 20 3523- fix(FTBFS): build: link OpenCL target with \`dep_clang` from top-level 3524 3525Karmjit Mahil (12): 3526 3527- freedreno: Enable the A735 3528- tu: Set \`TU_ACCESS_CCHE_READ` for transfer ops with read access 3529- freedreno: Define SP_DITHER_CNTL (0xA9AC) 3530- tu: Implement VK_EXT_legacy_dithering 3531- ir3: Use \`foreach_instr_safe` in ir3_shared_folding 3532- pvr: Fix \`barier`->`barrier` typo 3533- pvr: Use a pixel size of \`0` for invalid pbe accum formats 3534- pvr: Regenerate all descriptor program data sections 3535- pvr: Handle DISABLE_PIXELMERGE 3536- tu: Fix push_set host memory leak on command buffer reset 3537- tu: Fix potential alloc of 0 size 3538- nir: Fix \`no_lower_set` leak on early return 3539 3540Karol Herbst (101): 3541 3542- spirv: generate info for FunctionParameterAttribute 3543- spirv: initial parsing of function parameter decorations 3544- spirv: handle function parameters passed by value 3545- nak: allow clippy::not_unsafe_ptr_arg_deref lints 3546- nak: fix clippy::if_same_then_else error 3547- rusticl/platform: add perf debug option 3548- rusticl/memory: remove stale TODOs 3549- rusticl/memory: add a couple of performance warnings 3550- rusticl: move more kernel related things into kernel.rs 3551- rusticl/kernel: make some types and fields private 3552- rusticl/kernel: move setting of nir metadata into lower_and_optimize_nir 3553- rusticl/program: move attribute_str to the spirv module 3554- rusticl: use blob for all serialization 3555- rusticl: remove unsued serialization helpers 3556- rusticl/kernel: add type for the convert_spirv_to_nir return value 3557- rusticl/kernel: move internal_args into NirKernelBuild 3558- clc: force linking of spirvs with mismatching pointer types in signatures 3559- rusticl/spirv: do not add additional new lines when joining logs 3560- rusticl: fix clippy lint having bounds defined in multiple places 3561- rusticl/program: protect against 0 length in slice::from_raw_parts 3562- rusticl/api: protect against 0 length in slice::from_raw_parts 3563- rusticl/spirv: protect against 0 length in slice::from_raw_parts 3564- nouveau: handle realloc failure inside cli_kref_set 3565- rusticl: bump rust req to 1.76 3566- rusticl: use div_ceil 3567- rusticl: use next_multiple_of 3568- rusticl: use pointer byte offset API 3569- rusticl: replace some raw pointer casts 3570- rusticl: use ptr::from_ref and ptr::from_mut 3571- rusticl/icd: use ptr::addr_eq 3572- meson: centralize rust handling 3573- docs: move rust update policy 3574- meson: centralize bindgen version check 3575- meson: centralize checking for new enough meson for rust support 3576- nir: add load_global_size intrinsic 3577- rusticl/kernel: emit system values with the correct device size 3578- rusticl/kernel: handle load_global_size 3579- mesa: check for enabled extensions for \*UID enums 3580- nouveau/winsys: fix handling of NV_DEVICE_TYPE_IGP 3581- nouveau: use nv_devince_info and fill in PCI and type information 3582- nouveau: add nv_device_uuid 3583- nouveau: implement driver_uuid and device_uuid 3584- nvk: use nv_device_uuid 3585- nouveau: remove nouveau_check_for_uma 3586- nouveau: implement PIPE_PCI caps 3587- nouveau: use nv_device_info directly for dumping push buffers 3588- zink: lower 64 bit find_lsb, ufind_msb and bit_count 3589- zink: lower 8/16 bit alu ops vk spirv doesn't allow 3590- rusticl/kernel: properly respect device thread limits per dimension 3591- rusticl/device: Make supported 1Dbuffer formats a strict subset of 1D 3592- rusticl/memory: Fix memory unmaps after rework 3593- rusticl/image: take pitches into account when allocating memory for maps 3594- rusticl/image: properly sync mappings content for 1Dbuffer images 3595- rusticl/queue: add clSetCommandQueueProperty 3596- util/u_printf: do not double print format string with unused arugments 3597- rusticl/memory: fix sampler argument size check 3598- rusticl/kernel: move kernel size into Constant arg type 3599- rusticl/kernel: rename InternalKernelArg to CompiledKernelArg 3600- rusticl/kernel: remove CompiledKernelArg::size 3601- rusticl/kernel: make CompiledKernelArg::offset a u32 3602- rusticl/kernel: make KernelArg::offset and KernelArg::binding proper u32 values 3603- rusticl/kernel: rename InternalKernelArgType to CompiledKernelArgType 3604- rusticl/kernel: move serialization code of KernelArgType 3605- rusticl/kernel: big kernel arg rework 3606- rusticl/kernel: add KernelArgType::is_opaque 3607- rusticl/kernel: move CompiledKernelArg decleration above the impl block 3608- rusticl/kernel: simplify adding internal kernel args 3609- rusticl/mem: do not check against image base alignment for 1Dbuffer images 3610- rusticl: do not use CL vector types in bindings and code 3611- ac/llvm: fix umul_high 3612- rusticl/device: limit CL_DEVICE_IMAGE_MAX_BUFFER_SIZE more aggressively 3613- vtn: ignore volatile on functions for now 3614- clc: add support for more image related extensions 3615- compiler/types: Add multisample vimage/vtexture types 3616- nir: Support multisampled images in lower_read_only_images_to_tex() 3617- vtn, nir: handle OpImageQueryLevels on images 3618- vtn: mark ImageMipmap as supported 3619- rusticl: reuse PipeContext 3620- zink: cache generated vendor and device name inside zink_screen 3621- rusticl/mesa: Return CStr for device and vendor names. 3622- clc: fix compilation error with llvm-20 3623- rusticl/kernel: move dead_var options into constant space 3624- rusticl/platform: add env variable to disable kernel variants 3625- rusticl/mesa: add more workgroup_Size functions to NirShader 3626- rusticl/mesa: implement Clone for NirShader 3627- rusticl/kernel: move assign_locations into CompiledKernelArg 3628- rusticl/kernel: split up nir compilation into various steps 3629- rusticl/kernel: track if arg is dead inside CompiledKernelArg 3630- rusticl/kernel: add CompilationResult to hold compilation artifacts 3631- rusticl/kernel: add optimized Kernel variant 3632- rusticl: add nir debugging option 3633- rusticl: allow devices to be enabled by default 3634- docs: document gallium-rusticl-enable-drivers 3635- asahi: implement get_cl_cts_version 3636- rusticl: allow asahi to be enabled by default 3637- zink: add CL CTS result 3638- radeonsi: move si_compute::global_buffers to si_context 3639- nv/codegen: Do not use a zero immediate for tex instructions 3640- nvc0: return NULL instead of asserting in nvc0_resource_from_user_memory 3641- rusticl/kernel: fix kernel variant selection 3642- vtn: handle struct kernel arguments passed by value 3643 3644Kenneth Graunke (48): 3645 3646- intel/brw: Don't force g1's live range to be the entire program 3647- intel/brw: Delete fs_reg_alloc::discard_interference_graph() 3648- intel/brw: Record that SHADER_OPCODE_SCRATCH_HEADER uses g0 3649- intel/brw: Record g0 as live for sends with send_ex_desc_scratch set 3650- intel/brw: Only force g0's liveness to be the whole program if spilling 3651- intel/brw: Mark all UBO access with a direct buffer index as speculative 3652- intel/brw: Delete the brw_fs_opt_peephole_select() pass 3653- intel/brw: Delete the brw_fs_opt_dead_control_flow_eliminate() pass 3654- intel/brw: Replace predicated break optimization with a simple peephole 3655- intel/brw: Set appropriate types for 16-bit sampler trailing components 3656- intel/brw: Drop image_{load,store}_raw_intel handling 3657- intel/brw: Pass opcode to brw_swsb_encode/decode 3658- intel/brw: Fix Xe2+ SWSB encoding/decoding for DPAS instructions 3659- intel: Fix bad align_offset on global_constant_uniform_block_intel 3660- intel/brw: Fix OOB reads when printing instructions post-reg-alloc 3661- intel/brw: Print blocks in brw_print_instructions_to_file() 3662- intel/brw: Add a file parameter to idom_tree::dump() 3663- intel/brw: Switch from LSC CMASK opcodes to regular LOAD/STORE 3664- intel/brw: Get rid of the lsc_msg_desc_wcmask helper 3665- intel/brw: Drop misguided sign extension attempts in extract_imm() 3666- intel/brw: Fix extract_imm for subregion reads of 64-bit immediates 3667- intel/brw: Use NUM_BRW_OPCODES in can_omit_write() check 3668- intel/brw: Use size_written for NoMask instructions in is_partial_write 3669- intel/brw: Rename lsc_aop_for_nir_intrinsic to "op" instead of "aop" 3670- intel/brw: Handle load/stores in lsc_op_for_nir_intrinsic() 3671- intel/brw: Introduce new MEMORY_*_LOGICAL opcodes 3672- intel/brw: Expose functions to convert LSC enums to strings 3673- intel/brw: Pretty-print memory logical opcodes 3674- intel/brw: Lower MEMORY_OPCODE_*_LOGICAL to LSC messages 3675- intel/brw: Lower MEMORY_OPCODE_*_LOGICAL to HDC messages 3676- intel/brw: Switch load_num_workgroups to the new memory intrinsic 3677- intel/brw: Switch to emitting MEMORY_*_LOGICAL opcodes 3678- intel/brw: Switch load_ubo_uniform_block_intel over to memory intrinsics 3679- intel/brw: Delete old-style surface and A64 message opcodes 3680- intel/brw: Don't include sync.nop in INTEL_DEBUG instruction counts 3681- nir: Don't generate single iteration loops to zero-initialize memory 3682- intel/brw: Fix spill/fill count for load/store_scratch in SIMD32 3683- intel/brw: Delete Gfx7-8 code from emit_barrier() 3684- intel/brw: Make a ubld temporary in emit_barrier() 3685- intel/brw: Fix register and builder size in emit_barrier() for Xe2 3686- intel/brw: Delete more Gfx8 code from brw_fs_combine_constants 3687- intel/brw: Use whole 512-bit registers in constant combining on Xe2 3688- brw: Optimize 16-bit texture fetches later 3689- intel/brw: Allow immediates in the BFE instruction on Gfx12+ 3690- brw/validate: Return an error for Align16 access mode on Icelake+ 3691- brw/emit: Fix align16 3src subregister encodings for HF types 3692- intel: Set shader_spilling_rate=11 in intel_clc 3693- brw: Fix try_rebuild_source's ult32/ushr handling to use unsigned types 3694 3695Konrad Dybcio (1): 3696 3697- freedreno: Add initial A621 support 3698 3699Konstantin (5): 3700 3701- util: Handle more formats in get_plane_(width|height) 3702- vtn: Remove dead shader_call_data from all RT stages 3703- radv: Handle instruction encodings > 8 bytes when splitting disassembly 3704- radv: Handle repeated instructions when splitting disassembly 3705- meson: Allow building lavapipe without specifying llvmpipe 3706 3707Konstantin Seurer (28): 3708 3709- llvmpipe: Improve PIPE_QUERY_PIPELINE_STATISTICS accuracy 3710- lavapipe: Fix box sorting during ray traversal 3711- aco: print s_delay_alu INSTSKIP>3 correctly 3712- gallivm: Use variables for loop exec mask handling 3713- gallivm: Skip inactive branches 3714- radv: Synchronize shader dumping 3715- llvmpipe: Use derivative intrinsics 3716- nir: Introduce nir_debug_info_instr 3717- spirv: Emit nir_debug_info_instr for OpLine 3718- nir/print: Add a helper for generating debug info 3719- nir/opt_loop: Fix handling else-breaks in merge_terminators 3720- gallium,st/mesa: Add and set pipe_image_view::is_2d_view_of_3d 3721- lavapipe: Implement VK_EXT_image_2d_view_of_3d with sparse textures 3722- lavapipe: Do not adjust imageGranularity for different block sizes 3723- radv: Work around broken terrain in Warhammer III 3724- radv: Initialize sqtt state before meta state 3725- radv: Workaround apps using ray tracing when it is unsupported 3726- lavapipe: Fix report_ray_intersection affecting terminated rays 3727- lavapipe: Do not return in report_ray_intersection 3728- radv: Fix report_ray_intersection affecting terminated rays 3729- lavapipe: Implement clustered reductions 3730- lavapipe: Implement VK_KHR_shader_subgroup_rotate 3731- Revert "lavapipe: Do not return in report_ray_intersection" 3732- lavapipe: Set minAccelerationStructureScratchOffsetAlignment to 8 3733- radv/meta: Do not pass NULL to vk_texcompress_astc_finish 3734- nir/print: Fix the alignment of 8-bit definitions 3735- radv: Remap 10 and 12 bit formats to 16 bit formats 3736- radv: Move ac_addrlib to the physical device 3737 3738Kovac, Krunoslav (4): 3739 3740- amd/vpelib: DPP starting changes 3741- amd/vpelib: MPC refactoring HW registers 3742- amd/vpelib: Refactor MPC registers 3743- amd/vpelib: Reuse existing float to reg format conversion 3744 3745Lars Harrison (3): 3746 3747- Provide error message on unhandled VK structs 3748- Always initialize gfxstream vulkan 3749- Fix coherent memory allocation to use device 3750 3751Lars-Ivar Hesselberg Simonsen (12): 3752 3753- panfrost: Move ForEachMacros into panfrost 3754- pan/genxml: Add pan_pack_nodefaults() 3755- panfrost: Fix near/far depth clip 3756- panfrost: Enable support for depth clamping 3757- panvk: Fix paths that assume a fragment shader 3758- panvk: Fix use of cs_sync64_wait in cmd_wait_event 3759- panvk: Map device memory for PANVK_DEBUG=trace 3760- panvk: Use the passed Vertex Attribute divisor 3761- panvk: Only set index buffer size for DrawIndexed 3762- panvk: Add get_tiler_idvs_flags 3763- panvk: Split panvk_cmd_draw into prepare_draw 3764- panvk: Add support for Draw[Indexed]Indirect 3765 3766Leder, Brendan Steve (4): 3767 3768- amd/vpelib: Fix output_ctx gamma curve cache bug 3769- amd/vpelib: Fix backend bug for multiple instances 3770- amd/vpelib: Generalize visual confirm handling 3771- amd/vpelib: Update chip headers 3772 3773Lepton Wu (1): 3774 3775- egl/android: Fix wrong pipe format for RGB_565 3776 3777Lin, Ricky (4): 3778 3779- amd/vpelib: Added JFIF format to RGB output side 3780- amd/vpelib: Use VPE_IP_LEVEL_1_0 for VPE IP 6.1.3 3781- amd/vpelib: Remove extra collaborate sync commands in IB 3782- amd/vpelib: Increase the CD field in vpe descriptor programming 3783 3784LingMan (1): 3785 3786- nak/hw_tests: Use f32::to_bits() 3787 3788Lionel Landwerlin (139): 3789 3790- anv: split instance from anv_device.c 3791- anv: split physical_device from anv_device.c 3792- anv: split events from anv_device.c 3793- anv: split buffer from anv_device.c 3794- anv: split sampler from anv_device.c 3795- anv: split buffer view from anv_image.c 3796- anv: split image view from anv_image.c 3797- isl: account for protection in base usage checks 3798- anv: properly flag image/imageviews for ISL protection 3799- anv: propagate protected information for blorp operations 3800- anv: stop using 3DSTATE_WM::ForceThreadDispatchEnable 3801- anv: fix check on pipeline mode to track buffer writes 3802- u_trace: add mako exception printouts 3803- u_trace: rework tracepoint argument declaration 3804- u_trace: remove timestamp reference in allocations 3805- u_trace: add support for indirect data 3806- anv/hasvk: add indirect tracepoint arguments 3807- u_trace: add csv output 3808- util/debug: update parse_enable_string to deal with +all/-all 3809- anv: better signal new frames to utrace 3810- vulkan/runtime: allow null/empty debug names 3811- anv: reuse object string for RMV token 3812- anv: prevent asserts with debug printf in internal shaders 3813- anv: remove some unused includes 3814- anv: reuse cs_prog_data pointer 3815- intel-clc: missing printf lowering 3816- anv: add missing MEDIA_STATE_FLUSH for internal shaders 3817- anv: ensure max_plane_count is at least 1 3818- genxml: unify some bits between Gfx8/Gfx11/Gfx12.5 3819- intel/mi_builder: enable control flow API on Gfx9+ 3820- anv/blorp: force CC_VIEWPORT reallocation when programming 3DSTATE_VIEWPORT_STATE_POINTERS_CC 3821- anv: limit some state dirtying after blorp/simpler-shaders 3822- brw/rt: fix ray_object_(direction|origin) for closest-hit shaders 3823- intel/nir: remove load_global_const_block_intel intrinsic 3824- vulkan/runtime: fix GetBufferMemoryRequirements2 for maintenance4 3825- anv: fix extended buffer flags usages 3826- anv: remove unused macro 3827- anv: only set 3DSTATE_CLIP::MaximumVPIndex once 3828- anv: add assert to detect problematic instruction merges 3829- anv: reuse common pipeline state for compute push allocations 3830- anv: simplify loading driver internal constants 3831- anv: move lowering of descriptor intrinsics to apply_layout 3832- nir: remove unused intel intrinsics 3833- brw: remove unused prog_data field 3834- anv: optimize CLIP::MaximumVPIndex setting 3835- anv: move conditional render predicate after gfx_flush_state 3836- anv: don't miss workaround for indirect draws 3837- anv: optimize STATE_BYTE_STRIDE emission 3838- anv: explicitly disable BT pool allocations at device init 3839- anv: always use workaround_address, not workaround_bo 3840- anv: order data in wa_bo to leave wa_addr last 3841- nir/divergence: add missing load_constant_base_ptr 3842- brw: switch mesh/task URB fence prior to EOT to GPU 3843- brw: add a comment what