1Mesa 25.0.0 Release Notes / 2025-02-19 2====================================== 3 4Mesa 25.0.0 is a new development release. People who are concerned 5with stability and reliability should stick with a previous release or 6wait for Mesa 25.0.1. 7 8Mesa 25.0.0 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 25.0.0 implements the Vulkan 1.4 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: 96a53501fd59679654273258c6c6a1055a20e352ee1429f0b123516c7190e5b0 mesa-25.0.0.tar.xz 25 SHA512: 7f5b6674c40b6c8dcab7934512ff754b40a6a8a466422c90236f614d322033d4d465307ddcd983f9f3afb1310e132ec3186a085d261c95493a0c460b2ec59ce8 mesa-25.0.0.tar.xz 26 27 28New features 29------------ 30 31- cl_khr_depth_images in rusticl 32- Vulkan 1.4 on radv/gfx8+ 33- VK_KHR_dedicated_allocation on panvk 34- VK_KHR_global_priority on panvk 35- VK_KHR_index_type_uint8 on panvk 36- VK_KHR_map_memory2 on panvk 37- VK_KHR_multiview on panvk/v10+ 38- VK_KHR_shader_non_semantic_info on panvk 39- VK_KHR_shader_relaxed_extended_instruction on panvk 40- VK_KHR_vertex_attribute_divisor on panvk 41- VK_KHR_zero_initialize_workgroup_memory on panvk 42- VK_KHR_shader_draw_parameters on panvk 43- VK_KHR_shader_float16_int8 on panvk 44- VK_KHR_8bit_storage on panvk 45- VK_EXT_4444_formats on panvk 46- VK_EXT_global_priority on panvk 47- VK_EXT_global_priority_query on panvk 48- VK_EXT_host_query_reset on panvk 49- VK_EXT_image_robustness on panvk 50- VK_EXT_pipeline_robustness on panvk 51- VK_EXT_provoking_vertex on panvk 52- VK_EXT_queue_family_foreign on panvk 53- VK_EXT_sampler_filter_minmax on panvk 54- VK_EXT_scalar_block_layout on panvk 55- VK_EXT_tooling_info on panvk 56- depthClamp on panvk 57- depthBiasClamp on panvk 58- drawIndirectFirstInstance on panvk 59- fragmentStoresAndAtomics on panvk/v10+ 60- sampleRateShading on panvk 61- occlusionQueryPrecise on panvk 62- shaderInt16 on panvk 63- shaderInt64 on panvk 64- imageCubeArray on panvk 65- VK_KHR_depth_clamp_zero_one on RADV 66- VK_KHR_maintenance8 on radv 67- VK_KHR_shader_subgroup_rotate on panvk/v10+ 68- Vulkan 1.1 on panvk/v10+ 69- VK_EXT_subgroup_size_control on panvk/v10+ 70- initial GFX12 (RDNA4) support on RADV 71 72 73Bug fixes 74--------- 75 76- radeonsi: regression with running DaVinci Resolve under rusticl since 666a6eb871d5dec79362bdc5d16f15915eb52f96 77- [ANV][LNL] - Black Myth: Wukong (2358720) - Corruption is visible near the edge of water. 78- [ANV][LNL] - Hogwarts Legacy (990080) - Pixelated corruption is visible when looking out at the water. 79- radv/video/h265: pps.flags.transform_skip_enabled_flag = 1 randomly hangs GPU 80- [ANV][LNL] - Steel Rats (619700) - Game crashes after opening logos play before reaching main menu 81- nvk: Implement host-only descriptors 82- Gnome-shell Wayland fails to start with segfault at modifier-less driver 83- [ANV][LNL] - DYNASTY WARRIORS: ORIGINS (2384580) - Dithered transparency has vertical bands. 84- AMD Radeon R9 270 randomly causes video playback applications to crash with "amdgpu: The CS has been rejected" 85- Rendering issues on GravityMark with RadeonSI ACO 86- i915: multiple tests assert with tgsi_ureg.h:893: ureg_swizzle: Assertion \`reg.File != TGSI_FILE_NULL' failed. 87- shaders/closed/steam/deus-ex-mankind-divided/260.shader_test fails NIR validation 88- shaders/closed/steam/deus-ex-mankind-divided/260.shader_test fails NIR validation 89- panvk : vk_pipeline_cache_object_deserialize: Assertion \`reader.current == reader.end && !reader.overrun' failed. 90- 46a8d5e7ef61735416d0c54886a7a9930621ae2c causes a permission denied spam 91- [BUILD] Build Failure: Implicit Function Declaration 'timespec_sub_saturate' (loader_wayland_helper.c) 92- intel genX_acceleration_structure: missing dependency to bvh/header.spv.h 93- KHR_subgroup glsl parsing broken 94- intel: add config options to disable ELK compiler bits 95- a618: godot-tps-gles3-high trace reproducible flakes 96- radv: mesh shader depth-only rendering is broken 97- anv: Enable VK_FORMAT_A4R4G4B4_UNORM_PACK16_EXT for Android 15 98- Using a buffer allocated on a rx 6800XT for scanout on a Ryzen 7950X results in glitches 99- Systemfreeze from mesa version 1:24.3.0-1-x86_64 and above with Chromium and derivatives [and more or less all other graphic related things] 100- msm_kgsl.h:560:21: error: expected ‘:’, ‘,’, ‘;’, ‘}’ or ‘__attribute__’ before ‘*’ token 101- [radeonsi] VC1 hardware decoding over vaapi outputs green screen 102- consecutive glDrawPixels do not reflect a changed pixel mapping 103- Crashing while Processing Shaders in Marvel Rivals on Mesa 24.3.2 & Mesa 24.3.3 104- Assertion \`nir_cf_node_get_function(&block->cf_node)->structured' failed 105- r300: Conditional jump or move depends on uninitialised value in Xnine.mova test 106- anv: Mesh shaders with two OpSetMeshOutputsEXT instructions are not supported 107- hasvk: apps crash since "intel/compiler: Remove usage of variable length arrays" 108- nir_validate should check metadata 109- anv: vkcube(pp) segfault in multi-GPU config, apparent vkCreateSwapchainKHR failure 110- anv,regression: Black square artifacts in Fenyx Rising on BMG 111- [anv] Cyberpunk visual corruption on BMG 112- [ANV][LNL] - Cyberpunk 2077 (1091500) - Flickering mesh during benchmark. 113- Intel Arc A770: Crosshair in THE FINALS renders too large 114- 3d render issues in Chromium after 1:24.3.1-3 update over 1:24.2.7-1 of mesa package 115- intel/compiler: Out of bounds read in brw_eu_compact.c 116- intel/compiler: Out of bounds read in brw_eu_compact.c 117- egl,dri2: Segfault when running wayland clients on non-default GPU 118- anv,regression: Visual glitches in Ghost of Tsushima on BMG 119- anv, regression: Resident Evil 2 d3d12 freezes in main menu on a Arc b580 120- radeonsi: fails to build with libc++ 121- Random mesa crashes in kwin_wayland on a 6600XT 122- enc->enc_pic.enc_pic_order_cnt_type always zero even if pic->pic_order_cnt_type non-zero that application set 123- [anv] Visual corruption in Cyberpunk on LNL and BMG 124- [anv] Borderlands 3 visual corruption on BMG 125- [ANV] LNL triangle corruption on clothing in HogwartsLegacy-trace-dx12-1080p-ultra 126- Intel: Dark graphical glitches on cars and characters on Disney Speedstorm 127- Regression in VA-API decoding 128- freedreno: fails to build with Android NDK 27c 129- hk_cmd_draw.c:3471:32: error: expression in static assertion is not constant 130- anv/gfx12: Enable non-zero fast clears for non-FCV CCS_E 131- gen12: 5% regression in factorio 132- 32-bit: error: format ‘%lx’ expects argument of type 133- regression;bisected;FTBFS: commit b13e2a495e9e3da56add7d852ca01b2cd7eef52d breaks x86_32 mesa build 134- glxext.c: error: 'struct glx_screen' has no member named 'frontend_screen' 135- regression;bisected;FTBFS: commit ae76a6a04596bfdbd37bab165bc5f2a5ff60d389 breaks x86 mesa build 136- Can't allocate dpb buffer on firefox 137- Segmentation fault resetting a query pool used to get BLAS properties 138- libvulkan_lvp link fails if glslangValidator is not installed 139- lvp acceleration structure broken on \`main` but not on \`staging/24.x` 140- radv: warning that "radv is not a conformant Vulkan implementation" on Navi 32 141- [anv][UHD630] DXVK 2.5 - 2.5.2 with DXVK_HUD=compiler or DXVK_HUD=fps freezes the game or the entire system (Works without compiler/fps HUD, DXVK 2.4.1 works fine) 142- Licenses seems incomplete/misleading 143- anv: Symbol clash in intel_batch_decoder build when expat not available 144- glcts failures on LNL/BMG 145- Lavapipe vulkan 1.4 support? 146- d3d12 vaapi: thread safety issues 147- anv: Missing textures and glitches in It Takes Two (game) 148- [anv][bisected] GravityMark segfault when enabling u-trace on RT workload 149- features.txt does not have a Vulkan 1.4 section despite some drivers already supporting the new version 150- Black screen bug that only affects AMD 151- Failure to correctly decode H.264, possibly specific to use of array output view 152- X1-85: Portal 2: Bottom of portal gun disappears 153- X-Plane 12: Prop disc rendering regression 154- Errors when enumerating devices create incorrect expecations 155- Resident evil 3 remake hanging - f8b584d6 regression 156- R6700XT: QP value doesn't affect output when using CQP rate control w/ H264/H265 VAAPI encoders 157- Bug in Mesa headers: \`error: redefinition of typedef 'GLsync'` 158- nak: Crash when starting The First Descendant 159- [r300] Regression in f424ef18010 breaks wayland on RS480M 160- anv: Missing text in Age of Mythology Retold on a Arc b580 161- RustiCL: and Clover broken with 9b7ea720c93 (!32713 (merged)) 162- nvk: Artifact Classic crash at loading screen 163- radeonsi VAAPI - vc-1 interlaced decoding garbled on Polaris 164- VDPAU AV1 hardware decoding broken for Mesa 25.0.0-devel 165- mesa: st_glsl_to_nir call to nir_opt_fragdepth might not be valid with MSAA 166- rusticl: warning: pointers cannot be transmuted to integers during const eval 167- rusticl: warning: pointers cannot be transmuted to integers during const eval 168- X1-85: Half Life 2 water rendering artifacts 169- crash on video playback 170- anv: Allow buffer compression for vkd3d by default? 171- anv: bellwright needs force_vk_vendor=-1 %command% to launch 172- [anv] Possible regression from !31269 173- Up to 60% perf drop in SynMark DrvRes benchmark 174- Memory leak on closing and re-opening X11 windows 175- SIVPE errors on GPU-based screen recording (Radeon 890M) 176- d3d12: va-api: build failure regression since 24.3.0-rc1 with MinGW GCC and clang 177- anv: Marvel Rivals XeSS crash, game needs force_vk_vendor=-1 env variable 178- anv: \`MESA: warning: INTEL_HWCONFIG_MIN_GS_URB_ENTRIES (2) != devinfo->urb.min_entries[MESA_SHADER_GEOMETRY] (0)` 179- aco: two nir_shader_clock are miss optimized to one for GFX12 180- aco: opengl buffer blit test fail when using aco on GFX12 181- aco: nir_ddx/ddy v_interp optimization does not work on GFX12 182- VAAPI b_depth 2 causes "manage_dpb_before_encode UVD - Failed to find ref0" error 183- regression;bisected;FTBFS: commits 37d47913437e2e9f72283ea8bffce00efc40fce2 and e67e44522f4f5de4fcde53ad0fb75e396ef31f52 breaks x86 mesa build 184- anv: Enable storage image compression on TGL 185- zink: zink_create_quads_emulation_gs doesn't write primitive ID 186- DZN/DXIL doesn't validate GTK shaders 187- black screen and "Failed to add framebuffer" error in wayland compositors when not filtering dmabuf formats with ccs modifiers on intel graphics when upgrading to mesa 24.3.0 188- nir: nir_opt_if_merge_test fails validation with NIR_DEBUG=validate_ssa_dominance 189- radv: Vulkan AV1 video decode glitches 190- radv: support RGP captures for purely compute pipelines 191- regression;bisected: c49a71c03c9166b0814db92420eadac74cbc4b11 leads to artifacts if on top of launched game (in full screen mode) show list running apps (Hold Alt + Tab) 192- !32067 broke piglit "spec\@egl_khr_create_context\@no-error context gl" 193- Intel: Re-enable bo cache in iris driver (Xe2) 194- [amdgpu][regression] GPU Hang/Reset Triggered by Several Applications 195- ANV: X4 Foundations crashes with vkAllocateDescriptorSets -12 196- About twenty vulkan-samples cases will crash caused by the same error while running on PanVK 197- Firestorm crashes on startup with Mesa 24.3 198- anv: Use-after-free detected by AddressSanitizer while running dEQP-VK 199- GPU process crash via WebGPU shader - UAF in mesa gcm_schedule_early_instr at src/compiler/nir/nir_opt_gcm.c:477 200- radv: DCC causes glitches in Red Dead Redemption 2 201- A5xx rendering issues with firefox 202- [ANV][Regression] Broken rendering in Flycast + Per-Pixel Alpha Sorting 203- [TGL][anv] Performance regression in Dota 2 replay 204- vtn: OpTypeStruct in kernel parameters trigger assertion in glsl_types.h 205- anv: Assertion failure in \`dEQP-VK.image.extended_usage_bit_compatibility.image_format_list.s8_uint_optimal_transfer_src_bit` 206- 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 207- [AMD RX 6700 XT] Artifacts while upscaling games in fullscreen mode 208- Distorted pixelated graphics with Radeon RX 7900 XT with some games 209- Total War Warhammer 2 Graphical Glitch 210- Glitching artifacts in tile shaped patterns on 6700 XT, when using upscaled fullscreen game on labwc 211- anv: Page fault when using MTL simulator in dEQP-VK.ray_tracing_pipeline.data_spill.report_intersection.float32 212- mesa_cache_db.c:316:33: error: call to undeclared function 'mremap' 213- [trunk] shaders fail hard in openmw after cbfc225e2bda2c8627a4580fa3a9b63bfb7133e0 214- u_perfetto.h:33:9: error: unknown type name 'clockid_t'; did you mean 'clock_t'? 215- brw_fs_opt_copy_propagation incorrectly handles size changes of uniforms 216- RADV Command buffer reuse doesn't reinitialize is_secondary 217- Virgl:Qcom sa8155 GL_MAX_FRAGMENT_SHADER_STORAGE_BLOCKS/GL_MAX_VERTEX_SHADER_STORAGE_BLOCKS is too small to run antutu benchmark apk 218- nouveau paraview msaa corruption 23.1 bisected regression 219- mesa fails to build due to missing SPV_ENV_UNIVERSAL_1_6 symbol 220 221 222Changes 223------- 224 225Aaron Ruby (6): 226 227- meson: Remove experimental from gfxstream driver build 228- gfxstream: Some cleanup in manual entrypoints 229- gfxstream: Remove VK_HOST_CONNECTION macro 230- gfxstream: Fix unused variable warnings in ResourceTracker.cpp 231- vulkan/util: Add c99_compat.h inclusion for cpp 'restrict' compatibility 232- gfxstream: Remove internal vk_util.h and vk_struct_id.h entirely 233 234Adam Jackson (2): 235 236- docs/envvars: Remove mention of IRIS_ENABLE_CLOVER 237- docs/envvars: Combine WGL sections 238 239Alejandro Piñeiro (1): 240 241- docs/features: mark VK_EXT_scalar_block_layout as supported for vc7+ 242 243Aleksi Sapon (9): 244 245- draw: primitive ID is per-patch 246- llvmpipe: spec\@arb_tessellation_shader\@execution\@gs-primitiveid-instanced is fixed 247- zink: spec\@arb_tessellation_shader\@execution\@gs-primitiveid-instanced is fixed 248- draw: front-face injection must check geometry shader primitive type 249- llvmpipe: PointCoord is offset when multisampling is enabled 250- meson: fix finding Python on Windows 251- llvmpipe: fix lp_test_arit on Windows 252- llvmpipe: LLVM v2f32 trunc/floor/ceil/nearbyint generates optimal x86 code since at least version 8 253- llvmpipe: disable anisotropic filtering for non-2D textures 254 255Alyssa Rosenzweig (206): 256 257- nir/opt_algebraic: optimize patterns from Skia 258- nir/opt_algebraic: add more 64-bit patterns 259- nir/opt_algebraic: add another 64-bit pattern 260- nir: add amul flag 261- nir: add late_lower_int64 option 262- nir: add ilea_agx/ulea_agx opcodes 263- nir/builder: use amul over ishl on agx 264- nir/opt_algebraic: don't lower amul if requested 265- nir/lower_uniforms_to_ubo: use amul 266- rusticl: respect late_lower_int64 267- agx: vectorize SSBOs 268- agx: model IC dispatch 269- agx: fix bfeil timing 270- hk: reduce max SSBO size 271- libagx: promote math to use AGX address mode 272- agx: rewrite address mode lowering 273- agx: change int conversion test 274- agx: add pseudo for signext 275- agx: optimize signext+iadd 276- agx: fold zext into int sources 277- agx: add tests for sign/zero-extend propagate 278- agx: fix atomics in tess count shaders 279- hk: don't advertise impossible modifiers 280- agx: optimize signext imad 281- agx: fuse iadd+large shift into imad 282- agx: make imad+ishl rules actually work 283- hk: drop assert 284- hk: fix meta shader name 285- libagx: fix cl warning 286- libagx: drop branch 287- libagx: drop dead code 288- libagx: vectorize triangle def'n 289- libagx: drop Clockwise 290- libagx: simplify index patch expression 291- libagx: don't key unroll to index size 292- libagx: fix unroll kernel constant qualifier 293- libagx: drop silliness in restart kernel 294- agx: fuse also 8-bit address math 295- asahi: extract agx_get_num_cores 296- asahi: correct core count, max freq 297- asahi: fix a2c with sample shading, harder 298- asahi: assert/cse resource valid 299- asahi: don't take compiled_shader in agx_build_internal_usc 300- asahi: drop dead param 301- asahi: factor out more compiled shader 302- asahi: move agx_gather_device_key 303- util: add u_tristate data structure 304- panfrost: switch to u_tristate 305- agx: make needs_g13x_coherency a tri-state 306- nir/lower_convert_alu_types: use intrinsics_pass 307- nir/conversion_builder: avoid redundant uint->uint clamp 308- nir/opt_algebraic: optimize convert_uint_sat(ulong) 309- nir: add names to function parameters 310- nir/print: print function signature 311- nir/print: annotate entrypoints 312- nir/print: print parameter names in calls 313- vtn: gather function parameter names 314- vtn: use rzalloc in bindgen 315- vtn: use named parameters in bindgen 316- vtn: preserve name, is_return in bindings 317- nir: split off some definitions for OpenCL 318- compiler: make glsl_sampler_dim available to CL 319- nir/lower_system_values: add ID to 32-bit lowering 320- nir: add nir_fixup_is_exported pass 321- vtn: introduce vtn_bindgen tool 322- libagx: switch to vtn_bindgen 323- libagx: move out of lib/ 324- libagx: DCE 325- asahi: drop dead ACCESS 326- asahi,agx: move texture lowering into the compiler 327- asahi: drop desc align alloc 328- asahi/decode: disasm 3D helper progs 329- asahi/clc: drop getopt 330- agx: vectorize scratch access 331- agx: gather workgroup size 332- asahi,hk: reenable rgb32 buffer textures 333- hk: generalize internal launch 334- hk: expose missing eds3 feature 335- hk: handle mismatching colour vs z/s dimensions 336- hk: implement EXT_depth_bias_control 337- hk: be robust against invalid MSAA inputs 338- hk: do not increment GS queries for passthru GS 339- hk: use common wg size 340- hk: add cmd buffer to hk_cs 341- hk: dce 342- libagx: fix return type 343- libagx: don't export vertex_id_for_top 344- asahi/genxml: fix 0 encoding for groups 345- asahi/genxml: fix 128-bit in CL path 346- asahi/genxml: optimize out masking with shr 347- asahi/genxml: define missing macros 348- asahi: add XML for cdm stream link with return 349- asahi: refmt 350- vtn: ignore SpvFunctionParameterAttributeSret 351- nir/pack_bits: handle 8-bit vec8 -> 64-bit 352- nir: add nir_lower_calls_to_builtins pass 353- asahi/clc: switch to nir_lower_calls_to_builtins 354- nir: add nir_foreach_entrypoint macros 355- nir: add workgroup size to functions 356- vtn: plumb through OpEntryPoint 357- vtn: gather workgroup size in libraries 358- nir: add nir_function::pass_flags 359- nir: add nir_remove_entrypoints helper 360- nir: add nir_lower_constant_to_temp helper 361- nir: add helpers for precompiled shaders 362- asahi,vtn: precompile kernels 363- libagx: increase wg size for query copy 364- asahi: crash on fault 365- hk: fix incorrect index size translate 366- hk: fix z bias perf regression 367- hk: implement hack for layered no attachments 368- hk: clarify bounds check calculations 369- agx: disable bounds check optimization 370- agx: reduce preamble/main alignment 371- asahi: drop dead pool stuff 372- asahi: don't leak rodata 373- hk,asahi,libagx: unify a bit of code 374- asahi: drop dead 375- asahi: fix page size alignment 376- asahi: fix u_blitter related leaks 377- asahi: label individual pools 378- asahi,hk: mmap BO on first use 379- asahi: add more asserts around bo add 380- asahi: fix agx_batch_add_bo 381- asahi: add =bodump debug help 382- asahi: fix agxdecode memory mapping 383- hk: implement timestamps 384- hk: claim 1.4 385- zink: fix gl_PrimitiveID reads with quads 386- nir/search_helpers: handle bcsel in is_only_used_as_float 387- nir/opt_algebraic: optimize sign bit manipulation 388- nir/opt_load_store_vectorize: match amul like imul 389- nir,asahi: make argument alignment configurable 390- mesa_clc: add depfile support 391- libagx: switch to depfile support 392- libagx: remove redundant source files 393- vulkan: rename depth bias graphics states 394- vulkan: bump layer api versions 395- nir: add printf_abort intrinsic 396- nir/lower_printf: allow fixed address 397- nir/lower_printf: lower aborts 398- nir/lower_printf: use unsigned math 399- nir/lower_printf: use 64-bit math 400- util/printf: be robust against truncated buffers 401- util/printf: add context-ful helpers 402- vulkan: add vk_check_printf_status helper 403- nir/lower_point_size: skip non-var derefs 404- clc: plumb cl_khr_subgroup_ballot 405- libcl: add a common header for CPU/GPU stuff 406- libcl: add VkDraw(Indexed)IndirectCommand definitions 407- util/bitpack_helpers: make partially CL safe 408- asahi: allow c23 extensions 409- asahi/clc: remap __FILE__ 410- asahi,hk: wire up printf, abort 411- agx: implement halts 412- libagx: drop pointless helper 413- libagx: port to common libcl.h 414- compiler: use libcl.h for CL 415- compiler: add mesa_prim_has_adjacency helper 416- asahi: use mesa_prim_has_adjacency 417- nir: add lower_scratch_to_var pass 418- compiler/glsl_types: add glsl_get_word_size_align_bytes 419- agx: optimize scratch access 420- radeonsi: use mesa_prim_has_adjacency 421- asahi: fix mmap'ing imported BOs 422- hk,libagx: move hk_draw to the gpu 423- asahi: use common draw 424- libagx: add missing agx_vdm_return 425- agx: add more 8-bit address fusing rules 426- asahi: reformat 427- agx: match another address pattern 428- libagx: move index size helpers to the gpu 429- libagx: refactor index buffer code 430- libagx: factor out load/store_index 431- hk: use index buffer overflow check 432- hk: factor out hk_draw_as_indexed_indirect 433- hk,libagx: accelerate index buffer robustness 434- hk,libagx: handle adjacency without a GS 435- libagx,hk: handle pipeline stats queries without a GS 436- libagx: use designated initializers 437- hk: avoid compiling unneeded VS->GS variants 438- hk: fix primitive restart dirty tracking 439- glsl: fix glsl_get_word_size_align_bytes 440- nir: pass a callback to nir_lower_robust_access 441- nir/lower_robust_access: fix robustness with atomic swap 442- libagx: add agx_barrier enum 443- nir,asahi,hk: add barrier argument to MESA_DISPATCH_PRECOMP 444- intel: set max_buffer_size to nir_lower_printf 445- nir/lower_printf: drop null check 446- nir/lower_printf: drop default max buffer size 447- nir,util: move printf serializing into util 448- util: add u_printf_hash helper 449- util/u_printf: add singleton implementation 450- util/u_printf: allow printing from singleton 451- nir/lower_printf: add option to hash format strings 452- nir/lower_printf: support dynamic buffer size 453- nir: add nir_lower_printf_buffer pass 454- agx: defer printf address lowering 455- nir/lower_printf: drop static buffer addr lowering 456- util,vulkan,asahi,hk: hash format strings 457- nir/lower_robust_access: do not preserve control flow 458- nir: fix O(N^2) behaviour in nir_remove_dead_variables 459- meson: project-wide fs = import('fs') 460- clc,libagx: drop --in for mesa_clc 461- clc,libagx: automatically set lang version 462- nir/serialize: strip function names names 463 464Antonino Maniscalco (1): 465 466- nir,zink,asahi: support passing through gl_PrimitiveID 467 468Antonio Ospite (53): 469 470- ci/deqp: replace local android patches with upstream solution 471- docs/android: update docs/android.rst after libgallium_dri updates 472- docs/android: improve documentation about building llvmpipe for Android 473- docs: remove leftover mention of meson dri3 option 474- ci/android: unset compiler env vars in debian/android_build.sh 475- ci/android: add a script to build LLVM libraries for Android 476- ci/container: remove S3_JWT_FILE when container_job_trampoline.sh exits 477- ci: set GIT_COMMITTER_DATE in a locale-agnostic format 478- ci/deqp: refresh some patches to apply on top of recent VK-GL-CTS 479- ci/deqp: cherry-pick fixes for building GL and GLES deqp on Android 480- ci/deqp: enable building testlog tools on Android too 481- ci/deqp: collect the mustpass lists also for the android target 482- ci/android: fix problem with deqp version file when building for Android 483- ci/android: build deqp for DEQP_API=VK 484- ci/android: build llvmpipe driver for Android by forcing llvm fallback 485- ci/android: don't copy the DRI drivers which are not needed anymore 486- ci/android: restart all services after copying the new mesa libraries 487- ci/android: handle premature exit of .gitlab-ci/cuttlefish-runner.sh 488- ci/android: update version of cuttlefish host tools 489- ci/android: add sudo to EPHEMERAL deps for debian/x86_64_test-android.sh 490- ci/android: get custom cuttlefish images from the S3 491- ci/android: make cuttlefish-runner.sh more robust against different Android images 492- ci/android: better separate host and guest mesa artifacts 493- ci/android: use a custom kernel when launching cuttlefish 494- ci/android: fix warning when using chown 495- ci/android: fix result dir for Android guest execution of deqp-runner 496- ci/android: don't call cuttlefish-host-resources script 497- ci/android: reorder PATH and LD_LIBRARY_PATH values to clarify priority 498- ci/android: also copy mesa vulkan libraries to the Android guest 499- ci/android: update list of deqp files pushed to the guest system 500- ci/android: use a native adb connection 501- ci/android: set XDG_CACHE_HOME and pass --shader-cache-dir to deqp-runner 502- ci/android: use a /data/deqp subdirectory on guest to store dEQP files 503- ci/android: set VK_DRIVER_FILES before launching cuttlefish 504- ci/android: add ci rules to test llvmpipe on Android 505- ci/android: add ci rules to test venus on Android 506- ci/android: upgrade DEBIAN_TEST_ANDROID_TAG 507- ci/android: fix meson C++ cross-compiler argument detection 508- ci/android: update ANDROID_NDK and ANDROID_SDK_VERSION 509- ci/android: use ANDROID_SDK_VERSION when building deqp components 510- ci/android: use ANDROID_SDK_VERSION for debian-android job too 511- ci/android: rename variable ANDROID_NDK to ANDROID_NDK_VERSION 512- docs/android: bump suggested platform-sdk-version to 34 513- freedreno/meson: remove C++ cross-build arguments HACKs 514- freedreno/meson: sort list of options passed to get_supported_arguments() 515- ci/android: update CUTTLEFISH_BUILD_NUMBER 516- ci/android: define an INSTALL var for the source of mesa artifacts 517- ci/android: improve handling of expectation files 518- ci/android: fix pulling results from Android device 519- ci/android: post-process testlog XML and create a junit.xml 520- ci/android: pass --max-fails to deqp-runner in cuttlefish-runner.sh 521- ci/android: pass --allow-downgrades when installing cuttlefish host tools 522- ci/android: stop pushing libglapi.so since it's not available anymore 523 524Arseny Kapoulkine (1): 525 526- radv: On GFX11, use box sorting heuristic based on ray flags 527 528Arvind Yadav (1): 529 530- amd: Add amdgpu userqueue IOCTL functions 531 532Asahi Lina (16): 533 534- asahi: Add pipe bind flags to resource debug 535- asahi: Add PIPE_BIND_SHARED to imported resources 536- asahi: Extract agx_decompress_inplace() 537- asahi: Introduce batch->feedback to disable compression in PBE 538- asahi: In-place decompress shared resources for feedback loops 539- hk: Add virtio implicit sync support 540- hk: Fix DRM modifier selection for compressed surfaces 541- hk: Enable missing swapchainMaintenance1 support 542- asahi: Use 64bit size fields 543- hk: Bump up max buffer size 544- asahi: UAPI update to add GET_TIME & cleanup 545- asahi: Fix agx_gpu_time_to_ns & implement DRM_ASAHI_GET_TIME 546- asahi: UAPI update to add support for user timestamp buffers 547- asahi: Add timestamp buffer ops 548- asahi: Virt UABI update 549- asahi: hk: Enable timestamps for virt 550 551Autumn Ashton (1): 552 553- radv/video: Fix bitstreamStartOffset including dstBufferOffset 554 555Bas Nieuwenhuizen (1): 556 557- util/perf: Fix some warnings. 