1Mesa 22.2.0 Release Notes / 2022-09-21 2====================================== 3 4Mesa 22.2.0 is a new development release. People who are concerned 5with stability and reliability should stick with a previous release or 6wait for Mesa 22.2.1. 7 8Mesa 22.2.0 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 22.2.0 implements the Vulkan 1.3 API, but the version reported by 16the apiVersion property of the VkPhysicalDeviceProperties struct 17depends on the particular driver being used. 18 19SHA256 checksum 20--------------- 21 22:: 23 24 b1f9c8fd08f2cae3adf83355bef4d2398e8025f44947332880f2d0066bdafa8c mesa-22.2.0.tar.xz 25 26 27 28New features 29------------ 30 31- WGL_ARB_create_context_robustness 32 33- d3d12 ARB_robust_buffer_access_behavior 34 35- VK_EXT_robustness2 for lavapipe 36 37- VK_EXT_image_2d_view_of_3d on RADV 38 39- zink and d3d12 GL_EXT_memory_object_win32 and GL_EXT_semaphore_win32 support 40 41- vertexAttributeInstanceRateZeroDivisor support for lavapipe 42 43- panfrost Valhall support (conformant OpenGL ES 3.1 on Mali-G57) 44 45- VK_EXT_primitives_generated_query on RADV 46 47- VK_EXT_non_seamless_cube_map on RADV, ANV, lavapipe 48 49- VK_EXT_border_color_swizzle on lavapipe, ANV, turnip, RADV 50 51- VK_EXT_shader_module_identifier on RADV 52 53- variablePointers on lavapipe 54 55- VK_EXT_multisampled_render_to_single_sampled on lavapipe 56 57- VK_EXT_shader_subgroup_vote/ballot on lavapipe 58 59- EGL_KHR_context_flush_control for all drivers 60 61- GLX_ARB_context_flush_control for all drivers 62 63- VK_EXT_attachment_feedback_loop_layout on RADV 64 65- None 66 67 68Bug fixes 69--------- 70 71- turnip/a660: dEQP-VK.spirv_assembly.instruction.terminate_invocation.terminate.subgroup_ballot failure due to "movmsk" reordered with "kill" 72- [radeonsi] Graphical corruption in Valve Source-based games 73- Updates to 22.2 RCs cause artifacts on nouveau and blank screen on VirtIO 74- nouveau: corruption on NVAC after switching to the NIR backend 75- zink: GL_TEXTURE_LOD_BIAS vs mipLodBias, and GTF-GL46.gtf21.GL3Tests.texture_lod_bias.texture_lod_bias_all failure on Turnip 76- Gallium OpenGL megadriver fails to build if LLVM is enabled but not required by any component built 77- Abort due to "Unknown intrinsic: intrinsic store_deref" (firefox, radeonsi) 78- turnip: dEQP-VK.subgroups.vote.graphics.subgroupallequal_*int regression on a635 79- regression from !17949 80- r600/sfn: xonotic shows a big performance regression when using the NIR backend with sb disabled 81- r600/sfn: register allocation fails for a number of piglits and deqp tests 82- DG2 corruption in gnome desktop from F36 83- radeonsi: error allocating sparse buffers.. bug? 84- [BISECTED/radeonsi] Flickering triangles with wayfire cube 85- ../src/util/u_queue.c:218: util_queue_fence_destroy: Assertion \`fence->signalled' failed for iris when disabled UTIL_FUTEX_SUPPORTED 86- turnip: gfxbench aztec ruins crashes on zink 87- glsl: textureGatherOffset param broken 88- v3dv: android: VTS big regression after 316728a55bc8fe398be1ac2796a22f8c91fb886c 89- radeonsi: driver crashes with kwin-effects-cube 90- [Turnip] Final Fantasy X: Incorrect hair rendering 91- turnip: gfxbench car chase crash on zink 92- [r300] - osu! - Shader optimizations cause black and glitchy rendering 93- [Bisected] Microsoft CLC: build failure with MSVC and Dshared-llvm=disabled with AMDGPU target available 94- anv: invalid returns from vkCreateDescriptorPool 95- [Bisected][RadeonSI] Blender crashes on startup 96- turnip: depth/stencil regressions 97- llvmpipe: assert in KHR-Single-GL45.enhanced_layouts.xfb_all_stages 98- Poor performance/utilization with Splitgate 99- freedreno / mesa 22.1.3 crashes Firefox Nightly 100- [Bisected][RADV] Rendering issues and performance drop with GFX10 on RPCS3 101- dzn: Build failure when it's the only driver built - 22.2-devel regression 102- d3d12: Unresolved external symbols from Version.lib 103- turnip: Geometry shader never invoked if there is no EmitVertex and gl_Position is not written 104- Trine 4 causing GPU HANG: ecode 