Gfx12.5 URB fences 3844- anv: fix utrace compute timestamp reads on Gfx20 3845- iris: fix utrace compute end timestamp reads on Gfx20 3846- brw: align spilling offsets to physical register sizes 3847- anv: selectively disable binding table usage on Gfx20 3848- brw: remove (load|store)_raw_intel 3849- brw: fix TGM messages to use cmask lsc opcodes 3850- clc: find opencl headers from the installed llvm/clang location 3851- brw: use a builder of the size of the physical register for uniforms 3852- brw: fix vecN rebuilds 3853- brw: fix virtual register splitting to not go below physical register size 3854- anv: fix missing tracking for alpha-to-coverage runtime changes 3855- anv: Only flush render target cache when detecting RT changes 3856- iris: ensure null render target for specific cases 3857- brw: move null_rt control up a layer 3858- brw: disable null_rt only if color output does not affect other outputs 3859- anv: add missing pipeline instance multiplier 3860- brw: make sampler message emission more generic 3861- brw: fix mask componentation for 16-bit sampler returns 3862- brw: remove EOT handling from sampler messages 3863- intel/decoder: constify functions not modifying instructions/fields 3864- intel/decoder: add filter feature 3865- intel/decoder: split state tracking handlers from printing ones 3866- intel/decoder: decode the 8 BLEND_STATEs 3867- anv: avoid setting up a null RT unless needed 3868- anv: implement VK_KHR_dynamic_rendering_local_read 3869- anv: expose VK_EXT_pipeline_protected_access 3870- zink: avoid host transfer usage with sparse 3871- blorp: convert fast clear color for unsupported formats 3872- anv: limit render target cache flushing due to color output remapping 3873- isl: fix comment typo 3874- isl/tests: add a space in memcpy test error printouts 3875- isl/tests: fixup multi-tile testing 3876- isl/tests: fix coordinates for comparison scan 3877- isl/tests: rename ytile to be consistent with other names 3878- isl/tests: bump coordinate types to 32bits 3879- isl/tests: use modulo for more variability per pixel 3880- isl/tests: fix inclusive coordinate checking 3881- isl/tests: add TileX memcpy testing 3882- isl/tests: rename tile_(width|height) variables 3883- isl/tests: split linear & tiled buffer sizes 3884- anv: limit 22018402687 to impacted platforms 3885- anv: consolidate pre/post draw workaround in helpers 3886- anv: optimize WA 16011107343/22018402687 3887- isl: remove duplicated copy for tileX/TileY 3888- isl/tests: add more coordinates for full tiles testing 3889- isl/tests: rename span variable to xt_sub_range_alignment 3890- isl: Tile W memcpy support 3891- iris: delete stencil mapping support 3892- isl: add support of aux disable bit on hiz 3893- anv: use stage mask to deduce cs/pb-stall requirements 3894- brw: remove unused prototype 3895- brw: delay printf lowering 3896- brw: avoid clashing nested loop indices 3897- brw: remove rebuild single element special case 3898- brw: fix mesh fence emission 3899- brw/nir: rework inline_data_intel to work with compute 3900- anv: stop using a binding table entry for gl_NumWorkgroups 3901- elk: Don't apply discard_if condition opt if it can change results 3902- isl: fix range_B_tile end_tile_B value 3903- isl/tests: add range_B_tile test 3904- anv: allow subresource queries on non-linear images 3905- anv: drop non host memory types for host-transfer on non-rebar 3906- anv: disable pat compression for host images 3907- anv: wrap binding address setting 3908- anv: add a host map of image for host image copy usage 3909- anv: factor out sanitizing mmap offset code 3910- anv: add VK_EXT_host_image_copy support 3911- anv: fix binding table entry count for compute shaders 3912- anv: fix missing inline parameter emission 3913- anv: reemit push constants on pipeline changes 3914- anv: use UINT32_MAX to be consistent 3915- anv: avoid companion usage on RCS 3916- anv: avoid L3 fabric flush in pipeline barriers 3917- anv: update some of the indirect invalidations 3918- vulkan/runtime: fix allocation failure handling 3919- anv: fix even set/reset on blitter engine 3920- anv: add texture cache inval after binding pool update 3921- anv: fix indentation 3922- anv: move pipe control debug to anv_util.c 3923- anv: fix extent computation in image->image host copies 3924- anv: update shader descriptor resource limits 3925- brw: allocate physical register sizes for spilling 3926- anv: fix incorrect aspect flag for depth/stencil formats 3927- anv: fix missing push constant reallocation 3928- anv: prevent access to destroyed vk_sync objects post submission 3929 3930Louis-Francis Ratté-Boulianne (18): 3931 3932- gallium: add PIPE_CAP_ASTC_DECODE_MODE and expose extension 3933- panfrost: add support for ASTC decode mode extension 3934- pan/bi: implement nir_intrinsic_load_draw_id 3935- panfrost: use special DrawID register on CSF-based GPUs 3936- panfrost: properly lower DrawID sysval on v9 GPUs 3937- panfrost: initialize all UBO and vertex buffer descriptors 3938- pan/cs: only flush block instrs if there is a pending \`if` block 3939- pan/cs: add block to handle registers backup in exception handler 3940- pan/genxml: enable decoding for SET_EXCEPTION_HANDLER 3941- pan/genxml: small fix to FINISH_FRAGMENT tracing 3942- panfrost: add debug tracing for CSF context initialization 3943- panfrost: create a backend-specific hook for \`emit_fbds` 3944- panfrost: add debug flag to enable extra CS checks 3945- panfrost: add support for incremental rendering in CSF 3946- panfrost: properly update data size of AFBC-packed resources 3947- panfrost: disable CRC for AFBC-packed resources 3948- panfrost: check the right discard property when selecting CRC target 3949- panfrost: properly align CRC buffer size for prefetching 3950 3951Lu Yao (1): 3952 3953- ac/radeonsi: compute htile for tile mode RADEON_SURF_MODE_1D on GFX6-8 3954 3955Lucas Fryzek (13): 3956 3957- llvmpipe: Implement EGL_ANDROID_native_fence_sync 3958- lavapipe: Implement VK_KHR_external_*_fd 3959- venus/ci: Update skip tests to prevent timeouts 3960- lp: only map dt buffer on import from dmabuf 3961- egl/surfaceless: Add error checking to create image 3962- llvmpipe: Add android platform integration 3963- vulkan/runtime: Use DMA_BUF to import anb 3964- lavapipe: Add android platform integration 3965- docs/android: Add llvmpipe instructions 3966- lavapipe: Add AHB extension 3967- drisw: Copy entire buffer ignoring damage regions 3968- egl/dri/wl: Move swrast damage region from put to swap 3969- lp: Only close udmabuf handle if its valid 3970 3971Lucas Stach (18): 3972 3973- etnaviv: drm: use COARSE clock for BO cache timing 3974- etnaviv: drm: use COARSE clock for timeouts when possible 3975- etnaviv: hwdb: add COMPUTE_ONLY cap 3976- etnaviv: properly set PIPE_CAP_GRAPHICS 3977- etnaviv: emit SAMPLER_LOG_SIZE on sampler state changes 3978- etnaviv: wire up missing gl_VertexID bits 3979- etnaviv: allow shader machine code dumps in release builds 3980- etnaviv: singlethread shader variant compiles when dumping shaders 3981- etnaviv: emit all PA shader attributes 3982- etnaviv: support more VS outputs on halti5 GPUs 3983- etnaviv: fix total varying count assertion 3984- etnaviv: validate number of VS outputs against GPU limit 3985- etnaviv: limit number of varyings to fit into VS outputs 3986- gbm: mark surface buffers as explicit flushed 3987- etnaviv: re-emit uniforms on sampler view changes when txs is used 3988- etnaviv: flush shader caches when UBO content is modified 3989- etnaviv: Update headers from rnndb 3990- etnaviv: set PE_COLOR_FORMAT_OVERWRITE when no color target is active 3991 3992Luigi Santivetti (4): 3993 3994- pvr: fix calculation for textures z position fractional part 3995- pvr: fix when to emit ppp state uniform base address 3996- pvr: for fragment stages only do not insert a barrier 3997- pvr: really free memory in subpass render init 3998 3999M Henning (15): 4000 4001- nak: Don't mix up two types of barrier 4002- nak: Rename num_barriers to num_control_barriers 4003- nouveau/codegen: Switch to derivative intrinsics 4004- nv50/peephole: merge/split can swap word order 4005- nv50_ir_from_nir: Use getFile() more often 4006- nv/codegen: Use nir_lower_mem_access_bit_sizes 4007- nv/codegen: Limit MemoryOpt to input/output/const 4008- nv/codegen: Vector load ssbo/global/shared/scratch 4009- nv/codegen: Vector store ssbo/global/shared/scratc 4010- nv/codegen: Modify nir pass order 4011- nv/codegen: Use nir_opt_load_store_vectorize 4012- nv/codegen: Stop overaligning var_function_temp 4013- nak: Phi coalescing via biased register coloring 4014- nvk/cmd_buffer: Pass count to set_root_array 4015- nvk: Fix invalidation of NVK_CBUF_TYPE_DYNAMIC_UBO 4016 4017Maaz Mombasawala (4): 4018 4019- svga: Validate surface during copy check 4020- svga: Ensure an active context exists for a vmw_screen 4021- svga: Introduce userspace managed surfaces 4022- svga: Support older define commands for userspace surfaces 4023 4024Mahmood - Zer0xFF (1): 4025 4026- d3d12: Fix Xbox GDK Build 4027 4028Marcin Radomski (7): 4029 4030- Add missing includes 4031- Enable VK_EXT_image_drm_format_modifier in cerealgenerator.py 4032- Enable VK_EXT_image_drm_format_modifier on Linux guests 4033- vkAllocateMemory: support DRM format modifiers for dmabuf memory 4034- vkGetPhysicalDeviceImageFormatProperties2: support drm modifiers 4035- vkCreateInfo: support VkImageDrmFormatModifierListCreateInfoEXT 4036- Add mesa3d_platforms Soong config flag 4037 4038Marek Olšák (189): 4039 4040- nir/opt_algebraic: use fmulz for fpow lowering to fix incorrect rendering 4041- nir: add nir_intrinsic_load_per_primitive_input, split from io_semantics flag 4042- nir/opt_varyings: improve convergent input handling to fix data corruption 4043- amd: expose nir_io_mix_convergent_flat_with_interpolated 4044- ac,radeonsi: set 16-bit flags in io_options optimally 4045- nir/opt_vectorize_io: optionally don't vectorize IO with different types 4046- ac/surface/gfx12: fix setting tile_swizzle 4047- radeonsi/gfx12: fix a GPU hang due to an invalid packet with window rectangles 4048- radeonsi: ensure TC_L2_dirty is set if we don't sync after internal SSBO blits 4049- radeonsi: don't set TC_L2_dirty in si_query_hw_get_result_resource 4050- radeonsi: fix buffer coherency issues on gfx6-8,12 due to missing PFP->ME sync 4051- radeonsi/gfx12: fix register programming to fix GPU hangs 4052- radeonsi/gfx11: prepare for CWSR 4053- radeonsi: program OREO_MODE optimally on gfx11.x and gfx12 4054- radeonsi/gfx12: fix VS output corruption with streamout 4055- ac/surface/gfx12: turn off HiZ for pre-production samples 4056- radeonsi: strengthen the condition determining OREO_MODE 4057- nir: add ACCESS_KEEP_SCALAR, preventing vectorization 4058- util: move util_lower_clearsize_to_dword here 4059- radeonsi: reject insert/extract opcodes in si_vectorize_callback 4060- radeonsi: test more alignment cases in si_test_dma_perf 4061- radeonsi: add correctness tests for the clear/copy_buffer compute shader 4062- radeonsi: minor changes at the beginning of si_compute_clear_copy_buffer 4063- radeonsi: implement optimized unaligned clear/copy_buffer compute shader 4064- radeonsi: align waves to 256B clear/copy area for the clear/copy_buffer shader 4065- ac/nir: import the clear/copy_buffer compute shader from radeonsi 4066- ac/nir: adjust performance-related decisions for clear/copy_buffer shader 4067- ac/nir: adjust gfx11 tuning for the compute blit 4068- amd: update addrlib 4069- ac/nir/tess: don't allocate LDS for HS inputs that are passed via VGPRs 4070- radeonsi: pass TCS inputs_read mask to LS output lowering on GFX9 + monolithic 4071- ac: fix WAVES_PER_SH value for gfx12 4072- ac/nir/meta: move the "skip compute if no DCC image stores" condition to common 4073- ac,radeonsi: update comments related to the L2 cache, use "L2", not "TC" 4074- radeonsi/gfx12: remove CP DMA workarounds because CP DMA is never used on gfx12 4075- radeonsi/gfx12: disallow DCC for protected content 4076- radeonsi/gfx11: disable RB+ when blending 4077- radeonsi: use better OREO_MODE programming 4078- radeonsi: tweak si_test_dma_perf for better experience 4079- radeonsi: add a new PM4 helper radeon_event_write 4080- radeonsi: remove SI_CONTEXT_VGT_STREAMOUT_SYNC, emit it directly 4081- radeonsi: add ACQUIRE_MEM, RELEASE_MEM PWS packet helpers 4082- radeonsi: add si_cp_acquire_mem helper and clean up its usage for gfx6-9 4083- radeonsi: add gfx10+ support into si_cp_acquire_mem 4084- radeonsi: use the correct cs for SQTT barriers for gfx9 4085- radeonsi: add si_cp_pfp_sync_me 4086- radeonsi: remove unused size parameter from get_cache_policy 4087- radeonsi: don't use get_cache_policy in si_launch_grid_internal_ssbos 4088- radeonsi: move get_cache_policy into si_cp_dma.c 4089- radeonsi: don't use SI_COHERENCY_NONE in si_compute_shorten_ubyte_buffer 4090- radeonsi: don't use SI_COHERENCY_CB_META 4091- radeonsi: don't use SI_COHERENCY_CP 4092- radeonsi: remove all SI_COHERENCY_* flags except SI_COHERENCY_SHADER 4093- radeonsi: remove enum si_coherency 4094- radeonsi: don't use si_get_flush_flags in si_launch_grid_internal_ssbos 4095- radeonsi: move enum si_cache_policy and si_get_flush_flags into si_cp_dma.c 4096- radeonsi: remove L2_STREAM enum 4097- radeonsi: replace enum si_cache_policy with cp_dma_has_L2() 4098- radeonsi: inline si_get_flush_flags 4099- radeonsi: split si_launch_grid_internal to start preparing for barrier rework 4100- radeonsi: move compute barrier calls up one level 4101- radeonsi: move barrier code out of si_launch_grid_internal_ssbos/images 4102- radeonsi: skip no-op si_compute_blit() 4103- radeonsi: only set need_check_render_feedback if binding textures for PS 4104- radeonsi: split image saving/binding/restoring from si_launch_grid_internal_images 4105- radeonsi: do init/deinit for multi-dispatch compute blits only once 4106- radeonsi: don't flag both VS and PS partial flush in si_barrier_before_internal_op 4107- radeonsi: remove CP DMA code for GDS & L2 prefetch in the clear/copy_buffer path 4108- radeonsi: use si_barrier_{before,after}_internal_op for CP DMA 4109- radeonsi: remove SI_OP_SYNC_CPDMA_BEFORE, always sync CP DMA 4110- radeonsi: fold si_improve_sync_flags logic into si_barrier_before_internal_op 4111- radeonsi: fix waiting for PS to finish if there are no CB/DB bindings on gfx6-8 4112- radeonsi/gfx12: fix shader uploads via CP DMA 4113- ac/nir/meta: tune clear/copy_buffer performance for gfx6-10.3 4114- radeonsi: fix broken tessellation in Unigine Heaven 4115- radeonsi: remove SI_OP_CS_IMAGE 4116- radeonsi: inline si_launch_grid_internal_images without the barriers 4117- radeonsi: move barriers out of si_launch_grid_internal_ssbos 4118- radeonsi: remove unnecessary barriers from clear_buffer_rmw and clear_dcc_msaa 4119- radeonsi: add barrier helpers for simple internal buffer ops 4120- radeonsi: move barriers out of si_compute_clear_copy_buffer & si_cp_dma_* 4121- radeonsi: move barriers out of si_clear_buffer 4122- radeonsi: move barriers out of si_copy_buffer 4123- radeonsi: move barriers out of si_compute_shorten_ubyte_buffer 4124- radeonsi: remove SI_OP_SKIP_CACHE_INV_BEFORE 4125- radeonsi: remove SI_OP_SYNC_AFTER 4126- radeonsi: always pass flags=0 into si_barrier_after_internal_op/simple_buffer_op 4127- radeonsi: always wait for idle before get_query_result_resource 4128- radeonsi: remove unused SI_OP_IS_NESTED 4129- radeonsi: remove SI_OP_SYNC_BEFORE* flags 4130- radeonsi: change the flags parameter of si_compute_blit to bool fail_if_slow 4131- radeonsi: remove unused flags and user_flags params from clear/copy functions 4132- radeonsi: enable conditional rendering for si_compute_shorten_ubyte_buffer 4133- radeonsi: replace SI_OP_CS_RENDER_COND_ENABLE with bool render_condition_enable 4134- radeonsi: remove SI_CP_DMA_CLEAR_METHOD, call si_cp_dma_clear_buffer directly 4135- radeonsi: remove unused buffer_subdata fallback from si_clear_buffer 4136- radeonsi: rename "cache_flush" -> "barrier" 4137- radeonsi: move most barrier code into a new file si_barrier.c 4138- radeonsi: update remaining comments related to the L2 cache, use "L2", not "TC" 4139- radeonsi: rename TC_L2_dirty -> L2_cache_dirty 4140- radeonsi: check sctx->flags inside si_emit_barrier_direct 4141- radeonsi: rewrite si_memory_barrier to generate fewer barrier flags 4142- radeonsi: rename si_context::flags -> barrier_flags 4143- radeonsi: move initial framebuffer barrier code into si_barrier.c 4144- radeonsi: restructure fb_barrier code in si_launch_grid 