558 559Benjamin Cheng (4): 560 561- ac/vcn: allow sq signature package to be skipped 562- radv/video: support event for pre-VCN4 encode queues 563- radv/video: support event for pre-VCN4 decode queues 564- radv/video: enable by default on vcn2/3 with latest fw 565 566Benjamin Lee (36): 567 568- panvk: inherit sample count in secondary cmdbufs 569- nir: clamp small W in nir_lower_viewport_transform 570- nir: document order requirement for nir_lower_viewport_transform 571- panvk: refactor fbinfo into a temp var in get_tiler_desc 572- panvk: treat provoking vertex as dynamic state 573- panvk: set provoking vertex in fbinfo 574- panvk: advertise VK_EXT_provoking_vertex 575- nir: handle arbitrary per-view outputs in nir_lower_multiview 576- nir: document index semantics in nir_lower_multiview 577- nir: treat per-view outputs as arrayed IO 578- nir: add option to use compact view indices 579- panvk: implement multiview support 580- panvk: only clear enabled views 581- panvk: disable position fifo optimization when multiview enabled 582- panvk: advertise multiview support on v10+ 583- panvk: add note about pan_lower_store_component requirements 584- nir: update docs for nir_get_io_arrayed_index_src 585- panvk: set uses_sample_shading NIR flag when sample shading is forced 586- panvk: fix sample position when sample shading is disabled 587- panvk/csf: fix alpha-to-coverage 588- panfrost: add intrinsic to load frag coord at a barycentric 589- panfrost: add nir pass to lower noperspective varyings 590- panfrost: collect noperspective varyings in shader info 591- panvk: pass noperspective_varyings sysval as a push constant 592- panfrost: add pass to lower noperspective varyings to a constant 593- panvk: use static noperspective when statically linking VS and FS 594- panfrost: factor FS shader key into a helper function 595- panfrost: specialize VS on FS interpolation qualifiers 596- panvk: handle sample mask writes on 1-sample targets 597- panvk: remove load_multisampled_pan sysval 598- panfrost/va: add FLUSH instruction 599- panfrost/va: implement fquantizetf16 ftz 600- panvk: disable round_to_nearest_even for NEAREST-filtered samplers 601- panfrost: remove incorrect usage of MALI_PIXEL_KILL_STRONG_EARLY 602- panfrost: fix hang by using MALI_PIXEL_KILL_WEAK_EARLY in color preload 603- panfrost: remove is_blit flag 604 605Benjamin Otte (1): 606 607- vulkan/wsi: Support alpha swapchains on win32 608 609Benjamin ROBIN (1): 610 611- util/disk_cache: Do not try to delete old cache if cache is disabled 612 613Bo Hu (5): 614 615- gfxstream: snapshot: avoid double boxing dispatchable handle 616- gfxstream: snapshot: DescriptorSet allocate and update 617- gfxstream-guest: update offset to correct value 618- update decoder.py to clean up un-used ApiCallInfo 619- remove the mReconstructionMutex in load 620 621Boris Brezillon (103): 622 623- panvk: Enable CI on G610 624- pan/ci: Move g610-vk jobs to post-merge CI 625- panvk: Change the prototype of panvk_select_tiler_hierarchy_mask() 626- panvk: Kill unused fields in panvk_cmd_graphics_state 627- panvk: Move the panvk_cmd_graphics_state definition to panvk_cmd_draw.h 628- panvk: Move panvk_cmd_compute_state to a common place 629- panvk: Move is_dirty() to panvk_cmd_draw.h and rename it 630- panvk: Don't link the VS and FS shaders on v10 631- panvk: Sanitize the driver-internal dirty state tracking 632- panvk: Move common gfx bits to a new source file in the common dir 633- panvk: Cache the fs_required() result 634- panvk/csf: Fix a wait-LS operation in finish_cs() 635- panvk/cs: Poison cmdbuf registers when PANVK_DEBUG=cs is set 636- panvk/ci: Update CI expectations to have a green CI 637- panfrost: Increase AFBC body alignment requirement on v6+ 638- panfrost: Add a helper to expose the maximum effective tile size 639- panfrost: Add the concept of render block 640- panfrost: Add support for AFBC(split) 641- panfrost: Advertise support for AFBC(32x8,sparse,split) 642- pan/decode: Flush the dump file before crashing 643- panvk/csf: Keep a cache of the CS reg file at the panvk_queue level 644- panvk/csf: Fix cross command buffer render pass suspend/resume 645- panvk/csf: Explain why the tiler is set to 0xdeadbeefdeadbeef 646- panvk: Fix panvk_plane_index() for D32_SFLOAT_S8_UINT 647- pan/cs: Add cs_exception_handler_ctx 648- pan/cs: Align exception handlers with NOPs 649- pan/cs: Add dynamic save_reg to exception handler 650- pan/cs: Add block macro for exception handler 651- panvk/csf: Fix register overlap in issue_fragment_jobs() 652- pan/cs: Return the dump region size when an exception handler is defined 653- pan/cs: Return exception handler size/address 654- panfrost: Add cs_exception_handler_def() to the ForEachMacros list 655- panvk/csf: Use the information returned by cs_exception_handler_def() 656- panfrost: Use the handler size returned by cs_exception_handler_def() 657- panvk: Filter out input-attachment usage on non renderable formats 658- pan/decode: Untangle CS disassembling and interpretation 659- pan/decode: s/interpret_ceu/interpret_cs/ 660- pan/decode: Rename pandecode_cs() into pandecode_interpret_cs() 661- pan/decode: Add a helper to print CS binaries without interpreting them 662- pan/decode: Provide a helper to print messages outside of the decoding path 663- pan/cs: Add a LOAD_IP pseudo instruction 664- pan/cs: Add an event-based tracing mechanism 665- panvk/csf: Use event-based CS tracing 666- panvk/csf: Don't disable SIMULTANEOUS_USE when tracing is enabled 667- panvk: Add a flag to force SIMULTANEOUS_USE 668- pan/texture: Move the plane info retrieval logic to a helper function 669- pan/texture: Stop passing the view format around 670- pan/texture: s/index/plane_index/ in panfrost_emit_plane() 671- pan/texture: Stop passing a layout to panfrost_emit_plane() 672- pan/texture: Pass pan_image_section_info around 673- nir: Let nir_lower_texcoord_replace_late() report progress 674- panfrost: s/NIR_PASS_V/NIR_PASS/ 675- panfrost: Use nir_shader_intrinsics_pass() for the line_smooth lowering pass 676- panvk: s/NIR_PASS_V/NIR_PASS/ 677- pan: s/NIR_PASS_V/NIR_PASS/ 678- panvk: Move the descriptors preparation out of CreateImageView() 679- vk/meta: Pass depth/stencil attachments only when a clear is requested 680- panvk: Ignore the view aspects when dealing with depth/stencil attachments 681- pan/cs: Fix cs_builder allocation failure robustness 682- panvk: Wrap our descriptor lowering passes in NIR_PASS() 683- panvk: Stop using magic values for the sysval push constant offset/range 684- panvk: Automate sysval access from NIR shaders 685- panvk: Lower dynamic push_constant loads in desc_copy logic 686- panvk: Lower load_push_constant with dynamic offset to global loads 687- pan/bi: Get rid of bi_lower_load_push_const_with_dyn_offset() 688- panvk: Don't define push_constant range/base when we don't have to 689- pan/indirect: Don't use .base to pass the push_constant offset 690- pan/mi: Don't pretend we support push constants 691- pan/bi: Disallow non-zero .{range,base} on load_push_constant instructions 692- pan/bi: Fix mem_access_size_align_cb() for push constants 693- panvk: Don't lower load_base_vertex 694- panvk: Fix first_vertex/base_instance types 695- pan: Don't pretend we support load_{vertex_id_zero_base,first_vertex} 696- panvk: Don't lower load_blend_const_color_rgba 697- panvk: Factor-out the sysvals initialization logic 698- panvk: Pass a cmdbuf to blend_emit_descs() 699- panvk: Pack push constants 700- panfrost: Kill the mali_ptr typedef 701- panfrost: Kill the uXX typedefs 702- panfrost: Move MALI_EXTRACT_INDEX to pan_format.h 703- panfrost: Move MAX_{MIP_LEVELS,IMAGE_PLANES} to pan_texture.h 704- panfrost: Kill panfrost-job.h 705- panvk: Don't invalidate the viewport on cull mode updates 706- panvk/jm: Fix depth clipping with small viewport depth range 707- panvk: Fix an alignment issue on x86 708- panvk: Fix panvk_priv_mem_bo() on 32-bit platforms 709- panfrost/ci: Add panvk and panfrost to the debian-x86_32 job 710- pan/genxml: s/PAN_PAN_HELPERS_H/PAN_PACK_HELPERS_H/ 711- pan/genxml: Include pan_pack_helpers.h instead of copying it 712- pan/genxml: Generate MALI_XXX_PACKED_T macros 713- panfrost: Fix instanced draws when attributes have a non-zero divisor 714- pan/cs: Fix the tracepoint register dump loops 715- pan/cs: Allow undefined value if condition=always in cs_branch_label() 716- pan/cs: cs_{break,continue} are not for_each macros 717- panvk/csf: Make all sync operations on the CSG scope 718- panvk/csf: Use cs_sr_reg64() instead of cs_reg64() when setting the OQ pointer 719- panvk/csf: Rework the occlusion query logic to avoid draw flushes 720- panvk/csf: Fix add_memory_dependency() for input attachment access 721- panvk/csf: Add a knob to force texture cache invalidation on RUN_FRAGMENT 722- panvk: Don't clobber registers if the render pass was suspended 723- pan/decode: Fix the blend_count mask 724- panvk/csf: Don't free the resources twice when init_render_desc_ringbuf() fails 725- panvk: Initialize device virtual address space after the VM creation 726 727Brad Smith (1): 728 729- util: Support elf_aux_info() on OpenBSD arm and ppc 730 731Brian Paul (2): 732 733- svga: add svga_resource_create_with_modifiers() function 734- svga: fix printing 64-bit value for 32-bit build 735 736Caio Oliveira (90): 737 738- intel/executor: Fix exec_size in \@read macro for Xe2 739- intel/brw: Add test for combining SWSB dependencies in SENDs 740- intel/brw: Allow extra SWSB encodings for Xe2 741- intel/common: Properly dispose resources in mi_builder tests 742- intel/common: Prepare mi_builder tests to support Xe KMD 743- intel/common: Implement Xe KMD in mi_builder tests 744- intel/common: Enable mi_builder test for PTL 745- intel/brw: Add SHADER_OPCODE_BALLOT 746- intel/brw: Add SHADER_OPCODE_QUAD_SWAP 747- intel/brw: Omit type and region in payload sources when printing IR 748- intel/brw: Use <V,W,H> notation for FIXED_GRF and ARF source when printing IR 749- intel/executor: Enable PTL 750- intel/brw: Fix decoding of cond_modifier and saturate in EU validation 751- intel/brw: Fix SWSB output when printing IR 752- intel/brw: Dump IR after lower scoreboard pass 753- util/ra: Remove unimplemented function declaration 754- intel/brw: Add is_control_source for the new subgroup ops 755- mr-label-maker: Rules for intel/executor 756- intel/brw: Enable EU validation and compaction tests for PTL 757- intel/brw: Dump errors when brw_assemble() fails EU validation 758- intel/compiler: Use #pragma once instead of header guards 759- intel/brw: Remove overloads for brw_print_instruction/s functions 760- intel/brw: Consider if SEND is gather variant when setting ex_desc 761- intel/brw: Add TGL_PIPE_SCALAR value 762- intel/brw: Add assembly support for ARF scalar register 763- intel/brw: Add validation for ARF scalar register 764- intel/executor: Add example using scalar register and send gather 765- intel/brw: Skip some regioning EU validation for Vx1 and VxH modes 766- intel/brw: Extract format enum in EU validation code 767- intel/brw: Add validation for some Xe2 register regioning restrictions 768- intel/brw: Add some tests for new Xe2 register regioning restrictions 769- intel/brw: Add SHADER_OPCODE_READ_FROM_CHANNEL and LIVE_CHANNEL 770- intel/brw: Disallow cmod in some cases of ARF scalar as destination 771- intel/brw: Use variable instead of manually count the passes 772- intel/brw: Rename brw_inst.h to brw_eu_inst.h 773- intel/brw: Rename brw_inst to brw_eu_inst 774- intel/brw: Rename brw_compact_inst to brw_eu_compact_inst 775- intel/brw: Rename brw_inst_bits/set_bits to brw_eu_inst_bits/set_bits 776- intel/brw: Rename brw_inst_* helpers to brw_eu_inst_* 777- intel/brw: Rename brw_compact_inst_* helpers to brw_eu_compact_inst_* 778- intel/brw: Gather brw_reg related implementations in brw_reg.cpp 779- intel/brw: Add missing call to invalidate analysis 780- intel/brw: Move two NIR passes to brw_nir.c 781- gallium/meson: Ensure all needed sym_config are set. 782- intel/brw: Remove 'fs' prefix from passes filenames 783- intel/brw: Remove 'fs' prefix from passes and related functions 784- intel/brw: Add missing bits in 3-src SWSB encoding for Xe2+ 785- intel/brw/xe2+: Do not use $.dst or $.src SWSB annotations in SENDs 786- intel/compiler: Use INFINITY spill cost to represent no_spill 787- util: Add operator new[] to linear context helper declarations 788- intel/compiler: Use linear allocator for ACP trees in copy-prop 789- intel/brw: Remove uses of VLAs 790- intel/elk: Add ELK_MAX_MRF_ALL for static allocating arrays 791- intel/elk: Remove uses of VLAs 792- intel/elk: Fix typo in assertion 793- util/ra: Move less used data out of ra_node 794- util/ra: Don't store a pointer to graph per ra_node 795- util/ra: Bump the initial size of adjacency lists 796- util/ra: Don't store a pointer to a ra_regs per ra_reg 797- intel/brw: Rename brw_fs_validate to brw_validate 798- docs: Update syntax on Performance tips page 799- intel/brw: Rename brw_fs_generator.cpp to brw_generator.cpp 800- intel/brw: Add brw_generator.h header 801- intel/brw: Rename fs_generator to brw_generator 802- intel/brw: Add missing cases to flags_written() 803- intel/brw: Remove extra wrapping around fs_visitor in tests 804- intel/brw: Rename brw_fs_builder.h to brw_builder.h 805- intel/brw: Rename fs_builder to brw_builder 806- intel/brw: Stop using namespace for brw_builder 807- intel/brw: Move a few builder helpers to brw_builder.h/cpp 808- intel/brw: Move shuffle_from_32bit_read implementation to brw_builder 809- intel/brw: Apply conventions to lower_src_modifiers helper 810- intel/brw: Rename brw_fs_reg_allocate.cpp to brw_reg_allocate.cpp 811- intel/brw: Remove 'fs' prefix from reg alloc code 812- intel/brw: Rely on existing helper for dispatch width of geometry stages 813- intel/elk: Fix wrong destination to memset 814- intel/brw: Use brw prefix for some schedule instructions identifiers 815- intel/brw: Use brw prefix instead of namespace in dynamic_msaa_flags() 816- intel/brw: Remove unused enum 817- intel/executor: Fix typo when copying result into Lua table 818- intel/tools: Use idep_libintel_common in meson 819- intel/tools: Add helpers for decoder_init/disasm 820- intel/tools: Merge libaub into libintel_tools 821- intel: Add meson option -Dintel-elk 822- intel/brw: Add scoreboard support for scalar register 823- intel/brw: Plumb through generator whether SEND is gather variant 824- intel/brw: Add SHADER_OPCODE_SEND_GATHER 825- intel/brw: Add lowering for SHADER_OPCODE_SEND_GATHER 826- intel/brw: Use SHADER_OPCODE_SEND_GATHER in Xe3 827- intel/brw: Fallback to SEND from SEND_GATHER if possible 828 829Caleb Callaway (2): 830 831- docs: Intel GPU performance tips 832- docs: clarify ASPM performance tips 833 834Casey Bowman (1): 835 836- vulkan/screenshot-layer: Add region command option 837 838Caterina Shablia (9): 839 840- pan/bi: fix a typo 841- pan/va: fix WMASK packing 842- pan/bi: handle read_invocation 843- pan/bi: handle ballot, ballot_relaxed and as_uniform 844- pan/bi: lower some subgroup intrinsics 845- pan/bi: lower the rest of subgroup ops using nir_lower_subgroups 846- pan/bi: add a MEMORY_BARRIER pseudo-instruction 847- pan/bi: handle barriers with SUBGROUP scope 848- panvk: enable subgroupSizeControl 849 850Chen, Phoebe (1): 851 852- amd/vpelib: Refactor YUV format check 853 854Chia-I Wu (69): 855 856- panvk: ensure res table is restored after meta 857- panvk: add memory mmap/munmap helpers 858- panvk: do not leak mapped memory 859- panvk: update CI expectations 860- panvk: add get_subqueue_stages 861- panvk: rework collect_cache_flush_info 862- panvk: rework collect_cs_deps 863- panvk: always skip frag->tiler subqueue wait 864- panvk: skip frag subqueue self-wait within a render pass 865- panvk: skip tiler subqueue self-wait within a render pass 866- panvk: improve should_split_render_pass 867- panvk: fix a missing cache invalidation 868- panvk: update expectations for G610 869- vulkan: include host write in expanded dst access flags 870- panvk: add normalize_dependency 871- panvk: improve VK_QUEUE_FAMILY_EXTERNAL support 872- panvk: add support for VK_EXT_queue_family_foreign 873- panvk: fix base_workgroup_id sysval 874- ci: update the comment on MESA_VK_ABORT_ON_DEVICE_LOSS 875- panvk: report queue lost timely when PANVK_DEBUG=sync 876- panvk: implement check_status on v10+ 877- panvk: no need to map IB internally on valhall 878- panvk: clang-format issue_fragment_jobs 879- panvk: fix frag_completed for layered rendering 880- panvk: minor clean up to prepare_blend 881- panvk: fix dirty check for prepare_blend 882- panvk: expand top-of-pipe and bottom-of-pipe 883- panvk: use u_foreach_bit to loop over mask bits 884- panvk: fix vs image support 885- panvk: add panvk_queue_submit_init 886- panvk: add panvk_queue_submit_init_storage 887- panvk: add panvk_queue_submit_init_waits 888- panvk: add panvk_queue_submit_init_cmdbufs 889- panvk: add panvk_queue_submit_init_signals 890- panvk: add panvk_queue_submit_ioctl 891- panvk: add panvk_queue_submit_process_signals 892- panvk: add panvk_queue_submit_process_debug 893- panvk: clean up panvk_queue_submit 894- panvk: move pandecode_next_frame a bit earlier 895- panvk/csf: fix SIMULTANEOUS_USE gpu faults 896- panvk/csf: fix subqueue ctx memory pool 897- panvk: use cs_tracing_ctx::enabled for exception handler 898- panvk: add u_trace_context to panvk_device 899- panvk: define cmdbuf begin/end tracepoints 900- panvk/csf: add CS_REG_SCRATCH_COUNT 901- panvk/csf: add u_trace to panvk_cmd_buffer 902- panvk/csf: add vk_sync to panvk_queue 903- panvk/csf: flush and process trace events for one-time cmdbufs 904- panvk/csf: flush and process trace events for all cmdbufs 905- panvk: improve C++ compat for perfetto 906- panvk: add u_trace perfetto support 907- panvk: silence a perfetto init warning 908- vulkan: add vk_device_get_timestamp 909- vulkan: add common GetPhysicalDeviceCalibrateableTimeDomainsKHR 910- vulkan: add common GetCalibratedTimestampsKHR 911- anv: use common calibrated timestamp support partially 912- hasvk: use common calibrated timestamp support 913- radv: use common calibrated timestamp support 914- tu: use common calibrated timestamp support 915- nvk: use common calibrated timestamp support 916- hk: remove calibrated timestamp support 917- panvk: no need to zero availability on query create 918- panvk: no need to check query count on query create 919- panvk: no need to zero results on query reset 920- panvk/csf: no need to sb wait on query begin 921- panvk/csf: no need to sb wait on query end 922- panvk/csf: no need to sb wait on query copy 923- panvk/csf: no need to flush caches after query copy 924- panvk/csf: add a comment on query synchronization 925 926Christian Gmeiner (20): 927 928- broadcom/common: Make v3d_device_info.h usable for C++ 929- v3d: Move v3d_ioctl(..) to src/broadcom/common 930- v3dv: Switch to v3d_ioctl(..) 931- v3d: Move v3d_X(..) to src/broadcom/common 932- v3dv: Switch to v3d_X(..) 933- broadcom: Add perfcount library 934- v3d: Switch to use libbroadcom_perfcntrs 935- v3dv: Switch to use libbroadcom_perfcntr 936- etnaviv: blt: Add DBG(..) why blt usage was not possible 937- etnaviv: rs: Add DBG(..) why blt usage was not possible 938- v3d: Sync v3d_drm.h with drm-misc-next 939- broadcom: Add perfetto data source 940- pps: Add support for v3d ds 941- perfetto: Add v3d data sources to system.cfg 942- perfetto: Add v3d data sources to gpu.cfg 943- docs: Update perfetto with the latest status 944- etnaviv: isa: Support src2 for texld 945- etnaviv: isa: Support src2 for texldb and texldl 946- egl/meson: Specify which symbols to export 947- v3dv: Add some CPU tracepoints 948 949Christopher Michael (5): 950 951- v3d: Add check to see if v3d supports cpu_queue 952- v3d: Add check to see if v3d supports multisync 953- v3d: Add support for timestamp queries 954- v3d: Add support for time elapsed queries 955- v3d: Add support for PIPE_QUERY_TIMESTAMP_DISJOINT 956 957Collabora's Gfx CI Team (5): 958 959- Uprev Piglit to eebe1b555f51dbb702f696d08ad5ae8153bcdcdd 960- Uprev Piglit to d04d6fff00849a2a8e29ef3251c6ca04a2f68dc7 961- Uprev Piglit to 468221c722481c470e6a23760b914c33143c2af6 962- Uprev Piglit to 4c0fd15fd956ec70c5509bedee219d602b334464 963- Uprev Piglit to 631b72944f56e688f56a08d26c8a9f3988801a08 964 965Connor Abbott (55): 966 967- vulkan/runtime: Add driver callbacks for BVH building 968- vulkan/runtime,radv: Add shared BVH building framework 969- vulkan/runtime,radv: Add shared BVH building framework 970- ir3: Fix reload_live_out() in shared RA 971- tu: Add Vulkan 1.4 features and properties 972- tu: Expose Vulkan 1.4 on a7xx 973- tu: Move queue-related code to a new file 974- tu: Refactor the submit path 975- tu/kgsl: Make wait_timestamp_safe() return VkResult 976- tu/knl: Move u_trace fence handling to generic code 977- tu: Rename bo_list to submit_bo_list 978- util/dynarray: Add macro for appending an array 979- tu: Make userspace RD dump generic 980- freedreno/fdl: Make tiled r8g8 images have 4k alignment 981- tu: Re-enable tiled non-ubwc R8G8 images 982- freedreno/fdl: Fix 3d mipmapping height alignment 983- freedreno/fdl, tu: Make mutable part of the image layout 984- freedreno/fdl: Don't enable r8g8 special case for mutable images 985- freedreno/fdl, tu: Allow swaps with mutable tiled images 986- tu: Allow UBWC with images with swapped formats. 987- vk/bvh: Fix clang build error with turnip 988- ir3: Allow collect sources to be undef 989- ir3: Support assembling/disassembling ray_intersection and resbase 990- ir3: Plumb through two-dimensional UAV loads 991- ir3: Plumb through ray_intersection intrinsic 992- tu: Implement cmd_fill_buffer_addr internal function 993- tu: Implement buffer_write_cp 994- freedreno: CP_SCRATCH_WRITE exists on a7xx too 995- freedreno: Add new a7xx CP_REG_RMW and CP_REG_TO_SCRATCH fields 996- freedreno/a7xx: Document partial workgroup register 997- tu: Stop emitting HLSQ_CS_KERNEL_GROUP_* 998- tu/a7xx: Emit HLSQ_CS_LAST_LOCAL_SIZE dynamically 999- tu: Implement unaligned dispatches 1000- tu: Add common define for maxTexelBufferElements 1001- tu: Create meta device 1002- freedreno: Introduce ray tracing features 1003- tu/kgsl: Bump uapi header 1004- tu: Plumb through raytracing fuse 1005- tu: Move fd_dev_info() before name generation 1006- tu: Display when raytracing is disabled in device string 1007- tu: Support VK_KHR_acceleration_structure 1008- tu: Support VK_KHR_ray_query 1009- tu: Expose VK_KHR_ray_tracing_maintenance1 1010- tu, ir3: Implement a750 RT workaround 1011- ir3: Use nir_split_struct_vars for temporaries 1012- vk/bvh: Add default stubs for unsupported entrypoints 1013- anv: Delete acceleration structure stubs 1014- radv: Delete acceleration structure stubs 1015- tu: Use image view format for sysmem resolves 1016- tu: Handle non-identity GMEM swaps when resolving 1017- tu: Handle non-identity GMEM swaps for input attachments 1018- tu, freedreno: Write PC_DGEN_SU_CONSERVATIVE_RAS_CNTL 1019- tu: Stop setting binning fields on a7xx 1020- tu: Support VK_EXT_conservative_rasterization on a7xx 1021- tu: Add missing assignment to shared_viewport 1022 1023Constantine Shablia (23): 1024 1025- panvk: move samplerAnisotropy in the order it appears in struct definition 1026- panvk: enable shaderInt64 1027- panvk: elaborate the comment on the maxMemoryAllocationCount limit 1028- panvk: adjust maxSamplerAllocationCount limit 1029- nir: introduce instance_index system value 1030- nir: lower INSTANCE_{ID,INDEX} to an offset load_instance_{index,id} respectively 1031- Revert "nir: lower INSTANCE_{ID,INDEX} to an offset load_instance_{index,id} respectively" 1032- Revert "nir: introduce instance_index system value" 1033- panvk: replace vkGetBufferMemoryRequirements2 with vkGetDeviceBufferMemoryRequirements 1034- panvk: never prefer or require dedicated allocation for buffers 1035- panvk: never require dedicated allocation for images 1036- panvk: add panvk_image_init helper 1037- panvk: implement vkGetDeviceImageMemoryRequirements 1038- panvk: enable shaderInt8, VK_KHR_8bit_storage and VK_KHR_shader_float16_int8 1039- pan/util: sort files in meson.build 1040- panvk: order KHR extension enables alphabetically 1041- panvk/csf: use gfx_state_set_dirty instead of touching state directly 1042- pan,nir: introduce load_attribute_pan 1043- pan/bi: handle load_attribute_pan 1044- panvk: Fix base_{instance,vertex} handling 1045- panvk: lower drawid to zero 1046- panvk: enable shaderDrawParameters 1047- panvk: enable drawIndirectFirstInstance 1048 1049Corentin Noël (6): 1050 1051- virgl: Propagate the GL_MAX_stage_SHADER_STORAGE_BLOCKS for each stage 1052- virgl: Simply loop over the resources to figure-out if it is already added 1053- virgl: Update virgl_hw.h from virglrenderer 1054- virgl: Use MAX_SAMPLERS instead of MAX_SHADER_SAMPLER_VIEWS 1055- virgl/ci: Remove screen size arguments 1056- virgl/ci: Re-enable virgl-traces 1057 1058Daniel Schürmann (49): 1059 1060- aco/ra: set Pseudo_instruction::scratch_sgpr to SCC if it doesn't need to be preserved 1061- aco/ra: use bitset for sgpr_operands_alias_defs 1062- aco/ra: explicitly assign scratch SGPR for linear phis 1063- aco: remove Pseudo_instruction::tmp_in_scc 1064- aco/insert_NOPs: implement vector-based RegCounterMap as replacement for VGPRCounterMap 1065- aco/insert_NOPs: use RegCounterMap as replacement for the CounterMap implementation 1066- aco/insert_NOPs: add early exit to handle_valu_partial_forwarding_hazard_instr 1067- aco/print_asm: allow for empty blocks with arbitrary offsets 1068- aco/assembler: constify assembly functions 1069- aco/assembler: Actually insert s_inst_prefetch instructions when aligning blocks for loops 1070- aco/assembler: change ctx.loop_header to uint32_t instead of Block* 1071- aco/assembler: chain branches instead of emitting long jumps 1072- aco: remove definition from SOPP branch instructions 1073- aco: remove definition from Pseudo branch instructions 1074- aco/assembler: Don't emit target basic block index when chaining branches 1075- aco/print_ir: don't print disconnected empty blocks 1076- aco/optimizer_postRA: set branch()->never_taken if exec is constant non-zero 1077- aco: move try_optimize_branching_sequence() to postRA optimizations 1078- aco/jump_threading: remove branch sequence optimization 1079- aco: move branch lowering optimization into separate file 'aco_lower_branches.cpp' 1080- aco/lower_branches: remove edges between blocks if there is no direct branch 1081- ac/lower_ngg: Fix collecting buffer offsets from 4 lanes on gfx12 1082- ac/lower_ngg: move break blocks after loop in streamout code generation for gfx12/ACO 1083- ac/lower_ngg: move readlane into break blocks in streamout code generation for gfx12/ACO 1084- nir/divergence: change nir_has_divergent_loop() to return true only for divergent breaks 1085- aco/jump_threading: don't remove loop preheaders 1086- aco/assembler: Find loop exits using the successor's loop nest depth 1087- aco: consider s_cbranch_exec* instructions in needs_exec_mask() 1088- aco/lower_branches: do eliminate_useless_exec_writes_in_block() during branch lowering. 1089- aco/lower_branches: implement try_remove_simple_block() in lower_branches() 1090- aco: move try_merge_break_with_continue() to lower_branches() 1091- aco/lower_branches: allow for non-fallthrough loop exits in try_merge_break_with_continue() 1092- aco: delete aco_jump_threading.cpp 1093- aco/lower_branches: stitch linear blocks if there is exactly one successor with one predecessor 1094- nir/from_ssa: only consider divergence if requested 1095- Revert "nir: add nir_clear_divergence_info, use it in nir_opt_varyings" 1096- aco/insert_NOPs: refactor VALUReadSGPRHazard detection 1097- aco/insert_NOPs: implement VALU -> VALU case for VALUReadSGPRHazard on GFX12 1098- nir/loop_analyze: only iterate loop header phis in compute_induction_information() 1099- nir/loop_analyze: remove nir_loop_variable::in_if_branch and nir_loop_variable::in_nested_loop 1100- nir/loop_analyze: remove nir_loop_variable::in_loop 1101- nir/loop_analyze: directly record induction variables into nir_loop_info 1102- nir/loop_analyze: don't initialize nir_loop_variable separately 1103- nir/loop_analyze: replace nir_loop_variable array with hash table 1104- nir/loop_analyze: insert only induction vars into hash map 1105- nir/loop_analyze: ignore terminating induction variable in guess_loop_limit() 1106- nir/loop_analyze: re-use the same nir_loop_variable struct before and after the increment 1107- nir/loop_analyze: store nir_loop_induction_variable hash table in loop_info 1108- nir/loop_analyze: stack-allocate loop_info_state 1109 1110Daniel Stone (22): 1111 1112- ci: Don't run Meson tests in critical-path jobs 1113- ci: Slash ASan and UBSan build coverage 1114- ci: Give much more time to ASan and UBSan jobs 1115- ci: Let rootfs builds run for 2 hours (!) 1116- pipe_loader: Fix pipe_i915 with the dynamic loader 1117- ci: Disable Werror on wrapped subprojects 1118- ci: Remove obsolete compiler-wrapper 1119- ci: Move build containers above test containers 1120- ci/fedora: Install which into build image 1121- ci: Define LLVM_VERSION as a container property 1122- ci: Require LLVM_VERSION to be set explicitly 1123- ci/debian: Upgrade Debian images to LLVM 19 1124- ci: Fix dependency on lint job 1125- ci: Fix kernel section nesting 1126- ci: Move dEQP message into section 1127- ci: Pass build targets to dEQP CMake 1128- ci: Don't build Vulkan for GL dEQP 1129- ci: Trim down VVL external builds 1130- ci: Capture Ninja log 1131- ci: Only build Perfetto in build-test jobs 1132- ci: Only build what we use for testing jobs 1133- ci: Move r300/nine/nvk builds out of critical path 1134 1135Danylo Piliaiev (31): 1136 1137- ir3/parser: Print the line where parsing error occurred 1138- nir/nir_opt_offsets: Do not fold load/store with const offset > max 1139- freedreno/registers: Define Fragment Shading Rate registers 1140- ir3,tu: Add support for Fragment Shading Rate and plumb it into Turnip 1141- tu/a7xx: Implement VK_KHR_fragment_shading_rate 1142- ir3/parser: Add fullnop and fullsync sections for debugging 1143- tu: Enable UBWC for 3D images without mipmaps 1144- freedreno/fdl: Pass fd_dev_info to fdl6_layout 1145- tu,freedreno: Enable linear mipmap tail for UBWC images 1146- tu: Disable fragmentShadingRateWithShaderSampleMask due to issues 1147- tu,ir3: Add workaround for reading shading rate on A7XX gen1,gen2 1148- tu: Handle cmdbuf and rp_blit flags of TU_DEBUG_STALE_REGS_FLAGS 1149- tu/perfetto: Always emit submission event and time it 1150- tu/perfetto: Add app and engine names to the command buffer tracepoint 1151- ir3: Make allocation of consts more generic and order independent 1152- ir3: Use generic consts alloc for driver params 1153- tu,ir3: Make push consts be able to start from higher than c0.x offsets 1154- ir3: Use generic const alloc for everything and call it once 1155- tu: Allocate consts for driver params as early as possible 1156- tu: Do not re-calculate static blend LRZ state 1157- freedreno/regs: Set correct shr for GRAS_LRZ_BUFFER_PITCH.ARRAY_PITCH 1158- tu: Fix LRZ for arrayed depth 1159- tu: Handle 8x MSAA for LRZ 1160- freedreno,tu: Unify LRZ layout calculations 1161- tu: Track at which draw call LRZ is disabled 1162- tu: Do not disable LRZ for whole RP if it is disabled in RP 1163- ir3: Consider const alloc alignment in free space size calcs 1164- tu: Fix stale A7XX_GRAS_LRZ_CNTL2 in 3d blits or !valid lrz case 1165- tu/a7xx: Always have depth/stencil in corresponding resolve groups 1166- tu: Get correct src view when storing gmem attachment 1167- tu: Handle mismatched mutability when resolving from GMEM 1168 1169Dave Airlie (9): 1170 1171- nir/functions: force inlining for barriers. 