12:1:849f7804, in vkr-ring-7 105- turnip: GPU fault on clear 106- intel_nullhw.c:179:28: error: passing 'const void \*const' to parameter of type 'void \*' discards qualifiers [-Werror,-Wincompatible-pointer-types-discards-qualifiers] 107- macOS undefined symbol _glapi_tls_Context build error 108- Build on macOS regressed 109- Regression: virgl: The switch to NTT breaks "Total War: Warhammer" 110- macOS's LIST_ENTRY macro conflicts with ours 111- src/util/disk_cache_os.c:234:21: error: too many arguments provided to function-like macro invocation 112- [RadeonSI] Teamfortress 2 crashing frequently with mesa 22.1.1 113- Unity game segfaults often 114- Alien Isolation segfaulting in radeonsi_dri.so starting with Mesa 22.1.0 115- turnip: dEQP-VK.subgroups.ballot_broadcast.framebuffer.* has failures on older deqp-vk 116- !17062 broke XTS xts5/Xlib9/XDrawImageString(16) with Xephyr using glamor on llvmpipe 117- DG2 full BAR broken by low BAR? 118- crocus: SSBOs regressed 119- Homerun Clash game triggers assert in compiler 120- radv: line rasterization mode affects multisampling 121- turnip: caselist fail 122- vtn: mediump_16bit_alu and compatibility 123- radv: dynamic vertex input failure 124- navi21\\radv\\amdgpu Halo Infinite screen corruption and then hang 125- ci/freedreno: a618_vk 1/7 timeouts 126- Windows OpenGL32.dll without shared-glapi currently broken 127- anv: KHR-GL46.tessellation_shader.single.xfb_captures_data_from_correct_stage fails on TGL 128- perfetto: crash on context teardown / driver unload 129- panfrost shader classification is wrong in some cases 130- kopper: the latest awfulness exposed by piglit glx-multi-window-single-context 131- android: r600 building error after commit 79ca456b 132- Error compiling gallium-nine on i686 using musl libc 133- [anv][regression][bisected][bsw,gen9atom] dEQP-VK.memory.mapping.dedicated_alloc failing on bsw and gen9atom 134- anv: GTF-GL46.gtf32.GL3Tests.packed_pixels.packed_pixels_pbo failure 135- r600: Tesseract does not render properly with the new NIR backend 136- radv: some zink cts fails 137- anv: ICL hiz issue 138- [RADV, bisected] addrlib changes cause Navi CTS regressions 139- r300: Segfault in shaders/closed/steam/kerbal-space-program/1017.shader_test 140- Get rid of debug_assert() 141- nine: docs? 142- d3d12: Gallium d3d12 backend performance is slow 143- [RADV][bisected] after commit 124848bf9efa1cc5d8a3cdfc654e0c03d57540e3 the games which use vulkan API not starting 144- freedreno: skqp instability from GPU hangs 145- ci/a618: intermittent network failure 146- turnip: EXT_primitives_generated_query broken 147- Godot crashes with RX5500m 148- intel: Some Alan Wake shader fails validation due to bad register usage in EOT message 149- r600/sfn: Add support for fmulz 150- radv: failure in image.sample_texture.*_compressed_format* 151- radeonsi: Hang on RX 5500 XT on latest mesa git 152- blob_align assert triggers 153- radv: Enabling support for instance extensions inside vkCreateInstance does not work 154- zink: rendering artifacts in Portal 2 155- a630: AmnesiaTDD trace gpu faults 156- freedreno build issue - FAILED: src/freedreno/decode/crashdec.txt 157- segfaults under X11 since the Ubuntu update to 22 158- anv: renderdoc explosion on icl 159- turnip: EXT_depth_clip_enable broken 160- lava: test result parsing fail 161- turnip: LRZ fast clear 162- [ANV] GTA IV flickering 163- [TGL] 25-100% perf impact on SynMark2 with 2GB dynamic pool size 164- [Crocus] - Minecraft with Forge and JourneyMap - Graphical glitches 165- intel - Minecraft char bar is broken after Mesa upgrade 166- [wine] RAGE: texture problems 167- "Unsupported modifier, resource creation failed." on stderr 168- [RADV/ACO] Star Citizen Flickeriing Lights 169- anv: query regressions 170- ir3: Infinite loop in scheduler when splitting 171- ci/lava: "Listened to connection" spam 172- ci: docs build fail 173- turnip: 3d copy/clear/... ops should not be counted by occlusion query 174- anv: sampleLocationsEnable ignored in VkPipelineSampleLocationsStateCreateInfoEXT 175- turnip: wrong border color for