4145- radeonsi: move CB synchronization into si_fb_barrier_after_rendering 4146- radeonsi: move DB synchronization into si_fb_barrier_after_rendering 4147- radeonsi: add si_fb_barrier_before_rendering 4148- radeonsi: rename SI_CONTEXT_* flags to SI_BARRIER_* flags 4149- radeonsi: consolidate code around unsetting barrier_flags in emit_barrier 4150- radeonsi: remove an obsolete comment about SMEM stores 4151- radeonsi: skip CB_META/DB_META events when they are redundant on gfx10-11 4152- radeonsi: deduplicate how GLM flush flags are set for gfx10-11 4153- radeonsi: merge the CB/DB conditional blocks in gfx10_emit_barrier 4154- mesa,glsl,gallium: add GL_OVR_multiview 4155- radeonsi: check and update compute_is_busy in get_reduced_barrier_flags 4156- radeonsi: don't sync VS and PS if they are idle 4157- radeonsi: count VS/PS/CS/L2 flushes in get_reduced_barrier_flags 4158- radeonsi: deduplicate code emitting VGT_FLUSH/PIPELINESTAT events 4159- radeonsi: don't sync CS and PS before rendering if there are no FBO attachments 4160- radeonsi: remove CB sync after FMASK and DCC decompression 4161- radeonsi: wait for idle after end_query in si_test_blit_perf 4162- radeonsi: don't pad esgs_vertex_stride if it's 0 4163- radeonsi: adjust GFX12 checks in si_compute.c 4164- radeonsi: clean up and make corrections to si_create_fmask_expand_cs 4165- radeonsi/aco: fix asm dumps to debug output via radeonsi_debug_disassembly=true 4166- radeonsi: move si_execute_clears barrier code into separate functions 4167- radeonsi: move barriers out of si_execute_clears 4168- radeonsi: execute clears at resource allocation using compute instead of gfx 4169- radeonsi: remove barriers around clears using aux_context.compute_resource_init 4170- radeonsi: don't use VS/PS/CS partial flushes if we use a TS event 4171- radeonsi: clean up set_log_context code for all aux contexts 4172- radeonsi: use ACO on GFX11.5 with LLVM 18 or older to work around GPU hangs 4173- radeonsi: don't insert any barrier after the copy for PIPE_MAP_READ 4174- radeonsi: rename hw_level -> view_level 4175- radeonsi: remove the make_texture_descriptor indirect function call 4176- nir/opt_shrink_vectors: shrink memory loads, not just IO 4177- ac: make sure VEGA20 and MI200 version ranges don't overlap with other chips 4178- ac/nir: set .image_dim and .image_array for all opcodes 4179- ac/llvm: use LLVM processor gfx942 for GFX940 when it's available 4180- ac/llvm: don't use the 64-bit umul_hi workaround with LLVM 19.1 4181- ac/gpu_info: print 32bpp modifiers 4182- st/mesa: copy some TES shader info fields to TCS 4183- nir/opt_vectorize_io: fix skipped output vectorization if inputs were vectorized 4184- nir/opt_vectorize_io: fix stack buffer overflow with 16-bit output stores 4185- nir/opt_load_store_vectorize: add entry::num_components 4186- nir: add hole_size parameter into the vectorize callback 4187- nir: reject unsupported component counts from all vectorize callbacks 4188- nir/opt_load_store_vectorize: allow overfetching, merge overfetched loads 4189- nir/opt_load_store_vectorize: allow a 4-byte hole between 2 loads 4190- nir/opt_load_store_vectorize: vectorize load_smem_amd 4191- gallium/u_threaded: fix crash in tc_create_image_handle due to resource == NULL 4192- radeonsi: set the valid buffer range for bindless image buffers 4193- nir/opt_varyings: assign locations of no_varying IO for TCS outputs only 4194- nir/opt_varyings: pack TCS inputs with cross-invocation access together 4195- ac/nir: rewrite ac_nir_lower_ps epilog to fix dual src blending with mono PS 4196- nir: add shader_info::tess::tcs_same_invocation_inputs_read(_indirect) 4197- mesa_interface: remove unused stuff 4198- mesa_interface: replace opaque __DRIscreen with struct dri_screen everywhere 4199- mesa_interface: replace opaque __DRIcontext with struct dri_context everywhere 4200- mesa_interface: replace opaque __DRIdrawable with struct dri_drawable everywhere 4201- mesa_interface: remove redundant __DRI*Rec types 4202- mesa_interface: rename __DRIconfig to struct dri_config 4203- mesa_interface: rename __DRIimage to struct dri_image 4204- nir: add nir_gather_tcs_info, new gathering/analysis pass 4205- nir: rename load_cull_small_primitives -> triangles, add load_cull_small_lines 4206- nir: rename load_cull_small_primitive_precision -> triangle, add line_precision 4207- nir: add cull_triangles\_, cull_lines\_ prefixes to viewport_xy_scale_and_offset 4208- ac/nir,radeonsi: use load_cull_small_lines_enabled_amd 4209- ac/nir,radeonsi: use load_cull_small_line_precision_amd 4210- ac/nir,radeonsi: use load_cull_line_viewport_xy_scale_and_offset_amd 4211- radeonsi: rewrite how small prim precision is passed to culling code 4212- radeonsi: simplify util_rast_prim_is_lines_or_triangles 4213- radeonsi/gfx11: fix Z corruption for Blender 4214- amd: move Tonga and Iceland TC-compat HTILE workarounds to ac_gpu_info.c 4215- radeonsi/gfx12: fix AMD_DEBUG=nodcc not working 4216- glsl: lower IO in the linker if enabled, don't lower it later 4217- nir: rename nir_io_glsl_lower_derefs -> nir_io_has_io_intrinsics 4218- nir: print interp_mode better 4219- nir/lower_io: change INTERP_MODE_NONE to SMOOTH when NONE means SMOOTH 4220- nir: add nir_clear_divergence_info, use it in nir_opt_varyings 4221- gallium: set proper type for pipe_shader_state::ir::nir 4222- st/mesa: fix incorrect types of shader CSOs 4223- amd/ci: adjust stoney traces checksums 4224- glsl: fix accidentally disabling nir_opt_varyings for all drivers 4225- Revert "amd/ci: adjust stoney traces checksums" 4226- st/mesa: implement key->persample_shading for lowered IO 4227- nir/lower_clip_disable: handle non-scalar store intrinsics 4228- radeonsi: fix gl_FrontFace elimination when one side is culled 4229 4230Mark Burton (1): 4231 4232- gallivm: Fix compilation errors when using LLVM 13. 4233 4234Mark Collins (2): 4235 4236- freedreno/registers: Document TPL1_2D_SRC_CNTL register 4237- tu: Improve 2D buffer-to-image copies for A7XX 4238 4239Martin Krastev (12): 4240 4241- svga/ci: enable vmware farm 4242- svga/ci: change DNS server for vmware jobs 4243- svga/ci: triage unexpected piglit pass 4244- svga/ci: update FORCE_KERNEL_TAG 4245- svga/ci: triage unexpected piglit passes 4246- svga/ci: disable vmware-qemu-traces jobs 4247- svga/ci: disable vmware farm 4248- svga/ci: enable vmware farm 4249- svga/ci: triage piglit tessellation failures 4250- svga/ci: disable vmware farm 4251- svga/ci: enable vmware farm 4252- svga/ci: disable vmware farm 4253 4254Martin Roukala (né Peres) (9): 4255 4256- radv/ci: update the flakes lists 4257- zink/ci: update the flakes list for ga106 4258- radv/ci: document more vkcts flakes 4259- zink/ci: document more flakes 4260- radv/ci: run vkcts-navi31-valve pre-merge 4261- freedreno/ci: fix the stage of the a750 jobs 4262- radv/ci: document more vkcts flakes 4263- zink/ci: document more radv flakes 4264- zink/ci: bump the polaris10 timeout to 30 minutes 4265 4266Mary Guillemard (61): 4267 4268- panvk: Update CI fails 4269- docs: Add panvk in VK_EXT_private_data to features.txt 4270- panvk: Implement CmdPushDescriptorSetWithTemplate2KHR 4271- panvk: Fix image support in vertex jobs 4272- panvk: Emit viewport when state is NULL 4273- panvk: Implement CmdBindDescriptorSets2KHR, CmdPushConstants2KHR and CmdPushDescriptorSet2KHR 4274- panfrost: Add B8G8R8A8_SNORM in panfrost_pipe_format 4275- panvk: Pass attrib_buf_idx_offset to desc_copy_info 4276- panvk: Fix NULL deref on model name when device isn't supported 4277- panvk: Skip blend descriptors when no fragment shader is present 4278- bi: Add bi_foreach_ssa_dest 4279- bi: Use bi_foreach_src in bi_foreach_ssa_src 4280- bi: Do not mark tex ops as skip when dest is used by control flow 4281- bi: Use nir_opt_load_store_vectorize 4282- bi: Use nir_opt_shrink_stores 4283- bi: Add loop_header in bi_block 4284- bi: Rewrite dead code elimination 4285- panvk/ci: Update G52 fail list to represent current state 4286- panvk: Properly propagate helper invocations requirement 4287- panvk: Expose trivial google extensions 4288- panvk: Fix viewport calculation 4289- nak: Use F2FP for nir_op_pack_half_2x16_split on SM86+ 4290- panvk: Close batch in CmdBeginRendering if already open 4291- panvk: Manually copy dynamic state in meta_gfx 4292- panvk: Simplify meta ctx save/restore 4293- panvk: Save and restore vertex buffer in gfx_meta 4294- panvk: Remove unused vk_vertex_input_state in graphics_save_ctx 4295- panvk: Extend the descriptor lowering pass to support Valhall 4296- panvk: Extend the shader logic to support Valhall 4297- nak: Fix wrong encoding for F2FP cbuf and immediate forms 4298- panfrost: Update panfrost_drm.h 4299- panfrost: Update panthor_drm.h 4300- pan/kmod: Add timestamp uapi support 4301- panfrost: Move occlusion query allocation to create_query 4302- panfrost: Implement timer queries 4303- docs: Mark timer queries as done on panfrost 4304- panvk: Add atomic modifiers for r32 based formats 4305- panvk: Reject SNORM only for color attachment and blend 4306- panvk: Return proper buffer format feature flags 4307- pan/lib: Simplify pan_blitter 4308- panvk: Include new failures from VKCTS 1.3.9.1 4309- panvk: Ensure to clear dirty dynamic state in panvk_cmd_draw 4310- mr-label-maker: Add panvk 4311- pan/va: Define the LEA_BUFFER instruction 4312- panvk/ci: Update Mali-G52 status with full VKCTS 4313- panvk/ci: Update to run full CTS on G52 4314- panvk/ci: Make panfrost-g52-vk run at premerge 4315- egl: Support NV_context_priority_realtime 4316- panfrost: Update panthor_drm.h with latest 4317- pan/kmod: Add priority query uapi support 4318- panfrost: Implement context priority on v10 4319- panvk/ci: Increase panfrost-g52-vk fraction 4320- panvk: Cleanup rw_nc pool in panvk_device_cleanup_mempools 4321- hk: Implement global priority extensions 4322- panvk: Define primitive size for RUN_TILER/RUN_IDVS 4323- winsys/nouveau: Rework to use u_pipe_screen_lookup_or_create 4324- winsys/nouveau: Reformat to stop relying on tabs 4325- nir: Move atomic_op_to_alu to common code 4326- hk: Fill deviceUUID 4327- panvk: Ensure that render_info is not null in force_fb_preload 4328- bi: Execute nir_opt_algebraic after nir_lower_pack 4329 4330Matt Coster (3): 4331 4332- pvr: Fix ds subtile alignment NULL pointer dereference 4333- pvr: Fix reordering of sub-cmds when performing ds subtile alignment 4334- pvr: Use vk_buffer_view base 4335 4336Matt Turner (14): 4337 4338- intel/clc: Free parsed_spirv_data 4339- intel/clc: Free disk_cache 4340- intel/brw: Use REG_CLASS_COUNT 4341- intel/elk: Use REG_CLASS_COUNT 4342- docs: Drop references to LIBGL_DRIVERS_PATH 4343- util: Add ATTRIBUTE_OPTIMIZE(flags) 4344- util: Force emission of stack frame in stack unit test 4345- nir: Skip opt_if_merge when next_if has block ending in a jump 4346- nir/tests: Add tests for opt_if_merge 4347- anv: Set shader_spilling_rate=15 by default 4348- anv: Set shader_spilling_rate=11 4349- anv: Align anv_descriptor_pool::host_mem 4350- vulkan: Avoid pointer aliasing 4351- nir: Get correct number of components 4352 4353Mauro Rossi (2): 4354 4355- nvk: Fix regression observed on Kepler 4356- android: gbm: split the DRI backend into separate library 4357 4358Maíra Canal (2): 4359 4360- v3d, vc4: remove unused functions 4361- v3d: Don't use performance counters names array with an older kernel 4362 4363Michael Catanzaro (1): 4364 4365- Disk cache should consistently create parent directories when needed 4366 4367Michael Cheng (2): 4368 4369- anv: Remove extra hdc_flush from Perfetto 4370- anv: move trace logic to batch_emit_pipe_control_write 4371 4372Michel Dänzer (10): 4373 4374- Revert "radeonsi: remove CB sync after FMASK and DCC decompression" 4375- radeonsi: Revert to GLSL_SAMPLER_DIM_2D in si_create_fmask_expand_cs 4376- util/mesa-db: Make mesa_db_lock robust against signals 4377- util/mesa-db: Recreate files if header load or index update fails 4378- util/mesa-db: Reserve hash table for total number of index entries 4379- util/mesa-db: Use single read for whole index 4380- util: Use persistent array of index entries 4381- util/mesa-db: Further simplify mesa_db_compact 4382- Revert "util/mesa-db: Further simplify mesa_db_compact" 4383- Revert "util: Use persistent array of index entries" 4384 4385Michel Zou (1): 4386 4387- ac/gpu_info: Fix missing prototype mingw error 4388 4389Mike Blumenkrantz (389): 4390 4391- zink: move some flags from driver_workarounds to separate struct 4392- zink: hash in some driver workaround flags to shader cache 4393- zink: add lavapipe to can_do_invalid_linear_modifier list 4394- zink: use local screen variable in zink_prep_fb_attachment 4395- zink: use GENERAL layout for depth attachments on some drivers 4396- zink: use blake3 instead of sha1 for program cache 4397- gallium: install gallium-$version.so to libdir 4398- ci: prune dri from LD_LIBRARY_PATH 4399- dril: rework config creation 4400- llvmpipe: handle vma allocation failure 4401- llvmpipe: only use vma allocations on linux 4402- glx: delete tests 4403- st/interop: return unsupported without screen hooks 4404- dri: fix kmsro define 4405- glx: tabs -> spaces 4406- glx: delete __GLXDRIdisplay 4407- glx: simplify zink init variables 4408- glx: rework __glXInitialize 4409- glx: delete dri3_display::loader_extensions 4410- glx: move has_multibuffer to base glx_screen struct 4411- glx: move driwindows_display::event_base to driwindows_context 4412- glx: rework/simplify drisw zink handling 4413- glx: move dri2Hash to base glx_display 4414- glx: tweak dri2 init ordering 4415- glx: move dri2 init checks to separate function 4416- glx: move dri2 loader extensions to static vtable 4417- glx: delete all driver-specific display handling 4418- glx: use base screen has_multibuffer for drisw 4419- glx: move drisw dri3 handling for zink out to glxInitialize 4420- glx: consolidate some LIBGL_KOPPER_DISABLE checks 4421- glx: unhackify inferring zink from dri3 4422- Revert "vl/dri3: use loader's dri3 init code and delete everything else" 4423- loader: move some common dri3 functions out of dri3 loader 4424- meson: move glx subdir after gallium build 4425- loader: split out dri3 into subdir 4426- dri: declare IMAGE_DRIVER symbols PUBLIC 4427- glx: inline IMAGE_DRIVER functions 4428- dri: declare DRI_CORE functions public 4429- loader/dri3: inline DRI_CORE 4430- glx: inline DRI_CORE functions 4431- glx: inline createNewScreen3 4432- glx: delete __DRImesaCoreExtension usage 4433- glx: inline DRI_SWRAST 4434- dri: delete __DRImesaCoreExtensionRec::createNewScreen 4435- glx: inline DRI_DRI2 4436- dri: delete driCreateNewScreen2 4437- dri: declare DRI_KOPPER as PUBLIC 4438- glx: use local var for LIBGL_KOPPER_DISABLE check 4439- glx: switch kopper check for setting up swapinterval/bufferage functions 4440- glx: delete DRI_KOPPER 4441- glx: rename driswCopySubBuffer 4442- drisw: declare DRI_COPY_SUB_BUFFER as PUBLIC 4443- glx: inline DRI_COPY_SUB_BUFFER 4444- dri: delete __DRI_COPY_SUB_BUFFER 4445- glx: delete releaseTexBuffer 4446- glx: delete the only users of __DRItexBufferExtension::setTexBuffer 4447- dri: delete __DRItexBufferExtension::setTexBuffer 4448- dri: declare DRI_TEX_BUFFER functions public 4449- glx: delete check for texbuffer extension 4450- glx: inline DRI_TEX_BUFFER 4451- loader/dri3: delete texBuffer struct member 4452- glx: delete DRI_TEX_BUFFER handling 4453- kopper: increment lastStamp during invalidate 4454- dri: add kopper stubs to avoid build failures 4455- dri: make DRI2_RENDERER_QUERY interface PUBLIC 4456- glx: inline __DRI2_RENDERER_QUERY 4457- dri: delete __DRI2_RENDERER_QUERY interface 4458- glx: always expose GLX_ARB_context_flush_control 4459- dri: make DRI_FLUSH public 4460- kopper: reuse dri2 invalidate 4461- loader/dri3: inline DRI2_FLUSH 4462- glx: inline DRI2_FLUSH 4463- loader/dri3: delete DRI2_FLUSH remnants 4464- dri: unify DRI2_CONFIG_QUERY interfaces 4465- dri: make DRI2_CONFIG_QUERY public 4466- dri: move swapinterval functions from loader to dri frontend 4467- dri: remove __DRI2configQueryExtension param from swapinterval functions 4468- glx: inline DRI2_CONFIG_QUERY 4469- loader/dri3: delete DRI2_CONFIG_QUERY remnants 4470- dri: make DRI2_THROTTLE public 4471- glx: inline DRI2_THROTTLE 4472- dri: delete DRI2_THROTTLE interface 4473- dri: make DRI_INTEROP public 4474- glx: inline DRI_INTEROP 4475- glx: delete checks for interop support 4476- glx: delete more DRI_INTEROP checks 4477- glx: call dri interop functions directly 4478- glx: stop binding DRI_INTEROP 4479- glx: expose GLX_MESA_gl_interop for drisw 4480- glx: simplify DRI_IMAGE checks 4481- dri: make DRI_IMAGE public 4482- gallium: move loader_dri_create_image to dri frontend 4483- dri: use image loader function directly in dri_create_image_with_modifiers 4484- loader/dri3: inline DRI_IMAGE usage 4485- loader/dri: delete DRI_IMAGE remnants 4486- glx: delete DRI_IMAGE remnants 4487- loader/dri3: delete loader_dri3_extensions 4488- glx: move up glx_driver enum declaration 4489- glx: include src/gallium for apple 4490- dri: link with libloader 4491- dri: only link_whole once 4492- kopper: check swapchain size after possible loader image resize 4493- egl/dri2: delete unused function 4494- egl: inline DRI_CORE usage 4495- gbm: inline DRI_CORE usage 4496- egl/gbm: delete remnants of DRI_CORE 4497- dri: publicize dri_query_compatible_render_only_device_fd 4498- gbm: inline DRI_MESA 4499- egl: inline DRI_MESA 4500- gbm: delete DRI_MESA remnants 4501- egl: delete DRI_MESA remnants 4502- egl: inline DRI_KOPPER 4503- egl/gbm: remove DRI_KOPPER remnants 4504- dri: make driSWRastQueryBufferAge public 4505- egl: inline DRI_SWRAST 4506- egl/gbm: rip out DRI_SWRAST 4507- egl: collapse driCreateNewDrawable conditional 4508- egl: simplify a non-swrast conditional 4509- egl/drm: always expose EXT_buffer_age 4510- gbm: inline DRI_IMAGE_DRIVER 4511- egl/gbm: delete DRI_IMAGE_DRIVER remnants 4512- dri: store dmabuf import caps to dri_screen on create 4513- dri: check for dmabuf_import at top of dri2_from_dma_bufs 4514- gbm: inline dri2_from_dma_bufs and delete now-redundant check in dmabuf import 4515- egl: simplify MESA_drm_image enablement 4516- dri: make a few more functions public 4517- egl: inline simple DRI_IMAGE usage 4518- gbm: inline simple DRI_IMAGE usage 4519- egl: swap DRI_IMAGE checks for dmabuf/modifier support for driver check 4520- egl/android: check PIPE_CAP_NATIVE_FENCE_FD directly 4521- egl/wayland: delete check for DRI_IMAGE blitImage in drm init 4522- egl: inline DRI_IMAGE compression modifier handling 4523- egl: replace display_name checks with swrast checks 4524- egl: always check dri caps for EGL_WL_bind_wayland_display support 4525- gbm: check for dmabuf import on device creation 4526- gbm: use new dmabuf import flag to replace DRI_IMAGE checks 4527- gbm: always pass through images to unmap 4528- gbm: check for compression modifiers on device create 4529- gbm: use driver check for dmabuf export 4530- gbm: delete DRI_IMAGE remnants 4531- egl: move some extension enablement out DRI_IMAGE conditional 4532- egl: delete last DRI_IMAGE check 4533- egl: delete DRI_IMAGE remnants 4534- egl: delete local_buffers 4535- egl/x11: replace DRI_DRI2 check with fd check 4536- egl: delete DRI_DRI2 remnants 4537- gallium: make some sw screen create functions public 4538- pipe-loader: fix driconf memory management 4539- egl: inline CONFIG_OPTIONS usage 4540- egl: delete DRI_CONFIG_OPTIONS remnants 4541- egl: further simplify driver loading 4542- egl: add a bool to indicate swrast but not kms_swrast 4543- gbm: inline DRI_FLUSH 4544- gbm: delete DRI_FLUSH remnants 4545- egl: use a non-kms_swrast check for swapbuffers DRI2_FLUSH 4546- egl: use a swrast check for a swapbuffers DRI2_FLUSH case 4547- egl: delete a DRI2_FLUSH check in dri2 swapbuffers 4548- egl: replace another DRI2_FLUSH check with non-kms_swrast check in dri2_wait_client 4549- egl: inline DRI2_FLUSH usage 4550- egl: replace another DRI2_FLUSH check with a non-kms_swrast check in surface query 4551- egl: replace a DRI2_FLUSH check with swrast check in x11_copy_buffers 4552- egl/drm: replace a DRI2_FLUSH check with a non-kms_swarst check 4553- egl/wayland: replace a couple DRI2_FLUSH checks with non-kms_swrast checks 4554- egl: delete DRI2_FLUSH remnants 4555- dri: only add DRI_BUFFER_DAMAGE if the driver supports set_damage_region 4556- egl: simplify DRI_BUFFER_DAMAGE checks 4557- dri: make set_damage_region public 4558- egl: add a util to get the pipe_screen 4559- egl: use pipe_screen check for damage region, simplify buffer_damage checks 4560- egl: inline DRI_BUFFER_DAMAGE 4561- egl: delete DRI_BUFFER_DAMAGE remnants 4562- egl: always advertise KHR_context_flush_control 4563- egl: delete DRI_FLUSH_CONTROL remnants 4564- egl: inline DRI2_CONFIG_QUERY 4565- egl: delete DRI_CONFIG_QUERY remnants 4566- egl: inline DRI_TEX_BUFFER 4567- egl: delete DRI_TEX_BUFFER remnants 4568- dri: make DRI2_FENCE public 4569- egl: delete checks for DRI2_FENCE 4570- egl: simplify native fence check 4571- egl: don't check for create_fence_fd method when importing a fence 4572- egl: inline DRI2_FENCE 4573- egl: delete DRI2_FENCE remnants 4574- egl: always advertise MESA_gl_interop, inline calls 4575- egl: delete DRI2_INTEROP remnants 4576- dri: make DRI2_BLOB public 4577- egl: inline DRI2_BLOB 4578- egl: delete DRI2_BLOB remnants 4579- egl/android: replace DRI_MUTABLE_RENDER_BUFFER_DRIVER check with non-kms_swrast check 4580- egl: delete DRI_MUTABLE_RENDER_BUFFER_DRIVER remnants 4581- egl: move multibuffers check into platform_x11 4582- egl: simplify multibuffers check 4583- dril: always take the egl init path 4584- llvmpipe: init fd=-1 to avoid closing fd=0 on failure 4585- egl: fix zink init 4586- glx: replace a straggler DRI_CONFIG_OPTIONS usage 4587- dri: delete unused DRI_MESA queryCompatibleRenderOnlyDeviceFd method 4588- dri: delete __DRIdri2Extension interface 4589- dri: delete a couple unused functions 4590- dri: delete some dri_screen cruft 4591- dri: consolidate a bunch of dri_screen init code 4592- dri: delete all the extra dri2 extension handling 4593- egl/dri: move some screen accessors back to dri frontend 4594- glx: stop checking dri extensions 4595- glx: deduplicate and move driverName to base glx_screen struct 4596- glx: store the DRIscreen to the base glx_screen 4597- glx: set drisw glx_driver more usefully 4598- glx: use glx_driver to determine drisw extension enablement 4599- glx: add a glx_screen member to indicate GLX_EXT_texture_from_pixmap 4600- glx: unify extension binding 4601- dri: delete dri_screen::extensions accessor 4602- dri: delete dri_screen::extensions 4603- dri: fix kms_swrast screen fail 4604- dri: delete more dead code 4605- dri: pass through a type enum for creating screen instead of driver_extensions 4606- kopper: inline DRI_IMAGE usage 4607- egl: stop loading driver extensions 4608- glx: stop loading driver extensions 4609- gbm: stop loading driver extensions 4610- dri: massively prune the dri target 4611- dri: delete a bunch of unused interfaces 4612- egl/wayland: bail on zink init in non-sw mode if extension check fails 4613- egl: fix fd passing on init with zink 4614- zink: fix partial update handling 4615- egl: make dmabuf/modifier display bools a bit more precise/intelligible 4616- egl/glx: make detecting explicit modifier support more explicit 4617- dri: set __DRI_IMAGE_ERROR_BAD_PARAMETER if driver doesn't support dmabuf import 4618- egl: initialize dri error code variables 4619- kopper: reuse drisw_update_tex_buffer to delete some code 4620- dri: merge in loader_dri3 4621- kopper: reuse dri3 functions 4622- dri: rename and move handle_in_fence to dri_helpers 4623- kopper: reuse dri_image_fence_sync 4624- dri: break out get_pixmap_buffer 4625- kopper: reuse loader_dri3_get_pixmap_buffer 4626- glx: always set swapInterval in drisw 4627- egl: check x11 multibuffer support before screen create 4628- dri: pass has_multibuffers through from the loader 4629- dri: further collapse dri_screen creation 4630- dri: guard dmabuf support by loader's dmabuf support 4631- dri: collapse some modifier/dmabuf checks 4632- kopper: call kopper_create_drawable directly 4633- kopper: delete some redundant code 4634- dri: rework drawable creation 4635- glx: simplify kopper_get_buffer_age 4636- glx: flatten out getBufferAge a bit 4637- glx: move/add __DRIdrawable struct to __GLXDRIdrawable 4638- glx: use __GLXDRIdrawable more broadly instead of display type drawables 4639- glx: unify dri bind_context 4640- glx: unify dri unbind context 4641- glx: unify dri destroy context 4642- glx: tweak some drisw context create code 4643- glx: tweak some dri3 context create code 4644- glx: unify dri create_context_attribs 4645- glx: unify renderer query hooks 4646- glx/dri2: strdup driver name 4647- glx/dri3: strdup existing driverName instead of fetching it again 4648- glx: unify dri get_driver_name 4649- glx: unify dri screen vtable 4650- glx: unify bind_tex_image 4651- glx: delete private DRIscreen pointers 4652- glx: move driver_configs to base glx_screen 4653- glx: rework screen destroy 4654- glx: move base screen destroy to glx_screen_cleanup 4655- glx/dri3: delete driScreenRenderGPU 4656- glx: unify more extension setup 4657- glx: determine is_direct_capable from driver type 4658- glx: inline __GLXDRIscreen onto base glx_screen 4659- glx: unify dri screen init 4660- zink: bail on choose_pdev immediately if no devices are available 4661- st/pbo: reject vs/fs pbo ops if rowstride < width 4662- device_select: reorder CreateInstance to have shorter failure paths 4663- device_select: shortcut EnumeratePhysicalDevice* for count-only calls 4664- device_select: disable reordering with zink on xwayland 4665- zink: delete hackaround for disabling wsi extensions 4666- egl: add DRI_USE_INVALIDATE with non-sw kopper 4667- egl: really fix kopper fd passing 4668- egl/wayland: split out dmabuf extension stuff for kopper 4669- egl: pass real value for software to dri2_setup_device() 4670- egl/wayland: do display gpu fd dance for non-sw zink init 4671- zink: don't skip cbuf store ops if resolve is set 4672- tc: set resolve on renderpass info if blit terminates the renderpass 4673- egl/android: call dri2_display_create() like every other driver 4674- lavapipe: VK_EXT_pipeline_robustness 4675- device-select: block xserver+zink reordering via xcb 4676- ci: add a660 flake 4677- dril: add zink stub 4678- egl: unify LIBGL_KOPPER_DRI2 checks 4679- egl/x11: pretend kopper is software with LIBGL_KOPPER_DRI2 4680- llvmpipe: export dmabuf caps for kms_swrast 4681- zink: update profile with missing extensions 4682- revert part of 94e470a32d214052d89f519a07566d8098d05d28 4683- egl: use more precise conditional for passing fd through to dri screen create 4684- dril: use the super fallback path for software fallback 4685- vk: rename DGC feature to DGC-NV 4686- zink: create a surface for resolve resources 4687- zink: use tc renderpass optimizing to inline resolves with dynamic render 4688- tc: break out blit enqueue to separate function 4689- tc: optimize out tracked winsys resolves 4690- dril: also create double-buffered configs in swrast fallback 4691- meson: delete dri3 build option 4692- meson: require dri3 modifiers 4693- ci: fix trace update script 4694- ci: add --pipeline-url to trace update script 4695- zink: delete erroneous kopper assert 4696- zink: set resolve layout when inferring scanout resolves 4697- zink: stop leaking inferred resolve surfaces 4698- zink: fix sparse bo deallocation 4699- va: flush_resource before presentation 4700- vl/winsys: move dri3_get_screen_for_root to common 4701- va/vl: add kopper support 4702- gallium: add PIPE_CAP_MULTIVIEW 4703- zink: support GL_OVR_multiview 4704- mesa: support OVR_multiview2 4705- zink: support OVR_multiview2 4706- docs: OVR_multiview features 4707- lavapipe: VK_EXT_pipeline_protected_access 4708- lavapipe: VK_KHR_global_priority 4709- egl: replace fd check with swrast check for dri2 4710- llvmpipe: bump max point size to 256 4711- egl/x11: fix dri3 connect conditional for kopper without modifiers 4712- lavapipe: bump mipmapPrecisionBits 4713- zink: add screen create to match windows LUID 4714- vl: rename/ifdef the x11 kopper functionality 4715- vl: add Windows kopper support 4716- va: support zink on Windows 4717- mesa: fix sample count handling for MSRTT 4718- mesa: OVR_multiview_multisampled_render_to_texture 4719- anv: fix video profile lists 4720- trace: add get_device_reset_status 4721- gallium: delete pipe_screen::is_compression_modifier 4722- egl/kopper: hook up EGL_EXT_surface_compression on wayland 4723- zink: hook up compression control extensions 4724- zink: store compression rate info for formats 4725- zink: implement compression control 4726- zink: move x8 format function to zink_format.c 4727- zink: fix zink_format_is_voidable_rgba_variant() 4728- zink: add X format variants for 30-bit formats 4729- zink: add some other missing X format emulation variants 4730- zink: add adl flake 4731- vk/image: fix view creation for planar video aspects 4732- zink: only clamp samplerview aspect for zs formats 4733- zink: fix planar image creation with mutable 4734- zink: correctly handle disjoint image creation 4735- zink: slightly refactor image tiling inference 4736- zink: swap sampler conversion conditional with yuv check 4737- zink: stop creating useless sampler conversions 4738- zink: move some image ici init up 4739- zink: check HAVE_LIBDRM for xf86drm.h include 4740- zink: only use DISJOINT for planar images 4741- zink: zink_modifier_prop -> zink_modifier_props 4742- zink: wrap all access to format_props and modifier_props 4743- zink: init format props dynamically 4744- zink: fix some indentation 4745- zink: delete redundant HIC usage check 4746- zink: add HIC back if suboptimal check fails 4747- zink: simplify some confusing modifier code 4748- zink: stop passing modifier pointers internally 4749- zink: rework modifier selection (again) 4750- zink: rework HIC image creation checks 4751- vk/runtime: add EXT DGC layout handling 4752- lavapipe: EXT DGC 4753- ci: bump gl cts versions 4754- ci: bump VVL to current week 4755- zink: revert compression control handling 4756- vl: add pipe_video_buffer::flags to sync up with pipe_resource::flags 4757- util/vbuf: delete/fix broken incompatible stride calc 4758- zink: block all 2d view creation with sparse 4759- zink: block dmabuf fallback into optimal tiling 4760- zink: assert images aren't created with dmabuf export and optimal tiling 4761- zink: also init format props when getting modifier props 4762- zink: clamp out dmabuf exports from optimal tiling images 4763- zink: fix external_only reporting for dmabuf formats 4764- zink: block srgb with winsys imports 4765- anv: add VK_FORMAT_G10X6_B10X6R10X6_2PLANE_420_UNORM_3PACK16 to modifier exceptions 4766- gallium: rework vbuf alignment pipe caps 4767- util/vbuf: rename/consolidate some bools 4768- gallium/sw: make llvmpipe the "default" sw driver 4769- util/framebuffer: add viewmask compare for fb equal 4770- iris: assert that viewmask is 0 4771- gallium: delete duplicated viewmask member in draw info 4772- llvmpipe: expose GL multiview extensions 4773- zink: further improve image usage detection 4774- zink: rewrite the high-IQ image tiling selection loop 4775- device-select: only try wayland/x11 if the required vars are set 4776- va: fail context create if driver does not support video 4777- vdpau: fail context create if driver does not support video 4778- zink: stop leaking precompiled generated tcs 4779- zink: add VVL for RADV jobs 4780 4781Mike Lothian (1): 4782 4783- Revert "device_select: shortcut EnumeratePhysicalDevice* for count-only calls" 4784 4785Mitchell Kember (1): 4786 4787- [fxbug.dev] Migrate bug numbers 4788 4789Mohamed Ahmed (6): 4790 4791- nvk: Use stride in the explicit modifier case for linear images 4792- nil: Expose px_to_B helpers to C 4793- nil: Add level_layer_size_B() helper 4794- nvk: Add host copy functions 4795- nvk: Block off non-2D DRM format modifier images 4796- nvk: Enable VK_EXT_host_image_copy 4797 4798Nanley Chery (49): 4799 4800- anv: Drop flush from unused depth workaround 4801- intel/blorp: Use WA helpers for depth pipecontrol 4802- anv,iris: Use WriteImmediate instead of Z flush for WA 4803- anv: Batch MCS and CCS aux-op flushes 4804- anv: Add want_hiz_wt_for_image() 4805- iris: Add and use want_hiz_wt_for_res 4806- intel: Enable more LOD0 HIZ+CCS fast clears 4807- intel/blorp: Allow LOD0 fast-clears with HiZ WT 4808- intel: Adjust partial depth fast clear checks 4809- intel: Use a simpler workaround for HiZ WT fast-clears 4810- iris: Invalidate state cache for some depth fast clears 4811- intel: Add and use isl_get_sampler_clear_field_offset 4812- intel/blorp: Simplify depth clear value updates 4813- iris: Move a HIZ_CCS_WT fast-clear flush higher up 4814- intel: Move depth clear value writes to drivers 4815- intel/isl: Fix packing of SINT formats 4816- intel: Don't use HW clear color conversion on gfx11 4817- intel/blorp: Don't use clear color conversion