1172- v3dv: report correct error on failure to probe 1173- venus: handle device probing properly. 1174- vulkan: update to 302 headers for av1 encode 1175- lavapipe: fix beta build due to changes in AMDX ext 1176- radv/video: set max slice counts to 1 for h264/5 encode 1177- anv: add default av1 tables from media-driver 1178- genxml: add av1 fields 1179- anv: add initial support for AV1 decoding 1180 1181David (Ming Qiang) Wu (3): 1182 1183- frontends/va: adding PIPE_FORMAT_P012 1184- frontends/va: add PIPE_VIDEO_PROFILE_AV1_PROFILE2 1185- radeonsi/vcn: support 12bit YUV420 AV1 decoding 1186 1187David Heidelberg (14): 1188 1189- util: Drop 3Dnow optimisation leftovers 1190- util: Remove MMX/MMXext detection code 1191- util: Drop ancient Intel CPU detection 1192- util: drop XOP detection code 1193- llvmpipe: align with u_cpu_detect struct changes 1194- compiler/rust: drop duplicated bindgen check 1195- ci/freedreno: update Adreno 306 expectations 1196- ci/freedreno: increase Adreno 618 timeout to 1h 1197- docs: remove deprecated component list and licenses 1198- docs: Clarify project name and include Mesa3D 1199- docs: move license(s) to licenses directory 1200- c11: use SPDX-License-Identifier header 1201- licenses: add missing licenses 1202- drm-uapi: update licenses statement 1203 1204David Rosca (148): 1205 1206- radeonsi/vcn: Fix coding AV1 render size 1207- frontends/va: Add minus_1 to AV1 render_width/height 1208- gallium: Add PIPE_VIDEO_CAP_SKIP_CLEAR_SURFACE 1209- frontends/va: Support skip clear on surface creation 1210- frontends/vdpau: Support skip clear on surface creation 1211- radeonsi: Support PIPE_VIDEO_CAP_SKIP_CLEAR_SURFACE 1212- radeonsi/vcn: Stop clearing decode internal buffers 1213- radv/video: Fix H264 slice control 1214- radv/video: Fix HEVC slice control 1215- radv/video: Report correct encodeInputPictureGranularity 1216- radv/video: Avoid selecting rc layer over maximum 1217- radv/video: Use 64x16 alignment for HEVC encode 1218- radv/video: Override pic_init_qp_minus26 in PPS 1219- radeonsi/vcn: Use correct frame context buffer for preencode on VCN5 1220- radeonsi: Check all supported formats in si_vid_is_target_buffer_supported 1221- frontends/va: Create surfaces with correct fourcc for RT format 1222- frontends/va: Stop reallocating to prefered format in EndPicture 1223- frontends/va: Stop reallocating from progressive to interlaced in EndPicture 1224- frontends/va: Stop reallocating buffers for protected playback 1225- frontends/va: Stop reallocating according to JPEG sampling factor 1226- frontends/va: Check if target buffer is supported in EndPicture 1227- frontends/va: Stop reallocating buffers in EndPicture 1228- frontends/va: Use compositor blit with different number of planes 1229- frontends/va: Only use interlaced surfaces when progressive is not supported 1230- pipe: Remove video update_decoder_target 1231- radeonsi/vpe: Set correct surface swizzle mode 1232- radeonsi/vpe: Don't allow DCC surfaces 1233- frontends/va: Return correct pixel formats in surface attributes query 1234- frontends/va: Change default fourcc for RGB 10bit to X2R10G10B10 1235- gallium/vl: Implement rendering to 3-plane YUV formats 1236- gallium/vl: Don't support planar RGB as video format 1237- frontends/va: Enable 3-plane YUV formats as postproc output 1238- radeonsi/vcn: Support tiling for JPEG decode 1239- radv/video: Fix IB signature checksum 1240- radv/video: Always use setup reference slot when valid 1241- ac/surface: Add RADEON_SURF_VIDEO_REFERENCE 1242- radeonsi: Support PIPE_BIND_VIDEO_DECODE/ENCODE_DPB 1243- radeonsi/vcn: Create decode DPB surfaces with PIPE_BIND_VIDEO_DECODE_DPB 1244- radeonsi/vcn: Create encode DPB surfaces with PIPE_BIND_VIDEO_ENCODE_DPB 1245- frontends/va: Add support for VA_SURFACE_ATTRIB_MEM_TYPE_DRM_PRIME_3 1246- frontends/va: Store picture type for buffers in encode DPB 1247- radeonsi/vcn: Don't allow encoding H264 B-frame references 1248- frontends/va: Move mjpeg sampling_factor to pipe_mjpeg_picture_desc 1249- radeonsi/vcn: Remove code handling buffer_get_virtual_address failure 1250- radeonsi/vcn: Unmap bitstream buffer in radeon_dec_destroy 1251- radeonsi/vcn: Gracefully handle decode errors and report to frontend 1252- radeonsi/vcn: Make sure JPEG target buffer format matches sampling factor 1253- radeonsi/vcn: Cleanup JPEG supported formats 1254- radeonsi/vpe: Silence expected errors with unsupported output format 1255- gallium/vl: Add plane order for Y8_400 format 1256- gallium/vl: Fix plane order for IYUV format 1257- frontends/va: Stop converting formats in Put/GetImage 1258- radeonsi: Update minimum supported encode size for VCN5 1259- radeonsi/vcn: Align bitstream buffer to 128 when resizing 1260- radeonsi/uvd: Align bitstream buffer to 128 when resizing 1261- radeonsi/vcn: Enable write combine for decode 1262- radeonsi/vcn: Don't keep last fence 1263- radeonsi/vcn: Use local variable for destory fence 1264- pipe: Remove PIPE_DEFAULT_DECODER_FEEDBACK_TIMEOUT_NS 1265- frontends/va: Get AV1 decode subsampling_x/y 1266- radeonsi/vcn: Return error when decoding 12bit VP9 and 4:2:2/4:4:4 AV1 1267- frontends/va: Fix decoding VC1 interlaced video 1268- frontends/va: Don't allow Render/EndPicture without BeginPicture 1269- frontends/va: Don't allow EndPicture without calling driver begin_frame 1270- ac/parse_ib: Parse VCN IB_COMMON_OP_WRITEMEMORY 1271- radv/amdgpu: Set VCN version for ac_parse_ib 1272- frontends/va: Fix deinterlace filter 1273- radeonsi/vcn: Change required FW version for rc_per_pic_ex on VCN3 1274- radv/video: Fix DPB tier2 surface params 1275- radv/video: Use correct array index for decode target and DPB images 1276- radv/video: Remove dt_field_mode handling code 1277- radv: Fix sampling from image layers of video decode target 1278- ac/surface: Don't force linear for VIDEO_REFERENCE with emulated image opcodes 1279- frontends/va: Get buffer feedback with locked mutex in MapBuffer 1280- radeonsi/vcn: Use compute only context 1281- gallium/vl: Fix unbinding sampler views 1282- gallium/vl: Create sampler state also when gfx is not supported 1283- gallium/vl: Add rgba compute shader 1284- gallium/vl: Add param to create compute only vl_compositor 1285- gallium: Add param to create compute only multimedia context 1286- frontends/va: Use compute only context if driver prefers compute 1287- radeonsi/vcn: Fix crash when failing to allocate internal buffers 1288- frontends/va: Only report surface alignment when non-zero 1289- frontends/va: Allow creating DRM PRIME surfaces without surface descriptor 1290- frontends/va: Set csc matrix in PutSurface 1291- gallium/vl: Fix creating buffers with auxiliary planes 1292- radeonsi: Add radeon_bitstream and use it in radeon_vcn_enc 1293- radeonsi/vce: Remove support for FW 50 and older 1294- radeonsi/vce: Set more header params 1295- radeonsi/vce: Move dual pipe context to offset 0 of CPB 1296- radeonsi/vce: Use app DPB management 1297- radeonsi/vce: Support slice encoding 1298- radeonsi/vce: Support VBAQ 1299- radeonsi/vce: Support quality presets 1300- radeonsi/vce: Support min/max QP and max frame size 1301- radeonsi/vce: Support intra refresh 1302- radeonsi/vce: Support raw packed headers 1303- radeonsi/vce: Set input pic swizzle mode on GFX9 1304- radeonsi/vce: Cleanup 1305- radeonsi/uvd: Stop clearing decode internal buffers 1306- radeonsi/uvd: Optimize bitstream buffer resizing 1307- radeonsi/uvd: Set decode target swizzle mode on GFX9 1308- radeonsi/uvd_enc: Rework DPB allocation 1309- radeonsi/uvd_enc: Use app DPB management 1310- radeonsi/uvd_enc: Consider input surface size for padding 1311- radeonsi/uvd_enc: Support Pre-Encode 1312- radeonsi/uvd_enc: Support VBAQ 1313- radeonsi/uvd_enc: Support quality presets 1314- radeonsi/uvd_enc: Support slice encoding 1315- radeonsi/uvd_enc: Support intra refresh 1316- radeonsi/uvd_enc: Support temporal layer rate control 1317- radeonsi/uvd_enc: Support min/max QP and max frame size 1318- radeonsi/uvd_enc: Support dynamic rate control changes 1319- radeonsi/uvd_enc: Support raw packed headers 1320- radeonsi/uvd_enc: Set input pic swizzle mode on GFX9 1321- radeonsi: Enable implemented VCE/UVD encode features 1322- gallium/vl: Fix sampler view components for Y8_400 format 1323- gallium/vl: Add vl compositor layer mirror 1324- gallium/vl: Clear remaining planes in YUV conversion 1325- gallium/vl: Use matrix for scale and crop in cs compositor 1326- gallium/vl: Implement rotation and mirror in cs compositor 1327- frontends/va: Simplify format check in PutSurface 1328- frontends/va: Disable color conversion for luma-only source formats 1329- frontends/va: Stop using util_compute_blit 1330- frontends/va: Refactor vlVaPostProcCompositor to be usable outside processing 1331- frontends/va: Support rotation and mirror for processing 1332- frontends/va: Implement format conversions in PutImage/GetImage 1333- gallium/auxiliary: Remove util_compute_blit 1334- radeonsi: Fix reporting support for AV1 Profile2 1335- radeonsi/vcn: Fix AV1 coded size for VCN 5.0 1336- radeonsi: Report surface alignment for AV1 encode 1337- gallium/vl: Add compute shader deinterlace filter 1338- frontends/va: Stop using extra context for deinterlacing 1339- frontends/va: Implement QuerySurfaceStatus as SyncSurface with 0 timeout 1340- frontends/va: Don't flush before resource_get_handle 1341- frontends/va: Remove vlVaBuffer derived_image_buffer 1342- frontends/va: Add surface pipe_fence for vl_compositor rendering 1343- gallium/vl: Don't flush in vl_compositor yuv_deint and rgb_to_yuv 1344- frontends/va: Add context mutex 1345- frontends/va: Unlock driver mutex for SyncSurface/Buffer fence wait 1346- frontends/va: Fix decoding VC1 streams with multiple slices 1347- ac/vcn_dec: Fix AV1 film grain on VCN5 1348- radeonsi/video: Avoid stream handle duplicates in PID namespace 1349- frontends/vdpau: Set H264 chroma_format_idc 1350- radeonsi/vcn: Set correct chroma format for H264 decode 1351- radeonsi/uvd: Set correct chroma format for H264 decode 1352- radv/video: Fix setting balanced preset for HEVC encode with SAO enabled 1353- radv/video: Move IB header from begin/end to encode_video 1354 1355David Tobolik (2): 1356 1357- rusticl/style: use Arc::clone instead of .clone() 1358- rusticl/style: add util for conversion with err 1359 1360Deborah Brouwer (36): 1361 1362- freedreno/ci: add prefix for a630-vk-asan tests 1363- ci: Remove duplicate slash before $RESULTS_DIR 1364- ci/b2c: update RESULTS_DIR for .b2c-test jobs 1365- ci: add a tool to summarize a failed pipeline 1366- ci/pipeline_message: add unit tests for tool 1367- ci: move pipeline_summary tool to .marge/hooks 1368- ci: debian/x86_64_pyutils remove redundant rules 1369- ci: python-test rename artifacts 1370- ci: yaml-toml-shell-test: use pyutils container 1371- ci: separate python tests and artifacts 1372- ci: post gantt: use logging instead of print 1373- ci: add some static typing to the gantt scripts 1374- ci: make the gantt scripts available as modules 1375- ci: post gantt: add --marge-user-id option 1376- ci: post gantt: add --project-id option 1377- ci: post gantt: add pipeline-id to gantt filename 1378- ci: post gantt: ignore pipeline_summary message 1379- ci: gantt chart: include in-progress jobs 1380- ci: add --ci-timeout option for gantt scripts 1381- ci: add pytests for the gantt chart scripts 1382- ci: update token retrieval method for gantt charts 1383- ci: collapse yamllint and shellcheck sections 1384- ci: run-pytest.sh: allow script to run locally 1385- ci: add .flake8 linting to ci scripts and tests 1386- ci: update_traces_checksum: fix E501 line too long 1387- ci: update the pyutils container 1388- ci: stop using a venv for run-pytest.sh 1389- ci: set python version 3.11 for run-pytest.sh 1390- ci: pipeline_message: catch module loading errors 1391- ci: pipeline_message: improve job list formatting 1392- ci: pipeline_message: add test to parse error logs 1393- ci: pipeline_message: ignore \`error_type` errors 1394- ci: pipeline_message: ignore harmless build logs 1395- ci: pipeline_message: ignore \`generated` errors 1396- ci: pipeline_message: parse \`fatal` messages 1397- ci: pipeline_message: reset empty errors 1398 1399Derek Foreman (3): 1400 1401- vulkan/wsi/wayland: Fix time calculation 1402- vulkan/wsi/wayland: Avoid spurious discard event at startup 1403- vulkan/wsi/wayland: Move timing calculations to the swapchain 1404 1405Detlev Casanova (3): 1406 1407- ci/fluster/lava: Add fluster in LAVA rootfs 1408- ci/fluster: Add radeonsi-raven-vaapi-fluster jobs 1409- ci/deqp-runner: uprev from 0.20.2 to 0.20.3 1410 1411Dylan Baker (25): 1412 1413- VERSION: bump to 25.0 1414- docs: reset new_features.txt 1415- docs/release-calendar: update one more time for pushed back release 1416- docs: add release notes for 24.3.0 1417- docs/relnotes/24.3.0: Add SHA sums 1418- docs/release-calendar: remove 24.3 RC dates 1419- docs: Add calendar entries for 24.3 release. 1420- anv: advertise Vulkan 1.4 1421- anv: bump max number of push constants to 256 1422- anv: Add new Vulkan 1.4 features and properties 1423- anv: bump conformance version to 1.4 1424- maintainer-scripts: Bump Vulkan release version to 1.4 1425- docs: add release notes for 24.3.1 1426- docs: Add SHA sums for 24.3.1 1427- docs: update calendar for 24.3.1 1428- clc: Tell clang to track imported dependencies 1429- docs: add release notes for 24.3.2 1430- docs: Update checksums for 24.3.2 1431- docs: update calendar for 24.3.2 1432- docs/release-calendar: Move next release to January 2nd 1433- intel/tests: Fix coverity warning about possibly leaked memory 1434- intel/tests: Fix missing assignment of error condition 1435- docs: add release notes for 24.3.3 1436- docs: Add SHA sums to 24.3.3 release notes 1437- docs: update calendar for 24.3.3 1438 1439Eric Engestrom (139): 1440 1441- meson: bump spirv-tools version needed to v2022.1 1442- radeonsi/ci: add more flakes seen recently 1443- radv/ci: add more flakes seen recently 1444- broadcom/ci: add more flakes seen recently 1445- freedreno/ci: add more flakes seen recently 1446- ci: upgrade the fedora image from 38 to 41 1447- ci/build: drop "verify after bump to F39" as that did not help 1448- ci/build: add workaround for incorrect maybe-uninitialized error 1449- ci: move error handling functions at the end 1450- ci: use quiet alias for commands 1451- ci: make error handling quieter 1452- broadcom/ci: add flakes seen recently 1453- freedreno/ci: add flakes seen recently 1454- nvk+zink/ci: add flakes seen recently 1455- radv+zink/ci: add flakes seen recently 1456- ci: raise priority of release manager pipelines 1457- ci: reduce priority of nightly pipeline jobs from 50 to 45 1458- meson: move openmp block out of the middle of the x11 deps block 1459- meson: define only once the versions of the x11 deps 1460- radv/ci: document flakes seen recently 1461- broadcom/ci: document flakes seen recently 1462- nvk/ci: document flakes seen recently 1463- freedreno/ci: document flakes seen recently 1464- docs: update calendar for 24.2.7 1465- docs: add release notes for 24.2.7 1466- docs: add sha sum for 24.2.7 1467- turnip/ci: document regression 1468- ci/crosvm: remove noise inside deqp-runner output 1469- v3dv/ci: mark whole group as flaky 1470- docs: fix invalid expression in new pipe cap 1471- docs: fix invalid expression in teflon docs 1472- intel/ci: disable CML jobs because of networking issues 1473- intel/ci: add missing .intel-common-manual-rules to .{iris,crocus,i915g}-manual-rules 1474- ci/build: drop mold wrapper for \`ninja install` 1475- ci: drop override forcing ld to be gold (and forcing gold to be installed everywhere) 1476- ci: when installing mold, make its use automatic 1477- ci: bump image tags 1478- radeonsi/ci: drop two failures that are mysteriously fixed by using mold? 1479- ci/container: move deqp build section into the script itself 1480- ci/container: move apitrace build section into the script itself 1481- ci/container: move crosvm build section into the script itself 1482- ci/container: move deqp-runner build section into the script itself 1483- ci/container: move fossilize build section into the script itself 1484- ci/container: move gfxreconstruct build section into the script itself 1485- ci/container: move kdl build section into the script itself 1486- ci/container: move libclc build section into the script itself 1487- ci/container: move llvm-spirv build section into the script itself 1488- ci/container: move mold build section into the script itself 1489- ci/container: move ninetests build section into the script itself 1490- ci/container: move piglit build section into the script itself 1491- ci/container: move rust build section into the script itself 1492- ci/container: move vkd3d-proton build section into the script itself 1493- ci/container: move vulkan-validation build section into the script itself 1494- ci/container: move wayland build section into the script itself 1495- ci/container: add sections around the other build scripts 1496- ci/container: close debian_{setup,cleanup} sections 1497- ci/lava: add setup-test-env.sh to the rootfs 1498- ci/container: add section around strip-rootfs.sh 1499- ci: bump image tags 1500- zink+nvk/ci: fix deqp binary used for gles tests 1501- zink+radv/ci: fix deqp binary used for gles tests 1502- ci/deqp: move testlog-to-* tools to /deqp 1503- ci/deqp: only compress caselists when they exist 1504- ci/deqp: build testlog tools on android 1505- ci/deqp: fetch & checkout exactly the commit/tag/branch requested 1506- ci/deqp: avoid downloading 1.47 GiB multiple times 1507- ci/deqp: error out in case of invalid build API 1508- ci/deqp: build glcts in gles build, for gles*-khr tests 1509- ci/deqp: add build of \`main` branch 1510- ci/deqp: make sure the main commit is actually from the main branch 1511- ci/deqp: fully isolate deqp builds 1512- ci: bump image tags 1513- ci/container: setup sections in all image builds 1514- radv/ci: document regression of test_shader_sm66_is_helper_lane in 7469f99e...25b8f4f7 1515- meson: simplify logic a bit 1516- meson: drop unused variables 1517- meson: reuse variable 1518- meson/megadriver: s/_/-/ in an argument name to be consistent 1519- meson/megadriver: simplify setting common megadriver arguments 1520- meson/megadriver: support various lib suffixes 1521- ci/deqp: simplify paths since we are already in /deqp-$deqp_api/ 1522- ci/deqp: fix the "is this a build on main?" check 1523- ci/deqp: support having commit backports and local patches for main too 1524- ci/deqp: simplify generating the version description file 1525- ci/deqp: mention the deqp api in the version string 1526- ci/deqp: only print the commit list header when the list is not empty 1527- ci/lava: turn the $BUILD_VK check into a proper if block 1528- ci/deqp: add a deqp-vk build on the \`main` branch 1529- ci: bump image tags 1530- radv/ci: use deqp-vk-main in radv jobs 1531- docs: update calendar for 24.2.8 1532- docs: add release notes for 24.2.8 1533- docs: add sha sum for 24.2.8 1534- ci/meson: make meson wrap fallback list more readable 1535- ci/meson: add FORCE_FALLBACK_FOR variable for build jobs to use 1536- docs/release-calendar: add 25.0 branchpoint and RCs schedule 1537- docs/release-calendar: fixup sed fail 1538- docs/release-calendar: push the 25.0 branchpoint back by 2 weeks 1539- docs: update calendar for 24.3.4 1540- docs: add release notes for 24.3.4 1541- docs: add sha sum for 24.3.4 1542- docs/release-calendar: push back the 24.3.x releases by one week 1543- docs: update url to vulkan features & extensions 1544- anv,gfxstream,panvk,zink: update urls to vulkan docs 1545- radv,lvp: fix url to VkAabbPositionsKHR docs 1546- ci: make linker warnings fatal 1547- VERSION: bump for 25.0.0-rc1 1548- [25.0-only] hk: comment out dead variable 1549- .pick_status.json: Update to 5b856a741d6dc18d409a0c06ad6492cc3ee9a6bd 1550- .pick_status.json: Mark 0ee5015da4c386c0ef8b6ff12fd2bb34022d86a6 as denominated 1551- .pick_status.json: Update to e49df902b4c1b98569921d8b858e6e3855bf10e0 1552- .pick_status.json: Update to e192d7d615dec9c9c04447c4b9ab0244d6380944 1553- .pick_status.json: Mark 39969409f6fb60b21aea36be4d5424718fcc26b8 as denominated 1554- VERSION: bump for 25.0.0-rc2 1555- .pick_status.json: Update to fdaf7c7b9647874e66e79653050f9d0999dc9134 1556- docs/android: drop libglapi.so now that it's gone 1557- .pick_status.json: Mark 5f54beb30728f6510ce50071ddaef5f9157b16ef as denominated 1558- gfxstream: fix signedness of shifts 1559- gfxstream: drop dead variables 1560- gfxstream: use \`range` variable for its intended purpose 1561- gfxstream: mark unused variables as such 1562- .pick_status.json: Update to ee9edd46254884ab7fe6c96518e23d421d5f5344 1563- llvmpipe/tests: include math.h for INFINITY 1564- ci: don't run on tag pipelines 1565- ci: only trigger the CI for release managers when pushing to staging branch 1566- .pick_status.json: Update to 18f0807408425da11cb1d8cd1d73de369317440d 1567- .pick_status.json: Update to 30a3d567c8b996fde86b07d2bad018013a54ff44 1568- ci: run containers builds on staging branches 1569- .pick_status.json: Mark 13e987669ccee373948753e113e9ce7e9bdbef55 as denominated 1570- VERSION: bump for 25.0.0-rc3 1571- .pick_status.json: Update to e41438275e005bbb20fc9c8115d7d29343c292d8 1572- ci: debian-testing-ubsan is used by tests 1573- ci/yaml-toml-shell-py-test: don't run on post-merge pipelines 1574- ci/yaml-toml-shell-py-test: run on direct push pipelines 1575- .pick_status.json: Update to a9b6a54a8cce0aab44c81ea4821ee564b939ea51 1576- .pick_status.json: Update to 06d8afff640c66e51517bf4bebd2a58abb2fa055 1577- .pick_status.json: Update to 2361ed27f34774f0a73324915a9ddb57f43e112a 1578- .pick_status.json: Update to 56aac9fdecad0f7d335f82653832927486f07d44 1579- .pick_status.json: Update to 6b20b0658489afe745a28b8f09c57067e45b47f3 1580 1581Eric R. Smith (28): 1582 1583- util: rename PIPE_FORMAT_Y8_U8V8_422_UNORM 1584- dri, mesa: fix NV16 texture format 1585- egl, mesa: add support for NV15 and NV20 textures 1586- dri: fix NV15 and NV20 definitions to make sure they will be used 1587- panfrost: add panfrost support for NV15, NV16 and NV20 1588- panvk: fix depth bias calculation 1589- panfrost: add a perf warning when resources need to be converted 1590- panfrost: convert resources before binding them to images 1591- panfrost: check afbc status in panfrost_query_compression_modifiers 1592- mesa: when blitting between formats clear any unused components 1593- aux: add support for dumping the swizzle in pipe_blit_info 1594- mesa: update more drivers to handle pipe_blit_info swizzle_enable 1595- format: Add R8_G8B8_422_UNORM format 1596- panvk: update feature support 1597- panvk: split device and instance version numbers 1598- panvk: advertise version 1.1 support 1599- panfrost: fix read/write resource confusion in afbc_pack 1600- panfrost: fix potential memory leak 1601- panvk: fix fs_required() 1602- panfrost: apply DEPTH_STENCIL flag consistently 1603- panfrost: Allow ATEST input to be a FAU index 1604- panfrost: ensure sample_mask is written before color 1605- panvk: re-enable fragmentStoresAndAtomics for v10 1606- drm-uapi: update drm_fourcc.h to latest version 1607- panfrost: support MTK 16L32S detiling 1608- panfrost: avoid potential divide by 0 calculating timer_resolution 1609- panfrost: fix YUV center information for 422 1610- panfrost: fix backward propagation of values in loops 1611 1612Erico Nunes (2): 1613 1614- ci/lima: update piglit ci expectations 1615- ci/lima: enable again 1616 1617Erik Faye-Lund (134): 1618 1619- panvk: drop unused include 1620- panfrost: use mesa_log infra instead of stdio 1621- glx: avoid null-deref 1622- panfrost: use 64-bits for layout calculations 1623- panvk: set correct max extents for images 1624- panvk: support binding swapchain memory 1625- panvk: wire up swapchain image creation 1626- panvk: remove duplicate property 1627- panvk: implement sampleRateShading 1628- panvk: check for maxResourceSize-overflow in vkCreateImage 1629- panvk: document reason for maxResourceSize-limit 1630- docs: mark GL_ARB_shader_subroutine as always supported 1631- docs: mark GL_ARB_get_program_binary as always supported 1632- docs: update GL_OES_shader_image_atomic support 1633- docs: update GL_ARB_multi_draw_indirect support 1634- docs: refer to panfrost by version 1635- docs: fixup a few mistakes with panfrost 1636- docs: add missing panfrost extensions 1637- lima: fixup typo 1638- lima: add assert to validate list-lenght 1639- lima: avoid memleak on error 1640- panfrost: sanity-check alignment 1641- panvk: correct signedness of timestamps 1642- panvk: widen type before multiplying 1643- mesa/main: properly check for EXT_memory_object 1644- mesa/main: properly check for EXT_memory_object_fd 1645- mesa/main: properly check for EXT_memory_object_win32 1646- mesa/main: properly check for EXT_semaphore 1647- mesa/main: properly check for EXT_semaphore_win32 1648- st/mesa: check requirements for MESA_texture_const_bandwidth 1649- mesa: error-check GL_TEXTURE_TILING_EXT params 1650- panvk: report minmax-support for sampled formats 1651- panvk: expose KHR_dedicated_allocation 1652- vulkan/meta: plug a couple of memory leaks 1653- panvk: free preload-shaders after compiling 1654- panvk, nvk: spell width correctly 1655- panvk/ci: correct name of skips-file 1656- panvk/ci: remove duplicate skips 1657- panvk/ci: add some missing skips 1658- panvk/ci: update ci results for g610 1659- panvk/ci: add a few flakes 1660- panvk/ci: add a full panvk job 1661- panfrost: match 4-bit format order 1662- panfrost: add missing 4-bit formats 1663- panvk: expose EXT_4444_formats 1664- panvk/ci: update g52 results 1665- panvk/ci: update g610 results 1666- panvk: expose scalarBlockLayout 1667- panvk/ci: remove duplicate skips 1668- panvk/ci: update g52 results 1669- panvk/ci: update g52-vk-full job 1670- panvk: do not expose subgroup support 1671- panvk: disable imageCubeArray on bifrost 1672- panvk: soften the language around opt-in 1673- panvk: do not require opt-in for panvk on v10 1674- panvk/ci: correct timeouts as crash 1675- panvk/ci: fixup g52 skip sorting 1676- panvk/ci: add a few more g52 skips 1677- panvk: fixup bad indent 1678- panvk: only validate the push-sets that we update 1679- panvk: back out of vk 1.1 support 1680- panvk: make vk-version helper internal to source 1681- docs: add new panvk features 1682- panvk: fix image size for cube-arrays on bifrost 1683- Revert "panvk: disable imageCubeArray on bifrost" 1684- st/mesa: document ARB_texture_float quirk 1685- pan/cs: fix broken allocation-failure check 1686- panfrost: clean up mmap-diagnostics 1687- panfrost: report errors from panfrost_bo_mmap 1688- panfrost: handle mmap failures 1689- panfrost: handle NULL-batches 1690- panfrost: propagate cs_builder error instead of asserting 1691- panfrost: handle pool-allocation errors 1692- panfrost: handle errors allocating csf oom-handler 1693- panfrost: try to survive start-up alloc fails 1694- pan/ci: update t860 ci xfails 1695- panvk: drop fragmentStoresAndAtomics support for now 1696- vulkan: add vk_descriptor_type_is_dynamic helper 1697- v3dv: use vk_descriptor_type_is_dynamic 1698- turnip: use vk_descriptor_type_is_dynamic 1699- dozen: use vk_descriptor_type_is_dynamic 1700- panvk: use vk_descriptor_type_is_dynamic 1701- radv: use vk_descriptor_type_is_dynamic 1702- asahi: use vk_descriptor_type_is_dynamic 1703- turnip: use vk_descriptor_type_is_dynamic 1704- pvr: use vk_descriptor_type_is_dynamic 1705- panvk: use vk_descriptor_type_is_dynamic 1706- lavapipe: use vk_descriptor_type_is_dynamic 1707- anv: use vk_descriptor_type_is_dynamic 1708- hasvk: use vk_descriptor_type_is_dynamic 1709- dozen: use vk_descriptor_type_is_dynamic 1710- nvk: use vk_descriptor_type_is_dynamic 1711- panvk/ci: update expected failures 1712- docs: fixup broken markup 1713- docs: fixup link in radv docs 1714- docs/ci: treat warnings as errors 1715- docs: update panvk status 1716- panvk/ci: drop needless envvar 1717- Revert "panfrost: Disable CRC by default" 1718- pan/ci: update t760 checksum 1719- pan/ci: update opencl expectations 1720- docs/panfrost: document vulkan support 1721- docs: update panvk status 1722- docs/features: fixup panvk KHR_shader_draw_parameters-support 1723- pan/va: fix base-level for nir_texop_lod 1724- pan/ci: add some occasional flakes 1725- docs/features: add a few missing extensions 1726- docs/features: mark panfrost as supporting GL_OES_texture_view 1727- pan/ci: drop empty trailing variables-list 1728- panfrost: reuse tiler hierarchy mask selection from panvk 1729- panfrost: limit maximum texture size 1730- panfrost: do not artificially limit texture-sizes 1731- pan/midgard: use macros for mir_prev_op / mir_next_op 1732- pan/midgard: constify pointers 1733- pan/compiler: don't pass midgard_instruction by value 1734- panvk: expose subgroup operations 1735- panvk: expose vk1.1 on v10 hardware 1736- pan/bi: bump iter_count to 2000 1737- panvk: do not expose EXT_subgroup_size_control on bifrost 1738- panvk/ci: update expected failures 1739- panfrost: mark helper as static 1740- panfrost: handle allocation errors when afbc-packing 1741- panfrost: unify emit_tls and emit_fbd 1742- panfrost: propagate allocation scratchpad allocation errors 1743- panfrost: propagate errors from panfrost_batch_create_bo 1744- panfrost: in-place map/unmap shouldn't grow 1745- gallium/aux: do not assert on map-failures 1746- meson: build panvk by default on arm 1747- panvk: fix line-rasterization of bifrost 1748- panvk/ci: add back incorrectly removed crash 1749- pan/ci: add flaky tests to the flake-list 1750- pan/ci: add fail from llvm 19 upgrade 1751- panvk: correct number of read bytes for dynamic buffers 1752- panvk: report passing the VK CTS 1753 1754Ernst Persson (1): 1755 1756- intel/vulkan: Add bvh build dependency 1757 1758Evan (1): 1759 1760- amd/vpelib: Shaper Refactor 1761 1762Faith Ekstrand (27): 1763 1764- vulkan: Allow the same item to show up twice in core version <requires> 1765- vulkan: Add Vulkan 1.4 feature aliases 1766- treewide: Stop putting enum in front of Vulkan enum types 1767- vulkan: Update XML and headers to 1.4.303 1768- nvk: Increase push constant space to 256B 1769- nvk: No-op implement VK_KHR_global_priority 1770- nvk: Add new Vulkan 1.4 features and properties 1771- nvk: Advertise Vulkan 1.4 1772- nvk: Only support Vulkan 1.4 on Turing+ 1773- nvk: Move Vulkan 1.4 features to the 1.4 section 1774- nvk: Move Vulkan 1.4 properties to the 1.4 section 1775- nvk: Set a command buffer error if pushbuf alloc fails 1776- nvk: Call nir_opt_access 1777- nak: Use ldc.constant for load_global when CAN_REORDER is set 1778- nvk: Handle pCounterBuffers == NULL in Begin/EndTransformFeedback 1779- nvk: Fix scissor bounds 1780- nvk: Rename nvk_descriptor_set::mapped_ptr 1781- nvk: Respect VK_DESCRIPTOR_POOL_CREATE_HOST_ONLY_BIT_EXT 1782- nvk: Implement descriptorBufferPushDescriptors 1783- nvk: Pull shaders from the state command buffer in nvk_cmd_process_cmds() 1784- nvk: Handle shader==NULL in nvk_cmd_upload_qmd() 1785- nvk: Allow sparse loads on EDB buffers 1786- nak: Handle sparse texops with unused color destinations 1787- nvk: Use suld for EDB uniform texel buffers 1788- nvk: Align UBO/SSBO addresses down rather than up 1789- nak: Use suld.constant when ACCESS_CAN_REORDER is set 1790- nvk: Use suld.constant for EDB uniform texel buffers 1791 1792Felix DeGrood (6): 1793 1794- iris: Use vfg distribution mode = RR_STRICT for Xe2+ 1795- anv: Use vfg distribution mode = RR_STRICT for Xe2+ 1796- anv: allow compressed buffers types on vkd3d titles 1797- anv: remove unnecessary driconf entries for anv_enable_buffer_comp 1798- vk/overlay-layer: defer log creation to swapchain creation 1799- intel/perf: add new perf consts to support more metrics 1800 1801Feng Jiang (2): 1802 1803- virgl: Ensure that PIPE_SHADER_CAP_MAX_CONST_BUFFERS is less than PIPE_MAX_CONSTANT_BUFFERS 1804- radv/rt: Fix memleak in radv_init_header() 1805 1806Francisco Jerez (27): 1807 1808- intel/fs/xe2: Fix up subdword integer region restriction with strided byte src and packed byte dst. 1809- intel/brw/xe3+: Relax SEND EOT register assignment restrictions. 