Z24S8 176- aco_tests fail with llvm-15 177- Clang 14 issues compiling src/util/u_cpu_detect.c 178- eglExportDMABUFImageQueryMESA crashes on amdgpu gallium/drivers/radeonsi 179- crocus: GL_SELECT regressed on hsw 180- Random texture flickering in Minecraft 181- panfrost: Broken interaction between CRC and tile size 182- panfrost: allow_forward_to_be_killed is too conservative 183- APU downgraded to OpenGL 3.0 since UBO size change in radeonsi 184- [anv, bisected, regression] Pipeline statistics query segfaults 185- radv/NIR: Suboptimal codegen for AccessChain on buffer device addresses 186- Regression - vah264dec + glimagesink causes green taint on entire video with mesa>=22.0.0 187- va: AMD: Wrong buffer size of P010 Luma plane 188- 625b352f14b broke glcts on Navi1x 189- Use marketing_name for the Vulkan device string in RADV 190- Gallium driver crashes due to fd double close in iris driver 191- radv: Khronos ray tracing sample issues 192- r300: copy propagate does not correctly account for control flow 193- r300: Amnesia the Dark Descent slowdown since NIR loop unrolling 194- r300: some unused uniforms are not removed with the NIR varying linker (compared to old glsl varying linker) 195- Mesa currently FTBFS on x32 architecture 196- Runescape randomly crashes. malloc_consolidate(): unaligned fastbin chunk 197- tu: viewport assert 198- d3d12: point coord regression due to NIR IO changes 199- nir_lower_phis_to_regs_block: Assertion \`src->src.is_ssa' failed. 200- Redream emulator showing artifacts 201- Build failure on sparc 202- r600 video playback broken 203- r300: hundreds of tests fail since "clear color buffers using color from a constant buffer" 204- turnip: shader viewport handling 205- Signal 6 (Aborted) when opening GTK3 applications 206- radv: Conditional rendering on compute queue 207- anv: line failure 208- 32 bit build failure 209- d3d12 build regression: d3d12video.h: error: expected ‘;’ at end of member declaration 210- panfrost midgard - major issues with MelonDS emulator - not able to trace properly too 211- Intel/anv: The cloak is flickering in Batman:Arkham Knight 212- regression: "NIR based varying linker" breaks some xwayland/glamor applications with etnaviv 213- pvr: build failure on archs with unsigned long != uint64_t 214- turnip: dEQP-GLES31.functional.copy_image.non_compressed.viewclass_32_bits.* fails with angle 215- Nheko misrendering on RK3399 (Mali T860) 216- intel_pps_driver.cc:308:74: error: too many arguments to function ‘void intel_perf_query_result_accumulate_fields(intel_perf_query_result*, const intel_perf_query_info*, const void*, const void*, bool) 217- radv: no nir validation in radv_link_shaders 218- r300: deqp regressions from "glsl: enable the use of the nir based varying linker" 219- [crocus] glmark2 -b jellyfish rendering issues 220- radeonsi-stoney: Failing jobs during a CI bug 221- iris: skqp gles_atlastext flake 222- turnip: use-after-free in shader compilation 223- anv: EXT_primitives_generated_query fails 224- r300: vs predicate logic fails in vs-loop-complex-unroll-with-else-break.shader_test 225- [Intel/Vulkan] White blinking artifacts in Wolfenstein Youngblood 226- radv: Indirect unaligned dispatch 227- [radv] DCC causes artifacts in Senran Kagura Shinovi Versus 228- Radeon RX 5700XT crash when using prusa-slicer 229- glcts proxy_textures_invalid_samples tests assert on iris 230- turnip: gfxbench mh3.1 offscreen crash with ANGLE 231- dEQP-GLES31.functional.shaders.builtin_functions.integer.[i|u]mulextended.* fail 232- Radv leaks a descriptor set layout 233- radeonsi: KHR_debug thread safety 234- [radeon, rv370] Running piglit shaders@glsl-vs-raytrace-bug26691 test causes hard lockup & reboot 235- clover: Assertion \`NumContainedTys && "Attempting to get element type of opaque pointer"' failed 236- wsi/x11: Avoid using xcb_wait_for_special_event in FIFO modes causes memory leak in Apex Legends 237- v3d: port v3dv workaround for GFXH-1918 238- freedreno: fragile location packing can break cull/clip distance 239- turnip: dEQP-VK.spirv_assembly.instruction.compute.float_controls.fp16.generated_args.