on gfx12 4818- anv,iris: Skip tex invalidate for clear conversion 4819- anv,hasvk: Add and use set_image_clear_color() 4820- anv,hasvk: Always use BLORP_BATCH_NO_UPDATE_CLEAR_COLOR 4821- iris: Always use BLORP_BATCH_NO_UPDATE_CLEAR_COLOR 4822- iris: Skip some fast-clears even on color changes 4823- intel: Drop BLORP_BATCH_NO_UPDATE_CLEAR_COLOR 4824- anv: Pass the VkClearDepthStencilValue for clears 4825- anv: Add and use anv_image_hiz_clear_value() 4826- intel: Support any depth fast-clear value on Xe2 4827- intel/isl: Always set EnableUnormPathInColorPipe 4828- anv,iris: Pack depth pixels into initialized arrays 4829- intel/isl: Reduce miptail slot usage to allow CCS 4830- intel/isl: Disable 3D Ys/Yf miptails for CCS 4831- anv: Change params of anv_can_fast_clear_color_view 4832- anv: Move and rename anv_can_fast_clear_color_view 4833- anv: Move exec_ccs_op and exec_mcs_op higher up 4834- anv: Support fast clears in anv_CmdClearColorImage 4835- intel: Avoid no-op calls to anv_image_clear_color 4836- anv: Require compression for fast-clears on gfx20+ 4837- anv: Use image formats when copying to/from buffers 4838- intel/blorp: Choose some copy formats independently 4839- intel/blorp: Use original surface format for some copies 4840- anv: Prevent clear color modifier corruption with views 4841- anv: Delete stale comment for BLORP clear color addr 4842- anv: Add an array of view formats to anv_image 4843- anv: Prepare dmabufs for clear color arrays 4844- anv: Move code out of loop in anv_CmdClearColorImage 4845- anv: Refactor clear color loading functions 4846- anv: Access more colors in fast_clear_memory_range 4847- anv: Load fast clear colors more often 4848- anv: Allow more fast clear colors for layouts 4849 4850Neha Bhende (1): 4851 4852- dri: fix macro name check to detect svga driver 4853 4854Nikita Popov (1): 4855 4856- gallium: Don't pass avx512er and avx512pf features on LLVM 19 4857 4858Oskar Rundgren (1): 4859 4860- pvr: improve buffer copy format selection 4861 4862PODISHETTY KUMAR (1): 4863 4864- Revert "gfxstream: add VK_EXT_vertex_attribute_divisor" 4865 4866Patel, Utpal (2): 4867 4868- amd/vpelib: Add input pixel format support 4869- amd/vpelib: Add resource function hooks for checking support 4870 4871Patrick Lerda (8): 4872 4873- iris: fix indirect draw refcnt imbalance 4874- i915: fix vertex atan regression 4875- iris: fix iris_ensure_indirect_generation_shader() memory leak 4876- r600/sfn: fix class Shader object m_register_allocations memory leak 4877- r600/sfn: fix class Shader object last_alu_with_indirect_reg memory leak 4878- r600: fix spec ext_packed_depth_stencil getteximage 4879- glsl: fix gl_nir_validate_intrastage_interface_blocks() memory leak 4880- r600: fix sfn_nir_legalize_image_load_store cubearray behavior 4881 4882Paulo Zanoni (33): 4883 4884- intel/genxml: add the BLT and COMP_CTX0 versions of the TR-TT registers 4885- anv/trtt: make genX(init_trtt_context_state) a little more compact 4886- anv/trtt: check the return value of anv_trtt_init_context_state() 4887- anv/trtt: submit a separate batch in anv_trtt_init_context_state() 4888- anv/trtt: make all contexts have the same TR-TT programming 4889- anv/trtt: fix the process of picking device->trtt.queue 4890- anv/trtt: don't just crash when we can't find device->trtt.queue 4891- anv/xe: try harder when the vm_bind ioctl fails 4892- anv: don't expose the compressed memory types when DEBUG_NO_CCS 4893- anv: disable CCS for Source2 games on Xe2 4894- intel: fix compute SLM sizes on Xe2 and newer 4895- anv: be consistent regarding non-render engines on i915.ko 4896- isl: don't assert(num_elements > (1ull << 27)) 4897- anv: fix compute engines when using ANV_QUEUE_OVERRIDE 4898- anv/trtt: mark vk_sync_get_value()'s value as defined for Valgrind 4899- anv: remove duplicate pipe_control workaround 4900- anv: remove another copy of the texture cache pipe_control workaround 4901- anv/trtt: set every entry to NULL when we create an L2 table 4902- anv/trtt: extract anv_trtt_first_bind_init() 4903- anv/trtt: inline anv_trtt_init_queues_state() 4904- anv/trtt: ensure all L3 entries are NULL-bound during init 4905- anv/trtt: unset trtt->l3_addr if initialization fails 4906- anv/trtt: fix error handling when adding binds 4907- anv/trtt: don't submit empty batches when there are no binds to do 4908- anv/trtt: decrement trtt->timeline_val when submission fails 4909- anv/trtt: delay batch garbage collection 4910- anv/trtt: remove TODO comment regarding the reloc list 4911- anv/trtt: convert anv_trtt_bind arrays to util_dynarray 4912- anv/trtt: extract anv_trtt_first_bind_init_queue() 4913- anv/trtt: fix the creation of sparse buffers of size 2^32 on 32bit systems 4914- anv/trtt: remove useless VK_RESULT checks 4915- brw: don't emit instruction to add zero in spilling code 4916- brw: add a NOP in between WHILE instructions on LNL 4917 4918Pavel Ondračka (28): 4919 4920- r300: bias presubtract fix 4921- ttn: switch to derivative intrinsics 4922- r300: fix RGB10_A2 CONSTANT_COLOR blending 4923- mesa: implement GL_FRAMEBUFFER_BLEND query 4924- r300/ci: add more flakes 4925- ttn: use nir_shader_get_entrypoint in nir_shader_gather_info 4926- ttn: unconditionally call nir_shader_gather_info 4927- r300: remove nir_opt_load_store_vectorize 4928- r300/ci: update fails list with a recent fix 4929- r300: move HyperZ blacklist to driconf 4930- r300: use fake ADD instead of MOV for negative KIL source fixup 4931- r300: preserve NaNs and denormals for MIN/MAX/CMP/CND 4932- r300: get rid of unneeded CMP before KIL 4933- r300: remove gl_ClipVertex early 4934- r300/ci: share some common variables 4935- r300/ci: add deqp and piglit on RV410 4936- r300: add .clang-format file for the compiler 4937- r300: opt in to clang-format CI enforcement for the compiler 4938- r300/compiler: reformat using default mesa .clang-format rules 4939- r300/ci: update ci expectations after piglit uprev 4940- nir/nir_group_loads: reduce chance of max_distance check overflow 4941- r300/ci: add new RV410 flakes 4942- ci: disable ondracka farm 4943- Revert "ci: disable ondracka farm" 4944- r300: add switch to support IEEE and FF math opcodes 4945- r300: remove wrong Unigine Sanctuary driconf override 4946- r300: add driconf math mode override for Unigine Tropics and Oilrush 4947- r300/ci: fails update after recent piglit uprev 4948 4949Peyton Lee (1): 4950 4951- radeonsi: correct map and unmap function timing 4952 4953Pierre-Eric Pelloux-Prayer (26): 4954 4955- amd: use a valid size for ac_pm4_state allocation 4956- radeonsi: reject modifiers with DCC when NO_EXPORTED_DCC is used 4957- radeonsi: consider PIPE_BIND_LINEAR when filtering modifiers 4958- radeonsi: consider DBG(NO_TILING) when filtering modifiers 4959- egl,gbm,glx: fix log message spam 4960- radeonsi: don't always update shader coherency draw call counter 4961- gallium/ddebug: implement modifiers function 4962- gallium/ddebug: implement set_inlinable_constants 4963- gallium/ddebug: fix sampler_states handling 4964- gallium: add pipe_screen::get_driver_pipe_screen 4965- radeonsi: add si_screen helper 4966- kopper, dri: remove trace_screen_unwrap 4967- frontends/va: honor DRI_PRIME for VA_DISPLAY_WAYLAND 4968- radeonsi/tests: allow to run radeonsi-run-tests.py with AMD_DEBUG 4969- radeonsi: include the shader key in the sqtt pipeline hash 4970- radeonsi: check the pointer before adding an offset 4971- radeonsi/sqtt: don't store the offsets in the pipeline 4972- radeonsi/sqtt: use XXH64_update 4973- radeonsi/sqtt: don't leak the pipeline bo 4974- radeonsi/sqtt: don't store pipeline->bo in shader->bo 4975- radeonsi/gfx12: fill missing dcc tiling info 4976- radeonsi: fix radeon_canonicalize_bo_flags domain handling 4977- ac/surface: fix determination of gfx12_enable_dcc 4978- nir: skip offset=0 in nir_io_add_const_offset_to_base 4979- glsl: use nir_io_add_const_offset_to_base in gl_nir_opts 4980- radeonsi/ci: mark \*.tessellation_shader_tessellation.max_in_out_attributes as fixed 4981 4982Pohsiang (John) Hsu (1): 4983 4984- d3d12: fix incorrect memset in d3d12_video_encoder_references_manager_hevc 4985 4986Qiang Yu (29): 4987 4988- radeonsi: lower subgroup ops after wave size is known 4989- ac/llvm: build wqm for quad intrinsics only when fragment shader 4990- ac,radv,radeonsi: stop using quad vote any/all when llvm 4991- gallium: add caps for KHR_shader_subgroup 4992- mesa: add KHR_shader_subgroup extension 4993- mesa: implement KHR_shader_subgroup Get* values 4994- mesa: add spirv GroupNonUniform* capbilities support 4995- glsl: add KHR_shader_subgroup extensions 4996- glsl: add builtin variables for KHR_shader_subgroup 4997- glsl: remove unused builtin function macros 4998- glsl: use generic convertion code for some intrinsics 4999- glsl: add KHR_shader_subgroup_basic builtin functions 5000- glsl: prepare vote function args for subgroup usage 5001- glsl: add KHR_shader_subgroup_vote builtin functions 5002- glsl: prepare ballot function args for subgroup usage 5003- glsl: add KHR_shader_subgroup_ballot builtin functions 5004- glsl: add KHR_shader_subgroup_shuffle builtin functions 5005- glsl: add KHR_shader_subgroup_shuffle_relative builtin functions 5006- glsl: add KHR_shader_subgroup_arithmetic builtin functions 5007- glsl: add KHR_shader_subgroup_clustered builtin functions 5008- glsl: add KHR_shader_subgroup_quad builtin functions 5009- glsl: remove unused fields in gl_shader 5010- glsl: set shader_info.subgroup_size for KHR_shader_subgroup 5011- radeonsi: remove NULL check in si_determine_wave_size 5012- radeonsi: consider both stages to determine merged shader wave_size 5013- radeonsi: use wave64 for KHR_shader_subgroup enabled shader 5014- radeonsi: enable KHR_shader_subgroup 5015- nir: add skip_lower_packing_ops shader compile option 5016- ac: do not lower some ops in nir_lower_packing 5017 5018Rebecca Mckeever (7): 5019 5020- panvk: Move the VkEvent logic to the jm folder and make it per-arch 5021- panvk: Add VkEvent support to the CSF backend 5022- panvk/csf: Set and clear vb.dirty flag 5023- panvk/csf: Split tiler flags initialization into two steps 5024- panvk: Track VkFormats for depth and stencil attachments 5025- panvk: Move rendering info to panvk_rendering_state struct 5026- panvk/csf: Implement vkCmdExecuteCommands 5027 5028Renato Pereyra (1): 5029 5030- egl: Refactor init of CPU tracing in EGL for Android 5031 5032Rhys Perry (88): 5033 5034- nir/instr_set: combine XXH32 calls 5035- nir/instr_set: stop sorting phi sources 5036- nir/instr_set: hash tex sources commutatively 5037- aco/gfx11.5: workaround export priority issue 5038- aco: add tests for export priority issue 5039- aco/gfx11.5: skip dealloc_vgprs for stages with exports 5040- aco: add export instructions to should_form_clause 5041- aco: form export clauses 5042- aco: move s_setprio to before NGG exec initialization 5043- aco: consider exec empty after divergent continue then divergent break 5044- aco: add struct and helpers for exec potentially empty 5045- aco: combine DPP into v_cvt_f16_f32 5046- aco: fix validation of v_s\_ opcodes 5047- aco: don't transform v_interp_p2_f32 with constant into fma 5048- docs: update ACO_DEBUG documentation for scheduler options 5049- docs: update ACO_DEBUG documentation for perfwarn 5050- nir/opt_uniform_atomics: require block index metadata 5051- aco: split selection_control_remove into rarely_taken and never_taken 5052- aco: only remove branch jumping over SMEM/barrier if it's never taken 5053- aco: completely skip branches if they're never taken 5054- aco: set prefer_remove for gfx9- too 5055- aco: calculate indices from dominance tree 5056- aco: use dominance helpers 5057- aco: preserve SSA in try_eliminate_scc_copy 5058- aco: validate temporary reachability 5059- aco: split insert_wait_states into two 5060- aco: move insert_delay_alu to after insert_NOPs 5061- aco: forget valu delays after certain s_waitcnt_depctr/LDSDIR 5062- aco: ignore exec and literals when mitigating VALUMaskWriteHazard 5063- aco: also consider VALU reads for VALUMaskWriteHazard 5064- aco: don't consider sa_sdst=0 before SALU write to fix VALUMaskWriteHazard 5065- aco: check SALU writing lanemask later for VALUMaskWriteHazard 5066- aco: preserve bitsets after a lane mask is written 5067- aco/tests: add more VALUMaskWriteHazard tests 5068- aco/ra: fix sub-dword get_reg_specified in some cases 5069- aco/ra: use DefInfo for get_reg_specified 5070- aco: have get_subdword_definition_info update DefInfo 5071- aco: fix printing of d16 MIMG dmask 5072- nir/opt_if: fix fighting between split_alu_of_phi and peel_initial_break 5073- nir/opt_loop: skip peeling if the break is non-trivial 5074- nir/opt_loop: skip peeling if the loop ends with any kind of jump 5075- nir/tests: add some loop peeling tests 5076- nir: skip opt_loop_peel_initial_break if continue block only has phis 5077- radv/rt: align constant data by 64 when inlining shaders 5078- radv/rt: don't split array/struct payload variables 5079- aco: do not use inline constants for 16-bit pseudo scalar trancendentals 5080- aco/tests: update assembler tests for llvm 5081- aco: workaround hazards in emit_long_jump 5082- nir/opt_loop: rematerialize header block derefs in their use blocks 5083- nir/tests: test opt_loop_peel_initial_break with derefs in header block 5084- aco: stop using instructions in ra_ctx::vectors 5085- aco: create vector affinities for phi operands 5086- aco: fix is_vector_intact for GFX11 BVH 5087- nir/divergence_analysis: disable phi undef optimization by default 5088- aco: ensure phis uniformized by divergence analysis are SGPR 5089- aco: skip uniformization of certain merge phis 5090- radv: fix output statistic for fragment shaders 5091- radv: optimize VS input load components to constants earlier 5092- radv: use explicitly sized types for some radv_shader_info members 5093- ac/llvm: cast to integer after derivative intrinsics 5094- aco,radv,radeonsi: add aco_shader_info::ps::has_prolog 5095- aco: split CounterMap off from VGPRCounterMap 5096- aco: minor CounterMap::operator== fix 5097- aco: workaround VALUReadSGPRHazard 5098- aco/tests: add tests for VALUReadSGPRHazard 5099- nir: fix shfr constant folding with zero src2 5100- nir/algebraic: fix shfr optimization with zero src2 5101- aco: add waitcnt build helper 5102- aco: fix printing of block_kind_discard_early_exit 5103- aco: insert NOP before dealloc_vgpr in the insert_NOPs pass 5104- aco: don't emit early exit over dealloc_vgprs 5105- aco: wait for scratch stores to complete before dealloc_vgprs 5106- aco: don't byte align global VMEM loads if it might be unsafe 5107- nir/algebraic: add bit-size check to extract_u8 pattern 5108- aco/gfx11: fix v1b=p_extract(src, 0, 16, 0) 5109- aco: check for SDWA before applying extract to lshl/cvt_f32 5110- aco: disallow p_extract(,,32,) 5111- aco: handle SGPR limitations when applying extract 5112- aco: add a bit more p_extract/p_insert validation 5113- aco: shrink code size of some p_extract 5114- aco: allow applying sign-extended sel to p_extract more often 5115- aco: add and use apply_extract_twice helper 5116- aco: combine extracts with sub-dword definitions 5117- aco: apply extract to p_extract_vector 5118- aco: apply extract to v_cvt_f32_ubyte0 5119- nir/algebraic: fix iabs(ishr(iabs(a), b)) optimization 5120- nir/algebraic: check bit sizes in lowered unpack(pack()) optimization 5121- nir/lcssa: fix premature exit of loop after rematerializing derefs 5122 5123Rob Clark (80): 5124 5125- freedreno/a6xx: Implement reg stomper support 5126- freedreno/a7xx: Fix GRAS_UNKNOWN_80F4 writes 5127- freedreno/cffdec: Fix a7xx CP_EVENT_WRITE decoding 5128- tu/drm/virtio: Add missing a7xx case 5129- freedreno/drm: Handle a7xx case 5130- freedreno: Move GENX/CALLX magic to common 5131- freedreno: Extract out common UBWC helper 5132- freedreno: Extract out shared LRZFC layout helpers 5133- freedreno/a6xx: Allocate lrcfc when needed for direction tracking 5134- freedreno/a6xx: Refactor CP_EVENT_WRITE emit 5135- freedreno/a6xx: Rework CCU_CNTL emit for a7xx 5136- freedreno/a6xx: Initial a7xx support 5137- tu: Fix issues with 16k (or larger) page sizes 5138- freedreno/drm/virtio: Fix issues with 16k (or larger) page sizes 5139- tu: Random non-consequential fixes 5140- subprojects: perfetto uprev to v47.0 5141- egl: Fix surfaceless + modifiers 5142- gallium: Add option to not add version to libgallium filename 5143- freedreno/computerator: Use CHIP variant reg builders 5144- tu: Add helper to calculate layer address 5145- tu: Use CHIP variant reg