1810- intel/brw: Saturate shifted subgroup index to avoid reading past the end of register file. 1811- intel/brw: Use urb_read_length instead of nr_attribute_slots to calculate VS first_non_payload_grf. 1812- intel/brw/xe3+: Mask subgroup shuffle index to be within valid range to avoid VRT hangs. 1813- anv/gfx12.5: Request subgroup size 8 for RT trampoline shader. 1814- intel/brw: Allow specifying a required subgroup size for fragment shaders. 1815- intel/blorp: Specify a subgroup size requirement of 16 for fast clear or repclear shaders. 1816- intel/common/xe2+: Allow SIMD32 PS for all multisample cases. 1817- intel/brw/xe3: Define XE3_MAX_GRF. 1818- intel/brw/xe3: Extend regalloc sets to maximum Xe3 GRF size. 1819- intel/brw/xe3+: Bump number of SBID tokens for Xe3. 1820- intel/brw/xe3+: Disable round-robin allocation heuristic on Xe3+. 1821- intel/brw: Indent body of brw_compile_fs() not applicable to xe3+. 1822- intel/brw: Indent conditional block from brw_compile_fs() not applicable to Xe2+. 1823- intel/brw: Exit early from run_fs() if compilation failed before optimization loop. 1824- intel/brw/xe3+: brw_compile_fs() implementation for Xe3+. 1825- intel/brw/xe3+: Optimize CS/TASK/MESH compile time optimistically assuming SIMD32. 1826- intel/brw: Report number of GRF registers used in brw_stage_prog_data. 1827- intel/brw: Define ptl_register_blocks() helper. 1828- intel/genxml/xe3+: Update definitions for shader state setup. 1829- iris/xe3+: Set RegistersPerThread during shader state setup based on prog_data. 1830- intel/blorp/xe3+: Set RegistersPerThread during shader state setup based on prog_data. 1831- anv/xe3+: Set RegistersPerThread during shader state setup based on prog_data. 1832- anv/xe3+: Set RegistersPerThread for bindless shader dispatch. 1833- iris/xe3+: Enable VRT. 1834- anv/xe3+: Enable VRT. 1835 1836Frank Binns (2): 1837 1838- pvr: add TI j721s2 as a supported device 1839- pvr: add 36.53.104.796 (BXS-4-64) to the list of supported GPUs 1840 1841Friedrich Vock (15): 1842 1843- vulkan/rmv: Correctly set heap size 1844- vulkan/runtime/bvh: Set leaf_node_count for updates 1845- radv,driconf: Apply DOOM Eternal/idTech workarounds for Indiana Jones 1846- aco/lower_to_hw_instr: Check the right instruction's opcode 1847- radv/rt: Remove nir_intrinsic_execute_callable instrs in monolithic mode 1848- aco: Fix dead instruction/index handling for try_insert_saveexec_out_of_loop 1849- nir: Serialize all parameter attributes 1850- nir,vtn: Add return info to parameters 1851- nir: Add parameter divergence info 1852- vtn: Set parameter type in glsl_type_add_to_function_params 1853- nir: Add indirect calls 1854- nir: Apply passes to all functions 1855- nir: Add nir_instr_is_before helper 1856- nir: Free liveness info when invalidating metadata 1857- nir: Add indirect call optimizations 1858 1859GKraats (1): 1860 1861- i915g: fix glClearColor using a 1 byte color format 1862 1863Georg Lehmann (79): 1864 1865- radv: run copy prop before vectorizing 1866- nir/opt_16bit_tex_image: optimize extract half sources 1867- nir: add nir_def_all_uses_ignore_sign_bit 1868- pan/bi: use nir_def_all_uses_ignore_sign_bit 1869- aco: use nir_def_all_uses_ignore_sign_bit 1870- nir: handle fmul(a,a)/ffma(a,a,b) in nir_def_all_uses_ignore_sign_bit 1871- aco/gfx8: use ds_swizzle_b32 rotate mode 1872- nir: return def for debug info in nir_instr_def 1873- nir/instr_set: replace nir_instr_get_def_def with nir_instr_def 1874- nir/instr_set: support instrs with no def 1875- nir: cse terminate/demote 1876- nir/opt_undef: replace undef in a separate pass 1877- nir/opt_undef: use some nir helpers 1878- nir/opt_undef: keep undefs used by partial undef vectors 1879- nir/opt_undef: handle unpack/pack like mov/vec 1880- aco/isel: use undef Operands for p_create_vector created from nir vecs 1881- util: add BITSET_LAST_BIT_BEFORE 1882- nir/move_discards_to_top: single final iteration 1883- nir/move_discards_to_top: don't move across is_helper_invocation 1884- radv/ci: document test_shader_sm66_is_helper_lane as fixed 1885- freedreno/ci: update a630 KSP checksum 1886- nir/opt_intrinsic: rework sample mask opt with vector alu 1887- nir/opt_intrinsic: fix sample mask opt with demote 1888- radv: optimize sample mask comparisons 1889- aco/optimizer: label fcanonicalize like a copy if there is nothing to flush 1890- nir/opt_algebraic: optimize ffma(b2f, b2f, c) 1891- nir/opt_algebraic: optimize d3d9 ftrunc 1892- nir/opt_algebraic: optimize d3d9 ceil 1893- nir/opt_algebraic: mark a - ffract(a) as nan incorrect. 1894- radv: fix reporting mesh/task/rt as supported dgc indirect stages 1895- radv: rework vk_property initialization 1896- aco/gfx12: disable vinterp ddx/ddy optimization 1897- aco/gfx12+: do not use v_pack_b32_f16 to pack untyped data 1898- radeonsi/ci: add vangogh ubo fail 1899- zink: spec\@ext_framebuffer_multisample\@blit-mismatched-formats was fixed 1900- aco/gfx11+: use v_and_b32 to extract local id 0 1901- radv: track holes in the clip/cull masks 1902- nir: add constant clip/cull distance optimization 1903- radv: use nir_opt_clip_cull_const 1904- nir/uub: properly limit float support to 32bit 1905- nir: add unsigned upper bound support for f2i32 1906- nir: add unsigned upper bound support for fsat 1907- aco/gfx12: don't assume memory operations complete in order 1908- aco/ra: don't write to exec/ttmp with mulk/addk/cmovk 1909- aco/ra: disallow s_cmpk with scc operand 1910- aco/ra: don't write to scc/ttmp with s_fmac 1911- nir/opt_remove_phis: rematerialize equal alu 1912- nir/opt_algebraic: optimize min(max(a, b), a) 1913- nir: optimize unpacking 8bit values from a 64bit source 1914- aco/isel: skip and(exec) for top level demote_if/terminate_if 1915- aco: rename p_early_exit_if to if_not 1916- aco: allow p_exit_early_if_not with exec condition 1917- aco/insert_exec: exit shader using exec for top level discard 1918- aco: create v_cmpx with s_andn2(exec, v_cmp) 1919- nir: sink/move alu with two identical, non constant sources. 1920- amd: switch to FRONT_FACE_ALL_BITS(0) 1921- nir: add load_front_face_fsign 1922- amd: support load_front_face_fsign 1923- nir: add nir_alu_srcs_negative_equal_typed 1924- nir,amd: optimize front_face ? a : -a 1925- aco/optimizer: fix signed extract of sub dword temps with SDWA 1926- aco/insert_exec: reset top exec for p_discard_if 1927- radv: run peephole_select in optimize_nir_algebraic 1928- nir/peephole_select: allow load_vector/scalar_arg_amd 1929- aco: guard small_vector move/copy operator against self assignment 1930- aco: support less trivial component types in small_vec 1931- aco: implement some more std::vector functions for small_vec 1932- nir/opt_algebaric: convert fadd(a, a) to a * 2.0 1933- aco: update is_dual_issue_capable for gfx11.5+ 1934- aco/sched_ilp: continue open clauses 1935- aco/sched_ilp: add dependencies of later clause instrs more aggressively 1936- aco/sched_ilp: only remove WaW/WaR for inter clause dependencies 1937- aco/sched_ilp: reorder VINTRP 1938- aco/sched_ilp: new latency heuristic 1939- aco/sched_ilp: rename priority to wait_cycles 1940- aco/sched_ilp: use more realistic memory latencies 1941- aco/sched_ilp: base latency and issue cycles on aco_statistics 1942- nir: fix range analysis for frcp 1943- nir: fix frsq range analysis 1944 1945Gert Wollny (6): 1946 1947- virgl/vtest: take handle from host when using protocol version >=3 1948- virgl/vtest: When trying to use protocol 3 check host feature 1949- virgl/vtest: change interface of virgl_vtest_submit_cmd 1950- virgl/vtest: Add support for creating blob resources 1951- ci: Upref virglrenderer version 1952- radeon/evergreen: ensure equal sizes for depth-stencil npot textures 1953 1954Guilherme Gallo (9): 1955 1956- ci/lava: Set default exit code to 1 for failed jobs 1957- ci/lava: Improve exception handling for job failures 1958- ci/lava: Uprev freezegun 1959- ci/intel: Set HWCI modules for puff DUT 1960- ci/iris: Force UART for puff boards 1961- ci/iris: Rebalance iris-cml-deqp jobs 1962- ci/iris: Fix iris-cml-traces expectations 1963- ci/iris: Update iris-cml-deqp CI expectations 1964- ci/container: set up S3_JWT_FILE also for container jobs 1965 1966Gurchetan Singh (17): 1967 1968- util: add c++ guards to u_mm.h 1969- gfxstream: move isHostVisible function 1970- gfxstream: nuke android::base::SubAllocator 1971- gfxstream: use vulkan_lite_runtime 1972- gfxstream: nuke EntityManager.h include 1973- gfxstream: aemu: vendor it 1974- gfxstream: modify libaemu for Mesa use case 1975- gfxstream: guest: use internal version of AEMU headers + impls 1976- gfxstream: use canonical Mesa dependencies 1977- gfxstream: conditionals for using gfxstream::aemu 1978- gfxstream: delete qemu_pipe target 1979- gfxstream: for Android, look for the autogenerated files 1980- gfxstream: change output location 1981- gfxstream: remove abort() 1982- gfxstream: fix issues with VK1.4 build 1983- gfxstream: remove references to Fuchsia Goldfish 1984- gfxstream: fix some integration bugs 1985 1986Hans-Kristian Arntzen (11): 1987 1988- vulkan/wsi/wayland: Use X11-style image count strategy when using FIFO. 1989- radv: Fix missing gang barriers for task shaders. 1990- radv/winsys: Report VA mappings in bo_log too. 1991- radv: Add sparse mappings to radv_check_va.py. 1992- wsi/x11: Do not use allocation callbacks on a thread. 1993- wsi/wayland: Only use commit timing protocol alongside present time. 1994- wsi/wayland: Don't fallback to broken legacy throttling with FIFO 1995- wsi/wayland: Handle FIFO -> MAILBOX transitions correctly 1996- wsi/wayland: Remove unused present_mode member. 1997- wsi/wayland: Add forward progress guarantee for present wait. 1998- radv: Add radv_invariant_geom=true for Indiana Jones. 1999 2000Hsieh, Mike (1): 2001 2002- amd/vpelib: Refactor 3D LUT parameters 2003 2004Hyunjun Ko (10): 2005 2006- anv: define ANV_VIDEO_H264_MAX_DPB_SLOTS 2007- anv: Enable remapping picture ID 2008- anv: handle negative value of slot index for h265 decoding. 2009- intel/genxml: define MEMORYADDRESSATTRIBUTES for Gen12.5 with TILEF 2010- anv/video: Fix to return supported video format correctly. 2011- anv: calculate global parmeters correctly for AV1 decoding 2012- anv: support in-loop super resolution for AV1 decoding 2013- anv: fix to set default cdf buf correctly. 2014- anv: change bool to VkResult 2015- anv: Fix to set CDEF flter flag correctly for AV1 decoding 2016 2017Iago Toral Quiroga (15): 2018 2019- v3d: add a V3D_DEBUG option to force synchronous execution of jobs 2020- broadcom: handle double buffer on V3D 7.1 tile size calculations 2021- v3d: group tile spec into a struct inside the job 2022- v3d: save a pointer to the TILE_BINNING_MODE_CFG packet in the CL 2023- v3d: do tile state BO allocation later 2024- v3d: only enable double-buffer for jobs where it might make sense 2025- v3dv: add missing support for double-buffer on V3D 7.x 2026- v3d: drop blank line 2027- v3d: store size of qpu program for compiled shaders 2028- broadcom: add helpers for double-buffer heuristic 2029- v3d: use heuristic to enable double-buffer mode 2030- v3dv: use the double buffer heuristic helpers 2031- broadcom: move double-buffer heuristic helpers to the compiler 2032- v3dv: fix missing access bit flag when checking for texel buffer reads 2033- v3dv: fix crash on 32-bit builds 2034 2035Ian Romanick (57): 2036 2037- brw/emit: Add correct 3-source instruction assertions for each platform 2038- brw/copy: Don't copy propagate through smaller entry dest size 2039- brw/cse: Don't eliminate instructions that write flags 2040- brw/lower: Don't emit spurious moves to or from NULL register 2041- brw/opt: Always do copy prop, DCE, and register coalesce after lower_regioning 2042- brw/opt: Always do both kinds of copy propagation before lower_load_payload 2043- brw/build: Add scalar_group() helper 2044- brw/lower: Lower invalid source conversion to better code 2045- Fix copy-and-paste bug in nir_lower_aapoint_impl 2046- brw/lower: Don't "fix" regioning of broadcast 2047- brw: Use resize_sources several more places 2048- brw/build: Use SIMD8 temporaries in emit_uniformize 2049- brw/copy: Allow copy prop into src1 of broadcast 2050- nir/algebraic: Optimize some trivial bfi 2051- brw/algebraic: Fix ADD constant folding 2052- brw/algebraic: Fix MUL constant folding 2053- brw/emit: Fix typo in recently added ADD3 assertion 2054- brw/algebraic: Partial constant folding of ADD3 2055- brw/const: Allow mixing signed and unsigned immediate sources 2056- brw/copy: Don't try to be clever about ADD3 constant propagation 2057- brw: Emit immediate value for MAD in canonical position 2058- brw/copy: Commute immediates for MAD multiplicands 2059- brw/algebraic: Constant fold multiplicands of MAD 2060- brw/algebraic: Don't restrict MAD(a, b, 1) optimization to float32 2061- brw/const: Refactor checking whether an immediate source is allowed 2062- brw/const: Allow constants in integer MAD 2063- brw/const: Allow HF constants in MAD on Gfx11 2064- brw/const: Remove TODO that isn't allowed by the hardware 2065- brw/algebraic: Pull brw_constant_fold_instruction out of the switch statement 2066- brw/emit: Fix BROADCAST when value is uniform and index is immediate 2067- brw: Add devinfo parameter to fs_inst::regs_read 2068- brw: Basic infrastructure to store convergent values as scalars 2069- brw/lower: Allow uniform and scalar sources to many kinds of SEND 2070- brw/nir: Fix up handling of sources that might be convergent vectors 2071- brw/lower: Adjust source stride on DF is_scalar sources to MAD on Gfx9 2072- brw/lower: Properly handle UNIFORM globals address in lower_trace_ray_logical_send 2073- brw/emit: Allow scalar sources to HF math instructions on Xe2 2074- brw/nir: Prepare try_rebuild_source for scalar values 2075- brw/build: Prepare BROADCAST for scalar values 2076- brw/nir: Treat load_const as convergent 2077- brw/nir: Treat some load_uniform as convergent 2078- brw/nir: Treat load_workgroup_id as convergent 2079- brw/nir: Treat some ALU results as convergent 2080- brw/nir: Treat some load_ubo as convergent 2081- brw/nir: Treat load_inline_data_intel as convergent 2082- brw/nir: Treat load_reloc_const_intel as convergent 2083- brw/nir: Treat load_btd_{global,local}_arg_addr_intel and load_btd_shader_type_intel as convergent 2084- brw/nir: Treat load_*_uniform_block_intel as convergent 2085- brw/nir: Treat some resource_intel as convergent 2086- brw/nir: Eliminate nir_to_brw_state::uniform_values 2087- brw/nir: Don't try optimize around emit_uniformize 2088- brw/nir: Simplify get_nir_image_intrinsic_image and get_nir_buffer_intrinsic_index 2089- brw/nir: Treat some ballot as convergent 2090- brw/nir: Don't generate scalar byte to float conversions on DG2+ in optimize_extract_to_float 2091- iris: Add missing nir_metadata_preserve in iris_lower_storage_image_derefs 2092- crocus: Add missing nir_metadata_preserve in crocus_lower_storage_image_derefs 2093- brw/copy: Fix handling of offset in extract_imm 2094 2095Icenowy Zheng (4): 2096 2097- zink: do not set transform feedback bits when not available 2098- meson: prefer 'python3' to 'python' when finding python3 2099- zink: emit consts as uint only on IMG proprietary drivers 2100- zink: use lazy descriptors for IMG proprietary drivers 2101 2102Igor Torrente (2): 2103 2104- Zink: Add NVK to the non \`driver_workarounds.implicit_sync` list 2105- NVK: Enable RW DMA-BUF export 2106 2107Ivan Avdeev (1): 2108 2109- radv: add a flag to indicate ray tracing support 2110 2111Iván Briano (6): 2112 2113- intel/rt: fix ray_query stack address calculation 2114- intel/decoder: fix INTEL_DEBUG=bat 2115- anv: remove unused/misleading/wrong parameters from the RT trampoline 2116- vulkan: calculate remaining layers of 2d view of 3d image correctly 2117- anv: disable logic op for float/srgb formats 2118- hasvk: disable logic op for float/srgb formats 2119 2120James Hogan (3): 2121 2122- glsl: Expose gl_ViewID_OVR back to GLSL 1.30 2123- mesa: Fix multiview attachment completeness check 2124- mesa: Fix FramebufferTextureMultiviewOVR num_views check 2125 2126Janne Grunau (1): 2127 2128- panvk: Silence warning on incompatible DRM render devices 2129 2130Jason Macnak (3): 2131 2132- Simplify ApiInfo 2133- Pass VkSnapshotApiCallInfo-s through VkDecoderGlobalState 2134- Update VkDecoderSnapshot locking 2135 2136Jesse Natalie (4): 2137 2138- microsoft/compiler: Put holes in driver_location based on I/O variable sizes 2139- microsoft/clc: Initialize printf buffer for tests 2140- microsoft/compiler: Skip POS for io compaction 2141- microsoft/compiler: Update clip/cull split pass to handle clip/cull getting merged 2142 2143Jianxun Zhang (5): 2144 2145- anv,hasvk,genxml: Rename genxml files using verx10 2146- isl: Refactor WA 22015614752 2147- iris: Allow compression on multi-sampled stencil (xe2) 2148- isl: Allow CCS in more cases (xe2) 2149- isl: Move a CCS restriction in GFX 12.x 2150 2151Job Noorman (87): 2152 2153- ir3/ra: prevent moving source intervals for shared collects 2154- ir3,tu: include ir3 debug flags in shader hash key 2155- ir3,tu: filter debug flags included in the hash key 2156- ir3: fold shared movs into other movs 2157- nir: add ir3-specific bitwise triop opcodes 2158- nir/search: make is_only_used_by_iadd reusable 2159- nir/search: add is_only_used_by_{iand,ior} helpers 2160- ir3: fix backend support for bitwise triops 2161- ir3: add codegen for bitwise triops 2162- ir3: add pass to select bitwise triops 2163- ir3/isa: allow rpt6/rpt7 2164- ir3: add workaround for predication hardware bug 2165- nir/lower_subgroups: support unknown subgroup size 2166- ir3: use generic lowering for 64b scan/reduce 2167- ir3: remove unused ir3_nir_lower_64b_subgroups 2168- nir: add read_getlast_ir3 intrinsic 2169- ir3: add codegen for read_getlast_ir3 2170- ir3: add helper to get the subgroup size 2171- ir3: rename cluster_size to brcst_cluster_size 2172- nir/lower_subgroups: add extra filter data to options 2173- nir/lower_subgroups: disable boolean reduce when not supported 2174- ir3: add support for clustered subgroup reductions 2175- tu: advertise VK_SUBGROUP_FEATURE_CLUSTERED_BIT 2176- nir/lower_subgroups: add option to only lower clustered rotates 2177- ir3: lower clustered rotates to shuffles 2178- tu: advertise VK_SUBGROUP_FEATURE_ROTATE_CLUSTERED_BIT_KHR 2179- ir3: don't update builder cursor for IR3_CURSOR_AFTER_BLOCK 2180- ir3: add ir3_after_instr_and_phis helper 2181- ir3: use generic INSTR0 implementation for ir3_NOP 2182- ir3: refactor builders to use ir3_builder API 2183- ir3: reformat after refactoring in previous commit 2184- ir3: add reformatting commits to .git-blame-ignore-revs 2185- ir3/isa: fix conflict between stib.b and stsc 2186- ir3/isa: fix cat3-alt immed src 2187- ir3/isa: fix isaspec for sad.s32 2188- ir3: teach backend about sad 2189- ir3: add codegen for sad 2190- ir3/cp: only mark mad srcs as swapped when swap succeeded 2191- ir3/cp: extract common src swapping code 2192- ir3/cp: make try_swap_mad_two_srcs more generic 2193- ir3/cp: add support for swapping srcs of sad 2194- ir3/validate: print file/line info 2195- ir3,freedreno: remove binning outputs after vs ucp lowering 2196- ir3/cp: swap back correct srcs when swap failed 2197- ir3: always set wrmask for movmsk 2198- ir3: emit uniform iadd3 as two adds 2199- ir3: output early-preamble stat as integer 2200- ir3/ra: fix non-trivial collect detection 2201- ir3/ra: allocate shared collects dst over its srcs when possible 2202- ir3/parser: fix parsing integer as float 2203- ir3/a7xx: properly handle alias scope and type 2204- ir3/a7xx: disasm halfness of alias dst 2205- ir3/a7xx: implement and document unknown alias field 2206- ir3/a7xx: handle alias.rt dst 2207- ir3/a7xx: document alias.rt 2208- ir3/print: add support for alias 2209- ir3: teach backend about alias 2210- ir3: introduce alias goups 2211- ir3: add validation for alias 2212- ir3: add ir3_compiler::has_alias 2213- ir3: add support for alias.tex 2214- ir3: optimize alias register allocation by reusing GPRs 2215- ir3/legalize: insert (ss) to read consts after stc 2216- ir3/legalize: insert (sy) to read consts after ldc.k 2217- ir3/dce: support partial writes from collects 2218- ir3: add some preamble helpers 2219- ir3: make find_end a global helper 2220- tu,ir3: inform ir3 of dynamically remapped FS slots 2221- ir3: make shader output struct non-anonymous 2222- ir3: reuse ir3_find_output in ir3_find_output_regid 2223- tu: add chip param to tu6_emit_fs_outputs 2224- tu: add support for aliased render target components 2225- freedreno: add chip param to emit_fs_output 2226- freedreno: add support for aliased render target components 2227- ir3: add support for alias.rt 2228- ir3: disable alias.rt pre-a750 2229- ir3: account for inserted nops in delay calculation 2230- freedreno: move ForEachMacros into freedreno 2231- freedreno: remove unused entries from ForEachMacros 2232- freedreno: add missing entries to ForEachMacros 2233- ir3: schedule alias.rt at the end of the preamble 2234- ir3: rematerialize preamble defs in block dominated by sources 2235- ir3: add helper to calculate src read delay 2236- ir3: make delay slots a compiler property 2237- ir3/a7xx: update delays slots 2238- ir3/a7xx: enable delayed src2 read for all cat3 instructions 2239- ir3: fix emitting descriptor prefetches at end of preamble 2240 2241John Anthony (2): 2242 2243- panvk: Enable storageBuffer16BitAccess 2244- panvk: Enable VK_KHR_vertex_attribute_divisor 2245 2246Jordan Justen (6): 2247 2248- intel/dev: Add PTL 0xb0b0 PCI ID 2249- intel/dev: Split hwconfig warning check into hwconfig_item_warning() 2250- intel/dev: Split apply and check paths for hwconfig 2251- intel/dev: Don't process hwconfig table to apply items when not required 2252- intel/dev: Add intel_check_hwconfig_items() 2253- iris: Check that mem_fence_bo was created 2254 2255Jose Maria Casanova Crespo (9): 2256 2257- v3d: Enable Early-Z with discards when depth updates are disabled 2258- rpi4/ci: mark another flaky timeline_semaphore test 2259- rpi4/ci: another detected flaky timeline_semaphore test 2260- vc4/ci: fails udpate after last piglit uprev 2261- rpi4/ci: Increase timeout for rusticl jobs. 2262- v3d: Don't load/store if rasterizer discard is enabled 2263- v3d/ci: update rpi expectations by last piglit uprev 2264- v3d: Apply FBO resources invalidations on job creation 2265- Revert "ci: take igalia farm offline" 2266 2267Joshua Duong (1): 2268 2269- gfxstream: update auto-generated comments. 2270 2271José Roberto de Souza (16): 2272 2273- intel/dev/xe: Fix access to eu_per_dss_mask 2274- intel/dev/xe: Fix size of eu_per_dss_mask 2275- intel/genxml/xe2: Add STATE_SYSTEM_MEM_FENCE_ADDRESS instruction 2276- anv: Always create anv_async_submit in init_copy_video_queue_state() 2277- anv: Emit STATE_SYSTEM_MEM_FENCE_ADDRESS 2278- iris: Emit STATE_SYSTEM_MEM_FENCE_ADDRESS 2279- iris: Add support for damage region 2280- anv: Allow larger SLM sizes for task and mesh shader 2281- anv: Check VkResult of perf query batch buffer 2282- anv: Check VkResult main batch buffer before start companion batch buffer 2283- iris: Drop BO_ALLOC_COHERENT from iris_utrace_create_ts_buffer() 2284- iris: Rename BO_ALLOC_COHERENT to BO_ALLOC_CACHED_COHERENT 2285- anv: Return scanout PAT entry for scanout and external buffers in discrete GPUs 2286- anv: Allow WSI blit_src Image to be kept compressed when transitioning to VK_IMAGE_LAYOUT_PRESENT_SRC_KHR 2287- iris: Make sure a uncached heap is choosen for scanout and shared buffers when LLC is not available 2288- iris: Pick scanout PAT entry for scanout buffers 2289 2290Juan A. Suarez Romero (26): 2291 2292- util/format: nr_channels is always <= 4 2293- v3dv: remove unused assignments 2294- v3dv: fix BO allocation 2295- v3dv: free pointers on multisync error 2296- v3dv: ensure there is always a perfmon and counter 2297- broadcom/compiler: ensure offset source exists 2298- broadcom/compiler: fix fp16 conversion operations 2299- v3d: make v3d_flush_resource reallocate non-shareable resources 2300- vc4: ensure sharing tiled resources are of proper format 2301- v3d: fix BO allocation 2302- v3d: remove intermediate variable 2303- v3d: find linear modifier when required 2304- vc4: find linear modifier when required 2305- v3d/ci: clean some asan failures 2306- v3d: avoid 0-size variable length array 2307- v3dv: fix assigned value is garbage or undefined 2308- vc4: initialize variable 2309- v3dv: check requirements for USAGE_INPUT_ATTACHMENT 2310- freedreno: a2xx: fix maybe uninitialized variable 2311- radeonsi/vcn: fix maybe uninitialized 2312- v3d: fix format overflow error 2313- virgl: fix member access to a NULL pointer struct 2314- etnaviv: cast assertion 2315- ci/build: add ubsan build jobs 2316- broadcom/ci: add ubsan jobs for broadcom drivers 2317- ci: take igalia farm offline 2318 2319Jung-uk Kim (1): 2320 2321- FreeBSD: Disable support for "-mtls-dialect" for FreeBSD 2322 2323Juston Li (1): 2324 2325- util/cache_test: Fix racey Cache.List test 2326 2327Kai Wasserbäch (1): 2328 2329- fix(FTBFS): clc/clover: pass a VFS instance explicitly 2330 2331Karmjit Mahil (21): 2332 2333- tu: Fix push_set host memory leak on command buffer reset 2334- tu: Fix potential alloc of 0 size 2335- nir: Fix \`no_lower_set` leak on early return 2336- tu: Fix memory leaks on VK_PIPELINE_COMPILE_REQUIRED 2337- nir/algebraic: turn \`u{ge,lt} a, 1` to \`i{ne,eq} a, 0` 2338- nir,ir3: Add icsel_eqz 2339- nir: Fix the spelling of compare 2340- freedreno/rddecompiler: clang-format fix 2341- freedreno/rddecompiler: Fix some unsused function warnings 2342- ir3: Fix some Wsign-compare when compiling a generate-rd.cc 2343- util/idalloc: Fix util_idalloc_foreach() build issue 2344- util/idalloc: Minor refactor of util_idalloc_foreach() 2345- tu: Fix \`clear_values` leak 2346- tu: Fix FDM patchpoint memory leak 2347- tu: Fix leaking of some descriptor sets 2348- tu: Initialize tu_tiling_config even when tiling isn't possible 2349- tu: Free pre_chain patchpoint data 2350- util/simple_mtx: Add ASSERTED to parameter used only in an assert 2351- vulkan: Add inital vram-report-limit layer 2352- freedreno/replay: Define __user for msm_kgsl 2353- loader/wayland: Fix missing timespec.h include 2354 2355Karol Herbst (77): 2356 2357- nv/codegen: Do not use a zero immediate for tex instructions 2358- nvc0: return NULL instead of asserting in nvc0_resource_from_user_memory 2359- clover: drop support for nir drivers 2360- gallium: drop PIPE_SHADER_IR_NIR_SERIALIZED 2361- rusticl/kernel: fix kernel variant selection 2362- vtn: handle struct kernel arguments passed by value 2363- nir/lower_cl_images: lower scalar image_loads to vec4 2364- rusticl/mem: add restrictions for CL_DEPTH, CL_DEPTH_STENCIL and msaa images 2365- rusticl/image: fix clEnqueueFillImage for CL_DEPTH 2366- rusticl/device: advertize cl_khr_depth_images if supported 2367- rusticl: enable cl_khr_depth_images 2368- rusticl: check for overrun status when deserializing 2369- rusticl/kernel: convert name and type_name to Option<CString> 