* failures 240- Panfrost: wrong textures and lighting in "Thief" games in wine (not happening when using Software render) 241- [Bisected] r600: Big performance regression on some games 242- turnip: report corresponding drmFormatModifierPlaneCount for planar format 243- piglit ext_image_dma_buf_import.ext_image_dma_buf_import* regression 244- Huge amount of anon_inode:sync_file file descriptor created 245- enable warnings for GNU extensions we don't want 246- Grid Autosport via proton/vulkan: glitches on the grass on amdgpu, fine on amdgpu-pro 247- NIR validation failed after nir_opt_if 248- Etnaviv: The triangle is lost during linking together triangle strips with degenerate triangles 249- intel: integer_mad_hi / integer_mad_sat / integer_mul_hi produce invalid results 250- Aperture Desk Job rendering artifacts Intel TGL 251- intel: integer_sub_sat produces stack corruption / OpISubBorrow compilation issue 252- r300/rv530: "Black rendering" of Baldur's Gate Dark Alliance in Xbox emulator CXBX-R under wine with Gallium Nine and RADEON_DEBUG=use_tgsi 253- radv/aco: xfb bug 254- '../src/amd/common/ac_surface.c:1575: void ac_copy_dcc_equation(const struct radeon_info \*, ADDR2_COMPUTE_DCCINFO_OUTPUT \*, struct gfx9_meta_equation \*): assertion "dcc->equation.gfx9.num_bits <= ARRAY_SIZE(equation->u.gfx9.bit)" failed' 255- Occasional flicker corruption in Rage 2, e.g. after loading, with ACO on RX 5700 XT 256- Vulkan::Calling vkWaitForFences Timeout 257- turnip: bad gmem rendering with PUBG mobile on angle 258- Intel (CHT) - Uplink text rendering bugged out in Mesa 22.0 259- gen9atom gpu hang on dEQP-VK.spirv_assembly.instruction.graphics.float16.arithmetic_1 260- bad memory managment on panfrost RK3399 - cannot alocate more ram - fury unleashed 261- radv,aco: Regression with global atomics with negative offsets 262 263 264Changes 265------- 266 267Adam Jackson (32): 268 269- docs: Note EGL enum allocation for EGL_EXT_present_opaque 270- docs: Update supported drivers for 22.x 271- glx/kopper: Wire up a way for SwapBuffers to fail 272- wsi/x11: Avoid using xcb_wait_for_special_event in FIFO modes 273- kopper: Grow a swap interval API 274- glx/kopper: Enable GLX_EXT_swap_control etc. 275- egl/kopper: Hook up eglSwapInterval 276- wsi/x11: Hook up KHR_incremental_present 277- dri: Inline dri2InvalidateDrawable into its one caller 278- dri: Move driCopySubBufferExtension into drisw 279- dri: move driQueryRendererIntegerCommon near its one caller 280- dri: Move driC{reate,oncat}Configs near their only caller 281- dri: Move dri{Get,Index}ConfigAttrib near their vtable 282- dri: Remove what's left of utils.[ch] 283- glx/dri2: Require DRI2 >= 1.3 for working buffer invalidation 284- dri2: Require a loader with working buffer invalidation 285- zink: Print the VkResult if vkCreateInstance fails 286- glx/dri: Fix DRI drawable release at MakeCurrent time 287- mesa: Stub out GL_APPLE_object_purgeable 288- glx: Remove the remains of GLX_MESA_multithread_makecurrent 289- gallium/dri: Remove the memory of GLX_MESA_multithread_makecurrent 290- glx: Remove some can't-happen checks from protocolDestroyDrawable 291- glx: Remove unused __glXSetupForCommand from __glXIsDirect 292- glx: Remove excess screen parameter from CreateContext 293- glx/dri2: Stop tracking the protocol version in the display state 294- glx: Fix return values from __glXGetDrawableAttribute 295- glx: Fix drawable refcounting for naked Windows 296- kopper: Implement {EGL,GLX}_EXT_buffer_age 297- wsi/x11: Synchronously check for error from xcb_present_pixmap 298- mesa: Remove Purgeable fields from texture/buffer/renderbuffer objects 299- llvmpipe: Indentation fix 300- gallium/dri2: Enable {GLX_ARB,EGL_KHR}_context_flush_control 301 302Alejandro Piñeiro (22): 303 304- v3dv/pipeline: fix small comment typo 305- v3dv/pipeline_cache: adds check to skip searching for a entry 306- v3dv/pipeline_cache: add on disk cache hit stats 307- v3dv/pipeline: include pipeline layout on the pipeline sha1 308- v3dv/bo: unmap any bo that goes to the bo