builders 5146- freedreno/registers: Deprecate non-variant reg builders 5147- tu: Quiet vla warnings 5148- freedreno/a6xx: Fix arb_texture_view regression 5149- freedreno/decode: Dump register values on LRZ_CLEAR 5150- freedreno/a6xx: Fix LRZ 5151- freedreno: Re-enable LRZ for a7xx 5152- freedreno/a6xx: Cleanup setup_slices() 5153- nir/opt_loop: Don't peel initial break if loop ends in break 5154- freedreno/a6xx: Sysmem clear fixes 5155- freedreno/drm: Fix ring_heap flags 5156- freedreno/crashdec: Fix fault address handling 5157- freedreno/register: A couple pm4 updates 5158- freedreno/ir3: Fix GPU name in disasm test 5159- freedreno/ir3: Add half-FLUT cases 5160- freedreno/ir3: Fix cat1 parser ambiguity vs FLUTs 5161- freedreno/ir3: Add more cat1 float-immed cases 5162- freedreno/a6xx: Only emit VFD/PC_POWER_CNTL for a6xx 5163- freedreno/a6xx: De-open-code VFD_MODE_CNTL 5164- freedreno/a6xx: Cleanup WFIs around RB_CCU_CNTL 5165- freedreno/a6xx: Set bin size per bin 5166- freedreno/a6xx: Move PC_TESSFACTOR_ADDR emit 5167- freedreno/a6xx: Move static regs to preamble IB 5168- freedreno/a6xx: Move more state emit per-bin 5169- freedreno/a6xx: Emit CP_SET_AMBLE packets 5170- freedreno/drm: Add preemption support 5171- freedreno: Balance out u_blitter cb0 save/restore 5172- freedreno/a6xx: Add some missing a7xx bits 5173- freedreno: Assert we aren't writing to 0x0 5174- freedreno/decode: Fix UBO decode on a7xx 5175- freedreno/ir3: Avoid draw/grid time input iteration 5176- freedreno/ir3: Add assert about const emit 5177- freedreno/ir3+tu: Convert driver-params to structs 5178- freedreno/ir3: Indentation fix 5179- freedreno/a6xx: Fix double SP_MODE_CONTROL emit 5180- freedreno/a6xx: Move tess-bo emit 5181- freedreno/ir3: Track # of app UBOs 5182- freedreno/ir3: Fix need_driver_params for UBO case 5183- freedreno: Rework indirect compute param emit 5184- freedreno/a6xx: Add support to load driver-params via UBO 5185- freedreno/a6xx: Fix color_cache_size 5186- freedreno/a6xx: Don't open-code INVALID_REG 5187- freedreno/a6xx: Add VPC hardware workaround for a750 5188- freedreno/a6xx: Add missing GRAS_SU_DEPTH_CNTL 5189- freedreno/ir3: Create UBO variables for driver-UBOs 5190- nir/lower_amul: Fix ASAN error 5191- freedreno/a6xx: Support variable wg size 5192- ir3/ra: Better CL/kernel support 5193- freedreno/computerator: Make shader show up in devcore/etc 5194- freedreno/ir3: Do not propagate away a widening move 5195- ir3: Fix binning pass driver UBOs 5196- ir3: Don't lower VS driver params to UBO 5197- freedreno/a6xx: Don't try resolve blits 5198- freedreno/a6xx: Stop exposing MSAA image load/store harder 5199- freedreno: Fix tile-per-pipe debug overrides 5200- freedreno/a6xx: Fix MSAA depth on a6xx gen1 5201- freedreno/ci: Garbage collect some obsolete xfails 5202- util/primconvert: Avoid OoB with improbable draws 5203- freedreno/a6xx: Don't check dst coords 5204- freedreno/a6xx: Random whitespace fix 5205 5206Robert Mader (5): 5207 5208- egl: Stop requiring texture_3D_image for EGL 1.5 5209- perfetto: Add Panfrost data sources to system.cfg 5210- egl: Add more errors cases during context creation 5211- egl: Stop requiring EXT_create_context_robustness for EGL 1.5 5212- v3d: Support SAND128 base modifier 5213 5214Rohan Garg (23): 5215 5216- anv: use the WA infrastructure when emitting WA 16013994831 5217- anv: program a custom byte stride on Xe2 for indirect draws 5218- anv,iris: prefix the argument format with XI for a upcoming refactor 5219- anv: refactor indirect draw support into it's own function 5220- anv: dispatch indirect draws with a count buffer through the XI hardware on ARL+ 5221- anv: migrate indirect mesh draws to indirect draws on ARL+ 5222- vtn: handle SpvOpExtInstWithForwardRefsKHR for non semantic instructions 5223- anv: enable KHR_shader_relaxed_extended_instruction 5224- anv: prefetch samplers when dispatching compute shaders 5225- iris,anv: simplify and inline sampler count calculations 5226- anv: enable the VK_KHR_shader_relaxed_extended_instruction feature 5227- hasvk: enable VK_KHR_shader_relaxed_extended_instruction 5228- intel/compiler: version can never be above 11 due to the previous check 5229- intel/compiler: use the correct cache enum for loads and stores 5230- hk: enable VK_KHR_shader_relaxed_extended_instruction 5231- nvk: enable VK_KHR_shader_relaxed_extended_instruction 5232- radv: enable VK_KHR_shader_relaxed_extended_instruction 5233- tu: enable VK_KHR_shader_relaxed_extended_instruction 5234- v3dv: enable VK_KHR_shader_relaxed_extended_instruction 5235- lvp: enable VK_KHR_shader_relaxed_extended_instruction 5236- docs: update feature matrix for VK_KHR_shader_relaxed_extended_instruction 5237- intel/brw: lower math op regions for Xe2+ 5238- anv: Xe2+ doesn't need the special flush for sparse 5239 5240Rohit Athavale (1): 5241 5242- d3d12: Fix typo in copyright comments 5243 5244Roland Scheidegger (1): 5245 5246- llvmpipe: Fix type mismatch when storing residency info 5247 5248Romaric Jodin (1): 5249 5250- perfetto: fix perfetto counter following perfetto v46.0 5251 5252Roy Chan (3): 5253 5254- amd/vpelib: Optimize the CPU usage by caching all the LUT configs 5255- amd/vpelib: fix zero input handling 5256- amd/vpelib: Add documentation 5257 5258Ruijing Dong (8): 5259 5260- radeonsi/vcn: correct a typo in a variable 5261- radeonsi/vcn: add new function for obu_header 5262- frontends/va: check av1 enc hdr metadata 5263- radeonsi/vcn: input av1 hdr metadata 5264- frontends/va: rework VAConfigAttribEncPackedHeaders query 5265- radeonsi/vcn: add HDR metadata obu in av1enc 5266- frontends/va: reset roi number 5267- radeonsi/vcn: qp map IB package sent by default 5268 5269Russell Greene (2): 5270 5271- wsi/wayland/perfetto: supply presentation clock to perfetto 5272- perfetto: fix macos compile 5273 5274Sagar Ghuge (8): 5275 5276- intel/compiler: Ray query requires write-back register 5277- intel/compiler: Adjust trace ray control field on Xe2 5278- intel/compiler: Fix indirect offset in GS input read for Xe2+ 5279- anv: Reduce clear color state alignment to 64B 5280- iris: Reduce clear color state alignment to 64B 5281- intel: uncached L1 to fix memory barrier issue in RT shader 5282- anv: Track all the descriptor sets 5283- intel: Switch to COMPUTE_WALKER_BODY 5284 5285Sai Teja (2): 5286 5287- ci: Don't run vulkan jobs for gbm changes 5288- ci: Disable angle jobs for GL changes 5289 5290Sai Teja Pottumuttu (2): 5291 5292- docs: Add required dependencies to meson.rst 5293- docs: Match README installation guide with meson.rst 5294 5295Sam Lantinga (1): 5296 5297- util: Fixed crash in HEVC encoding on 32-bit systems 5298 5299Samuel Pitoiset (233): 5300 5301- radv: stop re-initializing HTILE for non-compressed to compressed transitions 5302- amd/drm-shim: add missing GC_11_5_0 family 5303- amd/drm-shim: add GFX1150 support 5304- ci: add GFX1150 support to radv-fossils 5305- radv: fix handling view index from device index 5306- radv: cleanup robustness with vk_pipeline_robustness_state_fill() 5307- radv: add a helper to set shader stage key robustness info 5308- radv: add support for dynamic vertex input state with DGC 5309- radv: allow VK_EXT_vertex_input_dynamic_state with DGC 5310- radv/meta: simplify radv_meta_blit2d_normal_dst() 5311- radv/meta: rework creating blit pipelines 5312- radv/meta: simplify initializing DCC comp-to-single pipelines 5313- radv/meta: remove unused parameter from some blit init functions 5314- radv/meta: create blit2d layouts on-demand 5315- radv/meta: create itob layouts on-demand 5316- radv/meta: create btoi layouts on-demand 5317- radv/meta: create itoi layouts on-demand 5318- radv/meta: create cleari layouts on-demand 5319- radv/meta: simplify initializing bufimage pipelines 5320- radv/meta: create DGC prepare pipeline on-demand 5321- aco: fix bogus assert in RT prolog on GFX11+ 5322- radv: add a small helper to determine the index type with DGC 5323- radv: pass a dgc_cmdbuf parameter to more dgc helpers 5324- radv: refactor the DGC helpers to determine cmdbuf size 5325- radv: add a small comment describing the DGC implementation 5326- radv/winsys: pad gfx and compute IBs with only one NOP 5327- radv: pad GFX preambles IBs with only one NOP 5328- radv: use a sized NOP packet for the DGC preamble 5329- radv/amdgpu: fix CS padding for non-GFX/COMPUTE queues 5330- radv: remove redundant PA_SU_PRIM_FILTER_CNTL in the GFX preamble 5331- radv: stop emitting DB_RENDER_OVERRIDE in the GFX preamble 5332- radv: stop emitting PA_SC_CLIPRECT_RULE in the GFX preamble 5333- radv: stop emitting DB_COUNT_CONTROL in the GFX preamble 5334- vk/render_pass: add Mesa-specific flag for dynamic rendering 5335- radv: use the Mesa-specifc dynamic rendering flag for meta operations 5336- radv: add radv_shader_info::ps::uses_fbfetch_output 5337- radv: add support for input attachment indices with DRLR 5338- radv: fix DRLR with subpass input attachments and feedback loops 5339- radv/amdgpu: do not check that a CS is aligned if no padding is added 5340- radv/amdgpu: assert that the DGC IB VA is correctly aligned 5341- radv: optimize NOPs padding with DGC 5342- radv/ci: update flakes lists for NAVI21/VANGOGH 5343- radv/ci: enable RADV_PERFTEST=transfer_queue on GFX9+ 5344- radv: remove RADV_DEBUG=nogsfastlaunch2 5345- radv: rework computing the DGC cmdbuf layout 5346- ac: add ac_gfx103_get_cu_mask_ps() 5347- ac,radeonsi,radv: add common GFX preambles 5348- radv: prepare for specialized DGC shaders 5349- radv: add a pointer to the DGC layout in dgc_cmdbuf 5350- radv: specialize dispatch DGC token 5351- radv: specialize draw DGC token 5352- radv: specialize index buffer DGC token 5353- radv: specialize pipeline DGC token 5354- radv: specialize VBO DGC token 5355- radv: simplify allocating push constants with DGC 5356- radv: specialize push constant DGC token 5357- radv: specialize indirect command layout stride for DGC 5358- radv: remove useless check about non-indexed draws and DGC 5359- radv: move emitting the compute pipeline with DGC 5360- radv: stop passing the upload offset to dgc_emit_bind_pipeline() 5361- radv: add a helper to store data to the DGC upload space 5362- radv: move emitting VBOs with DGC 5363- radv: specialize push constant stages with DGC 5364- radv: allow VK_EXT_legacy_vertex_attributes with DGC 5365- radv: fix emitting DGC indirect draws with drawid/base_instance 5366- radv: return early when the vertex input state is dynamic 5367- radv: only initialize per-attribute vertex input state when needed 5368- radv: add radv_vs_input_state::attrib_index_offset 5369- radv: bind the static vertex input state from the pipeline 5370- radv: rename radv_vs_input_state to radv_vertex_input_state 5371- radv: pass the vertex shader to radv_write_vertex_descriptors() 5372- radv: pass more VBO info as parameters to the DGC prepare shader 5373- radv: stop storing DGC info as part of the VBO descriptors 5374- radv: specialize dispatch initiator for compute/task shaders with DGC 5375- radv: store the indirect compute pipeline VA to dgc_cmdbuf 5376- radv: use compute pipeline metadata for non-indirect DGC pipelines 5377- radv: fix missing attribute offset with dynamic VS inputs and DGC 5378- radv: specialize one more VBO parameter with DGC 5379- radv: pass use_per_attribute_vb_descs in radv_dgc_params 5380- radv: add a helper to write a single vertex descriptor with DGC 5381- radv: add a helper to get VBO info for emitting descriptors 5382- radv: compute rsrc3 for VBO descriptors in NIR for DGC 5383- radv: use radv_vbo_info for emitting descriptors with DGC 5384- radv: emit all VBO descriptors in the prepare DGC shader 5385- radv: pass vb_desc_usage_mask in radv_dgc_params 5386- radv: stop relying on the VS for emitting VBO descriptors in DGC 5387- radv: fix allocating sparse descriptor buffers in the 32-bit addr space 5388- radv: disable shaders linking with ESO when nextStage of VS/TES isn't present 5389- Revert "radv: specialize push constant stages with DGC" 5390- vulkan: Update XML and headers to 1.3.294 5391- vulkan: skip the disk cache when disableInternalCache is true 5392- radv: rework helpers for serialize/deserialize shaders 5393- radv: store the SHA1 RT stage to the shaders cache 5394- radv: store the number of RT stages per pipeline to the shaders cache 5395- radv: store whether a RT pipeline is a library to the shaders cache 5396- radv: make pipeline hashing functions non-static 5397- radv: disable the in-memory cache when disableInternalCache is true 5398- radv: add initial support for pipeline binaries 5399- radv: add support for capturing pipeline binaries 5400- radv: add support for importing pipeline binaries 5401- radv: advertise VK_KHR_pipeline_binary 5402- radv: rework emitting indirect compute pipelines with DGC 5403- radv: cleanup some functions that emit shader user SGPRs 5404- radv: store the indirect descriptor sets VA to the descriptor state 5405- radv: rename radv_emit_descriptor_pointers() to radv_emit_descriptors_per_stage() 5406- radv: unify emitting non-indirect/indirect descriptor sets 5407- radv: make the helper that uploads indirect descriptors non-static 5408- radv: stop allocating upload space for indirect descriptors with DGC 5409- radv: use radv_upload_indirect_descriptor_sets() in DGC 5410- radv: update PGM register for TES+GS compiled separately with ESO 5411- radv: precompute more PGM registers for all stages 5412- radv: use radv_get_user_sgpr_loc() more 5413- radv,radeonsi: remove remaining occurrences of TCS epilog 5414- aco,radv,radeonsi: move has_epilog to the fragment shader info 5415- radv: merge radv_emit_epilog() with radv_emit_ps_epilog_state() 5416- radv/rt: skip shaders cache for pipelines created with the capture/replay flag 5417- radv: fix copying the pipeline binary key 5418- radv: compute shader hash for shaders created without internal cache 5419- radv: enable more properties with VK_KHR_maintenance5 5420- radv: fix lowering the view index to an input varying for FS 5421- radv,aco: fix legacy vertex attributes when offset >= stride on GFX6-7 5422- radv: fix lowering VS inputs when offset >= stride on GFX6-7 5423- radv: add missing cache flushes for DRLR feedback loops 5424- radv: update the stencil layout for DRLR feedback loops 5425- radv/ci: stop skipping few tests that hang only with LLVM 5426- radv: stop reporting VKCTS conformance on GFX11+ 5427- amd/addrlib: remove bogus assert in HwlComputeSlicePipeBankXor() 5428- radv/ci: stop skipping dEQP-VK.api.command_buffers.many_indirect_disps_on_secondary 5429- zink/ci: update expected list of failures since recent piglit uprev 5430- radv: stop emulating GS invocations for legacy GS on RDNA1-2 5431- radv: only export KHR_video_maintenance1 with KHR_video_queue 5432- radv: fix assigning mesh shader outputs when clip/cull distances are read in FS 5433- radv: emit all shader related user SGPR states in one place 5434- radv: make sure to re-emit shader query state when a task shader is bound 5435- radv: use only one user SGPR for all NGG state 5436- radv: rename shader_query_state to task_state 5437- radv: introduce dirty flags for shaders state 5438- aco: fix descriptor leaking when printing assembly with CLRX 5439- radv: move emitting some RT user SGPRs when the RT pipeline is emitted 5440- radv: move updating compute scratch for RT when stack size is emitted 5441- radv: do not keep executable info when compiling shaders for ESO 5442- vulkan: Update XML and headers to 1.3.296 5443- radv: simplify determining conformant products 5444- ac/surface: add RADEON_SURF_VIEW_3D_AS_2D_ARRAY for GFX9+ 5445- radv: implement 2D views of 3D images using 2D_ARRAY descriptors on GFX9+ 5446- radv: advertise sampler2DViewOf3D 5447- zink/ci: update expected list of failures for VANGOGH 5448- ci: uprev VKCTS to 1.3.9.2 5449- radv/amdgpu: do not use a constant value for the IB size in dwords 5450- radv/amdgpu: add assertions to check the IB size 5451- radv: use base mip level from vk_image_view 5452- radv: stop passing redundant parameters to radv_image_view_make_descriptor() 5453- radv: remove redundant assertions about image views 5454- radv: stop passing image create flags to radv_image_view_init() 5455- zink/ci: update expected list of failures on NAVI31 5456- radv: do not expose NV DGC extensions on GFX6-7 5457- radv: remove GFX6-7 DGC support completely 5458- radv: update image view extent for non-compressed views earlier on GFX10+ 5459- radv: fix image view descriptors for samplers on GFX6-8 5460- radv: remove RADV_MAX_DRM_DEVICES 5461- radv: squash radv_get_memory_fd() with radv_GetMemoryFdKHR() 5462- radv: remove RADV_THREAD_TRACE_TRIGGER completely 5463- zink/ci: update list of expected failures for NAVI10 5464- vulkan: fix merging bind sparse submits together 5465- radv: fix conditional rendering with DGC preprocessing on compute 5466- radv: fix generating the global key for pipeline binaries 5467- radv: move radv_compact_spi_shader_col_format() to radv_cmd_buffer.c 5468- radv: add a helper to bind the color output state 5469- radv: add PKT3_INDIRECT_BUFFER_BYTES in the DGC path 5470- radv: implement IB chaining for DGC when it's executed on compute 5471- radv/amdgpu: simplify cs_execute_ib() 5472- radv/amdgpu: remove unused code about external IBs in the submit path 5473- radv: rename 'gfx' to 'main' in the DGC path 5474- radv: fix returning non-zero captured address without binding 5475- radv: use app names instead of exec name for shader based drirc workarounds 5476- radv: stop recomputing the viewport xform for guarband/viewport 5477- radv: optimize breaking batch when CB_TARGET_MASK change 5478- radv: use radv_normalize_blend_factor() more 5479- radv do not force-disable hierarchical stencil testing 5480- radv: fix emitting DB_RENDER_OVERRIDE on GFX12 5481- radv: track more redundant DB related registers 5482- radv: regroup and emit all DS related states in the same function 5483- ci: uprev vkd3d-proton to 65b81403435576d882d9141ae3eb4a29373fba0e 5484- radv: fix enabling/disabling user sample locations 5485- ci: uprev vkd3d-proton to 59d6d4b5ed23766e69fe252408a3401d2fd52ce8 5486- radv: fix initializing the HTILE buffer on transfer queue 5487- radv: do not use MRT counters for images created for db capture&replay 5488- radv: capture shader statistics when RGP is enabled 5489- radv: fix emitting NGG culling state for ESO 5490- radv: simplify determining if a graphics pipeline uses NGG culling 5491- radv: do no emit PA_SC_CONSERVATIVE_RASTERIZATION_CNTL in the preamble on GFX12 5492- amd: do not emit PA_SU_PRIM_FILTER_CNTL in the common GFX preamble 5493- ci: uprev VKCTS to 1.3.10.0 5494- amd/descriptors: set fmask_tile_swizzle for TC-compat CMASK images on GFX8 5495- radv: set missing FMASK surface counters for MSAA MRTs 5496- radv: fix considering NGG culling for depth-only rendering 5497- radv: remove useless check about gl_Position as PS inputs for NGGC 5498- radv: simplify determining if dual-source blending is enabled 5499- Revert "radv: advertise sampler2DViewOf3D" 5500- Revert "radv: implement 2D views of 3D images using 2D_ARRAY descriptors on GFX9+" 5501- Revert "ac/surface: add RADEON_SURF_VIEW_3D_AS_2D_ARRAY for GFX9+" 5502- radv: implement VK_EXT_device_generated_commands 5503- radv: advertise VK_EXT_device_generated_commands on GFX8+ 5504- radv: track more redundant raster related registers 5505- radv: regroup and emit all raster related states in the same function 5506- ac/spm: do not abort when the SPM BO is too small 5507- radv: resize the SPM bo when it's too small 5508- radv: fix wrong index in radv_skip_graphics_pipeline_compile() 5509- ac/nir: cull triangles/lines when all W positions are zero/NaN 5510- radv: fix configuring the memory violation exception for the compute stage 5511- docs: add missing RADV_TRAP_HANDLER 5512- radv: add an option to configure the trap handler exceptions 5513- radv/ci: skip dEQP-VK.api.command_buffers.many_indirect_disps_on_secondary 5514- radv: add missing L2 non-coherent image case for mipmaps with DCC/HTILE on GFX11 5515- ac: add ac_gpu_info::has_trap_handler_support 5516- radv: check for has_trap_handler_support instead of asserting 5517- radv: only emit the TBA/TMA registers on GFX8 5518- radv: fix dumping the faulty shader detected by the trap handler on GFX9+ 5519- radv,aco: save SQ_WAVE_GPR_ALLOC from the trap handler 5520- radv: add a struct that describes the trap handler layout 5521- aco: fix reading registers from the trap handler shader 5522- aco: add support for the trap handler shader on GFX9-GFX10.3 5523- radv: set missing shader info values for the trap handler 5524- radv,aco: dump all SGPRS from the trap handler 5525- aco: fix validation for VOP1 instructions without any dest/src 5526- aco: clear the current wave exception in the trap handler 5527- aco: add support for the trap handler shader on GFX11 5528- aco: use inlined constant offsets for storing SGPRs in the trap handler 5529- radv: cleanup tools related resources when destroying logical device 5530- radv: save the trap handler report in the HOME directory 5531- radv: mark some GFX6-7 GPUs as Vulkan 1.3 conformant 5532- radv: fix ignoring src stage mask when dst stage mask is BOTTOM_OF_PIPE 5533- radv: add a new drirc option to disable DCC for mips and enable it for RDR2 5534 5535Satadru Pramanik (1): 5536 5537- Update lp_bld_misc.cpp to support llvm-19+. 5538 5539Scott Moreau (2): 5540 5541- egl: small refactor to avoid code duplication 5542- egl: hook up swap interval bits when using zink on wayland 5543 5544Semenov Herman (Семенов Герман) (2): 5545 5546- radv: fix memleaks in radv_sqtt_reloc_graphics_shaders() 5547- radv: fix memleaks in radv_init_shader_upload_queue() 5548 5549Serdar Kocdemir (22): 5550 5551- Reduce heap pressure on getPacketContents 5552- Use append to chain memory pointer info 5553- Add VK_EXT_debug_utils to host modules for codegen 5554- Add VK_EXT_validation_features to host modules 5555- Add support for VK_EXT_robustness2 5556- Add VK_EXT_metal_objects support 5557- Mark VK_MVK_moltenvk extension as supported on vulkan 5558- Add VK_EXT_debug_report extension support 5559- Remove direct usages of MoltenVK library functions 5560- Remove empty ifdef blocks after the codegen 5561- Show missing extension's name in the abort message 5562- Handle all extensions for structure size calculation 5563- Add extensions used in test applications 5564- Add error messages for unhandled extensions 5565- Add VK_EXT_external_memory_metal definitions 5566- gfxstream: Add VkPrivateDataSlot handle type 5567- gfxstream: Handle tmp folder explicitly on codegen 5568- gfxstream: Use KHR version of the line_rasterization extension 5569- Update decoder.py to use try_unbox on destroy calls 5570- gfxstream: Check metal extension for external memory 5571- gfxstream: Keep VK_EXT_line_rasterization for codegen 5572- gfxstream: Allow VK_KHR_line_rasterization 5573 5574Sergi Blanch Torne (20): 5575 5576- ci: disable Collabora's farm due to maintenance 5577- Revert "ci: disable Collabora's farm due to maintenance" 5578- ci: disable Collabora's farm due to maintenance 5579- Revert "ci: disable Collabora's farm due to maintenance" 5580- New DUT for Alder Lake 5581- New testing jobs anv-adl{,-full} 5582- New testing jobs anv-adl-angle{,-full} 5583- New testing jobs intel-adl-skqp 5584- New testing jobs zink-anv-adl{,-full} 5585- Uprev Piglit to e9ab30aeaed97b69868cf4d6d6a3f70f3b53c362 5586- ci: Disable gk20a-gles 5587- ci: disable Collabora's farm due to maintenance 5588- Revert "ci: disable Collabora's farm due to maintenance" 5589- ci: disable Collabora's farm due to maintenance 5590- ci: shellcheck requires better 'boolean or' syntax 5591- ci: run shellcheck when changes on scripts to check 5592- Revert "ci: disable Collabora's farm due to maintenance" 5593- WIP: Re-enable Comet Lake 5594- ci: disable Collabora's farm due to maintenance 5595- Nightly full job for a630-gles-asan 5596 5597Sergio Lopez (2): 5598 5599- hk: limit the number of free BOs in each cmd pool 5600- hk: allow overriding sysmem with an env var 5601 5602Sergiu (1): 5603 5604- Tweaking YCBCR conversion to return correct VK Format 5605 5606Shih, Jude (1): 5607 5608- amd/vpelib: Update Plane Descriptor Writer 5609 5610Sid Pranjale (3): 5611 5612- glx: add a missing nullptr check 5613- vulkan/util: add vk_format_has_float_depth() 5614- nvk: implement VK_EXT_depth_clamp_zero_one 5615 5616Sil Vilerino (49): 5617 5618- Revert "d3d12: Video Encode - Remove PIPE_VIDEO_PROFILE_MPEG4_AVC_BASELINE as not supported" This reverts commit d6bb4ddc638f3ee37fbbe066c631dad80aaeb2d3. Fixes: d6bb4ddc638 ("d3d12: Video Encode - Remove PIPE_VIDEO_PROFILE_MPEG4_AVC_BASELINE as not supported") 5619- pipe: Add PIPE_BIND_VIDEO_DECODE_DPB/PIPE_BIND_VIDEO_ENCODE_DPB 5620- d3d12: Implement pipe_video_codec.create_dpb_buffer for AOT resources 5621- d3d12: Allow passing custom pipe_resource creation template/placed resource to d3d12_video_buffer_create_impl 5622- d3d12: Implement pipe_video_codec.create_dpb_buffer for texture array resources 5623- frontend/va: VaSyncSurface encoder check for surface feedback 5624- d3d12: Implement get_feedback_fence 5625- d3d12: Video Encode H264 to use direct DPB from frontend 5626- d3d12: Video Encode H264 - Support direct mmco operations 5627- d3d12: Video Encode HEVC to use direct DPB from frontend 5628- d3d12: Video Encode HEVC - Use VPS information from frontend, specifically for vps_max_dec_pic_buffering_minus1 5629- d3d12: Rename d3d12_video_encoder_config_dirty_flag_sequence_info to d3d12_video_encoder_config_dirty_flag_sequence_header 5630- pipe: Add pipe_h264_enc_pic_control.temporal_id 5631- d3d12: H264, HEVC, AV1 Store multiple rate control states for multiple temporal layers 5632- vl/vl_win32_screen_create_from_d3d12_device: Allow winsys to be injected 5633- vl/null_sw_winsys: Add extern declaration for C++ usage 5634- d3d12: Remove usage of WindowFromDC when building without d3d12 graphics 5635- d3d12: Video Encode HEVC - Store L0/L1 active nums at time to encode them in the headers 5636- d3d12: Video Encode H264 - Store L0/L1 active nums at time to encode them in the headers 5637- d3d12: Video Encode H264 - Workaround for D3D12 validation bug requiring pRefPicMarkingOperationsCommands for IDR frames 5638- pipe: Add PIPE_H26X_MAX_REFERENCES 5639- d3d12: Remove max_references usage from frontend to reserve barriers array 5640- d3d12: Fix d3d12_video_encoder_get_current_max_dpb_capacity to report actual max slots number in DPB 5641- d3d12: Use d3d12_video_encoder_get_current_max_dpb_capacity in tex array pool instead of separate constant 5642- d3d12: Use pipe_h264_enc_seq_param.max_num_ref_frames instead of pipe_video_codec.max_references in SPS 5643- Fix definitions of PIPE_BIND_VIDEO_DECODE/ENCODE_DPB as they conflict with PIPE_BIND_SCANOUT and PIPE_BIND_SHARED 5644- d3d12: Fix setting of direct_8x8_inference_flag in the SPS 5645- d3d12: Plumb constrained_intra_pred_flag and transform_8x8_mode_flag from pipe_h264_enc_pic_control 5646- d3d12: Set HEVC GOP state before calling d3d12_video_encoder_negotiate_requested_features_and_d3d12_driver_caps 5647- d3d12: Support for on demand vps, sps, pps, aud headers requests from frontend 5648- pipe: Remove renew_headers_on_idr, insert_aud_nalu unused params 5649- d3d12: Add HEVC support for on demand AUD NALUS 5650- pipe: Add HEVC encode range extension caps, SPS and PPS params 5651- d3d12: Add support for HEVC 4:4:4 video encode using AYUV format 5652- d3d12: Add support for HEVC range extension header parameters 5653- ci: Bump DirectX-Headers and Agility SDK dependencies to 1.614.1 5654- d3d12: Support create_dpb_buffer for non-opaque allocations when decoder/encoder supports it 5655- pipe: Add temporal_id to pipe_h26x_enc_dpb_entry 5656- d3d12: Support writing H264_SEI_SCALABILITY_INFO header on demand 5657- d3d12: Support writing H264 temporal scalability prefix slice NAL on demand 5658- d3d12: Report H264/HEVC support for 4 max temporal layers 5659- d3d12: H264 Encode - Set SPS.gaps_in_frame_num_value_allowed_flag=1 when num_temporal_layers > 1 5660- d3d12: d3d12_video_encoder_references_manager_h264 to use FrameDecodingOrderNumber as h264Pic->slice.frame_num 5661- d3d12: Plumb H264/HEVC temporal_id from pipe params 5662- d3d12: Fix HEVC chroma_format_idc, separate_colour_plane_flag and conf_win_*_offset for HEVC 444 5663- d3d12: Wrap d3d12_video_encode_get_hevc_codec_support around #if VIDEO_CODEC_H265ENC 5664- d3d12: fix incorrect memset in d3d12_video_encoder_references_manager_hevc::get_current_frame_picture_control_data 5665- d3d12: Fix video encode metadata snapshot at frame submission, must include changes done by DPB manager begin_frame method 5666- d3d12: Improve video encode intra-refresh logging 5667 5668Simon Perretta (6): 5669 5670- pvr: Add missing feature check in compute cmd stream 5671- pvr: Rescope msaa_mode 5672- pvr: Add missing include for FALLTHROUGH 5673- pvr: Inline xxhash header functions 5674- pvr: rename pds unused reg constant 5675- pvr: Make Vulkan driver depend only on nir headers, not library 5676 5677SoroushIMG (4): 5678 5679- pvr: fix intermediate size when configuring mrt setup 5680- pvr: cleanup 2101010 int format handling 5681- pvr: fix msaa setup for resolve TQs 5682- pvr: fix texture address offset when base level >0 5683 5684Stéphane Cerveau (2): 5685 5686- anv: inline query for vulkan video encode 5687- anv: check that inline query pool is VK_NULL_HANDLE 5688 5689Surafel Assefa (1): 5690 5691- vulkan: MESA_VK_ENABLE_SUBMIT_THREAD=0 disables threaded submit 5692 5693Sushma Venkatesh Reddy (4): 5694 5695- intel/clflush: Utilize clflushopt in intel_invalidate_range 5696- intel/brw: Fix DEBUG_OPTIMIZER 5697- intel/brw: Move get_nir_def() higher to avoid UNDEF 5698- intel/brw: Handle 16-bit sampler return payloads 5699 5700Sviatoslav Peleshko (14): 5701 5702- anv: Assert ray query BO actually exists 5703- brw,elk: Fix opening flags on dumping shader binaries 5704- anv: Release correct BO in anv_cmd_buffer_set_ray_query_buffer 5705- anv: Add full subgroups WA for the shaders with barriers in Breaking Limit 5706- brw: Fix mov cmod propagation when there's int signedness mismatch 5707- mesa: Reset vbo attributes after flushing them to Current in glPopAttrib 5708- anv: Update XeSS workaround executable names for Satisfactory 1.0 5709- intel/brw: Don't apply discard_if condition opt if it can change results 5710- intel/brw/gfx9: Implement WaClearArfDependenciesBeforeEot 5711- intel/elk/chv: Implement WaClearArfDependenciesBeforeEot 5712- intel/brw: List all instructions that have BranchCtrl bit 5713- intel/brw_asm: Add BranchCtrl support 5714- intel/elk: List all instructions that have BranchCtrl bit 5715- intel/elk_asm: Add BranchCtrl support 5716 5717Tapani Pälli (34): 5718 5719- intel/dev: update mesa_defs.json from workaround database 5720- blorp: support new flag for setting cps dependency 5721- anv: fix a cmd_buffer reference in simple shader 5722- anv: refactor cmd_buffer_flush_gfx_runtime_state for dirty state 5723- anv: implement workaround for Wa_18038825448 5724- intel/dev: update mesa_defs.json from workaround database 5725- gbm: depend on libdrm indepedent of dri2 setting 5726- anv: move setting 3DSTATE_CLIP::MaximumVPIndex from loop 5727- anv: check if RT writes are happening for HasWriteableRT 5728- anv: set correct miplevel for anv_image_hiz_op 5729- anv: change existing ICL workaround to depend on BLEND_STATE 5730- iris: use correct enum for aux state on depth fast clear 5731- mesa: check for ARB_ES3_compatibility in format checks 5732- anv: initialize pixel struct to zero when setting clear color 5733- iris: initialize pixel struct to zero when setting clear color 5734- glsl: set a version to standalone compiler context 5735- iris: fix issues with memory object updates via glBufferSubData 5736- blorp: assert that color depth is not 96 for Wa_16021021469 5737- anv: force StackIDControl value for Wa_14021821874 5738- drirc/anv: force_vk_vendor=-1 for Faaast Penguin 5739- anv: use mi_builder in CmdBeginTransformFeedbackEXT 5740- intel/genxml: introduce L3 Fabric Flush for gfx12 5741- intel/ds: add L3 fabric flush support 5742- anv: add plumbing/support for L3 fabric flush 5743- iris: add plumbing/support for L3 fabric flush 5744- iris: add depth, DC and L3 fabric flush for aux map invalidation 5745- anv: add depth, DC and L3 fabric flush for aux map invalidation 5746- drirc/anv: force_vk_vendor=-1 for Silent Hill 2 5747- intel/dev: update mesa_defs.json from workaround database 5748- anv: set StackIDControlOverride_RTGlobals for 2 workarounds 5749- mesa: fix DXT1 support with EXT_texture_compression_dxt1 5750- iris: implement VF_STATISTICS emit for Wa_16012775297 5751- anv: implement VF_STATISTICS emit for Wa_16012775297 5752- anv/iris: add note about Wa_18039438632 for RT flush after SBA 5753 5754Tatsuyuki Ishi (2): 5755 5756- util/u_queue: Replace relative time wait hack with u_cnd_monotonic 5757- radv: Enable descriptorBufferCaptureReplay. 