2370- rusticl/mesa: make driver_name() return a &CStr 2371- rusticl/program: check if provided binary pointers are null 2372- rusticl: rework query APIs 2373- rusticl/api: add a write_len_only variant for writing API properties 2374- rusticl/api: add a write_iter variant for writing API properties 2375- rusticl/program: use write_len_only for CL_PROGRAM_BINARIES 2376- rusticl/program: use write_iter for CL_PROGRAM_DEVICES 2377- rusticl/program: pass the slice directly for CL_PROGRAM_IL 2378- rusticl/program: use write_len_only for CL_PROGRAM_IL 2379- rusticl/platform: pass the slice directly for CL_PLATFORM_EXTENSIONS_WITH_VERSION 2380- rusticl/api: use constant arrays instead of Vecs for queries 2381- rusticl/context: use write_iter for CL_DEVICES_FOR_GL_CONTEXT_KHR 2382- rusticl/proc: make generated entry points unsafe 2383- rusticl/api: mark get_info and get_info_obj as unsafe 2384- rusticl/util: add Properties::is_empty() and len() 2385- rusticl/util: add Properties::iter() 2386- rusticl/util: make Properties::props private 2387- rusticl/util: reimplement Properties over Vec of scalars 2388- rusticl/api: simplify CLProp implementation of Properties 2389- rusticl/api: use Properties for 0 terminated arrays consistently 2390- rusticl/util: make Properties::from_ptr unsafe 2391- rusticl/api: remove Option around Properties 2392- rusticl/util: rename Properties::from_ptr to new 2393- rusticl/util: fix duplicate key detection in Properties::new 2394- rusticl/platform: silence static_mut_refs warning 2395- rusticl/util: fix ptr_to_integer_transmute_in_consts warning 2396- rusticl: fix clippy::needless-lifetimes 2397- rusticl: fix clippy::doc-lazy-continuation 2398- rusticl/queue: add a life check to prevent applications dead locking 2399- rusticl: stop using system headers for CL and GL 2400- include: Update the OpenCL headers to latest 2401- rusticl/mesa: remove PipeTransfer::res 2402- rusticl/mem: remove mem_type argument from new_image 2403- rusticl/device: remove unused functions 2404- rusticl/mesa/context: use Default for pipe_grid_info initialization 2405- rusticl/mesa: add missing files to meson.build 2406- rusticl/queue: make QueueContext::dev public 2407- rusticl/mem: pass around QueueContext instead of PipeContext 2408- rusticl/mesa/resource: port to NonNull 2409- rusticl/device: fix CL_DEVICE_HALF_FP_CONFIG query 2410- rusticl/device: fix default device enumeration 2411- rusticl/kernel: take set kernel arguments into account for CL_KERNEL_LOCAL_MEM_SIZE 2412- rusticl/kernel: fix image_size of 1D buffer images 2413- rusticl/mesa: set take_ownership to true for set_sampler_views 2414- rusticl/mesa: add PipeSamplerView wrapper 2415- rusticl/mesa: use PipeSamplerView over the raw type 2416- rusticl/kernel: create the sampler views earlier 2417- rusticl/mem: add functions to create sampler and image views to Image 2418- rusticl/mesa: rework image and sampler view creation APIs 2419- rusticl/kernel: store memory arguments as Weak references 2420- rusticl/device: add unsynchronized mapping functions to helper context 2421- rusticl/mem: simplify is_svm implementation 2422- rusticl/mem: add Allocation type 2423- rusticl/mem: reimplement has_same_parent and rename it to backing_memory_eq 2424- rusticl/mem: rework last user of get_parent() and remove it 2425- rusticl/mem: add Allocation::is_user_alloc_for_dev 2426- rusticl/mem: use get_res_for_access instead of get_res_of_dev 2427- trace: copy pipe_caps 2428- trace: add get_compute_state_info 2429- rusticl/mem: set bind flags for gl imports 2430- rusticl/mesa: add PipeContext::device_reset_status 2431- rusticl/queue: check device error status 2432- rusticl/kernel: call nir_lower_variable_initializers earlier 2433- rusticl/mem: do not apply offset with in copy_image_to_buffer 2434 2435Kenneth Graunke (35): 2436 2437- brw: Fix emit_a64_oword_block_header UNIFORM -> VGRF copies 2438- brw: Fix try_rebuild_source's ult32/ushr handling to use unsigned types 2439- nir: Use load_global_constant for reorderable nir_var_mem_global access 2440- nir/algebraic: Reassociate fadd into fmul in DP4-like pattern 2441- brw: Drop image deref handling from brw_analyze_ubo_ranges 2442- brw: Drop "regular uniform" concept from UBO push analysis 2443- brw: Drop a few crocus references in comments 2444- brw: Use nir_combined_align in brw_nir_should_vectorize_mem 2445- brw: Only consider components read for UBO loads 2446- brw: Only consider components read for UBO push analysis 2447- brw: Simplify choose_oword_block_size_dwords() 2448- nir: Allow large overfetching holes in the load store vectorizer 2449- anv: Don't consider nir_var_mem_global for vectorizer robustness checks 2450- brw: Tune vectorizer conditions to allow overfetching with holes 2451- brw: Fix register unit calculation in SIMD32 LOAD_PAYLOAD lowering 2452- brw: Allow SIMD32 math instructions on Xe2 2453- brw: Combine convergent texture buffer fetches into fewer loads 2454- iris: Tune the BO cache's bucket sizes 2455- brw: Don't rely on SIMD splitting in opt_combine_convergent_txfs 2456- brw: Limit maximum push UBO ranges to 64 registers in the NIR pass. 2457- brw: Don't shrink UBO push ranges in the backend 2458- brw: Delete pull constant lowering 2459- brw: Delete assign_constant_locations and push_constant_loc[] 2460- brw: Fix vectorizer hole_size condition after signedness change 2461- nir: Add a nir_def_first_component_read() helper 2462- brw: Add more safeguards against misaligned OWord Block messages 2463- brw: Skip fetching unread leading components of UBO loads 2464- brw: Make get_nir_src_imm() usable for non-32-bit-sizes. 2465- brw: Skip unnecessary work for trivial emit_uniformize of IMMs 2466- brw: Skip unread leading/trailing components in convergent block loads 2467- brw: Add a new MEMORY_MODE_CONSTANT option 2468- brw: Allow CSE of MEMORY_MODE_CONSTANT loads 2469- brw: Align and combine constant-offset UBO loads in NIR 2470- brw: Always use MEMORY_LOAD for load_ubo_uniform_block_intel intrinsics 2471- brw: Fix Xe2 spilling code to limit to SIMD32 rather than SIMD16 2472 2473Kevin Chuang (3): 2474 2475- anv: Implement encode shader to fit in ANV BVH 2476- anv: Add INTEL_DEBUG for bvh dump and visualization tools 2477- anv/bvh: Dump BVH synchronously upon command buffer completion 2478 2479Kevron Rees (1): 2480 2481- anv, drirc: Add workaround to speed up Spiderman reg allocation 2482 2483Konstantin (5): 2484 2485- nir/lower_non_uniform_access: Group accesses using the same resource 2486- radv/printf: Guard against helper invocations 2487- radv: Do not overwrite VRS rates when doing fast clears 2488- vulkan/meta: Add a pipeline cache 2489- vulkan: Fix the argument order of update_as 2490 2491Konstantin Seurer (39): 2492 2493- util: Fix some brackets in util_dynarray\_.*_ptr 2494- nir: Add missing access flags to print_access 2495- radv: Lower non-uniform access after vectorization 2496- amd: Add ac_shader_debug_info 2497- aco: Handle nir_debug_info_instr 2498- aco: Pass debug information to the driver 2499- radv: Add a helper for accessing the shader binary 2500- radv: Store debug info inside radv_shader 2501- radv: Dump nir shaders before compiling 2502- nir: Add a first_line parameter to gather_debug_info 2503- nir: Do not gather source locations for phis 2504- radv: Add RADV_DEBUG=nirdebuginfo 2505- gallivm: Add float operation behavior flags to lp_type 2506- gallivm: Preserve -0 and nan 2507- lavapipe: Implement VK_KHR_shader_float_controls2 2508- gallivm: Use an accurate log2 implementation for lodq 2509- lavapipe: Implement VK_KHR_compute_shader_derivatives 2510- radv: Fix encoding empty acceleration structures 2511- llvmpipe: Disable anisotropic filtering for explicit lod 2512- llvmpipe: Use a simpler and faster AF implementation 2513- llvmpipe: Remove unused AF code 2514- llvmpipe: Move max_anisotropy to static sampler state 2515- lavapipe: Advertise vulkan 1.4 2516- meson: Require glslangValidator when building lavapipe 2517- lavapipe: Check the pool type in handle_reset_query_pool 2518- meson: Include the loader subdir when building lavapipe 2519- gallivm: Take helper invocations into account when skipping branches 2520- nir/print: Print less unused shader info 2521- nir/tests: Improve shader creation 2522- nir/tests: Add a helper for comparing a shader against a string 2523- nir/tests: Add reference shaders 2524- nir: Add a test runner 2525- nir/print: Do not print trailing spaces after preds/succs 2526- docs: Add documentation for NIR unit testing 2527- llvmpipe: Fix half-pixel sample offset with AF 2528- llvmpipe: Avoid a crash when using 5 coords with AF 2529- radv/rmv: Use radv_rmv_log_resource_destroy more 2530- radv/meta: Stop using strings for meta keys 2531- gallivm: Remove loop limiting 2532 2533Koo, Anthony (1): 2534 2535- amd/vpelib: Add system event logging 2536 2537Lars-Ivar Hesselberg Simonsen (26): 2538 2539- panvk: Set fs.multisampled sysval for v10+ 2540- panvk: Add frag->frag barrier before resolve 2541- panvk: update expectations for G610 2542- pan/genxml: Fix decode of exception_handler 0x0 2543- pan/cs: Add mask support for reg_perm 2544- panvk: Build cmd_fb_preload on explicit fb_info 2545- panvk: Add incremental rendering support on v10+ 2546- panfrost: Disable AFRC texture/sampler reswizzle 2547- panvk: Disable AFBC for mutable formats on v7 2548- panfrost: Only allow AFBC(RGB) and AFBC(BGR) on v7 2549- panfrost: Limit reswizzle to AFBC formats 2550- panfrost: Decouple reswizzling from texture build 2551- panfrost: Standardize naming of sampler reswizzle 2552- panvk: Remove ZS texture_swizzle_replicate_x 2553- panvk: Fix descriptor decode 2554- panvk: Fix valgrind issue in nir_lower_descriptors 2555- panvk: Fix valgrind issue in panvk_compile_shaders 2556- pan/genxml: Fix vertex_packet Attribute on v9+ 2557- panvk: Use LD_VAR[_IMM] + ADs for varyings 2558- panvk: Limit AD allocation to max var loads in v9+ 2559- panvk: Use LD_VAR_BUF[_IMM] when possible 2560- panvk: Fix barriers in secondary cmdbufs w/o rp's 2561- panfrost: Do not evaluate_per_sample for non-MSAA 2562- Revert "panfrost: remove is_blit flag" 2563- Revert "panfrost: fix hang by using MALI_PIXEL_KILL_WEAK_EARLY in color preload" 2564- panvk: Set missing shader_modifies_coverage flag 2565 2566Leder, Brendan Steve (2): 2567 2568- amd/vpelib: Refactor OCSC and update missing check 2569- amd/vpelib: Move bg color 2570 2571Leonard Göhrs (1): 2572 2573- ci/lava: update lavacli from version 1.5.2 to 2.2.0 2574 2575Lina Versace (3): 2576 2577- anv: Sort extensions in enablement table 2578- anv: Update features.txt 2579- anv: Fix feature pipelineProtectedAccess 2580 2581LingMan (10): 2582 2583- mesa: Bump required Rust version to 1.78 2584- nak/hw_test: Use std::mem::offset_of!() 2585- compiler/rust: Use std::mem::offset_of!() 2586- mesa: Add rustfmt.toml 2587- rusticl: Use C-string literals 2588- rusticl: Use C-string literals for spirv extension names 2589- rusticl/cl_prop: Use C-string literals 2590- rusticl/core: Use C-string literals for XPlatManager::get_proc_address_func 2591- rusticl: Use C-string literals for NirShader::add_var 2592- rusticl: Use C-string literals for DiskCache::new 2593 2594Lionel Landwerlin (96): 2595 2596- anv: fix extent computation in image->image host copies 2597- anv: update shader descriptor resource limits 2598- anv: split generated draw flags from mocs/dword-count 2599- intel: make sure intel_wa.h can be included by opencl code 2600- anv: implement Wa_16011107343/22018402687 for generated draws 2601- brw: allocate physical register sizes for spilling 2602- anv: fix descriptor asserts 2603- anv: fix incorrect aspect flag for depth/stencil formats 2604- anv: fix missing push constant reallocation 2605- anv: prevent access to destroyed vk_sync objects post submission 2606- anv: track allocated descriptor pool sizes 2607- anv: indent driconf code 2608- anv: add a workaround for X4 Foundations 2609- anv: document the X4 Foundations workaround a bit more 2610- anv: move helpers out of genX_pipeline.c/anv_private.h 2611- anv: remove 3DSTATE_RASTER from pipeline 2612- anv: remove 3DSTATE_MULTISAMPLE from the pipeline 2613- anv: remove 3DSTATE_VF_STATISTICS from pipeline 2614- anv: pass anv_device to batch_set_preemption 2615- anv: rework vertex input helper 2616- anv: split vertex buffer emission in a different function 2617- anv: move gfx tracking values to anv_cmd_graphics_state 2618- anv: move tracking of tcs_input_vertices/fs_msaa_flags to hw state 2619- anv: split runtime flushing code for reuse 2620- brw: change fs_msaa flags checks to test compiled flag first 2621- brw: rename brw_sometimes to intel_sometimes 2622- brw: move barycentric_mode enum to intel_shader_enums.h 2623- brw: move fs_msaa_flags logic to intel_shader_enums.h 2624- fix 2625- Revert in correct commit "fix" 2626- anv: move primitive_topology to anv_gfx_dynamic_state 2627- anv: try to avoid using cmd_buffer in gfx runtime flushing 2628- anv: reuse device local variable in hw state emission 2629- anv: rework Wa_18038825448 to track state on anv_gfx_dynamic_state 2630- anv: avoid using cmd_buffer for TBIMR state computation 2631- anv: avoid using cmd_buffer for flushing runtime 2632- anv/iris: leave 4k alignments for clear colors with modifiers 2633- brw: use transpose unspill messages when possible 2634- anv: report formats supported by the common bvh framework 2635- anv: fix missing bindings valid dynamic state change check 2636- anv: set pipeline flags correct for imported libs 2637- vulkan: make acceleration structure debug markers virtual 2638- vulkan: add an enum for the build step 2639- vulkan: track encode step of the BVH building 2640- anv: add BVH building tracking through u_trace 2641- intel/decoder: fix COMPUTE_WALKER handling 2642- anv: document UBO descriptor range alignments 2643- blorp: use 2D dimension for 1D tiled images 2644- hk: fix timeline value type 2645- anv: fix index buffer size changes 2646- anv: limit the memcpy data for push constants 2647- vulkan/runtime: avoid emitting empty build_leaves 2648- anv: add tracepoints timestamp mode for empty dispatches 2649- anv: rework tbimr push constant workaround 2650- anv: ensure null-rt bit in compiler isn't used when there is ds attachment 2651- anv: use the correct MOCS for depth destinations 2652- intel: fix generation shader on Gfx9 2653- brw: introduce a new register type for the address register 2654- brw: use phys_nr() more in generation 2655- brw: split validation iteration into blocks 2656- brw: add infra to make use of the address register in the IR 2657- brw: add scheduler support for address registers 2658- brw: avoid having the scratch surface handle partially written 2659- brw: move final send lowering up into the IR 2660- brw: fix coarse_z computation on Xe2+ 2661- brw: handle load_printf_buffer_size intrinsic 2662- anv: handle printf buffer size relocations 2663- nir: make lower-level printf helper respect buffer size 2664- anv: update debug printf example code 2665- anv: remove print lowering 2666- blorp: disable PS shaders with depth/stencil HiZ ops 2667- brw: fix CSE with negation 2668- anv: don't look at pipelines to figure out CPS values 2669- compiler: add VARYING_BIT_PRIMITIVE_COUNT 2670- anv/Wa_18019110168: copy the primitive count writes 2671- anv/brw: rework primitive count writing 2672- libcl: add MIN2/MAX2 macros 2673- libcl_vk: add some vulkan enums/structures for DGC 2674- spirv: build vtn_bindgen for Anv/Iris 2675- brw/elk: move internal kernel parsing out of intel_clc 2676- meson: build mesa_clc for Anv/Iris 2677- intel/cl: switch to SPIRV as shader storage 2678- meson: rework mesa-clc=system handling 2679- intel: rework CL pre-compile 2680- meson: required SPIRV-Tools LLVM workaround on LLVM17+ 2681- intel: fix dependency for internal CL shaders 2682- anv: use flags for format capabilities 2683- anv: pass physical device to format helpers 2684- anv: add a drirc to disable border colors without format 2685- anv: expose A4B4G4R4_UNORM_PACK16 support with CBCWF is disabled 2686- anv: dirty pipeline & push constants after internal CS shaders 2687- anv: reduce alignment for small heaps 2688- brw: fixup scoreboarding for find_live_channels 2689- anv,driconf: Add sampler coordinate precision workaround for Dynasty Warriors 2690- anv: disable VF statistics for memcpy 2691- anv: ensure Wa_16012775297 interacts correctly with Wa_18020335297 2692 2693Lorenzo Rossi (1): 2694 2695- nvk: fix preprocess buffer alignment 2696 2697Louis-Francis Ratté-Boulianne (3): 2698 2699- panfrost: Split up allocation and packing of tiler descriptor 2700- panfrost: Select the effective tile size as part of pan_fb_info 2701- panfrost: Re-emit texture descriptor if the data size has changed 2702 2703Lu Yao (1): 2704 2705- zink: fix decomposed_attrs val error when zink_vs_key->size is 4 2706 2707Lucas De Marchi (1): 2708 2709- intel/tools: Fix Xe KMD error dump parser 2710 2711Lucas Stach (26): 2712 2713- etnaviv: drm: properly handle BO list member 2714- etnaviv: drm: assert mutual exclusivity between cache and zombie list 2715- etnaviv: drm: use list_first_entry 2716- etnaviv: stall after RS/BLT operation when draw_stall debug option is enabled 2717- etnaviv: Update headers from rnndb 2718- etnaviv: add debug switch to disable texture descriptor usage 2719- etnaviv: fix polygon offset for 24bpp depth buffers 2720- ci/etnaviv: drop gl-1.4-polygon-offset fail 2721- etnaviv: isa: fix typo in SRC2_USE map 2722- etnaviv: Update headers from rnndb 2723- etnaviv: clean up component use setting in linker 2724- etnaviv: fix flatshading 2725- etnaviv: emit full varying component use 2726- ci/etnaviv: drop GC2000 flat shading fails 2727- etnaviv: split dummy RT backing store from reloc 2728- etnaviv: fix rendering without vertex buffers/attributes 2729- ci/etnaviv: drop failures caused by missing vertex attributes 2730- etnaviv: fix polygon offset disable 2731- etnaviv: memcpy varying setup from stack 2732- etnaviv: emit varying interpolation state on halti5 2733- etnaviv: fix flatshading on halti5 GPUs 2734- etnaviv: only emit used PA_SHADER_ATTRIBUTES states 2735- etnaviv: track TS flushed status as bool 2736- etnaviv: dynamically partition the constant memory in unfied uniform mode 2737- etnaviv: allow more constants in unified uniform mode 2738- etnaviv: hwdb: fix lookup of GC3000 in i.MX6QP 2739 2740Lukas Lipp (1): 2741 2742- wsi: Fix wrong function name for lvp wsi metal surface 2743 2744M Henning (6): 2745 2746- nvk/cmd_buffer: Pass count to set_root_array 2747- nvk: Fix invalidation of NVK_CBUF_TYPE_DYNAMIC_UBO 2748- nvk: Remove params for dirty_cbufs_for_descriptors 2749- nvk: Fix two typos in comments 2750- nvk: Fix uninitialized var warnings in host_copy 2751- nak/hw_runner: Skip copy call for empty buffer 2752 2753Manuel (1): 2754 2755- gfxstream: Avoid repeated functionality 2756 2757Manuel Dun (4): 2758 2759- gfxstream: Using DETECT_OS_ANDROID from util instead of __ANDROID__ 2760- gfxstream: Using DETECT_OS_FUCHSIA from util instead of __Fushsia__ 2761- gfxstream: Using DETECT_OS_LINUX from util instead of __linux__ 2762- Gfxstream: Initial mingw "compilable" Windows version of mesa/gfxstream 2763 2764Marc Herbert (5): 2765 2766- docs: add "apt-get build-dep" and "dnf buildep" 2767- docs: cross-compile: add useful "apt" and "dnf" builddep commands 2768- docs: show how to use ccache when cross-compiling 2769- docs: show which pkg-config Fedora uses for cross-compilation 2770- docs: move cross c*_args from [properties] to [built-in options] 2771 2772Marek Olšák (353): 2773 2774- gallium/radeon: import libdrm_radeon source code, drop the dependency 2775- aco: remove unused TCS fields from aco_shader_info 2776- ac/nir: get pass_tessfactors_by_reg from nir_gather_tcs_info 2777- radeonsi: fix passing TCS wave ID from LS to HS for monolithic LS+HS 2778- radeonsi: don't overwrite info.tess._primitive mode when it can be correct 2779- radeonsi: get the value for load_tcs_primitive_mode_amd from shader info 2780- radeonsi: replace are_tessfactors_def_in_all_invocs with nir_gather_tcs_info 2781- radeonsi: reduce si_shader_key_ge::tes_prim_mode size to 2 bits 2782- radeonsi: remove unused function si_get_tcs_out_patch_stride 2783- radeonsi: don't set tess level outputs in patch_outputs_written unconditionally 2784- radeonsi: remove unused si_shader_info::output_readmask 2785- radeonsi: set \*outputs_written in scan_io_usage instead of later 2786- radeonsi: split outputs_written_before_tes_gs into ls_es_* and tcs_* masks 2787- radeonsi/ci: update navi31 failures 2788- glsl: add a helper for duplicated code calling nir_opt_varyings 2789- gallium: use struct nir_shader * type in finalize_nir instead of void * 2790- st/mesa: call pipe_screen::finalize_nir outside of st_finalize_nir 2791- gallium: add PIPE_CAP_CALL_FINALIZE_NIR_IN_LINKER 2792- st/mesa: add ST_DEBUG=xfb printing xfb info 2793- mesa: capture shaders to disk before invoking the linker 2794- nir/opt_varyings: add nir_io_always_interpolate_convergent_fs_inputs 2795- nir/opt_varyings: add nir_io_compaction_rotates_color_channels 2796- nir/opt_varyings: fix packing color varyings 2797- nir/opt_varyings: implement compaction without flexible interpolation 2798- nir/opt_varyings: don't count the cost of the same instruction multiple times 2799- radeonsi: fix buffer_size for emulated GS statistics 2800- radeonsi: fix an assertion failure in si_shader_ps with AMD_DEBUG=mono 2801- radeonsi: handle nir_intrinsic_component in kill_ps_outputs 2802- radeonsi: fix gl_FrontFace elimination when one side is culled 2803- radeonsi/ci: add options to test llvmpipe, softpipe, virgl, zink 2804- nir/print: print fb_fetch_output for variables 2805- nir/lower_pntc_ytransform: handle lowered IO 2806- nir/lower_clip: fixes for lowered IO without compact arrays 2807- nir/lower_clip: rewrite find_output to handle vec2/3 and make it readable 2808- nir/lower_fragcoord_wtrans: handle trimmed fragcoord loads 2809- nir/lower_two_sided_color: fix for lowered IO 2810- nir: add nir_io_semantics::fb_fetch_output_coherent 2811- nir: rename nir_io_glsl_opt_varyings to nir_io_dont_optimize and deprecate it 2812- nir: add nir_io_separate_clip_cull_distance_arrays to replace PIPE_CAP 2813- vc4/lower_blend: don't read non-existent channels 2814- nir: make use_interpolated_input_intrinsics a nir_lower_io parameter 2815- ac/surface: adjust HiZ enablement 2816- radeonsi: prepare for making SI_NGG_CULL_TRIANGLES/LINES VS only, rename them 2817- radeonsi: optionally return MESA_PRIM_UNKNOWN from si_get_input_prim 2818- radeonsi: rewrite/replace gfx10_ngg_get_vertices_per_prim 2819- radeonsi: return a better value for load_initial_edgeflags_amd 2820- radeonsi: clean up and rename gfx10_edgeflags_have_effect 2821- radeonsi: add helper si_shader_culling_enabled 2822- radeonsi: only compute and use min_direct_count on gfx7-8 2823- radeonsi: enable NGG culling for non-monolithic TES and GS 2824- radeonsi: don't use nir_io_dont_optimize because it's deprecated 2825- r300: don't lower sin/cos in finalize_nir 2826- nir/opt_varyings: use a hash table to make cloning SSA faster 2827- amd: import libdrm_amdgpu ioctl wrappers 2828- util,amd: add inlinable versions of drmIoctl/drmCommandWrite* 2829- nir: allow cloning indirect array derefs in nir_clone_deref_instr 2830- nir/lower_io_to_temporaries: fix interp_deref_at_* lowering 2831- radeonsi: don't call set_framebuffer_state in si_destroy_context 2832- radeonsi: handle a failure to create gfx_cs 2833- winsys/amdgpu: fix FD mismatch 2834- Revert "gbm: mark surface buffers as explicit flushed" 2835- nir/lower_clip: don't set cursor to fix crashes due to removed instructions 2836- nir/lower_clip: separate code for IO variables and intrinsics 2837- nir/lower_clip: set clip_distance_array_size outside of create_clipdist_vars 2838- nir/lower_clip: convert nir_lower_clip_gs to nir_shader_intrinsics_pass 2839- nir/lower_clip: implement ClipVertex lowering for GS + lowered IO correctly 2840- vc4: lower clip planes in st/mesa 2841- nir/opt_varyings: always call remove_dead_varyings in init_linkage 2842- nir/opt_varyings: add a default callback for varying_estimate_instr_cost 2843- nir/opt_varyings: replace options::lower_varying_from_uniform with a cost number 2844- nir/algebraic: use is_used_once in a few iand/ior patterns 2845- nir/algebraic: optimize (a & b) & (a & c) ==> (a & b) & c 2846- nir/algebraic: optimize (a | b) | (a | c) ==> (a | b) | c 2847- nir/algebraic: optimize (a & b) | (a | c) => a | c, (a & b) & (a | c) => a & b 2848- gallium: replace PIPE_SHADER_CAP_INDIRECT_INPUT/OUTPUT_ADDR with NIR options 2849- st/mesa: replace EmitNoIndirectInput / EmitNoIndirectOutput with NIR options 2850- util/bitset_test: test the return value of BITSET_TEST_RANGE_INSIDE_WORD better 2851- util/bitset: add BITSET_GET_RANGE_INSIDE_WORD 2852- nir/linking_helpers: don't promote interpolated varyings to flat 2853- nir/opt_varyings: remove redundant conditions from a while loop 2854- nir/opt_varyings: fix compaction with sparse indirect FS inputs 2855- nir/opt_varyings: count the number of unused components for compaction correctly 2856- nir/opt_varyings: fix max_slot for color varying compaction 2857- nir/opt_varyings: make top-level compaction code for TES, TCS, GS separate 2858- nir/opt_varyings: change try_move_postdominator param to nir_instr type 2859- amd,zink: remove options.varying_estimate_instr_cost callbacks 2860- nir/opt_varyings: propagate indirect uniform/UBO loads into the next shader 2861- nir/opt_varyings: add inter-shader code motion for uniform/UBO indexing 2862- nir/opt_varyings: fix getting deref variables for sysvals 2863- nir/opt_varyings: remove rare dead output stores after inter-shader code motion 2864- nir/opt_varyings: fix compile failures in the disabled PRINT code 2865- amd/ci: add piglit failures due to a overzealous test 2866- nir/lower_io_passes: lower indirect IO for TCS 2867- radeonsi: pass cull face state via user SGPRs for shader culling 2868- radeonsi: revert to always returning true for load_cull_any_enabled_amd 2869- radeonsi: try to fix Navi14 regression in debug builds 2870- radeonsi: don't compute total_direct_count in si_draw if it's unused 2871- radeonsi/ci: handle glinfo errors better 2872- radeonsi/ci: stop using a global flakes list, only use a per-chip flakes list 2873- radeonsi/ci: remove most flakes and some skips, update navi31 failures 2874- radeonsi/ci: remove --slow 2875- radeonsi/ci: update navi31 failures 2876- r600: fix a constant buffer memory leak for u_blitter 2877- ac/lower_ngg: improve streamout code generation for gfx12/ACO to match LLVM 2878- ac: update SPI_GRP_LAUNCH_GUARANTEE_* register values for gfx12 2879- ac/surface/gfx12: enable DCC 256B compressed blocks and reorder modifiers 2880- radeonsi/gfx12: set DB_RENDER_OVERRIDE based on stencil state 2881- radeonsi/gfx12: adjust HiZ/HiS logic 2882- ac/nir: reserve the first LDS vec4 for the HS tf0/1 group vote in TCS 2883- ac/nir: use s_sendmsg(HS_TESSFACTOR) to optimize writing tess factors for gfx11 2884- ac/nir: allow a TCS input to be available from both VGPRs and LDS 2885- ac,radv,radeonsi: enable TCS input reads from VGPRs for all compatible loads 2886- ac/nir: add new helpers for computing the TCS LDS/offchip size accurately 2887- radeonsi: remove unused parameter tcs_vgpr_only_inputs from si_get_nir_shader 2888- radeonsi: switch to the new TCS LDS/offchip size computation 2889- radv: switch to the new TCS LDS/offchip size computation 2890- ac/nir: call nir_gather_tcs_info only once for RADV 2891- nir/opt_varyings: set all IO types to float to facilitate full vectorization 2892- nir/opt_varyings: clear info->clip/cull_distance_array_size if relocated 2893- st/mesa: don't use nir_opt_fragdepth because it's incorrect with MSAA 2894- mesa: set correct XFB prim mode for draw validation after resuming XFB 2895- mesa: fix printing _NEW_* flags 2896- gallium: pass XFB primitive mode to set_stream_output_targets 2897- st/mesa: add a pass that unlowers IO intrinsics to variables 2898- glsl,st/mesa: always lower IO for GLSL, unlower IO for drivers 2899- v3d: enable uniform expression propagation from outputs to the next shader 2900- ci: update fail lists and trace checksums 2901- virgl/ci: disable virgl-traces because it doesn't upload results 2902- radeonsi/ci: don't copy skips.csv to the results directory 2903- radeonsi/ci: update failures and flakes 2904- radeonsi: fix a gfx10.3 regression due to a gfx12 change 2905- radeonsi: kill Z and stencil PS outputs if depth or stencil is disabled 2906- radeonsi/gfx11: fix alpha-to-coverage + alpha-to-one used together 2907- radeonsi: fix alpha-to-coverage + alpha-to-one used together for gfx6-10.3 2908- radeonsi: implement nir_opt_frag_depth using kill_z instead of the NIR pass 2909- radeonsi: eliminate shader code computing killed Z/S/samplemask PS outputs 2910- radeonsi: make NGG streamout output primitive type known at compile time 2911- radeonsi/gfx12: fix DrawTransformFeedback(stream != 0) 2912- radeonsi/gfx12: tune streamout performance 2913- radeonsi: make nir->info and si_shader_info::base identical 2914- radeonsi: remove some uses of enum pipe_shader_type 2915- radeonsi: make si_init_shader_args static 2916- radeonsi: call si_init_shader_args in si_get_nir_shader 2917- radeonsi: use nir->info instead of sel->info.base 2918- radeonsi: disable luminance alpha formats on gfx6 2919- radeonsi,radv: fix incorrect min_esverts for NGG subgroup calculation 2920- ac: remove unused code 2921- ac/llvm: remove unused code 2922- radeonsi/ci: update failures 2923- radeonsi: fix a TCS regression 2924- radeonsi: switch si_get_blitter_vs to IO intrinsics 2925- radeonsi: remove unused code 