cache 309- v3dv/format: no need for GetPhysicalDeviceFormatProperties 310- v3dv/format: Add support for VK_KHR_format_feature_flags2 311- nir: get res binding using component 0, instead of asssumig an uint 312- v3dv/pipeline: call nir_lower_explicit_io after first nir optimization loop 313- v3dv/pipeline: expand nir_optimize, drop st_nir_opts 314- v3dv/device: add comment on depthClamp feature 315- broadcom: add no_merge_jobs debug option 316- broadcom: document which debug options are specific for a given driver 317- broadcom/compiler: wrap nir_convert_to_lcssa with NIR_PASS_V 318- broadcom/compiler: use NIR_PASS for nir_lower_vars_to_ssa at v3d_optimize_nir 319- nir: call nir_metadata_preserve at nir_remove_unused_io_vars 320- v3dv/pipeline: use nir_shader_instructions_pass when lowering layout 321- broadcom/compiler: removed unused function 322- broadcom/compiler: make several passes to return a progress 323- v3d/v3dv: use NIR_PASS(_ 324- vc4/v3d: stop adding NORAST when SHADERDB debug option is used 325- vc4/v3d: restore calling debug_get_option_vc4/v3d_debug 326 327Alessandro Astone (5): 328 329- kopper: Only include XCB headers on supported platforms 330- Android.mk: Cleanup mesa3d-lib 331- Android.mk: Make mesa3d-lib work with absolute path meson outputs 332- Android.mk: Generate the dummy source in local-generated-sources-dir 333- Android.mk: Intermediate output paths may already be absolute 334 335Alexander Kanavin (1): 336 337- swrast_kms: use swkmsDRI2Extension instead of driDRI2Extension 338 339Alexey Bozhenko (2): 340 341- intel/compiler: fix singleton pointer coverity warning 342- spirv: fix OpBranchConditional when both branches are the same 343 344Alyssa Rosenzweig (454): 345 346- panfrost: Add 48-bit and 96-bit tiling routines 347- panfrost: Remove bpp check for tiling 348- panfrost: Only handle 2D in should_linear_convert 349- panfrost: Implement tiled 3D resource access 350- panfrost: Only tile for 2D dimensions 351- panfrost: Allow tiling all texture targets 352- panfrost: Assert alignment in tiling routine 353- panfrost: Document arguments to tiling routines 354- panfrost: Unit test u-interleaved tiling routines 355- docs/panfrost: Document u-interleaved tiling 356- panfrost: Rework dirty tracking for Valhall 357- panfrost: Dirty track batch masks 358- panfrost: Rework PAN_GPU_ID mechanism 359- docs/panfrost: Document drm-shim usage 360- meson: Ignore unused variables in release builds 361- pan/bi: Simplify bi_propagate_pass_flags 362- panvk: Respect line width 363- panvk: Consider primitive restart in index buffer walks 364- pan/bi: Implement fquantize2f16 365- panvk: Handle empty shaders gracefully 366- panvk: Lower shared memory 367- util: Add a generic worklist implementation 368- nir: Use u_worklist to back nir_block_worklist 369- pan/bi: Implement pack_uvec[24]_to_uint 370- panfrost: Add texture features enum to v9.xml 371- pan/blit: Compile blit shaders without IDVS 372- pan/blit: Generalize texture alignment 373- pan/blit: Prepare for Valhall port 374- pan/blit: Support v9 data structures 375- panfrost: Don't allocate storage for PSIZ on Valhall 376- panfrost: Compile libpanfrost for v9 377- panfrost: Add helper to emit UBOs 378- panfrost: Port uniform/UBO logic to Valhall 379- panfrost: Don't fix up alpha test on Bifrost 380- panfrost: Use common state emit for compute jobs 381- panfrost: Adapt compute job emit for Valhall 382- panfrost: Specialize rasterizer state for Valhall 383- panfrost: Simplify attribute format expression 384- panfrost: Specialize vertex elements for Valhall 385- panfrost: Specialize ZSA state for Valhall 386- panfrost: Specialize shader descriptors for Valhall 387- panfrost: Use dirty flags to emit v9 descriptors 388- panfrost: Generate Valhall Malloc IDVS jobs 389- panfrost: Compile for v9 390- panfrost: Kick off v9 support in the driver 391- panfrost: Advertise all textures in drm-shim 392- panfrost: Split pan_layout.c from pan_texture.c 393- panfrost: Always use 64-bit SD with strides 394- panfrost: Add afbc_superblock_{size, width, height} helpers 395- panfrost: Use panfrost_afbc_superblock_width 396- panfrost: Don't pretend to support multiplane AFBC 397- panfrost: Extract panfrost_afbc_is_wide helper 398- panfrost: Use pan_block_size in layout calculation 399- panfrost: Unify panfrost_block_size paths 400- panfrost: Unify paths through image_layout_init 401- panfrost: Remove unused dev argument 402- panfrost: Unit test block size queries 403- panfrost: Simplify how image_layout_init is called 404- panfrost: Add a test for pan_image_layout_init 405- asahi: Workaround broken GLSL compiler 406- agx: Implement some shader-db stats 407- agx: Note that RA proceeds in dominance-order 408- agx: Track write registers more accurately 409- agx: Match order for designated initializers 410- agx: Use correct types for some IR enums 411- agx: Wrap compiler header in extern "C" 412- agx: Add unit test infrastructure 413- agx: Use pseudo ops for mov/not/and/xor/or 414- agx: Implement simple copyprop 415- agx: Dynamically allocate agx_instr->src 416- agx: Add p_split pseudoinstruction 417- agx: Add phi pseudo instruction 418- agx: Add a hash table for vector extracts 419- agx: Add helper to emit combines 420- agx: Add helper to emit splits 421- agx: Use a transfer graph for parallel copies 422- agx: Lower p_split after RA 423- agx: Optimize p_split(kill) specially 424- agx: Emit splits for intrinsics 425- agx: Use extract helper for tex internally 426- agx: Remove else optimization 427- agx: Use a dynarray for predecessors 428- agx: Add predecessor index helper 429- agx: Mark the logical ends of blocks 430- agx: Add agx_after_block_logical helper 431- agx: Introduce worklist infrastructure 432- agx: Adapt liveness analysis for SSA 433- agx: Make DCE dumber 434- agx: Translate phi nodes 435- agx: Copy prop into phi nodes 436- agx: Don't lower phis in NIR 437- agx: Lower phi nodes to parallel copies 438- agx: Split up RA from post-RA lowering 439- agx: Always use hash table for extracts 440- agx: Unit test parallel copy lowering 441- agx: Remove identity moves 442- agx: Add validation pass 443- agx: Update RA comment 444- agx: Remove has_liveness 445- agx: Remove nir_register support 446- agx: Lower mediump flat shading 447- panvk: Ignore point size for !points 448- panvk: Use correct point size limits 449- panfrost: Only flip point sprites for GL 450- panvk: Streamline no shader RSD case 451- panvk: Gate rasterization on !discard, not fs req 452- panvk: Emit fragment RSDs even with no shader 453- panfrost: Remove ancient unused code 454- panfrost: Call tgsi_to_nir earlier 455- panfrost: Don't subclass pipe_compute_state 456- panfrost: Use s->info.stage instead of tgsi stage 457- panfrost: Simplify shader key architecture 458- panfrost: Refactor variant selection code 459- panfrost: Refactor variant rebind code 460- panfrost: Lower point sprites on Bifrost 461- panfrost: Lower user clip planes 462- panfrost: Flip point coords in hardware 463- pan/va: Generalize message? check for asm 464- pan/va: Add missing TABLE (SFU) instructions 465- pan/va: Mark NOT instructions as deprecated 466- pan/va: Add VAR_TEX_BUF_GRADIENT instruction 467- pan/va: Add non-IDVS VAR_TEX instructions 468- panvk: Lower global vars to local 469- panvk: Allocate stack for each thread 470- pan/bi: Handle texture offset + index 471- panvk: Lower image deref atomics 472- panvk: Only write depth when depth test enabled 473- pan/decode: Print row strides, not line strides 474- panvk: Remove unused layout structs 475- panvk: Report row_stride in GetImageSubresourceLayout 476- panfrost: Add helpers to work with legacy strides 477- panfrost: Use row stride for explicit layouts 478- panfrost: Rename away from bytes_per_pixel 479- panfrost: Use row stride to calculate slice size 480- panfrost: Adapt get_param for row strides 481- panfrost: Use row_stride even for linear resources 482- lima,panfrost: Use row stride for tiling routines 483- panfrost: Remove line_stride 484- panfrost: Unify row stride and AFBC row stride 485- panfrost: Unit test stride calculations 486- panfrost: Unit test "from legacy" helper 487- pan/bi: Rename bi_block->name to bi_block->index 488- pan/bi: Count blocks 489- pan/bi: Add u_worklist wrapper macros 490- pan/bi: Use bi_worklist in analyze_helper_requirements 491- pan/bi: Use bi_worklist for liveness 492- pan/bi: Use bi_worklist for post-RA liveness 493- pan/bi: Use worklist for scoreboard analysis 494- pan/bi: Use a dynarray for predecessors 495- nir: Don't set writes_memory for reading XFB 496- panfrost: Relax image check 497- pan/va: Serialize memory stores 498- pan/bi: Align spilled registers on Valhall 499- pan/bi: Add option to test spilling 500- pan/bi: Assert that blend shaders may not spill 501- panfrost: Correct XML for TLS 502- panfrost: Use emit_tls 503- panfrost: Use packed TLS on Valhall 504- panfrost/ci: Smoke test spilling 505- panfrost: Don't use VARYING_SLOT_TEX0 internally 506- panvk: Don't use VARYING_SLOT_TEX0 internally 507- pan/bi: Rework varying linking on Valhall 508- panfrost: Increase the maximum tiler heap size 509- mesa/st: Fix building tests on macOS 510- panvk: Stub pipeline cache using the common code 511- panvk: Conform viewport code to Vulkan spec 512- panvk: Call nir_opt_trivial_continues 513- pan/va: Add some whitespace to Valhall disassembly 514- pan/va: Add whitespace after disassembled branches 515- agx: Restore Valve copyright header 516- pan/bi: Don't analyze td on Valhall 517- pan/bi: Emit LEA_TEX on Valhall 518- pan/bi: Support image loads on Valhall 519- pan/bi: Preserve AXCHG.i32 destination 520- pan/bi: Handle shared atomic exchange on Valhall 521- pan/bi: Handle shared/scratch on Valhall 522- pan/bi: Support atomics on Valhall 523- pan/bi: Ensure the end NOP isn't eliminated 524- asahi: Fix hangs waiting on the notification queue 525- pan/bi: Update bi_count_write_registers for Valhall 526- pan/va: Mark more source sizes 527- pan/va: Lower split 64-bit sources 528- pan/va: Use 64-bit lowering for texturing 529- pan/bi: +JUMP can't read same-cycle temp 530- pan/bi: Lower phis to scalar 531- pan/bi: Use value-based interference with LCRA 532- pan/bi: Add source/destination counts 533- pan/bi: Add collect and split instructions 534- pan/bi: Add bi_before_block cursor 535- pan/bi: Lower split/collect before RA 536- pan/bi: Fix mov and pack_32_2x16 537- pan/bi: Expand MAX_DESTS to 4 538- pan/bi: Create COLLECT during isel 539- pan/bi: Lift split/collect cache from AGX 540- pan/bi: Simplfy BLEND emit 541- pan/bi: Emit collect and split 542- pan/va: Use split for 64-bit lowering 543- pan/va: Don't use bi_word in FAU unit test 544- pan/bi: Scalarize bi_lower_swizzle 545- pan/bi: Scalarize bi_opt_cse 546- pan/bi: Scalarize modifier propagation 547- pan/bi: Scalarize copyprop 548- pan/bi: Remove bi_word and bi_word_node 549- pan/bi: Simplify register precolouring in the IR 550- pan/bi: Remove liveness metadata tracking 551- pan/bi: Don't propagate discard 552- pan/bi: Optimize split of collect 553- pan/bi: See through splits for var_tex fusion 554- pan/bi: Validate preload constraints are satisfied 555- pan/bi: Validate vector widths 556- pan/va: Handle 2-src blend in lower_split_src 557- pan/va: Remove DISCARD.f32 destination 558- pan/va: Use ^ instead of ` to indicate last-use 559- asahi: Run lower_blend for all fragment shaders 560- asahi: Use z24_in_z32f lowering 561- asahi: Validate depth/stencil formats 562- asahi: Don't store to unbound render targets 563- asahi: Handle uncompressed Z32F depth buffers 564- asahi: Pass depth/stencil clear values to firmware 565- asahi: Flush on partial clears 566- asahi: Handle 3rd deflake address 567- asahi: Account for every byte of the cmdbuf 568- asahi: Identify more depth pipeline fields 569- asahi: Dump "unknown 4a" structure 570- asahi: Add XML used for visibility tests 571- asahi: Relax Draw "command" check 572- asahi: Note unknown field 573- asahi: Identify depth bias array 574- asahi: Identify "set depth bias index" field 575- asahi: Identify depth bias enable bit 576- asahi: Add a depth bias structure 577- asahi: Implement polygon offset 578- asahi: Identify partial render pipeline 579- asahi: Sync some names