5758 5759Thomas Debesse (1): 5760 5761- meson: fix linking gallium with xcb-xfixes when using x11 5762 5763Thomas H.P. Andersen (8): 5764 5765- nvk: EXT_post_depth_coverage 5766- nak: special case PhiDsts as not uniform 5767- docs: update for new extensions in nvk 5768- nvk: do not overwrite test cases 5769- dirconf: add a common vk_zero_vram 5770- nvk: handle driconf for zeroing vram 5771- driconf: set vk_zero_vram driconf for Path of Exile 5772- driconf: set vk_zero_vram driconf for X4 Foundations 5773 5774Thomas Wagner (4): 5775 5776- llvmpipe: Fix external memory object reference 5777- llvmpipe: Unmap DMA buffer to release all resources 5778- llvmpipe: properly save llvmpipe_memory_allocation in memory_object 5779- llvmpipe: Remove unused struct member 5780 5781Thong Thai (1): 5782 5783- ci: partially emulate cdna devices using lower image opcodes 5784 5785Timothy Arceri (48): 5786 5787- nir: create validate_tex_src_texture_deref() helper 5788- nir: add nir_tex_src_{sampler,texture}_deref_intrinsic 5789- glsl: make use of new tex src deref intrinsic 5790- nir: set disallow_undef_to_nan for legacy ARB asm programs 5791- glsl: fix glsl to nir support for lower precision builtins 5792- glsl: always copy bindless sampler packing constructors to a temp 5793- nir: make static assert more flexible 5794- nir/glsl: set deref cast mode during function inlining 5795- glsl: drop double support checks in helper 5796- glsl: store has implicit conversions bools in gl_shader 5797- glsl: apply implicit matching rules when linking 5798- glsl: fix function overload resolution at link time 5799- ci: bump piglit version 5800- nir/tests: make add_loop_terminators more flexible 5801- nir/tests: add basic terminator merge test 5802- ci: disable gimark trace 5803- mesa/glsl: add nir fields to the gl shader structs 5804- glsl: add gl_nir_validate_intrastage_interface_blocks() 5805- glsl: make gl_nir_validate_intrastage_arrays() more flexible 5806- glsl: add support for converting global instructions to NIR 5807- util/hash_table: add _mesa_string_hash_table_create() helper 5808- nir: add max_ifc_array_access field to vars 5809- nir: add nir_function_impl_clone_remap_globals() 5810- glsl: move _mesa_glsl_can_implicitly_convert() to linker_util.cpp 5811- nir: add function in/outs to variable modes 5812- nir: serialize nir_parameter type 5813- nir: store variable mode in nir_parameter 5814- nir: add implicit_conversion_prohibited field to nir_parameter 5815- glsl: add gl_nir_link_function_calls() 5816- glsl: add gl_nir_linker_size_arrays() 5817- nir: add nir_intrinsic_deref_implicit_array_length intrinsic 5818- nir: handle wildcard array deref 5819- glsl: add intrastage shader linking helpers for nir linker 5820- glsl: add lower_derivatives_without_layout() helper 5821- glsl: set blake3 hash in standalone scaffolding 5822- glsl: switch to a full nir based linker 5823- glsl: remove now unused linker code 5824- glsl: fix gl_{Clip,Cull}Distance error messages 5825- glsl/mesa: move uniform related shader structs to shader_types.h 5826- glsl: move resource_name_updated() to linker_util.cpp 5827- glsl/mesa: remove ir_uniform.h 5828- glsl: move error and warning helpers to util file 5829- glsl: move shader cache lookup call to st 5830- glsl: rename link_shaders() -> link_shaders_init() 5831- glsl: remove linker.cpp 5832- nir/glsl: set cast mode for image during function inlining 5833- nir/glsl: set deref cast mode for blocks during function inlining 5834- glsl/nir: fix function cloning at link time 5835 5836Timur Kristóf (22): 5837 5838- radv: Add gs/hs_inputs_read field for linked LS and ES. 5839- ac/nir/tess: Add tcs_inputs_read to LS output lowering. 5840- ac/nir/tess: Don't emit VS outputs that aren't read by TCS. 5841- ac/nir/esgs: Add gs_inputs_read to ES output lowering. 5842- ac/nir/esgs: Don't emit ES outputs that aren't read by GS. 5843- ac/nir: Add helper for I/O location mapping. 5844- ac/nir/tess: Map linked LS/HS I/O based on TCS input mask. 5845- ac/nir/esgs: Map linked ES/GS I/O based on GS input mask. 5846- ac/nir: Remove ac_nir_calc_io_offset function. 5847- radv: Stop assigning linked driver locations. 5848- nir: Remove unused nir_assign_linked_io_var_locations. 5849- ac/nir/tess: Simplify calculation of HS output LDS offset. 5850- ac/nir: Shorten the name of ac_nir_calc_io_offset_mapped. 5851- ac/nir/tess: Remove no_inputs_in_lds. 5852- ac/nir: Move varying cost functions from radeonsi to common code. 5853- radv: Add NIR XFB IO intrinsic info to legacy stages not just NGG. 5854- radv: Add new linking step and use nir_opt_varyings. 5855- radv: Determine linked I/O info after nir_opt_varyings. 5856- nir/opt_algebraic: Add various bitfield extract patterns. 5857- radv: Flush L2 cache for non-L2-coherent images in EndCommandBuffer. 5858- radv: Increase NGG culling PS param limit to 12 on GFX10. 5859- radv: Enable NGG culling by default on GFX10. 5860 5861Tom Cherry (1): 5862 5863- Do not pass std::shared_ptr when not transfering ownership 5864 5865Tomeu Vizoso (2): 5866 5867- etnaviv/ml: Fix includes 5868- etnaviv/nn: Fix use of etna_core_info 5869 5870Tomson Chang (2): 5871 5872- amd/vpelib: Fix color adjustment change detect hole 5873- amd/vpelib: Only update cached adjustment when it is valid 5874 5875Trigger Huang (1): 5876 5877- mesa: Fix AMD performance monitor implementation 5878 5879Utku Iseri (1): 5880 5881- st/mesa: clamp reported max lod bias 5882 5883Valentine Burley (72): 5884 5885- tu/rmv: Fix missing image bind logging for WSI images 5886- tu: Implement VK_EXT_device_address_binding_report 5887- tu/kgsl: Remove unused variable 5888- docs/features: Update features.txt for Turnip 5889- lavapipe: Block more formats from getting blit feature flags 5890- vulkan/format: Translate two 420_UNORM formats properly 5891- tu, fdl: Drop custom mapping for 1-plane YCbCr formats 5892- tu: Drop tu_vk_format_to_pipe_format helper function 5893- vulkan/wsi: Refactor can_present_on_device 5894- tu: Always report that we can present on kgsl 5895- freedreno/ci: Update expectations for Adreno 660 5896- freedreno/ci: Re-enable Adreno 660 for Vulkan pre-merge jobs 5897- android: Extract version from llvm-project instead of hardcoding it 5898- tu: Remove redundant bits override from lower_tex_ycbcr 5899- ci: Backport two fixes tor the Vulkan CTS 5900- freedreno/ci: Update expectations 5901- freedreno/ci: Skip tests that might time out during pre-merge jobs 5902- freedreno/ci: Move and document last remaining failures 5903- freedreno/ci: Re-enable Adreno 630 for Vulkan nightly jobs 5904- tu: Expose VK_KHR_workgroup_memory_explicit_layout 5905- freedreno/ci: Wire up skips for a630_vk_full 5906- nvk: Advertise VK_EXT_legacy_vertex_attributes 5907- llvmpipe: Only use udmabuf with libdrm 5908- vulkan, radv: Add new common vk_format_get_plane_width/height helpers 5909- tu: Use vk_format_get_plane_width/height to get the plane dimensions 5910- tu: Use vk_format_get_plane_count for tu6_plane_count 5911- util: Remove Vulkan-only formats from get_plane_width/height 5912- tu: Fix VK_EXT_extended_dynamic_state3 feature 5913- tu: Simplify VK_EXT_sample_locations SampleCounts assignment 5914- freedreno/ci: Uprev kernel for a630 runner 5915- freedreno/ci: Skip timing out test on a630 5916- freedreno/ci: Uprev kernel to 6.11 5917- freedreno/ci: Update expectations after Piglit uprev 5918- turnip/ci: Remove fixed test from a660 xfails 5919- freedreno/ci: Document some a630 EGL flakes 5920- ci: Add a new build-only job without libdrm 5921- freedreno/devices: Fix A740v3 from Quest 3 5922- freedreno/devices: Unify magic_regs for A740 and A32 5923- freedreno/devices: Document an alias for FD644 5924- egl: Fix missing libdrm check 5925- ci: Build zink and llvmpipe in debian-no-libdrm 5926- freedreno/ci: Document new flake 5927- tu: Streamline setting YCbCr feature bits 5928- tu: Be more consistent with using vk_format helpers 5929- tu: Use vk_format_get_plane_count instead of special casing 5930- tu: Fix conditional check in tu_image_view_init 5931- tu: Use vk_format_get_plane_format 5932- tu: Use existing helpers in tu_clear_blit 5933- tu: Introduce tu_aspects_to_plane helper function 5934- tu: Unify zeroing bufferFeatures 5935- tu: Set some feature bits earlier 5936- tu: Use v2 format feature flags 5937- tu: Update comments about the blob driver 5938- tu: Add an assert to tu6_plane_index 5939- util/format: Add new 10-bit P010 RGB/planar formats 5940- util/format: Add new 12-bit P012 RGB/planar formats 5941- lavapipe: Remap 10 and 12 bit formats to 16 bit formats 5942- nvk: Remap 10 and 12 bit formats to 16 bit formats 5943- hk: Remap 10 and 12 bit formats to 16 bit formats 5944- vulkan: Fix incorrect bpcs value for padded formats 5945- tu: Fix incorrect bpcs value for padded formats 5946- vulkan/format: Properly translate 10-bit and 12-bit formats 5947- vulkan/format: Update vk_format_from_pipe_format with additional formats 5948- ci: Add global ANGLE skips for its waiver 5949- freedreno/devices: Inline a690 quirk 5950- freedreno/devices: Document common name for a635 speedbins 5951- freedreno/devices: Enable 64-bit atomics on a735 and a740v3 5952- freedreno/devices: Establish a7xx sub-generations 5953- ir3: Don't lower to LCSSA before calling nir_divergence_analysis() 5954- intel/ci: Fix Alder Lake's configuration 5955- anv/ci: Remove additive blending fails on ADL 5956- lavapipe: Block YCbCr formats from getting blit feature flags 5957 5958Vignesh Raman (19): 5959 5960- ci/deqp-runner: unset CC for arm32 cross-compilation 5961- ci: uprev deqp-runner from 0.18.0 to 0.20.0 5962- ci: enable Mali V10 testing 5963- ci: use v6.11-rc5 kernel for Mali V10 testing 5964- ci: move mtl-fw.json to .gitlab-ci directory 5965- ci/deqp-runner: pass EXTRA_CARGO_ARGS 5966- ci: include exit code in test result output 5967- ci/lava: handle non-zero exit codes 5968- ci/lava: update unit tests 5969- ci/bare-metal: handle non-zero exit codes 5970- ci: add Gen10 Arm Mali firmware to rootfs 5971- ci/image-tags: add DEBIAN_PYUTILS_TAG 5972- ci/lava: set exit code in exception case 5973- ci: prepare-artifacts-python: copy if src dir exist 5974- ci: build KERNEL_IMAGE_BASE in before_script 5975- ci: move rules from rules-anchors to workflow 5976- ci/crosvm: Use default value for CROSVM_GALLIUM_DRIVER 5977- ci: rename FORCE_KERNEL_TAG to EXTERNAL_KERNEL_TAG 5978- Revert "ci: disable Collabora's farm due to maintenance" 5979 5980Visan, Tiberiu (2): 5981 5982- amd/vpelib: set the same range for clr adj 5983- amd/vpelib: Remove TODO comments and legacy check(#421) 5984 5985Vitaliy Triang3l Kuzmin (2): 5986 5987- r600: Add FMT_1_REVERSED to r600_formats.h 5988- r600: Add missing formats to r600_colorformat_endian_swap 5989 5990Vlad Schiller (3): 5991 5992- pvr: Handle VK_STRUCTURE_TYPE_EXPORT_MEMORY_ALLOCATE_INFO 5993- pvr: Handle VK_STRUCTURE_TYPE_IMAGE_FORMAT_LIST_CREATE_INFO 5994- pvr: Minor code refactor 5995 5996Vldly (1): 5997 5998- freedreno: Fix resource tracking on repeated map with discard 5999 6000Víctor Manuel Jáquez Leal (1): 6001 6002- frontends/va: Don't return P010/P016 as surface formats when encoding 6003 6004WANG Xuerui (2): 6005 6006- meson: Force use of LLVM ORCJIT for hosts without MCJIT support 6007- meson: Additionally probe -mtls-dialect=desc for TLSDESC support 6008 6009Wang, Kyle (1): 6010 6011- amd/vpelib: Add missing copyrights 6012 6013X512 (2): 6014 6015- egl/haiku: fix double free of BBitmap 6016- egl/haiku: fix synchronization problems, add missing header 6017 6018Yahan Zhou (30): 6019 6020- Update vk.xml and code generator 6021- Fix testAndroidBaselineProfile2021Support for host swiftshader 6022- Add support for descriptor with inline uniform block 6023- Implement vkQueueSubmit2 in the same way as vkQueueSubmit 6024- Add some vk 1.3 extension strings 6025- Create sync device before calling goldfish_sync_queue_work 6026- Parse more ext structs in vkAllocateMemory and vkCreateBuffer 6027- Disable compatible handle types for semaphore timeline 6028- Fix dEQP-VK.wsi.android.maintenance1 6029- VkImageCreateInfo should ignore queeu family indices in some situation 6030- Use old en/decoding for vkQueueSignalReleaseImage 6031- Add VK_EXT_vertex_attribute_divisor support for guest ANGLE 6032- Declare AHB blob unsupported in Vulkan 6033- Add VK_EXT_fragment_density_map for Chrome 6034- Add snasphot support for vk coherent memory 6035- Snapshot vk image content in common situation 6036- Better track color buffer usage in Vulkan 6037- Make it not crash during vk snapshot load 6038- Fix corner cases where vk dependency graph breaks 6039- Handle dependency by VkMemoryDedicatedAllocateInfo 6040- Snapshot VkBuffer 6041- Add snapshot for command buffer 6042- Fix handle create mismatch for Vulkan snapshot 6043- Add snapshot support for vkBindImageMemory2 6044- Fix Vulkan snapshot with -guest-angle 6045- Revert^2 "Add snapshot support for vkBindImageMemory2" 6046- [Vulkan Snapshot] tryLock when an API might create extra handles 6047- Signal the fence after color buffer copying 6048- Revert "Signal the fence after color buffer copying" 6049- [Vulkan Snapshot] Add dependency graphics pipeline -> render pass 6050 6051Yao Zi (1): 6052 6053- panvk: Link with --build-id explicitly 6054 6055YaoBing Xiao (1): 6056 6057- vulkan/x11: use xcb_connection_has_error to check for failue 6058 6059Yinjie Yao (2): 6060 6061- radeonsi/vcn: Rename transform_skip_disabled and remove hardcoded value for VCN5 6062- radeonsi/vcn: Add support for QVBR rate control mode for VCN5 6063 6064Yiwei Zhang (8): 6065 6066- venus: only request ring thread prio for TLS ring 6067- venus: clarify wsi image ownership 6068- anv: improve vma usage for descriptor buffer 6069- ci: re-enable venus ci 6070- venus: fix a race condition between gem close and gem handle tracking 6071- Revert "meson: disallow Venus debug + LTO build via GCC" 6072- venus: workaround cacheline overflush issue on Intel JSL 6073- venus: avoid over-caching sfb cmds 6074 6075Yunhyeok "Yune" Choi (1): 6076 6077- glx: Getting rid of the double assignment in \`__glXWireToEvent`. 6078 6079Zack Rusin (1): 6080 6081- svga: Redo the way generated files are handled 6082 6083Zan Dobersek (11): 6084 6085- freedreno: fix a7xx perfcntr countables 6086- ir3: some 8-bit subgroup intrinsics must execute as 16-bit instructions 6087- nir/lower_vars_to_scratch: calculate threshold-limited variable size separately 6088- ir3: lower relevant 8-bit ALU ops in nir_lower_bit_size 6089- ir3: use fully-functional dp4acc when available 6090- ir3: indicate possible dword straddle for any multi-component pvtmem access 6091- ir3: improve validation, display for ldp instructions 6092- tu: enable shaderInt8 support 6093- tu: use instance indices in RD dump filenames 6094- freedreno: sanitize device names for config name usage 6095- zink: fix bo_export caching 6096 6097Zhang He (1): 6098 6099- iris, crocus: fix a typo and break comment line correctly 6100 6101antonino (10): 6102 6103- panfrost: rename \`PAN_GPU_INDIRECTS` 6104- panfrost: refactor draw logic to support native draw indirect 6105- panfrost: implement indirect draw for CSF-based GPUs 6106- pan/cs: add helpers to emit contiguous csf code blocks 6107- pan/cs: add helpers to generate csf loop 6108- panfrost: implement multi draw indirect for CSF-based GPUs 6109- panfrost: advertise \`PIPE_CAP_MULTI_DRAW_INDIRECT` 6110- mesa/main: expose \`EXT_multi_draw_indirect` 6111- pan/cs: add support for nop instruction 6112- pan/genxml: fix SET_EXCEPTION_HANDLER instruction definition 6113 6114bbhtt (1): 6115 6116- pipe_loader_drm: Fix virtgpu_drm header path 6117 6118cheyang (1): 6119 6120- Android15 support gralloc IMapper5 6121 6122chyyran (1): 6123 6124- spirv_to_dxil: add missing SPIR-V capabilities 6125 6126dawnhan (1): 6127 6128- vulkan/android: Add missing AHB formats 6129 6130itycodes (1): 6131 6132- intel: Fix a typo in intel_device_info.c:has_get_tiling 6133 6134llyyr (1): 6135 6136- vulkan/wsi/wayland: fix suboptimal flag being ignored with explicit sync 6137 6138nyanmisaka (4): 6139 6140- intel/pci_ids: Update ADL device names 6141- intel/pci_ids: Update RPL device names 6142- intel/pci_ids: Update TGL device names 6143- intel/pci_ids: Update DG1 device names 6144 6145quic_lkondred (1): 6146 6147- freedreno: Add support for Adreno 663 GPU 6148 6149sergiuferentz (2): 6150 6151- Adding support for VK_EXT_external_memory_host extension 6152- gfxstream: VulkanBatchedDescriptorSetUpdate toggled on caps on Guest 6153 6154yinjiyao (2): 6155 6156- frontends/va: check hevc enc hdr sei 6157- radeonsi/vcn: add HDR sei in hevc enc 6158