2926- amd: update addrlib 2927- radeonsi: fix a front face regression (crash) 2928- nir/opt_load_store_vectorize: make hole_size signed to indicate overlapping loads 2929- radv: reduce maxGeometryShaderInvocations to 32 2930- ac/nir: handle disabled PS VGPRs in ac_nir_load_arg_at_offset 2931- amd: lower load_pixel_coord in NIR 2932- amd: lower load_frag_coord in NIR 2933- amd: lower load_local_invocation_id in NIR 2934- amd: lower load_first_vertex/base_instance/draw_id/view_index in NIR 2935- amd: lower load_invocation_id in NIR 2936- amd: lower load_sample_id in NIR 2937- amd: lower load_sample_pos in NIR 2938- amd: lower load_frag_shading_rate in NIR 2939- amd: lower load_front_face in NIR 2940- ac,radeonsi: move load_vector_arg flags to common code 2941- amd: lower load_barycentric_pixel/centroid/sample in NIR 2942- amd: lower load_barycentric_at_offset in NIR 2943- amd: lower load_gs_wave_id_amd in NIR 2944- amd: lower load_vertex_id/instance_id and overwrite_vs_arguments in NIR 2945- radeonsi: don't return 0 from si_get_max_workgroup_size 2946- ac/nir: extract a load_subgroup_id lowered helper 2947- amd: lower load_local_invocation_index in NIR 2948- amd: lower load_subgroup_invocation in NIR 2949- amd: lower load_tess_rel_patch_id/primitive_id/tess_coord and overwrite.. in NIR 2950- ac/llvm: remove already lowered cases 2951- ac/nir: lower more loads in ac_nir_lower_intrinsics_to_args instead of drivers 2952- ac/nir: clean up ac_nir_lower_indirect_derefs 2953- ac/nir: add helper ac_nir_load_arg_upper_bound 2954- ac/nir: set arg_upper_bound_u32 for vs_rel_patch_id 2955- ac/nir: split local_invocation_ids to 3 separate VGPR inputs 2956- ac/nir: set upper ranges for range analysis while lowering system values 2957- radeonsi: lower sysval intrinsics as late as possible 2958- amd: optimize atomics before lowering intrinsics 2959- radeonsi: use nir_opt_sink 2960- radeonsi: use nir_opt_move 2961- vulkan: silence an unused variable warning 2962- llvmpipe: silence an unused result warning 2963- util/disk_cache: silence unused result warnings 2964- nir: set nir_io_semantics::num_slots to at least 1 in build helpers 2965- nir: set src_type and dest_type to float implicitly for IO build helpers 2966- nir: don't set num_slots/src/dest_type/write_mask when they're set automatically 2967- nir: flip the early exit condition in nir_lower_io_temporaries 2968- nir: remove redundant option linker_ignore_precision 2969- nir: use IO intrinsics in nir_lower_bitmap 2970- nir: use IO intrinsics in nir_lower_drawpixels 2971- mesa: remove unused PROGRAM_SYSTEM_VALUE 2972- mesa: remove unused PROGRAM_WRITE_ONLY 2973- st/mesa: fold st_translate_prog_to_nir into prog_to_nir 2974- st/mesa: run DCE before st_unlower_io_to_vars 2975- st/mesa: use IO intrinsics in st_nir_lower_fog 2976- st/mesa: use IO intrinsics in st_nir_lower_position_invariant 2977- st/mesa: switch ATI_fs to IO intrinsics 2978- st/mesa: unlower IO for internal shaders if needed 2979- st/mesa: switch Z/S DrawPixels shaders to IO intrinsics 2980- st/mesa: switch GL_SELECT shader to IO intrinsics 2981- st/mesa: switch st_nir_make_passthrough_shader to IO intrinsics 2982- st/mesa: switch st_pbo_create_vs and st_pbo_create_gs to IO intrinsics 2983- st/mesa: switch PBO create_fs to IO intrinsics 2984- st/mesa: switch st_nir_make_clearcolor_shader to IO intrinsics 2985- st/mesa: don't use nir_copy_var 2986- st/mesa: recompute IO bases for ARB_vp/fp 2987- glsl: remove unused code 2988- glsl: fix corruption due to blake3 hash not being set for nir_opt_undef 2989- radeonsi: ignore PIPE_RESOURCE_FLAG_TEXTURING_MORE_LIKELY for TC-compatible HTILE 2990- radeonsi: simplify and fix enable_tc_compatible_htile_next_clear logic 2991- radeonsi: re-enable non-TC-compatible HTILE for write-only Z/S 2992- mesa: switch ARB_vp/fp to IO intrinsics 2993- mesa: switch fixed-func fragment program to IO intrinsics 2994- nir/algebraic: use is_used_once for comparison patterns 2995- nir/algebraic: add and improve pack/unpack patterns 2996- nir/algebraic: optimize pack_split(unpack(a).x, unpack(a).y) -> a 2997- radeonsi: fix a perf regression due to slow reply from GEM_WAIT_IDLE for timeout=0 2998- radeonsi: always use RADEON_USAGE_DISALLOW_SLOW_REPLY 2999- ac: update ATOMIC_MEM definitions 3000- ac/nir: sort xfb info to facilitate vectorization of xfb stores 3001- ac/nir: vectorize streamout stores for legacy pipeline optimally 3002- ac/nir/ngg: vectorize streamout stores for NGG optimally 3003- ac/nir/ngg: fold so_vertex_index * so_stride into immediate offset 3004- ac/nir/ngg: export positions after streamout to improve performance 3005- ac,radeonsi: scalarize overfetching loads 3006- radeonsi: lower descriptors sooner to allow vectorizing descriptor loads 3007- amd: vectorize SMEM loads aggressively, allow overfetching for ACO 3008- radeonsi: don't set BREAK_PRIMGRP/WAVE_AT_EOI when tessellation is disabled 3009- radeonsi: only set BREAK_PRIMGRP/WAVE_AT_EOI when TES/GS need PrimID sysval after TES 3010- radeonsi/gfx12: enable alt_hiz_logic 3011- radeonsi/gfx12: set DIS_PG_SIZE_ADJUST_FOR_STRIP after shader compilation 3012- radeonsi/gfx12: use ACO if LLVM is 19 or older 3013- radeonsi/gfx12: use ACO for streamout because it's faster 3014- mesa: rework enablement of force_gl_names_reuse 3015- mesa: enable GL name reuse by default for all drivers except virgl 3016- ac/nir: remove broadcast_last_cbuf because it can be deduced from NIR 3017- ac/nir: split ac_nir_lower_ps into 2 passes 3018- nir: add barycentric coordinates src to load_point_coord_maybe_flipped 3019- ac: use Z_EXPORT_FORMAT=32_AR for Z + Alpha mrtz exports 3020- ac/llvm: lower vector load_const in NIR 3021- ac/llvm: remove the low-optimizing compiler option 3022- radeonsi: add si_screen::use_aco to shader cache key to fix shader cache failures 3023- radeonsi: remove unused variables from si_shader_context (LLVM) 3024- radeonsi: make many shader functions static or move them to .c files 3025- radeonsi: remove unused functions 3026- nir: add next_stage param to nir_slot_is_varying & nir_remove_sysval_output 3027- Revert "ac/llvm: enable wqm for ac_build_quad_swizzle from ac_build_fs_interp_mov" 3028- nir: add a pass that moves output stores to the end of the shader 3029- st/mesa: move VS & TES output stores to the end before unlowering IO 3030- mesa: switch fixed-func vertex program to IO intrinsics 3031- st/mesa: assert that all incoming shaders use lowered IO 3032- st/mesa: remove dead/no-op code due to IO being always lowered 3033- glsl: remove dead code due to IO being always lowered 3034- glsl: simplify nir_lower_io_to_temporaries logic 3035- nir: remove dead code due to IO being always lowered in st/mesa 3036- st/mesa: inline st_finalize_nir_before_variants 3037- nir: remove handling IO variables from passes used by st/mesa 3038- gallium/u_threaded: move tc_batch_execute after all call functions 3039- gallium/u_threaded: make the execute function table private 3040- gallium/u_threaded: use TC_END_BATCH to terminate the loop 3041- gallium/u_threaded: replace the function table with a switch and direct calls 3042- gallium/u_threaded: inline all tc_call functions 3043- gallium/u_threaded: sort cases in batch_execute by their occurrence 3044- zink/ci: skip KHR-Single-GL46...SizedDeclarationsPrimitive due to random timeout 3045- dri: put shared-glapi into libgallium.*.so 3046- glapi: stop using the remap table 3047- glapi: remove the remap table 3048- loader: improve the existing loader-libgallium non-matching version error 3049- glapi: rename exported symbols so as not to conflict with old libglapi 3050- freedreno/ci: skip a dmat3 div test timing out 3051- radv: don't call ac_nir_lower_ps_early 3052- ac/nir: optimize front_face in ac_nir_lower_ps_early 3053- ac/nir: lower sample_pos in ac_nir_lower_ps_early 3054- ac/nir: lower barycentric_at_offset/sample in ac_nir_lower_ps_early 3055- ac/nir: lower fbfetch_output in ac_nir_lower_ps_early 3056- ac/nir: return progress from ac_nir_lower_ps_early 3057- ac/nir: return progress from ac_nir_lower_ps_late 3058- ac/nir: handle FRAG_RESULT_COLOR with dual src blending in ac_nir_lower_ps_early 3059- ac/nir: switch passes to use nir_shader_intrinsics_pass 3060- ac/nir: drop 16x EQAA support from ac_get_ps_iter_mask 3061- ac/nir: clamp vertex color outputs in the right place 3062- radeonsi: sample shading state fixes 3063- ac,aco,radeonsi: replace SampleMaskIn with 1 << SampleID if full sample shading 3064- ac/nir: simplify force_*_sample_interp options in ac_nir_lower_ps_early 3065- ac/nir: simplify force_*_center_interp options in ac_nir_lower_ps_early 3066- ac/nir: optimize barycentric_at_sample(sample_id) in ac_lower_ps_early 3067- ac/nir: optimize frag_coord <-> pixel_coord in ac_nir_lower_ps_early 3068- ac/nir: eliminate sample_mask_in without MSAA in ac_nir_lower_ps_early 3069- ac/nir: cosmetic stuff for ac_nir_lower_ps 3070- aco: implement replacing frag_coord with pixel_coord in PS prolog 3071- aco: simplify how broadcast_last_cbuf is implemented in PS epilog 3072- aco: implement replacement of sample_mask_in with helper_invocation in PS prolog 3073- ac/nir: compute ddx/ddy for barycentric_at_offset at the beginning of shaders 3074- ac/nir: lower sample_pos to load_sample_positions_amd when frag_coord is center 3075- nir/opt_varyings: handle user barycentrics 3076- mesa: enable GL name reuse for virgl 3077- radeonsi: disallow compute queues on Raven/Raven2 due to hangs 3078- ac/nir: clamp vertex color outputs in the right place 3079- radeonsi: get sample positions from user SGPRs instead of memory 3080- radeonsi: fix PS prolog not counting used fragcoord VGPRs correctly 3081- radeonsi: implement replacing frag_coord with pixel_coord at draw time 3082- radeonsi: don't set the alpha ref user SGPR if alpha test doesn't use it 3083- radeonsi: simplify how broadcast_last_cbuf is implemented for PS epilogs 3084- radeonsi: use load_pixel_coord for polygon stipple lowering 3085- radeonsi: remove si_nir_kill_ps_outputs and use ac_nir_lower_ps_early instead 3086- radeonsi: add load_polygon_stipple_buffer_amd instead of using si_shader_args 3087- radeonsi: call si_init_gs_output_info in si_get_nir_shader 3088- radeonsi: add si_nir_shader_ctx holding parameters from si_get_nir_shader 3089- radeonsi: call si_nir_late_opts unconditionally 3090- radeonsi: set the "first" parameter of si_nir_opts correctly 3091- radeonsi: simplify how the NIR name of shader variants is modified 3092- radeonsi: cosmetic changes in get_nir_shader 3093- radeonsi: reorder NIR passes in get_nir_shader (part 1) 3094- radeonsi: reorder NIR passes in get_nir_shader (part 2) 3095- radeonsi: reorder NIR passes in get_nir_shader (part 3) 3096- radeonsi: split and restructure get_nir_shader 3097- radeonsi: get LS+HS and ES+GS together in get_nir_shader instead of separately 3098- radeonsi: set uses_vmem_load/sampler in get_nir_shaders 3099- radeonsi: move/rewrite PS color input gathering for shader variants 3100- radeonsi: use barycentrics from load_point_coord_maybe_flipped 3101- radeonsi: lower indirect indexing sooner 3102- radeonsi: move spi_ps_input_config functions up 3103- radeonsi: split si_fixup_spi_ps_input_config 3104- radeonsi: get SPI_PS_INPUT_ENA from shader variant NIR for ACO 3105- radeonsi: minor restructuring of si_llvm_compile_shader 3106- radeonsi: verify that SPI_PS_INPUT_ENA from LLVM is equal to ACO 3107- radeonsi: remove ac_shader_config from si_shader_part 3108- radeonsi: precompute COMPUTE_PGM_RSRC3 3109- radeonsi: set SHARED_VGPR_CNT for compute for ACO 3110- radeonsi: set SHARED_VGPR_CNT for gfx shaders for ACO 3111- radeonsi: gather PS inputs from shader variant NIR 3112- radeonsi: don't set BASE in si_nir_lower_ps_color_input 3113- radeonsi: remove si_shader_info code that is no longer needed 3114- radeonsi: implement replacement of sample_mask_in with helper_invocation 3115- radeonsi: ignore pipe_rasterizer_state::force_persample_interp 3116- radeonsi: fix interpolateAt* with non-GL4 ARB_sample_shading 3117- radeonsi/ci: add more gfx11 flakes 3118- radeonsi: set gl_FragCoord to pixel center to fix GLCTS failures 3119- radeonsi: validate BITSET_TEST_RANGE_INSIDE_WORD assertion at compile time 3120- radeonsi: remove SI_TRACKED__UNUSED_GAP 3121- radeonsi: dead code removal and move some code out of headers 3122- radeonsi: remove redundant divergence analysis and smem flagging 3123- radeonsi: remove an incorrectly defined modifier 3124- winsys/amdgpu: disable DCC for gfx12 when using AMD_FORCE_FAMILY 3125- ac/fake_hw_db: deobfuscate GPU name strings 3126- gallium,st/mesa: allow reporting compile failures from create_vs/fs/.._state 3127 3128Mark Collins (5): 3129 3130- util: Add file modification notifier utility 3131- tu/util: Support toggling TU_DEBUG options at runtime 3132- tu/lrz: Check for TU_DEBUG(nolrz) late 3133- freedreno/docs: Document TU_DEBUG_FILE 3134- util/u_debug: Ignore newlines in \`parse_*_string` 3135 3136Martin Krastev (7): 3137 3138- svga/ci: enable vmware farm 3139- svga/ci: set vmware piglit job parallelism to 2 3140- svga/ci: triage piglit failures 3141- svga/ci: update svga/ci KERNEL_TAG 3142- svga/ci: drop FDO_CI_CONCURRENT to 1 3143- svga/ci: disable vmware farm 3144- svga/ci: enable vmware farm 3145 3146Martin Roukala (né Peres) (39): 3147 3148- zink/ci: document new-ish vangogh flakes 3149- ci: disable mupuf's farm 3150- Revert "ci: disable mupuf's farm" 3151- ci: disable mupuf's farm 3152- Revert "ci: disable mupuf's farm" 3153- freedreno-ci: document more a618-gl flakes 3154- freedreno-ci: document a a750-gl flake 3155- turnip/ci: document the a750-vkcts expectations 3156- turnip/ci: bump the vkcts a750 timeout by 15 minutes 3157- turnip/ci: skip a vkd3d test that causes a GPU hang on a750 3158- nvk/ci: update the ga106 expectations 3159- zink/ci: update the nvk-ga106 expectations 3160- zink/ci: update the radv expectations 3161- radv/ci: update the vkcts expectations 3162- ci/test: make the .b2c-${arch}-test-* jobs provide a default b2c 3163- ci/tests: de-duplicate the b2c version between architectures 3164- ci/test: uprev to b2c v0.9.14 3165- freedreno/ci: use the default b2c 3166- r300/ci: use the default b2c 3167- i915g/ci: use the default b2c version 3168- ci/b2c: modernize the job description to use run_* 3169- ci/b2c: run the machine registration check before the test container 3170- radeonsi/ci: update the vangogh expectations 3171- radeonsi/ci: run on ACO changes 3172- radeonsi/ci: run a fraction of glcts-vangogh in pre-merge 3173- ci/init-stage2: use the common scripts from the build artifact 3174- ci/b2c: use the runner description rather than ID 3175- ci/b2c: allow defining a boot watchdog 3176- freedreno/ci: use the boot watchdog to ensure the a750 boots 3177- zink/ci: update nvk expectations 3178- zink/ci: update RADV expectations 3179- radeonsi/ci: update the vangogh expectations 3180- ci/b2c: allow jobs to select a file in the dtb url 3181- ci/b2c: allow using another initrd that contains firmware 3182- freedreno/ci: uprev the a750 kernel to msm-next 3183- ci: fix the artifact name 3184- zink/ci: use the debian-built-testing for nvk 3185- ci/b2c: fix the S3 artifact for amd64 manual vk/gl 3186- turnip/ci: re-introduce the \`multiviewport` flakes 3187 3188Mary Guillemard (56): 3189 3190- agx: Add support for EGL_NV_context_priority_realtime 3191- panfrost: Report default value for GROUP_PRIORITIES_INFO in drm-shim 3192- pan/kmod: Expose medium priority on panfrost 3193- panvk: Implement global priority extensions 3194- panvk: Advertise VK_EXT_tooling_info 3195- panvk: Advertise VK_KHR_shader_non_semantic_info 3196- panvk: Advertise VK_KHR_shader_relaxed_extended_instruction 3197- panvk: Implement VK_KHR_zero_initialize_workgroup_memory 3198- bi: Execute nir_opt_algebraic after nir_lower_pack 3199- panvk: Implement VK_EXT_sampler_filter_minmax for v10 3200- panvk: Only flag rw_nc pool as uncached on v10+ 3201- panvk: Take rasterization samples into account in draw 3202- panfrost: Remove faulty assert in cs_loop_conditional_* 3203- panvk: Wire occlusion queries to internals 3204- panvk: Implement occlusion queries for JM 3205- panvk: Implement occlusion queries for CSF 3206- panvk: Expose precise occlusion queries 3207- panvk: Advertise VK_EXT_host_query_reset 3208- panvk: Enable depthClamp and depthBiasClamp 3209- panvk: Enable shaderInt16 3210- panvk: Advertise VK_KHR_index_type_uint8 3211- panvk: Advertise VK_KHR_map_memory2 3212- panvk: Disable integer array indices clamping 3213- panvk: Advertise VK_EXT_image_robustness 3214- panvk: Advertise VK_EXT_pipeline_robustness 3215- panvk: Call vk_free on queue array instead of vk_object_free 3216- panvk: Use vk_zalloc for queue array allocation 3217- panvk: Update Mali-G52 CI baseline 3218- panvk: Add a nightly job for Mali-G52 3219- nak: Fix 8-bit selection for vectors 3220- nak: Simplify 16-bit vector selection to not use try_from 3221- meson: Add mesa-clc and install-mesa-clc options 3222- meson: Add precomp-compiler and install-precomp-compiler options 3223- asahi: Remove unneeded dependencies for asahi_clc 3224- util/bitpack_helpers: Use UINT64_MAX instead of ~0ULL 3225- util/bitpack_helpers: Make fixed packs CL safe 3226- nir,agx: Allow nir_precomp_print_blob to print a static array 3227- libcl: Respect NDEBUG for assert 3228- panforst: Update ForEachMacros 3229- pan/genxml: Move pack_header to an external file 3230- libcl: Add VkQueryType and VkQueryResultFlagBits definitions 3231- pan/genxml: Switch unpack to use uint32_t 3232- pan/genxml: Emit struct details before pack function 3233- pan/genxml: Move [un]pack internals to use packed structs 3234- pan/genxml: Enforce explicit packed types on pan_[un]pack 3235- pan/genxml: Switch pan_section_ptr to cast to packed type 3236- pan/genxml: Switch [un]pack codegen to macros 3237- pan/genxml: Switch __gen_unpack to macros 3238- panfrost: Fix group priorities in drm-shim 3239- panfrost: Fix PROGRESS_LOAD destination register 3240- pan/bi: Properly encode LEA_BUF_IMM 3241- pan/bi: Remove shift lanes invalid encodings 3242- pan/bi: Fix invalid CLPER encoding 3243- pan/bi: Use 2D dimension with TEX_FETCH with CUBE on Valhall 3244- pan/decode: Fix indirect branch calculation for 64-bit 3245- panvk: Disallow unknown GPU models early in physical device init 3246 3247Matt Turner (16): 3248 3249- anv: Align anv_descriptor_pool::host_mem 3250- vulkan: Skip memcpy() call if passed null pointers 3251- anv: Protect memcpy/memset/qsort calls against NULL arguments 3252- anv: Avoid null ptr dereference 3253- intel: Avoid unaligned pointer access 3254- vulkan: Avoid pointer aliasing 3255- nir: Get correct number of components 3256- intel/decoder: Avoid duplicate symbols when expat is not available 3257- brw: Avoid reading past the end of \`p->store` 3258- brw: Pass brw_codegen to next_offset 3259- brw: Bounds check access to \`p->store` 3260- brw: Pass number and sizeof separately to calloc 3261- elk: Avoid reading past the end of \`p->store` 3262- elk: Pass brw_codegen to next_offset 3263- elk: Bounds check access to \`p->store` 3264- elk: Pass number and sizeof separately to calloc 3265 3266Matthew Brost (1): 3267 3268- anv/xe: Bind queue per anv_queue 3269 3270Mauro Rossi (4): 3271 3272- nvk/android: Avoid building error in nak bindings 3273- nvk/android: Advertise Vulkan 1.1 for Android 12L and lower 3274- nvk/android: Add support for ANDROID_native_buffer 3275- android: remove shared-glapi building rules 3276 3277Maíra Canal (3): 3278 3279- v3dv: Check multiple DRM primary nodes before picking the display fd 3280- v3dv: delete \`v3dv_debug.h` 3281- v3dv: use Mesa log infrastructure instead of using stderr 3282 3283Mel Henning (27): 3284 3285- nak: Fix two warnings of elided_named_lifetimes 3286- gallium/winsys/nouveau: Don't mark the api PUBLIC 3287- nak: Add nak_nir_mark_lcssa_invariants 3288- compiler/rust/bitset: Fix the bitset iterator 3289- compiler/rust: Fix running tests 3290- compiler/rust/bitset: Add a basic test 3291- compiler/rust/bitset: Removed unused start param 3292- compiler/rust/bitset: Make BitSetIter private 3293- compiler/rust/bitset: impl FromIterator 3294- compiler/rust/bitset: Remove impl Not 3295- compiler/rust/bitset: Add a lazy expression API 3296- compiler/rust/bitset: Take a stream in union_with 3297- nak: Migrate liveness to new bitset expression api 3298- compiler/rust/bitset: Don't expose words 3299- compiler/rust/bitset: Test next_unset() 3300- nak: Add ShaderModel::hw_reserved_gprs() 3301- nak: Add gpr_limit_from_local_size 3302- nir_validate: Handle unstructured control flow 3303- nak: lower_load_ssbo_descriptor modifies cf 3304- nir: Update num_blocks in sort_unstructured_blocks 3305- nvk: Fix an assertion in nvk_slm_area_ensure 3306- nak: Return VK_ERROR_UNKNOWN on assertion failure 3307- nak: Fix a spelling error 3308- nak/opt_copy_prop: Fix IAdd3 overflow check 3309- nak/opt_copy_prop: Add force_alu_src_type 3310- nak/opt_copy_prop: Force alu src for IAdd2X/IAdd3X 3311- driconf: force_vk_vendor on Deep Rock Galactic+NVK 3312 3313Mi, Yanfeng (2): 3314 3315- anv:Fix memory grow calculation overflow issue 3316- anv:increase instruction heap to 3Gb 3317 3318Michael Cheng (2): 3319 3320- anv : Add tracepoint for as_build 3321- intel : Expose Shader hashes for utrace and Perfetto 3322 3323Michel Dänzer (4): 3324 3325- Revert "util/mesa-db: Further simplify mesa_db_compact" 3326- Revert "util: Use persistent array of index entries" 3327- Revert "winsys/amdgpu: fix FD mismatch" 3328- winsys/amdgpu: Always use amdgpu_device_get_fd for aws->fd 3329 3330Michel Zou (1): 3331 3332- ac/gpu_info: Fix missing prototype mingw error 3333 3334Mike Blumenkrantz (38): 3335 3336- zink: restrict implicit feedback loop detection using miplevels/layers 3337- mesa: use default params for clearbuffer functions 3338- zink: rework query result checking 3339- zink: use internal map flag for qbos 3340- glsl: make gl_ViewID_OVR visible to all shader stages 3341- glsl: enable OVR_multiview if OVR_multiview2 is enabled 3342- lavapipe: stop storing texture handle for samplers 3343- vk/sampler: split out sampler init from create 3344- lavapipe: split out sampler init from create 3345- lavapipe: split out bda descriptor function params from struct 3346- lavapipe: fix bitmask type for sampler updating 3347- lavapipe: move workgraph lowering up and delete pipeline param 3348- lavapipe: unsupport NV_device_generated_commands 3349- lavapipe: stop using pipeline layouts in some places 3350- lavapipe: handle VK_REMAINING_ARRAY_LAYERS with HIC 3351- lavapipe: fix 3D->2D blitting 3352- lavapipe: abort on unsupported depth copy ops 3353- lavapipe: support zs<->color copies 3354- lavapipe: maintenance8 3355- zink: enable maintenance8 3356- glsl: plumb num_views down to shader_info::view_mask 3357- zink: fix viewport detection when switching last stage shaders 3358- zink: add radv ci fail 3359- zink: disable shader objects when viewmask is set 3360- zink: fix replacing incompatible pipelines 3361- egl: never select swrast for vmwgfx 3362- zink: deduplicate VkDevice and VkInstance 3363- aco: exclude novalidateir from codegen flags 3364- zink: check for bound gfx stages before dereferencing 3365- zink: add zink_resource_reference() util function 3366- zink: refcount needs_present resource 3367- ci: mark radv-raven-traces-restricted with allow_failure 3368- zink: emit SpvCapabilityDemoteToHelperInvocation for IsHelperInvocation 3369- zink: also refcount needs_present from frontbuffer flush 3370- zink: guard rebar check against fallback heap detection 3371- radv: fix error reporting for VkExternalMemoryTypeFlagBitsKHR 3372- zink: only enable unsynchronized_texture_subdata with HIC 3373- zink: never try to oom flush during unsync texture upload 3374 3375Mike Lothian (1): 3376 3377- gallium/radeon: Fix r600_pci_ids.h include 3378 3379Mykhailo Skorokhodov (1): 3380 3381- drirc/anv: force_vk_vendor=-1 for Bellwright 3382 3383Nanley Chery (22): 3384 3385- anv: Support non-0/1 sRGB fast-clear colors on gfx9 3386- anv: Store fast-clear colors with the view swizzle 3387- anv: Drop fast-clear value conversion check 3388- intel/blorp: Assert 3D Ys fast-clear restriction 3389- intel/isl: Allow CCS on 3D 64bpp+ Tile64 3390- intel: Allow CCS on 3D surfaces for gfx120 3391- intel/isl: Fix DecompressInL3 assignment on gfx12.5 3392- anv: Enable storage accesses with modifiers on gfx12+ 3393- anv: Enable more storage compression on gfx12+ 3394- anv: Only consider R32 image formats as supporting atomics 3395- anv: Allow compressed memtypes with default buffer types 3396- anv: Slow clear if fast-clear cost is not mitigated 3397- iris: Reduce fast-clear post-amble flushes 3398- iris: Use L3 Fabric flush in fast-clear post-amble on TGL 3399- anv: Reduce fast-clear post-amble synchronization 3400- anv: Use L3 Fabric flush in fast-clear post-amble on TGL 3401- anv: Drop bpc check for non-zero fast clears 3402- Revert "anv: turn off non zero fast clears for CCS_E" 3403- anv: Inline can_fast_clear_with_non_zero_color 3404- anv: Allow more single subresource fast-clears with FCV 3405- anv: Drop can_fast_clear_with_non_zero_color() 3406- anv: Limit slow clear heuristic to ACM and prior 3407 3408Patrick Lerda (8): 3409 3410- r600: fix the evergreen sampler when the minification and the magnification are not identical 3411- r600: restructure r600_create_vertex_fetch_shader() to remove memcpy() 3412- r600: ensure that the last vertex is always processed on evergreen 3413- r600: evergreen stencil/depth mipmap blit workaround 3414- r600: reverse fix spec ext_packed_depth_stencil getteximage 3415- winsys/radeon: fix radeon_winsys_bo_from_handle() related race condition 3416- r600: fix r600_init_screen_caps() has_streamout issue 3417- r600: fix r600_init_shader_caps() has_atomics issue 3418 3419Paulo Zanoni (3): 3420 3421- brw: don't forget the base when emitting SHADER_OPCODE_MOV_RELOC_IMM 3422- brw: don't read past the end of old_src buffer in resize_sources() 3423- brw: increase brw_reg::subnr size to 6 bits 3424 3425Pavel Ondračka (27): 3426 3427- r300: group KIL for R300/R400 3428- r300: run nir_opt_algebraic in the backend 3429- r300: always transform sin/cos input for fs 3430- r300/ci: update RV410 CI expectations 3431- ci: bring back some i915g testing 3432- i915/ci: update CI expectations 3433- r300: disable ATI2N textures on R400 3434- r300: disable microtiling for scanout buffers 3435- r300/ci: update CI expectations 3436- r300: fix uninitialized use in transform_vertex_ROUND 3437- nir: add support for clamping in nir_lower_tex_shadow 3438- etnaviv: always clamp shadow sampler comparison reference value 3439- r300: fix presubtract assert 3440- r300: move shadow lowering to NIR 3441- r300: reswizzle some shadow texture calculations to use w channel 3442- r300: delete backend shadow lowering code 3443- r300: use ssa-like form for gl_FragCoord transformation 3444- r300: add some more nir cleanup compiler passes 3445- r300: use ssa-like form for backend texture lowering 3446- r300: don't allocate fs registers when translating from NIR 3447- r300: get rid of the register rename pass 3448- r300: get rid of some texture fixups 3449- r300: remove support for register arrays from nir_to_rc 3450- r300: fix memory leak in contant remaping 3451- ci: fix debian-build-testing BUILDTYPE 3452- i915/ci: use debian-build-testing instead of debian-testing 3453- i915: rework shader compile failures reporting 3454 3455Peyton Lee (5): 3456 3457- frontends/va: add support for VAProcColorStandardExplicit 3458- frontends/va: add support for VAProcColorStandardExplicit 3459- frontends/va: function process_frame has return value 3460- radeonsi/vpe: optimize software functions 3461- radeonsi/vpe: add destroy_fence function 3462 3463Philipp Zabel (11): 3464 3465- teflon: Use correct convolution params struct 3466- teflon: Mark dilated convolutions and fused activation as not supported 3467- teflon: Support fused ReLU activation 3468- etnaviv/nn: Enable fused ReLU activation 3469- teflon: Add is_signed parameter to ml_subgraph_invoke and ml_subgraph_read_output 3470- etnaviv/nn: Add support for signed 8-bit tensors 3471- teflon/tests: prep test executor for signed convolutions 3472- teflon/tests: Enable int8 tests 3473- etnaviv/ml: Create combined input tensors for addition first 3474- teflon: Reject per-axis quantization 3475- teflon: Support fused ReLU6 activation via output saturation 3476 3477Pierre-Eric Pelloux-Prayer (40): 3478 3479- radv: set info->family_overridden when RADV_FORCE_FAMILY is used 3480- ac/surface: add flags to surface metadata 3481- radeonsi: refuse to import texture with family_overriden being set 3482- ac: rename ac_surface_test_common -> ac_fake_hw_db 3483- ac: add 'polaris12' gpu to ac_fake_hw_db 3484- ac: switch AMD_FORCE_FAMILY handling to using ac_fake_hw_db 3485- radeonsi/tests: update expected results 3486- ac/perfcounter: fix buffer overflow 3487- dri: Remove unused function 3488- radeonsi/gfx12: disable display dcc for front buffer rendering 3489- radeonsi: disable DCC for PIPE_BIND_USE_FRONT_RENDERING 3490- glx: return BadMatch for invalid reset notification strategy 3491- ac/nir: remove prim_stride_ret arg from ngg_build_streamout_buffer_info 3492- radeonsi: use bytes units in streamout 3493- DEPENDENCY: ac/llvm: fix sparse code handling 3494- radeonsi: fallback to util_blitter_draw_rectangle 3495- radeonsi/tests: update results 3496- gl/spirv: update subgroup_size if GroupNonUniform is used 3497- amd: move all uses of libdrm_amdgpu to ac_linux_drm 3498- amd: amdgpu-virtio implementation 3499- ac/virtio: disable userptr and local buffers 3500- ac/virtio: disable timeline syncobj support 3501- radeonsi: enable virtio native context support 3502- radv: enable virtio native context support 3503- radv/virtio: disable syncobj timeline support 3504- ac/virtio: add virtio-only AMDGPU_GEM_CREATE flag 3505- radeonsi, radv, virtio: use AMDGPU_GEM_CREATE_VIRTIO_SHARED 3506- radeonsi: clear the debug callback on ctx destroy 3507- ttn: init source_blake3 and name from tgsi_shader_info 3508- ac/llvm: add wqm param to ac_build_quad_swizzle 3509- ac/llvm: enable wqm for ac_build_quad_swizzle from ac_build_fs_interp_mov 3510- radeonsi: do not use std::max 3511- glx: fix glx-create-context-invalid-es-version 3512- dri: use _checked variants of xcb requests 3513- dri: deal with ARGB1555 3514- egl/wayland: validate dri_screen_display_gpu before use 3515- amd: add ac_drm_device_get_cookie 3516- radeonsi: use ac_drm_device_get_cookie 3517- radeonsi: update si_need_gfx_cs_space upper bound 3518- radeonsi: disable dcc when external shader stores are used 3519 3520Qiang Yu (81): 3521 3522- ac/surface/tests: support all block sizes 3523- ac/surf: add more modifiers to gfx12 supported list 3524- radeonsi: disable use_gfx12_xfb_intrinsic when use ACO 3525- util/blake3: add _mesa_blake3_from_printed_string 3526- radeonsi: add AMD_FORCE_SHADER_USE_ACO for debug 3527- nir: do not generate b2i64 when driver want to lower it 3528- aco: enable gfx12 support for radeonsi 3529- radeonsi: fix unigine heaven crash when use aco on gfx8/9 3530- aco: fix voffset missing when buffer store base >=4096 3531- radeonsi: fix OpenCL shader compile fail 3532- ac/nir: lower access for shared and scratch memory 3533- ac,radv: move ac_nir_lower_bit_size_callback to common place 3534- radeonsi: fix OpenCL piglit tests fails when using ACO 3535- radeonsi: replace ac_nir_lower_subdword_loads 3536- ac: remove ac_nir_lower_subdword_loads 3537- radeonsi: fix global access ACO compile fail when OpenCL 3538- radeonsi: enable ACO by default for pre-GFX10 GPUs 3539- radeonsi: unify disk cache id no matter use_aco or not 3540- gallium: add pipe_caps struct definition 3541- gallium: add u_init_pipe_screen_caps 3542- asahi: add agx_init_screen_caps 3543- crocus: add crocus_init_screen_caps 3544- d3d12: add d3d12_init_screen_caps 3545- etnaviv: add etna_init_screen_caps 3546- freedreno: add fd_init_screen_caps 3547- i915: add i915_init_screen_caps 3548- iris: add iris_init_screen_caps 3549- lima: add lima_init_screen_caps 3550- llvmpipe: add llvmpipe_init_screen_caps 3551- nouveau/nv30: add nv30_init_screen_caps 3552- nouveau/nv50: add add nv50_init_screen_caps 3553- nouveau/nvc0: add nvc0_init_screen_caps 3554- panfrost: add panfrost_init_screen_caps 3555- r300: add r300_init_screen_caps 3556- r600: add r600_init_screen_caps 3557- radeonsi: add si_init_screen_caps 3558- softpipe: add softpipe_init_screen_caps 3559- svga: add svga_init_screen_caps 3560- tegra: init screen caps 3561- v3d: add v3d_init_screen_caps 3562- vc4: add vc4_init_screen_caps 3563- virgl: add virgl_init_screen_caps 3564- zink: add zink_init_screen_caps 3565- nine: change cap macros to use pipe_caps access 3566- egl,gallium,glx: replace dri_get_screen_param with pipe_caps access 3567- mesa/st: enable extension use pipe_caps access 3568- egl,gallium,gbm,mesa: replace get_param with pipe_caps access 3569- gallium,mesa: replace get_paramf with pipe_caps access 3570- rusticl: use pipe_caps access 3571- asahi: remove agx_get_param and agx_get_paramf 3572- crocus: remove crocus_get_param and crocus_get_shader_paramf 3573- d3d12: remove d3d12_get_param and d3d12_get_paramf 3574- etnaviv: remove etna_screen_get_param and etna_screen_get_paramf 3575- freedreno: remove fd_screen_get_param and fd_screen_get_paramf 3576- i915: remove i915_get_param and i915_get_paramf 3577- iris: remove iris_get_param and iris_get_paramf 3578- lima: remove lima_screen_get_param and lima_screen_get_paramf 3579- llvmpipe: remove llvmpipe_get_param and llvmpipe_get_paramf 3580- nouveau/nv30: remove nv30_screen_get_param and nv30_screen_get_paramf 3581- nouveau/nv50: remove nv50_screen_get_param and nv50_screen_get_paramf 3582- nouveau/nvc0: remove nvc0_screen_get_param and nvc0_screen_get_paramf 3583- panfrost: remove panfrost_get_param and panfrost_get_paramf 3584- r300: remove r300_get_param and r300_get_paramf 3585- r600: remove r600_get_param and r600_get_paramf 3586- radeonsi: remove si_get_param and si_get_paramf 3587- softpipe: remove softpipe_get_param and softpipe_get_paramf 3588- svga: remove svga_get_param and svga_get_paramf 3589- tegra: remove tegra_screen_get_param and tegra_screen_get_paramf 3590- v3d: remove v3d_screen_get_param and v3d_screen_get_paramf 3591- vc4: remove vc4_screen_get_param and vc4_screen_get_paramf 3592- virgl: remove virgl_get_param and virgl_get_paramf 3593- zink: remove zink_get_param and zink_get_paramf 3594- gallium: remove get_param and get_paramf 3595- docs,src: replace doc and comments for PIPE_CAP with pipe_caps 3596- gallium,mesa: remove uint surffix from pipe_caps 3597- radeonsi: remove si_screen.max_texel_buffer_elements 3598- etnaviv: remove min/max_texture_gather_offset init 3599- lavapipe: fix min_vertex_pipeline_param 3600- gallium: fix ddebug and noop screen caps init 3601- radeonsi: fix has_non_uniform_tex_access info 3602- radeonsi: fix GravityMark corruption when use aco 3603 3604Rebecca Mckeever (14): 3605 3606- panvk: Use vk_image::drm_format_mod instead of pan_image::layout.modifier 3607- panvk: Replace tab with spaces 3608- panvk: Enable multiplane images and image views 3609- pan/texture: s/pan_image_view_get_zs_image/pan_image_view_get_zs_plane/ 3610- pan/texture: s/pan_image_view_get_rt_image/pan_image_view_get_color_plane/ 3611- pan/texture: Accept holes in the pan_image_view::planes array 3612- pan/desc: Pass an image to pan_force_clean_write_rt() 3613- pan/desc: Add a pan_image_view_get_s_plane() helper and use it 3614- panvk: Support D32_S8 as a multiplanar format 3615- pan/format: Use HW version to determine siting for YUV 422 formats 3616- pan/texture: Only use plane_chroma_2p for chroma planes 3617- util/hash_table: Add _mesa_hash_table_u64_replace() 3618- panvk: Allow a 32-bit binding value in desc id key and use 64-bit keys 3619- panvk: Fix assertion in is_disjoint() 3620 3621Rhys Perry (72): 3622 3623- nir: add more intrinsics to nir_intrinsic_can_reorder 3624- nir/algebraic: optimize bcsel(ieq(b, 0), a, shift(a, b)) 3625- nir/algebraic: optimize ushr(a, ishl(iand(b, 3), 3)) 3626- ac/nir: add ACCESS_CAN_REORDER to lowered load_global_constant 3627- aco: optimize nir_op_shfr with <32 src1 3628- nir,aco,ac/llvm: add nir_op_alignbyte_amd 3629- nir_lower_mem_access_bit_sizes: support 64-bit offsets 3630- nir_lower_mem_access_bit_sizes: add nir_mem_access_shift_method 3631- nir_lower_mem_access_bit_sizes: pass access to callback 3632- nir_lower_mem_access_bit_sizes: support load_constant 3633- aco,ac/nir: flag loads to use smem in NIR 3634- radv,ac/nir: lower sub-dword loads using nir_lower_mem_access_bit_sizes 3635- aco: remove load byte_align 3636- radv,ac/nir: split global access using nir_lower_mem_access_bit_sizes 3637- nir/algebraic: fix iabs(ishr(iabs(a), b)) optimization 3638- nir/algebraic: check bit sizes in lowered unpack(pack()) optimization 3639- nir/lcssa: fix premature exit of loop after rematerializing derefs 3640- glsl/list: add comments above foreach macros 3641- glsl/list: add and use helpers in foreach_list_typed macros 3642- glsl/list: remove parenthesis in foreach_list_typed macros 3643- glsl/list: remove underscores in foreach_list_typed macros 3644- nir/opt_move_discards_to_top: use nir_tex_instr_has_implicit_derivative 3645- nir: fix return value of nir_instr_move for some cases 3646- nir/opt_move_discards_to_top: remove recursion 3647- nir/opt_move_discards_to_top: update variable name 3648- nir/opt_move_discards_to_top: use nir_intrinsic_can_reorder 3649- nir/opt_move_discards_to_top: add more intrinsics to add_src_to_worklist 3650- nir/opt_move_discards_to_top: allow multiple discards to be moved 3651- nir/lcssa: use nir_intrinsic_can_reorder 3652- nir/algebraic: add ddxy to is_only_used_as_float 3653- nir/algebraic: add is_used_once to bcsel(, bcsel()) opts 3654- nir/algebraic: optimize more bcsel(, bcsel()) 3655- aco: add SSA repair pass 3656- aco: use repair pass for LCSSA workaround 3657- aco: require WQM after demote in control flow 3658- aco: skip code if exec is empty 3659- aco/tests: add tests for empty exec masks 3660- aco: don't use uniform continues if exec might be empty 3661- aco: make small_vec copyable 3662- aco: use small_vec in RegCounterMap 3663- nir/tests: fix SSA dominance in opt_if_merge tests 3664- aco/gfx12: insert wait between VMEM WaW 3665- aco: force linear for event_vmem_sample and event_vmem_bvh 3666- aco: don't CSE p_shader_cycles_hi_lo_hi 3667- radv: constant fold after lowering memory accesses 3668- radv: fix expanded push constant loads when all are inlined 3669- radv: skip loading unused push constants 3670- ac/nir: have ac_nir_lower_mem_access_bit_sizes preserve >128 bit SMEM 3671- nir: make load_helper_invocation non-reorderable 3672- nir/move_discards_to_top: don't move across more intrinsics 3673- nir: make ballot ALU and mbcnt_amd operations reorderable 3674- aco: fix max_workgroup_count[0] 3675- aco: decrease max_workgroup_size 3676- radv: increase maxComputeWorkGroupCount[0] 3677- aco/tests: fix skip_lines=True with remaining characters in matches 3678- aco/util: fix bit_reference::operator&= 3679- aco: use VOP3 v_mov_b16 if necessary 3680- v3dv: fix SSA dominance error 3681- microsoft/compiler: invalidate loop analysis in dxil_nir_lower_double_math 3682- microsoft/compiler: repair SSA in dxil_nir_split_tess_ctrl 3683- d3d12: fix phi handling in d3d12_lower_primitive_id 3684- d3d12: store only once in d3d12_emit_points 3685- nir: rerun loop analysis if the parameters change 3686- nir/loop_analyze: use a sparse array and stop indexing SSA defs 3687- nir/gcm: stop preserving nir_metadata_loop_analysis 3688- nir/liveness: stop requiring instr indices 3689- nir/validate: validate metadata 3690- nir/validate: preserve dominance during SSA validation 3691- nir/validate: validate ssa dominance by default 3692- radv: set has_image_bvh_intersect_ray for null winsys 3693- aco: don't use divergence information for most ALU defs 3694- nir/divergence: assume all instructions are loop invariant if no continues 3695 3696Rob Clark (11): 3697 3698- vdrm+tu+fd: Make cross-device optional 3699- freedreno/registers: Add GMU_CORE_FW_VERSION 3700- freedreno/a6xx: Align lrz setup with tu 3701- freedreno/a6xx: Add nolrzfc debug option 3702- freedreno/a6xx: Align lrz height to 32 3703- tu: Align lrz height to 32 3704- freedreno/a6xx: Use LATE_Z with OC + discard 3705- freedreno/a6xx: Fix timestamp emit 3706- ir3: Add preamble instr count metric 3707- freedreno/pps: Fix multiple counter collection runs 3708- tu: Fix raytracing query with vdrm 3709 3710Robert Mader (2): 3711 3712- v3d: Support SAND128 base modifier 3713- freedreno: Support offset query for multi-planar planes 3714 3715Rohan Garg (5): 3716 3717- intel/compiler: disable mesh autostrip for WA 16020916187 3718- iris: use CALLOC_STRUCT instead of calloc for readability 3719- isl: disable aux when creating uncompressed TileY/Tile64 surfaces from compressed ones 3720- anv: refactor choose_isl_tiling_flags to pass fewer arguments 3721- iris: assert that we're not exporting a TILE64 surface 3722 3723Roland Scheidegger (1): 3724 3725- llvmpipe: Fix overflow issues calculating loop iterations for aniso 3726 3727Roman Stratiienko (1): 3728 3729- v3dv/android: Suppress AHB-related log spam 3730 3731Ruijing Dong (2): 3732 3733- radeosi/vcn: enable EFC for VCN5.0+ when gfx >= 12 3734- radeonsi/vcn: center mv map buffer changed in vcn5.x 3735 3736Russell Greene (1): 3737 3738- perfetto: fix macos compile 3739 3740Sagar Ghuge (30): 3741 3742- anv: Enable MCS_CCS compression on Gfx12+ 3743- blorp: Use the calculated execution mask 3744- anv: Update include dir for anv_tests 3745- anv: Split GRL code path in separate file 3746- anv: Add header to track BVH data structures 3747- anv: Add shader to build BVH header 3748- anv: Add shader to copy acceleration structures 3749- anv: Implement cmd_fill_buffer_addr callback 3750- anv: Move update buffer code in helper 3751- anv: Implement write_buffer_cp callback 3752- anv: Implement flush_buffer_write_cp callbck 3753- anv: Implement cmd_dispatch_unaligned callback 3754- anv: Implement acceleration structure API 3755- anv: Add helper to copy data from src to dest anv_address 3756- intel: Use the common RT BVH framework 3757- intel/compiler: Extend nir_intrinsic_load_topology_id_intel for xe3 3758- intel/genxml: Drop morton walk field from Xe2 3759- intel/genxml: Update COMPUTE_WALKER_BODY 3760- intel: Use Morton compute walk order 3761- intel/genxml: Update SAMPLER_STATE structure 3762- anv: Switch to ANISOTROPIC_FAST filter mode 3763- iris: Switch to ANISOTROPIC_FAST filter mode 3764- intel: Set correct maxComputeSharedMemorySize for Xe3+ 3765- intel/genxml: Add coarse pixel related changes 3766- anv: Add pipelined coarse pixel state 3767- intel/genxml: Update URB related instructions and structures 3768- iris: Use 3DSTATE_URB_ALLOC_* instructions 3769- blorp: Use 3DSTATE_URB_ALLOC_* instructions 3770- anv: Use 3DSTATE_URB_ALLOC_* instructions 3771- intel/brw/xe3+: Don't compile SIMD32 if there is ray queries 3772 3773Sam Lantinga (1): 3774 3775- util: Fixed crash in HEVC encoding on 32-bit systems 3776 3777Samuel Pitoiset (241): 3778 3779- aco: cleanup using fixed registers in the trap handler shader 3780- aco: save/restore SCC in the trap handler shader 3781- aco: use scalar buffer stores for dumping SGPRS from the trap on GFX8 3782- aco: add a helper to dump SGPR to memory for the trap handler 3783- aco: fix storing SQ_WAVE_STATUS in the trap handler shader 3784- aco: declare phys regs for tba_hi/tma_hi 3785- radv,aco: dump m0 and exec from the trap handler 3786- vulkan/runtime: return same cmdbuf level from the command pool freelist 3787- docs: add missing documentation for RADV_DEBUG=psocachestats 3788- radv: remove unused parameter to radv_fill_nir_compiler_options() 3789- radv: dump the trap handler shader with RADV_DEBUG=dump_trap_handler 3790- aco: do not reorder s_trap instructions 3791- radv: cleanup printing SGPRS dumped from the trap handler 3792- radv,aco: dump more SQ_WAVE regs from the trap handler 3793- radv,aco: add a separate function to compile the trap handler shader 3794- aco: simplify postprocessing the trap handler shader 3795- radv,aco: use the trap handler layout struct while compiling the shader 3796- radv: fix the TMA descriptor size 3797- radv: compute the TMA BO size instead of using a constant 3798- radv,aco: save/restore overwritten VGPRs in the trap handler shader 3799- nir: add nir_intrinsic_debug_break instruction 3800- spirv: handle NonSemantic.DebugBreak to emit nir_debug_break() 3801- aco: emit nir_intrinsic_debug_break 3802- radv: emit nir_debug_break instructions when the trap handler is enabled 3803- radv: do not always invalidate L2 for GPUs with non-coherent RBs on GFX10+ 3804- radv: move the GFX11 special case for mips to radv_image_is_pipe_misaligned() 3805- radv: determine the first mip that is pipe misaligned on GFX10+ 3806- radv: use vk_image_view_subresource_range() when possible 3807- radv: pass the image subresource range to radv_{src,dst}_access_flush() 3808- radv: optimize the pipe misaligned L2 cache invalidation on GFX11 3809- aco: fix saving/restoring VGPRS in the trap handler on GFX9 3810- aco: use a 64-bit mov to save exec in the trap handler shader 3811- aco: add a new variant for vop1() with two operands 3812- aco: fix validation for v_movrels_b32 and friends 3813- aco: restore m0/exec before exiting the trap handler 3814- aco: use all invocations from the current wave in the trap handler 3815- aco: save/restore VGPRS on GFX8 in the trap handler shader 3816- aco: drop the second M0 operand for s_set_gpr_idx_on 3817- radv,aco: dump VGPRS from the trap handler shader 3818- radv: mark live invocations when dumping VGPRS with the trap handler 3819- radv: dump SPIR-V and NIR for the faulty shader detected with the trap 3820- radv: fix ignoring src stage mask when dst stage mask is BOTTOM_OF_PIPE 3821- radv: consider VK_PIPELINE_STAGE_2_NONE like BOTTOM_OF_PIPE 3822- radv: destroy meta resources properly when creating the device failed 3823- radv: add a helper to destroy a logical device 3824- radv: add a new drirc option to disable DCC for mips and enable it for RDR2 3825- radv,aco: dump LDS from the trap handler 3826- radv: remove VK_VALVE_descriptor_set_host_mapping 3827- radv: fix skipping on-disk shaders cache when not useful 3828- radv: mark VERDE (GFX6) as Vulkan 1.3 conformant 3829- radv: fix dumping debug/perftest options when there are holes 3830- radv: add a pipeline helper to skip shaders cache 3831- radv: fix dumping the trap handler shader disassembly 3832- radv: fix printing with RADV_DEBUG=psocachestats 3833- radv: only pass relevant stages when emitting DGC push constants 3834- radv: capture shader executable info at shader creation time 3835- radv: allow shaders caching with RADV_DEBUG=hang and the trap handler 3836- vulkan: add MESA_VK_TRACE_PER_SUBMIT 3837- radv: finish tools after cleaning meta resources 3838- radv: add new start/stop sqtt helpers for capturing with SQTT 3839- radv: add support for capturing RGP per-submit 3840- radv: add address binding report support for BOs imported with a fd 3841- radv: add address binding report support for BOs imported with a ptr 3842- radv: add a small helper to dump VM fault with the GPU hang report 3843- radv: dump address binding report with RADV_DEBUG=hang 3844- radv: try to detect use-after-free with address binding report 3845- zink/ci: skip one more modifier test on POLARIS10 3846- radv: promote VK_KHR_dynamic_rendering_local_read to core 1.4 API 3847- radv: promote VK_KHR_global_priority to core 1.4 API 3848- radv: promote VK_KHR_index_type_uint8 to core 1.4 API 3849- radv: promote VK_KHR_line_rasterization to core 1.4 API 3850- radv: promote VK_KHR_maintenance5 to core 1.4 API 3851- radv: promote VK_KHR_maintenance6 to core 1.4 API 3852- radv: promote VK_KHR_map_memory2 to core 1.4 API 3853- radv: promote VK_KHR_push_descriptor to core 1.4 API 3854- radv: promote VK_KHR_shader_subgroup_rotate to core 1.4 API 3855- radv: promote VK_EXT_pipeline_robustness to core 1.4 API 3856- radv: add new Vulkan 1.4 features/properties 3857- radv: advertise Vulkan 1.4 on GFX8+ 3858- radv: bump VKCTS conformance version to 1.4.0.0 for some GFX8+ GPUs 3859- radv/ci: mark few tests as expected failures 3860- ac/parse_ib: fix parsing SDMA CONSTANT_FILL packet 3861- ac/parse_ib: print VA for the SDMA CONSTANT_FILL/WRITE packets 3862- radv: fix stencil only copies of depth/stencil images with SDMA 3863- radv: enable DGC IES for compute with ESO 3864- radv: fix initializing HTILE when the image has VRS rates 3865- ci: update VKCTS main to a9f7069b9a5ba94715a175cb1818ed504add0107 3866- radv: remove redundant drirc for incorrect dual-source blending 3867- radv: add radv_disable_dcc_stores and enable for Indiana Jones: The Great Circle 3868- radv: only dump device name info on Linux with RADV_DEBUG=hang 3869- radv: dump the Mesa version with RADV_DEBUG=hang 3870- radv/meta: add missing vk_meta_device_finish() 3871- radv/meta: move vk_meta_device_init() to radv_device_init_meta() 3872- radv: disable alphaToOne except for Zink 3873- ac/nir: export alpha to MRTZ.a and one to MRT0.a for alpha-to-one on GFX11 3874- aco: export alpha to MRTZ.a and one to MRT0.a for alpha-to-one on GFX11 3875- radv: fix alpha-to-coverage with alpha-to-one when MRTZ is also exported 3876- radv: remove remaining discard to demote options 3877- radv: fix disabling DCC for stores with drirc 3878- radv: simplify determining some fragment shader info with epilogs 3879- radv: fix alpha-to-coverage with alpha-to-one without MRTZ 3880- Revert "radv: disable alphaToOne except for Zink" 3881- spirv: add an options to lower SpvOpTerminateInvocation to OpKill 3882- radv: add radv_lower_terminate_to_discard and enable for Indiana Jones 3883- radv: mark HAWAII (GFX7) as Vulkan 1.3 conformant 3884- radv: report same buffer aligment for DGC preprocessed buffer 3885- Revert "radv: fix creating unlinked shaders with ESO when nextStage is 0" 3886- radv/ci: fix expected list of failures for TAHITI 3887- radv: fix missing variants for the last VGT stage with shader object 3888- ci: uprev vkd3d-proton to c965c1351fd6915a65bb7f647319536252a24a93 3889- radv: fix capturing RT pipelines that return VK_OPERATION_DEFERRED_KHR for RGP 3890- radv: reorganize query code by adding separate begin/end helpers 3891- radv: remove dead code in radv_CmdCopyQueryPoolResults() 3892- radv: add few more query helpers for copying results 3893- radv: only enable emulated mesh/task shader queries on GFX10.3 3894- radv/nir: fix checking if task shader invocations query is enabled 3895- radv: fix getting the number of vertices per prim for the last VGT stage 3896- radv: rename GDS queries to emulated queries 3897- radv/nir: simplify lowering of query intrinsics 3898- radv: cleanup enabling the global BO list when BDA is used 3899- radv: check descriptor indexing features for enabling the global BO list 3900- radv: rework emitting SPI_SHADER_Z_FORMAT 3901- radv: rename color output state to fragment output state 3902- radv: add support for VK_PRIMITIVE_TOPOLOGY_META_RECT_LIST_MESA 3903- radv: use VK_PRIMITIVE_TOPOLOGY_META_RECT_LIST_MESA for meta pipelines 3904- radv: pass extra graphics pipeline create info using pNext 3905- radv/meta: rework creating meta pipelines for query resolves 3906- radv/meta: convert the copy/fill pipelines to vk_meta 3907- radv/meta: convert the copy VRS to HTILE pipelines to vk_meta 3908- radv/meta: convert the FMASK expand pipelines to vk_meta 3909- radv/meta: convert the FMASK copy pipelines to vk_meta 3910- radv/meta: convert the DCC retile pipelines to vk_meta 3911- radv/meta: convert the HTILE expand CS pipelines to vk_meta 3912- radv/meta: convert the DCC decompress CS pipelines to vk_meta 3913- radv/meta: convert the clear HTILE mask pipelines to vk_meta 3914- radv/meta: convert the DCC comp-to-single pipelines to vk_meta 3915- radv/meta: convert DGC pipeline layout to vk_meta 3916- radv/meta: convert the query resolve pipelines to vk_meta 3917- radv/meta: convert the image-to-buffer pipelines to vk_meta 3918- radv/meta: convert the buffer-to-image pipelines to vk_meta 3919- radv/meta: convert the image-to-image pipelines to vk_meta 3920- radv/meta: convert the clear image pipelines to vk_meta 3921- radv/meta: convert the compute resolve pipelines to vk_meta 3922- radv/meta: remove radv_meta_create_compute_pipeline() 3923- vulkan: add a new vk_meta option to use the rect list pipeline path 3924- vulkan: use the meta pipeline cache for graphics pipelines 3925- radv/meta: convert the HTILE expand GFX pipelines to vk_meta 3926- radv/meta: convert the HW resolve GFX pipelines to vk_meta 3927- radv/meta: convert the fast-clear GFX pipelines to vk_meta 3928- radv/meta: convert the blit GFX pipelines to vk_meta 3929- radv/meta: convert the clear GFX pipelines to vk_meta 3930- radv/meta: convert the resolve GFX pipelines to vk_meta 3931- radv/meta: use only one push constant range for blit2d pipelines 3932- radv/meta: convert the blit2d GFX pipelines to vk_meta 3933- radv/meta: remove unused radv_meta_create_xxx() helpers 3934- radv: fix destroying DGC pipelines 3935- radv: disable RT with LLVM completely 3936- radv/meta: remove a workaround for building accel structs with LLVM 3937- radv/meta: always initialize emulated etc2 on-demand 3938- radv/meta: move initializing emulated astc to radv_device_init_meta() 3939- radv/meta: stop initializing RT accel structs 3940- radv: fix adding the BO to cmdbuf list when emitting buffer markers 3941- radv/meta: fix loading the meta pipeline cache 3942- radv/meta: reduce length of some cache keys 3943- radv/meta: add radv_meta_get_noop_pipeline_layout() 3944- radv/meta: do not create redundant pipeline layout objects 3945- radv: disable logic op for float/srgb formats 3946- ac/descriptors: fix configuring NBC views on GFX12 3947- aco: fix VS prologs on GFX12 3948- radv: disable VRS coarse shading with 8x MSAA on GFX12 3949- radv: configure the VRS surface swizzle mode on GFX12 3950- radv: fix programming WALK_ALIGN8_PRIM_FITS_ST on GFX12 3951- radv: program DB_RENDER_OVERRIDE correctly on GFX12 3952- ac/nir: fix lowering subgroup ID for compute shaders on GFX12 3953- ac/nir: fix a comment typo in load_subgroup_id_lowered() 3954- ac/gpu_info: add cp_dma_use_L2 3955- radv: fix CP DMA clears/copies on GFX12 3956- aco: always use ds_bpermute for shuffle/rotate on GFX12 3957- radv: fix configuring the attribute ring size on GFX12 3958- radv: rename attr_ring to ge_rings 3959- radv: change the BASE_HI field for VGT_TF_MEMORY_BASE_HI on GFX12 3960- ac/surface: honor RADEON_SURF_PREFER_xxx_ALIGNMENT on GFX12 3961- radv: advertise VK_MESA_image_alignment_control on GFX12 3962- radv: fix emitting SPI_SHADER_GS_OUT_CONFIG_PS with NULL FS on GFX12 3963- radv: fail to initialize when the AMD GPU generation is unsupported 3964- radv: mark AMD CDNA as unsupported 3965- radv: add GFX12 support to the null winsys 3966- ac/nir: fix skipping streamout when no buffers are bound on GFX12 3967- vulkan: Update XML and headers to 1.4.305 3968- radv: promote VK_EXT_depth_clamp_zero_one to KHR 3969- radv: bump maxViewportDimensions to 32K on GFX12 3970- radv: add a helper to report if cooperative matrix is enabled 3971- zink/ci: add lists for RADV/GFX1200 3972- radv: remove duplicate definition of SQTT_BUFFER_ALIGN_SHIFT 3973- ac/sqtt: update programming SQTT on GFX12 3974- radv: add support for VkMemoryBarrierAccessFlags3KHR 3975- radv: adjust the source aspect for color to depth/stencil image copies 3976- radv: advertise VK_KHR_maintenance8 3977- radv: do not overallocate the number of exports for streamout on GFX12 3978- radv: fix transform feedback on GFX12 3979- radv: declare a new user SGPR for emulating queries on GFX12 3980- radv: lower emulated queries with global atomics on GFX12 3981- radv: allocate memory for the shader query buffer on GFX12 3982- radv: emit the shader buffer query VA on GFX12 3983- radv: use global atomics for generated/written primitives query on GFX12 3984- radv: re-emit streamout state for GFX12 when the user SGPR changes 3985- radv: exclude layer when recomputing FS input bases 3986- ac/cmdbuf: program SPI_SHADER_GS_MESHLET_CTRL to 0 in the GFX12 preamble 3987- radv: program COMPUTE_DISPATCH_INTERLEAVE on GFX12 3988- radv: add support for BO metadata on GFX12 3989- radv: add a new helper to set image BO metadata 3990- ac/gpu_info: add gfx12_supports_display_dcc 3991- radv: fix an assertion about DCC and modifier on GFX12 3992- radv: fix the number of drm modifier planes for DCC on GFX12 3993- ci: update VKCTS main to a9988483c0864d7190e5e6264ccead95423dfd00 3994- radv/ci: update descriptor buffer skipped tests 3995- radv: fix disabling logic op for srgb/float formats when blending is enabled 3996- radv: disable video support on GFX12 3997- radv: disable VK_KHR_cooperative_matrix on GFX12 3998- radv: fix programming pitches for LINEAR_SUB_WINDOW on GFX12 3999- radv: fix programming mip level for TILED_SUB_WINDOWS on GFX12 4000- radv/ci: add expected list of failures for GFX1200 4001- radeonsi: fix programming DCC for SDMA on GFX12 4002- radv: use stage instead of entrypoint to determine valid gfx stages 4003- docs: add a note about GFX12 (RDNA4) on RADV 4004- ac,radeonsi: add SDMA DCC tiling for GFX12+ 4005- ac/descriptors: allow to configure DCC for buffer descriptors 4006- radv/amdgpu: add support for AMDGPU_GEM_CREATE_GFX12_DCC 4007- radv/meta: add missing pipeline lookups 4008- radv/meta: stop using string keys also for DGC and query objects 4009- util/disk_cache: add a new helper to create a disk cache 4010- vulkan/runtime: allow to use a different disk cache 4011- radv: fix caching on-demand meta shaders 4012- radv: fix adding the BO to cmdbuf list when starting conditional rendering 4013- radv: fix fetching draw vertex data from counter buffers with transform feedback 4014- radv/meta: disable conditional rendering for fill/update buffer operations 4015- radv: fix adding the VRS image BO to the cmdbuf list on GFX11 4016- ac,radv,radeonsi: add new GFX12_DCC_WRITE_COMPRESS_DISABLE tiling flag 4017- ac/gpu_info: add gfx12_supports_dcc_write_compress_disable 4018- radv: add initial DCC support on GFX12 4019- radv: fix adding the BO for unaligned SDMA copies to the cmdbuf list 4020 4021Saroj Kumar (1): 4022 4023- ac/surface: fix missing NULL check in gfx12_select_swizle_mode() 4024 4025Sathishkumar S (1): 4026 4027- radeonsi/vcn: enable roi decode and rgb targets on JPEG_5_0_1 4028 4029Scott Moreau (1): 4030 4031- dri: Fix hardware cursor for cards without modifier support 4032 4033Serdar Kocdemir (4): 4034 4035- Change C style cast on extension structs 4036- Wrap queue related functions on codegen 4037- The BumpPool of VkStream is not freeAll'ed 4038- gfxstream: add VK_DRIVER_FILES to devenv 4039 4040Sergi Blanch Torne (6): 4041 4042- ci: disable Collabora's farm due to maintenance 4043- Revert "ci: disable Collabora's farm due to maintenance" 4044- ci: disable Collabora's farm due to maintenance 4045- Revert "ci: disable Collabora's farm due to maintenance" 4046- ci: disable Collabora's farm due to unexpected power cut 4047- Revert "ci: disable Collabora's farm due to unexpected power cut" 4048 4049Shashank Sharma (1): 4050 4051- amd: add new AMDGPU_INFO subquery for userqueue metadata 4052 4053Sil Vilerino (26): 4054 4055- vl/vl_winsys: Add missing include for function declaration 4056- u_dynarray.h: Fix warning C4267 conversion from 'size_t' to 'type', possible loss of data 4057- u_math.h: Change power of two assert to fix warning C4146: unary minus operator applied to unsigned type, result still unsigned 4058- src/gallium/auxiliary/util/u_draw.h: Fix C4244 'argument' : conversion from 'type1' to 'type2', possible loss of data 4059- util: Fix warning C4244 'argument' : conversion from 'type1' to 'type2', possible loss of data 4060- src/compiler: Fix warning C4244 'argument' : conversion from 'type1' to 'type2', possible loss of data 4061- src/compiler: Fix warning C4389: An == or != operation involved signed and unsigned variables. This could result in a loss of data. 4062- d3d12: Fix warning C4267 conversion from 'size_t' to 'type', possible loss of data 4063- d3d12: Fix warning C4244 'argument' : conversion from 'type1' to 'type2', possible loss of data 4064- d3d12: Fix warning C4389: An == or != operation involved signed and unsigned variables. This could result in a loss of data. 4065- d3d12: Fix warning C4018 signed/unsigned mismatch 4066- d3d12: Add offset limit check to d3d12_resource_from_memobj 4067- d3d12_bufmgr.cpp: Fix warning C4244 for x86 builds assign uint64_t to size_t 4068- util: cpu_detect.c Fix warning C5274: behavior change: _Alignas no longer applies to the type '<unnamed-tag>' (only applies to declared data objects) 4069- d3d12_video_encoder_bitstream_builder_h264: Fix warning C4244 for x86 builds assign uint64_t to size_t 4070- d3d12_resource: Fix warning C4244 for x86 builds assign uint64_t to uintptr_t 4071- d3d12_video_dec_h264: Fix warning C4244 uint64_t to size_t cast 4072- d3d12_video_dec_vp9.cpp: Fix warning C4244: 'argument': conversion from 'uint64_t' to 'const unsigned int', possible loss of data 4073- d3d12_video_dec_hevc.cpp: Fix warning C4244: 'argument': conversion from 'uint64_t' to 'const unsigned int', possible loss of data 4074- d3d12_video_proc.h/cpp: Fix warning C4244: 'argument': conversion from 'uint64_t' to 'const unsigned int', possible loss of data 4075- d3d12_video_enc_av1.cpp: Fix warning C4244: 'argument': conversion from 'uint64_t' to 'unsigned int', possible loss of data 4076- d3d12_video_enc_h264.cpp: Fix warning C4244: 'argument': conversion from 'uint64_t' to 'unsigned int', possible loss of data 4077- d3d12_video_enc_hevc.cpp: Fix warning C4244: 'argument': conversion from 'uint64_t' to 'unsigned int', possible loss of data 4078- d3d12_video_dec.h/cpp: Fix warning C4244: 'argument': conversion from 'uint64_t' to 'unsigned int', possible loss of data 4079- d3d12_video_enc.h/cpp Fix warning C4244: 'argument': conversion from 'uint64_t' to 'unsigned int', possible loss of data 4080- d3d12: Enable Warnings C4267, C4996, C4146, C4244, C4389, C4838, C4302, C4018 in src/gallium/drivers/d3d12 subtree 4081 4082Simon Perretta (70): 4083 4084- pvr: add initial pco stub/boilerplate 4085- pvr, pco: Add new compiler framework and shader gen stubs 4086- pco: add env debug option parsing 4087- pco: stubs for SPIR-V/NIR compilation options 4088- pvr: connect basic pco functions to the driver 4089- pvr: remove pipeline shader hard-coding support 4090- pvr: add device info and functions for calculating available temps 4091- pvr: add shader compilation stubs 4092- pvr: track pipeline flags 4093- pvr: add device info for additional iterator features 4094- pvr: fix GetInstanceProcAddr ubsan warning when _instance == NULL 4095- pvr: drop PVRX macro 4096- pco: suppress warning for functions passing structs 4097- pco: pygen stubs 4098- pco, pygen: enum emit support, define some enums and op/ref mods/types 4099- pco, pygen: define basic isa field types 4100- pco, pygen: define and emit isa instruction group header variant fields 4101- pco, pygen: isa instruction group header validation and encoding support 4102- pco, pygen: isa lower source definitions 4103- pco, pygen: isa upper sources definitions 4104- pco, pygen: isa internal source selector definitions 4105- pco, pygen: isa destination definitions 4106- pco, pygen: isa main alu ops 4107- pco, pygen: isa backend alu ops 4108- pco, pygen: isa bitwise alu ops 4109- pco, pygen: isa control alu ops 4110- pco, pygen: query bytes required for each variant 4111- pco, pygen: generate op and mod info 4112- pco: define data structures and basic builder implementation with ops 4113- pco: NIR translation and PCO IR pass boilerplate 4114- pco: printing and validation boilerplate 4115- pco, pygen: generate string representations of enum elements 4116- pco: basic instruction printing 4117- pco, pygen: move unnamed tuple structs into classes 4118- pco, pygen: add bitset support for op mods 4119- pco, pygen: common underscore replacement for op names 4120- pco: add verbose printing debug option 4121- pco, pygen: distinguish hw ops that are built directly into instruction groups 4122- pco, pygen: instruction to instruction group mapping, printing 4123- pco: additional ref functions 4124- pco: boilerplate nir lowering passes 4125- pco, pygen: add initial uvsw op boilerplate 4126- pco, pygen: add better exception messages 4127- pco: adjust align padding to be per-function instead of per-shader 4128- pco, pygen: support querying ref mods, if op/ref mods have been set 4129- pco: set up and tear down glsl type singleton with context 4130- pco, pygen: add support for instructions with variable srcs/dests 4131- pco, pygen: re-order some mods to match their evaluation order 4132- pco: print ranges of non-ssa refs with >1 channel, datatypes for immediates 4133- pco, pygen: drop unspecified bit sizes for references 4134- pco, pygen: add defs and mappings for common ops 4135- pco, pygen: restructure igrp alu components into arrays 4136- pco, pygen: amend bitfield assertion messages 4137- pco, pygen: isa ditr op 4138- pco, pygen: isa itrsmp op 4139- pco: initial implementation of translation and passes 4140- pco: add public print wrappers 4141- pco: vector component tracking, vector collation when ingesting NIR 4142- pco: re-indexing debug option and additional vector and component tracking 4143- pco: add mappings and translation for ditr 4144- pco: temporarily add hardcoded vs/fs I/O for testing, BXS-4-64 iteration support 4145- pco: add helpers for overriding ref chans and offsetting vals 4146- pco: vec coalescing improvement to register allocation 4147- pco: add opt subpass for propagating comps referencing hw regs 4148- pco: track the number of bytes encoded for each function 4149- pvr, pco: rewrite compiler/driver interface for vs & fs I/O 4150- pco: modifier propagation optimization, shared opt context boilerplate 4151- pco: initial validation boilerplate and SSA checks 4152- CODEOWNERS: update for new pco compiler tree 4153- pco: fix x86 build 4154 4155Simon Ser (6): 4156 4157- dri: revert INVALID modifier special-casing 4158- llvmpipe: handle llvmpipe_resource_map() errors 4159- dri: don't fetch X11 modifiers if we don't support them 4160- egl/wayland: only supply LINEAR modifier when supported 4161- egl/wayland: fallback to implicit modifiers if advertised by compositor 4162- gbm: fix get_back_bo() failure with gbm_surface and implicit modifiers 4163 4164Sonny Jiang (1): 4165 4166- radeonsi/vcn: Add vcn_5_0_1 support 4167 4168Tapani Pälli (21): 4169 4170- intel/dev: update mesa_defs.json from workaround database 4171- anv: utilize ray query bo per queue for Wa_14022863161 4172- anv: extend Wa_14017794102 with lineage Wa_14023061436 4173- isl: modify existing assert by allowing CCS_E aux usage 4174- intel/dev: update mesa_defs.json from workaround database 4175- intel/dev: lower amount of max gs threads for Wa_18040209780 4176- anv/android: always create 2 graphics and compute capable queues 4177- iris: allow bo cache for compressed bos on verx10 == 200 4178- drirc/anv: force_vk_vendor=-1 for Marvel Rivals 4179- intel/dev: update mesa_defs.json from internal database 4180- dri: remove GLsync typedef 4181- anv: handle mesh in sbe_primitive_id_override 4182- iris: initialize whole pipe_box struct for memcmp 4183- intel/compiler: take reg_unit size into account with ubo ranges 4184- anv: set dependency between SF_CLIP and CC_PTR states 4185- mesa/st: take pixelmaps in to account in drawpixels cache 4186- intel/dev: update mesa_defs.json from internal database 4187- isl: use workaround framework for Wa_1207137018 4188- mesa: enable GL_EXT_conservative_depth extension 4189- anv: tighten condition for changing barrier layouts 4190- anv: apply cache flushes on pipeline select with gfx20 4191 4192Thomas H.P. Andersen (2): 4193 4194- drirc/nvk: force_vk_vendor=-1 for Artifact Classic 4195- nvk: follow naming convention for devices 4196 4197Tim Huang (1): 4198 4199- amd: add GFX v11.5.3 support 4200 4201Tim Keller (1): 4202 4203- dril: Check for null config in dril_target.c 4204 4205Timothy Arceri (24): 4206 4207- glsl/nir: fix function cloning at link time 4208- glsl: fix compiler global temp collisions 4209- glsl: tidy up glsl_to_nir() params 4210- glsl: remove unused member 4211- Revert "glsl: Move ForceGLSLAbsSqrt handling to glsl-to-nir." 4212- glsl: remove more now unused params from glsl_to_nir() 4213- glsl: don't copy symbol table to shaders 4214- glsl: drop _mesa_glsl_copy_symbols_from_table() 4215- glsl: use symbol table directly for builtin functions 4216- glsl: drop unused symbol table from gl_shader 4217- glsl: disable function return lowering in glsl ir 4218- glsl: remove return lowering from glsl ir 4219- glsl: drop last remaining lower jump test 4220- glsl: remove now unused ir reader 4221- glsl: move _mesa_glsl_compile_shader() declaration 4222- glsl: remove glsl/program.h 4223- nir: allow loops with unknown induction var initialiser to unroll 4224- glsl: drop unused ir_equals.cpp 4225- glsl: drop unused array refcount code and tests 4226- glsl: drop opt_dead_code_local 4227- glsl: enable layout qualifier if OVR_multiview enabled 4228- glsl: fix num_views validation message 4229- glsl: fix num_views linker error 4230- glsl: fix return value for subgroupBallot() 4231 4232Timur Kristóf (109): 4233 4234- radv: Mark GS copy shaders as internal. 4235- radv: Add ability to dump shaders based on stage. 4236- aco: Separate options for printing IR and recording disassembly. 4237- radv: Separate option to dump NIR. 4238- radv: Separate option to print shader disassembly. 4239- radv: Separate option to dump backend IR. 4240- radv: Refactor RADV_DEBUG=shaders to be a combination of other options. 4241- radv: Slightly reword preoptir debug flag. 4242- radv: Also allow filtering SPIR-V dump per stage. 4243- radv: Set dump flags in a smarter way by default. 4244- amd: Rename GFX1103_R1/R2 to PHOENIX/2 4245- radv: Add a flush postamble on GFX6. 4246- radv: Don't flush at the end of each command buffer on GFX6. 4247- ac/nir/ngg: Don't emit dead code with dot_op. 4248- ac/nir/ngg: Trade 1 VALU shift for 2 SALU add. 4249- ac/nir/cull: Slightly refactor control flow for small primitive culling. 4250- ac/nir/ngg: Slightly refactor workgroup scan. 4251- ac/nir/ngg: Pass wg_repack_result as pointer instead of returning it. 4252- ac/nir/ngg: Workgroup scan over two bools. 4253- ac/nir/ngg: Implement optional primitive compaction. 4254- ac/nir/ngg: Remove erroneous NUW addition from workgroup scan. 4255- radv: Reorder potentially per-primitive FS builtins. 4256- radv: Slightly simplify potentially per-primitive FS inputs. 4257- radv, aco: Consolidate num_interp + num_prim_interp into num_inputs. 4258- radv: Emit SPI_PS_IN_CONTROL when emitting PS inputs on GFX10.3. 4259- radv: Remove now unused num_prim_interp from shader_info. 4260- radv: Use default 0 for undefined builtin PS inputs. 4261- radv: Only set NGG_DISABLE_PROVOK_REUSE for VS. 4262- ac/nir/ngg: Add ability to store primitive ID as per-primitive. 4263- radv: Reorder FS primitive ID input after layer and viewport. 4264- radv: Configure implicit VS primitive ID to be per-primitive. 4265- ac/nir/ngg: Use ac_nir_prerast_out in mesh shader lowering. 4266- ac/nir/ngg: Simplify updating mesh shader output info. 4267- ac/nir: Pass ac_nir_prerast_out to ac_nir_export_parameters. 4268- ac/nir: Pass ac_nir_prerast_out to ac_nir_export_position. 4269- ac/nir: Introduce ac_nir_store_parameters_to_attr_ring. 4270- ac/nir/ngg: Refactor VS/TES attribute ring stores. 4271- ac/nir/ngg: Refactor GS attribute ring stores. 4272- ac/nir/ngg: Refactor export_pos0_wait_attr_ring. 4273- ac/nir/ngg: Remove dead code for attribute ring stores. 4274- ac/nir/ngg: Move wait attr ring workaround for GS to better place. 4275- ac/nir/ngg: Move emitting GS vertex param exports to if. 4276- ac/nir/ngg: Refactor storing per-primitive primitive ID to attribute ring. 4277- ac/nir: Mark when pre-rast output is used as varying or sysval. 4278- ac/nir: Split GS output usage masks to varying and sysval masks. 4279- ac/nir: Only export positions when they are really system values. 4280- ac/nir: Only export parameters when they are actually varying. 4281- ac/nir: Only store params to attribute ring that are varying. 4282- aco: Update documentation 4283- radv: Add some documentation. 4284- radv: Implement FS layer ID input as a system value. 4285- Revert "nir/opt_varyings: Add workaround for RADV mesh shader multiview." 4286- ac/nir/ngg: Don't mark multiview layer output as varying. 4287- amd: Set lower_layer_fs_input_to_sysval in common code, not in drivers. 4288- radv: Rename layer_input to reads_layer in PS info. 4289- radv: Only print "testing use only" message on GFX12+. 4290- ac/nir: Move ac_nir_lower_bit_size_callback to ac_nir.c 4291- ac/nir: Move ac_nir_get_mem_access_flags to ac_nir.c 4292- ac/nir: Move ac_nir callback functions to ac_nir.c 4293- ac/nir: Move ac_set_nir_options to ac_nir.c 4294- ac: Stop including nir.h in ac_shader_util.h 4295- ac/nir: Rename emit_streamout to ac_nir_emit_legacy_streamout 4296- ac: Move ac_nir_config struct to ac_nir.h 4297- ac/nir: Move ac_nir_create_gs_copy_shader to separate file. 4298- ac/nir: Expose ac_nir_unpack_value in ac_nir_helpers.h 4299- ac/nir: Move ac_nir_lower_intrinsics_to_args to separate file. 4300- ac/nir: Move ac_nir_lower_legacy_vs to separate file. 4301- ac/nir: Move ac_nir_lower_legacy_gs to separate file. 4302- ac/nir: Move ac_nir_gs_shader_query declaration to ac_nir_helpers.h 4303- ac/nir: Move ac_nir_opt_pack_half to separate file. 4304- ac/nir: Move ac_nir_lower_mem_access_bit_sizes to separate file. 4305- ac/nir: Move ac_nir_lower_sin_cos to separate file. 4306- ac/nir: Move pre-rasterization related utilities in separate file. 4307- ac/nir: Rename ac_nir_lower_ngg_ms to ac_nir_lower_ngg_mesh. 4308- ac/nir: Move ac_nir_lower_ngg_mesh to separate file. 4309- ac: Move AC_HS_MSG_VOTE_LDS_BYTES to ac_shader_util.h 4310- ac: Stop including ac_nir.h from ac_shader_util.c 4311- ac/nir: Move all ac_nir_* files to a new folder. 4312- radv: Lower array derefs of vectors outside of shader linking. 4313- ac/nir/ngg: Mitigate NGG fully culled bug when GS output is compile-time zero. 4314- ac/nir/ngg: Mitigate attribute ring wait bug when primitive ID is per-primitive. 4315- aco: Move NGG pos export scheduling determination to drivers. 4316- ac/nir/ngg: Remove some superfluous variables from culling code. 4317- ac/nir/ngg: Add a few comments explaining some variables. 4318- ac/nir/ngg: Remove unused vs_output struct. 4319- ac/nir/ngg: Carve out ac_nir_ngg_alloc_vertices_and_primitives. 4320- ac/nir/ngg: Use ac_nir_ngg_alloc_vertices_and_primitives in mesh shader lowering. 4321- ac/nir/ngg: Carve out ac_nir_create_output_phis. 4322- ac/nir/ngg: Carve out NGG streamout code. 4323- ac/nir/ngg: Carve out ac_nir_repack_invocations_in_workgroup. 4324- ac/nir/ngg: Slightly refactor emitting vertex parameters. 4325- ac/nir/ngg: Add radeon_info to NGG lowering options. 4326- ac/nir/ngg: Add and use a has_attr_ring_wait_bug field to ac_gpu_info. 4327- ac/nir/ngg: Add and use a has_attr_ring field to ac_gpu_info. 4328- ac/nir/ngg: Add and use a has_ngg_fully_culled_bug field to ac_gpu_info. 4329- ac/nir/ngg: Add and use a has_ngg_passthru_no_msg field to ac_gpu_info. 4330- ac/nir/ngg: Use gfx_level from radeon_info. 4331- ac/nir/ngg: Remove gfx_level and family from NGG lowering options. 4332- ac/nir/ngg: Pass radeon_info to mesh shader lowering. 4333- ac/nir/ngg: Use has_attr_ring and has_attr_ring_wait_bug in mesh shader lowering too. 4334- ac/nir/ngg: Rework attribute ring wait workaround in VS/TES. 4335- ac/nir/ngg: Carve out ngg_gs_process_out_primitive. 4336- ac/nir/ngg: Carve out ngg_gs_process_out_vertex. 4337- ac/nir/ngg: Rework GS output code for better attribute ring handling. 4338- ac/nir/ngg: Remove now unused export_pos0_wait_attr_ring. 4339- ac/nir/ngg: Don't call has_input_primitive in GS lowering. 4340- ac/nir/ngg: Move GS lowering to separate file. 4341- radv, radeonsi: Disable early prim export on GFX11+. 4342- ac/nir/ngg: Use SALU to calculate which threads store to attribute ring in GS. 4343 4344Tomeu Vizoso (42): 4345 4346- etnaviv/ml: Fix includes 4347- etnaviv/nn: Fix use of etna_core_info 4348- etnaviv/ci: Add expectation files for the VIPNano-SI+ NPU 4349- etnaviv/ml: Rework the dumping of tensors 4350- etnaviv: Add script to decode weights in Huffman format 4351- etnaviv/ml: Split V7 coefficient encoding to a new file 4352- etnaviv/ml: Add encoding of coefficients for V8 4353- etnaviv/ml: Fix padding for convolutions in V8 4354- etnaviv/ml: Implement tiling for V8 4355- etnaviv/ml: Set two bits in the NN instruction for V8 4356- etnaviv/ml: Disable caching on V8 4357- etnaviv/ml: Fix reshuffle TP jobs on V8 4358- etnaviv/ml: Only reshuffle when needed on V8 4359- etnaviv/ml: Make use of the new depthwise support in V8 4360- etnaviv/ci: Update expectations for the NPU in the A311D 4361- etnaviv/ml: Zero out the NN config 4362- etnaviv/ml: Zero all BOs 4363- teflon: Support multiple graph inputs and outputs 4364- etnaviv/ml: Adapt to changes in teflon regarding multiple inputs 4365- etnaviv/ml: Support addition operations on V8 4366- teflon: Add files mentioned in the docs for image classification 4367- teflon/docs: Update performance measurements on LibreComputer Alta 4368- teflon/docs: Add i.MX8MP to list of supported NPUs 4369- teflon/docs: Clarify smoke test instructions 4370- teflon: Add tests for the YOLOX model 4371- teflon: Support tests with inputs with less than 4 dims 4372- teflon: Rename model tests so they aren't skipped by gtest-runner 4373- teflon: Don't crash when a tensor isn't quantized 4374- teflon/tests: Add support for models with float inputs and outputs 4375- teflon/tests: Also use the cache for models in the test suite 4376- etnaviv/ml: Specify which of the input tensors need transposing. 4377- etnaviv/ml: Fix in_image_slice in transposes when width != height 4378- etnaviv/ml: Take offsets into account in TP operations 4379- teflon: Add support for tensor split and concatenation operations 4380- etnaviv/ml: Add support for tensor split and concatenation operations 4381- teflon: Limit support for Add to two unpopulated tensors 4382- etna/ml: Write out the size of the requested tensor 4383- teflon: Add support for tensor padding operations 4384- etnaviv/ml: Add support for tensor padding operations 4385- teflon: Add support for FullyConnected 4386- teflon: Add tests for FullyConnected 4387- etnaviv/ml: Implement FullyConnected 4388 4389Valentine Burley (99): 4390 4391- amd/ci: Drop x86_64 suffix from job names 4392- amd/ci: Merge and convert Raven piglit testing 4393- amd/ci: Convert LAVA RADV jobs to deqp-runner suites 4394- amd/ci: Increase fraction for radeonsi-raven-piglit 4395- panfrost/ci: Turn redundant GLESCTS-full run into disabled Piglit job 4396- svga/ci: Convert to deqp-runner suite 4397- panfrost/ci: Convert to deqp-runner suite 4398- ci: Drop lava-piglit:(x86_64|arm64) definitions 4399- radv/ci: Convert Valve RADV jobs to deqp-runner suites 4400- turnip/ci: Bump the number of tests per group for a618 4401- turnip/ci: Bump the number of tests per group for a630 4402- turnip/ci: Bump the number of tests per group for a660 4403- turnip/ci: Decrease fraction for a630-vk-asan 4404- turnip/ci: Adjust some timeouts 4405- turnip/ci: Remove a630-vk-asan skip 4406- turnip/ci: Update expectations 4407- freedreno/ci: Drop redundant DEQP_VER 4408- turnip/ci: Ony increase hangcheck timer for spilling tests on a630 4409- lavapipe/ci: Convert lavapipe-vk-asan to a deqp-runner suite 4410- etnaviv/ci: Convert to deqp-runner suites 4411- softpipe/ci: Convert softpipe-asan-gles31 to a deqp-runner suite 4412- radv/ci: Use deqp-vk-main in Raven and Stoney RADV jobs 4413- turnip/ci: Enable ASan leak detection in a630-vk-asan 4414- ci/deqp: Remove non-suite support 4415- llvmpipe/ci: Move Piglit timeout inside the suite 4416- ci/deqp: Simplify conditional arguments 4417- ci/deqp: Add a DEQP_FORCE_ASAN option 4418- llvmpipe/ci: Actually enable ASan testing for llvmpipe-deqp-asan 4419- anv/ci: Fix GPU_VERSION configuration for anv-jsl and anv-jsl-full 4420- anv/ci: Bump the number of tests per group for ADL 4421- anv/ci: Bump the number of tests per group for JSL 4422- anv/ci: Bump the number of tests per group for TGL 4423- anv/ci: Re-enable TGL and JSL manual jobs 4424- anv/ci: Remove fails that are in .gitlab-ci/all-skips.txt 4425- anv/ci: Update expectations 4426- ci/lava: Use CI_JOB_TIMEOUT instead of separate variable 4427- ci/windows: Bump the number of tests per group 4428- ci/windows: Add a manual full job 4429- ci/windows: Update expectations 4430- turnip/ci: Update expectations 4431- ci/windows: Always include windows-msvc in scheduled pipelines 4432- panvk/ci: Move the fractions out of suites 4433- panvk/ci: Bump the number of tests per group for G52 4434- lavapipe/ci: Bump the number of tests per group 4435- lavapipe/ci: Update expectations 4436- venus/ci: Bump the number of tests per group 4437- venus/ci: Update expectations 4438- angle/ci: Update expectations 4439- zink/ci: Update expectations for ANV 4440- turnip/ci: Document flake 4441- lavapipe/ci: Update expectations 4442- lavapipe/ci: Re-enable lavapipe-vk-asan 4443- ci: Uprev vkd3d-proton to b121e6d746341e0aaba7663e3d85f3194e8e20e1 4444- virgl/ci: Disable virgl-iris-traces-performance 4445- virgl/ci: Migrate the two iris jobs to 1130g7-volteer 4446- anv/ci: Increase anv-tgl-angle parallelism to 2 4447- zink/ci: Migrate the two TGL traces jobs to 1130g7-volteer 4448- zink/ci: Increase zink-anv-tgl parallelism to 4 4449- ci: Add Valentine to the restricted traces access list 4450- freedreno/ci: Update a630-traces-restricted checksums 4451- zink/ci: Skip crashing trace in zink-anv-tgl-traces-restricted 4452- turnip/ci: Decrease the fraction on a660-vk-full 4453- ci: Fix trace update script reading GitLab token from default location 4454- pan/ci: Document some flakes 4455- android/ci: Allow specifying Vulkan driver in cuttlefish-runner.sh 4456- android/ci: Build ANV for Android 4457- freedreno/ci: Update expectations 4458- panfrost/ci: Revert to 6.6 kernel on G57 4459- amd/ci: Add lava-hp-x360-14a-cb0001xx-zork and use it for VA-API testing 4460- amd/ci: Run full radeonsi-raven-va job pre-merge 4461- freedreno/ci: Update expectations again 4462- turnip/ci: Bump the number of tests per group for a630-vk-asan 4463- anv/ci: Move a test to common anv-skips 4464- ci: Uprev VKCTS to 1.4.1.0 4465- pan/ci: Properly wire up DRIVER_NAME 4466- panvk/ci: Skip waived tests 4467- ci: Uprev VKCTS to 1.4.1.1 4468- ci: Skip broken PenumbraOverture trace for zink and freedreno 4469- zink/ci: Update checksum for Osmos trace on TGL 4470- anv/ci: Revert to 6.6 kernel on anv-jsl 4471- iris/ci: Decrease iris-glk-deqp paralellism 4472- panfrost/ci: Move panfrost-g52-piglit to nightly 4473- zink/ci: Increase zink-anv-adl parallelism 4474- turnip/ci: Increase a660-vk fraction 4475- freedreno/ci: Decrease a660-gl paralellism 4476- freedreno/ci: Disable a618-gl, a618-egl, and a618-piglit 4477- turnip/ci: Disable a630-vk 4478- freedreno/ci: Decrease a630-gl parallelism 4479- freedreno/ci: Re-enable some traces on a618 and disable a630-traces 4480- zink/ci: Increase parallelism of zink-tu-a618 4481- freedreno/ci: Don't automatically retry manual jobs 4482- freedreno/ci: Migrate a618-piglit-full to kingoftown 4483- amd/ci: Migrate amd-raven-skqp from lenovo-zork to hp-zork 4484- anv/ci: Decrease anv-jsl-angle parallelism 4485- virgl/ci: Skip flaky trace 4486- amd/ci: Increase amd-raven-skqp parallelism 4487- freedreno/ci: Document flakes 4488- venus/ci: Skip flaky test due to intermittent timeouts 4489- amd/ci: Revert to 6.6 kernel on Raven 4490 4491Vignesh Raman (6): 4492 4493- ci: Uprev crosvm 4494- ci: Force db410c to host mode 4495- ci: Uprev kernel to 6.13 4496- ci: update expectation files 4497- ci: export RESULTS_DIR in crosvm-script.sh 4498- ci: use CI_PROJECT_NAME for artifacts name 4499 4500Vinson Lee (4): 4501 4502- hk: Fix hk_ia_update arguments order 4503- vulkan: Add missing va_end 4504- intel/elk: Fix assert with side effect 4505- hk: Fix build error with static_assert 4506 4507Visan, Tiberiu (3): 4508 4509- amd/vpelib: patch to match shader (#456) 4510- amd/vpelib: remove luma offset (#459) 4511- amd/vpelib: fixed file headers for Palamida scan 4512 4513Vldly (1): 4514 4515- freedreno: Fix resource tracking on repeated map with discard 4516 4517Xaver Hugl (1): 4518 4519- vulkan/wsi: unset GAMMA_LUT, CTM and DEGAMMA_LUT when doing a modeset 4520 4521Yinjie Yao (3): 4522 4523- radeonsi/vcn: Indentation fix 4524- radeonsi/vcn: Fix compile warnings with previously uninitialized variables. 4525- radeonsi/vcn: Disable 2pass encode for VCN 5.0. 4526 4527Yiwei Zhang (4): 4528 4529- venus: enable VK_EXT_external_memory_acquire_unmodified if needed 4530- venus: use dedicated allocation for ANB image memory import 4531- venus: fix to handle pipeline flags2 from maint5 4532- venus: fix maintenance5 props init and create flags2 4533 4534Yogesh Mohan Marimuthu (25): 4535 4536- amd: update amdgpu_drm.h for new userq ioctl 4537- amd: include amdgpu_drm.h from mesa instead of system for ac_fake_hw_db.h 4538- winsys/amdgpu: add DOORBELL domain to bo 4539- winsys/amdgpu: add CLEAR_VRAM flag to zero vram when creating bo 4540- winsys/amdgpu: add userq helper functions 4541- ac/gpuinfo: add use_userq and AMD_USERQ variable 4542- winsys/amdgpu: call userq init and destroy functions 4543- ac: add new userq signal and wait packet id 4544- ac: add inherit vmid field to indirect buffer packet 4545- winsys/amdgpu: use bo_va_op_raw() function instead of bo_va_op() 4546- winsys/amdgpu: use timeline syncobj for userq vm operations 4547- winsys/amdgpu: destroy bo_fence_lock late in do_winsys_deinit() 4548- winsys/amdgpu: pass job fences to VM ioctl 4549- winsys/amdgpu: wait for vm syncobj before creating userq 4550- winsys/amdgpu: move noop and ib_bytes adjustment to cs_flush 4551- winsys/amdgpu: move legacy chunk init and submission to new function 4552- winsys/amdgpu: add userq cmd submission support in amdgpu_cs_submit_ib() 4553- winsys/amdgpu: don't add fence dependency of other queues for userq 4554- winsys/amdgpu: send hdp flush packet for userq 4555- winsys/amdgpu: keep has_local_buffers true for userq 4556- winsys/amdgpu: use VM_ALWAYS_VALID for all VRAM and GTT allocations 4557- ac/gpu_info: populate fw info using new fw info ioctl for userq 4558- winsys/amdgpu: ring doorbell before calling userq_signal ioctl 4559- winsys/amdgpu: use next_wptr as cache for userq 4560- winsys/amdgpu: ensure strict order in updating mqd wptr and doorbell 4561 4562You, Min-Hsuan (1): 4563 4564- amd/vpelib: fix coverity defects 4565 4566Zan Dobersek (8): 4567 4568- fd/pps: specify counter group for each countable 4569- fd/pps: provide derived counters on a7xx 4570- freedreno/registers: update RB_BLIT_INFO, RB_CCU_CNTL 4571- tu/a7xx: use concurrent resolve groups 4572- tu: ensure completion of generic-clear resolves for color, depth/stencil clears 4573- tu/a7xx: support 8x MSAA 4574- freedreno/registers: fix RBBM_PRIMCTR understanding and usage 4575- freedreno/a7xx: fix fd_lrzfc_layout 4576 4577Zhao, Jiali (1): 4578 4579- amd/vpelib: 420 and 422 Output Single Segment cositing support 4580 4581Zoltán Böszörményi (3): 4582 4583- features.txt: Add Vulkan 1.4 section 4584- docs/features: Mark VK_EXT_host_image_copy as implemented on Turnip 4585- docs/features: Mark more Vulkan 1.4 features as done for drivers 4586 4587duncan.hopkins (9): 4588 4589- glx: change \`#if` guard around \`dri_common.h` to stop missing 'driDestroyConfigs' symbol on MacOS builds. 4590- glx: ignore zink check for has_explicit_modifiers and DRI3 on MacOS. 4591- kopper: Add '#if' guard around \`loader_dri3_get_pixmap_buffer` to stop missing symbol on MacOS. 4592- glx: Guard some of the bind_extensions() code with the same conditions as \`glx_screen`s `frontend_screen` member. 4593- glx: Add back in \`applegl_create_display()` so the OpenGL.framework, on MacOS, pointer get setup. 4594- zink: MoltenVk has conditional VK_DYNAMIC_STATE_VERTEX_INPUT_BINDING_STRIDE support. 4595- zink: Avoid optimalDeviceAccess on MoltenVK when creating depth taregts. 4596- zink, kopper: Conitionally add VK_IMAGE_USAGE_INPUT_ATTACHMENT_BIT to swap chain imageUsage. 4597- zink: stop zink_set_primitive_emulation_keys producing geometry shaders on platforms that do not support them. 4598 4599liuqiang (2): 4600 4601- lavapipe: Resolved write to pointer after free 4602- d3d10umd: Modify comment 4603 4604nyanmisaka (1): 4605 4606- frontends/vdpau: Get AV1 decode subsampling_x/y 4607 4608sergiuferentz (1): 4609 4610- Use try_unbox in VkDescriptorBufferInfo 4611