with Project Zero 580- asahi: Add segment list pretty printer 581- asahi: Split unknown field in segment list header 582- asahi: Validate segment list length 583- asahi: Compute segment list length 584- asahi: Align segment list unknowns with Metal 585- asahi: Decode resource allocation responses 586- asahi: Decode resource allocation requests 587- asahi: Identify suballocation mechanism 588- asahi: Identify suballocated size 589- asahi: Drop some unknowns 590- asahi: Sync some more unknown fields 591- asahi: Improve render target dumping 592- asahi: Fix pipeline decoding 593- asahi: Handle RECT textures as 2D 594- asahi: Allocate a larger segment list 595- asahi: Guard against encoder overflows 596- asahi: Add wrap library 597- docs/asahi: Add initial driver docs 598- panfrost: Simplify the kmsro create path 599- panfrost: Don't allocate SHARED as scanout 600- pan/bi: Recoalesce tied operands after spilling 601- pan/bi: Schedule for pressure pre-RA 602- pan/bi: Allow CSEing LEA_BUF_IMM 603- panfrost: Fix crash with GALLIUM_HUD 604- panfrost: Extract afbc_stride_blocks helper 605- panfrost: Extract pan_afbc_row_stride helper 606- panfrost: Fix is_wide return type 607- panfrost: Add pan_afbc_tile_size helper 608- panfrost: Support tiled AFBC in stride helpers 609- panfrost: Align layouts to tiles of superblocks 610- panfrost: Add AFBC stride unit tests 611- panfrost: Add 3D texture layout unit test 612- panfrost: Calculate header_size based on row_stride 613- panfrost: Add a tiled 16x16 layout unit test 614- panfrost: Add 1x1 layout unit tests 615- panfrost: Fix AFBC flags on v6 616- panfrost: Put comment in correct #ifdef 617- panfrost: Handle AFBC Tiled 618- panfrost: Add helper checking tiled AFBC support 619- panfrost: Gate tiled AFBC on GPUs that support it 620- panfrost: Advertise 16x16 tiled AFBC 621- panfrost: Use tiled AFBC textures 622- pan/bi: Fuse result types 623- pan/bi: Test CMP result_type optimization 624- panfrost: Allow FPK on Valhall 625- panfrost: Set overdraw_alpha[01] flags 626- panfrost: Set allow_rotating_primitives 627- pan/bi: Don't allow spilling coverage mask writes 628- pan/bi: Add VAR_TEX fusing unit test 629- pan/va: Fill in missing src_flat16 enum 630- pan/bi: Make LD_VAR w=format instead of w=vecsize 631- pan/bi: Model Valhall source formats 632- pan/bi: Interpolate varyings at 16-bit 633- pan/bi: Extend bi_scoreboard_state for finer tracking 634- pan/bi: Add slot to bi_instr 635- pan/bi: Preserve flow control for non-psiz variant 636- pan/bi: Add ASSERT_SHADER_EQUAL macro 637- pan/bi: Add a trivial ctx->inputs for unit tests 638- pan/bi: Extract bit_block helper 639- pan/bi: Export bi_block_add_successor 640- pan/bi: Export helper termination analysis 641- pan/bi: Print flow control on instructions 642- pan/va: Add pass to insert flow control 643- pan/va: Do scoreboard analysis 644- pan/va: Terminate helper threads 645- pan/va: Add flow control merging pass 646- pan/va: Assign slots roundrobin 647- pan/va: Respect assigned slots 648- pan/va: Unit test flow control insertion 649- pan/va: Unit test flow control merging 650- pan/bi: Use flow control lowering on Valhall 651- pan/mdg: Scalarize with 64-bit sources 652- pan/mdg: Remove uppercase write masks 653- pan/mdg: Use "<<" instead of "lsl" 654- pan/mdg: Only print masked components of swizzle 655- pan/mdg: Only print 2 sources for ALU 656- pan/mdg: Only print 1 source for moves 657- pan/mdg: Distinguish SSA vs reg when printing IR 658- pan/mdg: Print outmods when printing IR 659- panfrost: Fix XML for AFBC header on v9 660- panfrost: Pass arch instead of dev into afbc_format 661- panfrost: Add pan_afbc_compression_mode on Valhall 662- panfrost: Support rendering to AFBC on Valhall 663- panfrost: Handle texturing from AFBC on Valhall 664- panfrost: Allow 2D AFBC on Valhall 665- panvk: Use vk_image_subresource_*_count for clears 666- pan/bi: Respect swizzles in nir_op_pack_64_2x32_split 667- pan/bi: Allow vec6 for collects 668- pan/bi: Align 64-bit register sources 669- pan/bi: Rework Valhall register alignment 670- pan/bi: Fix clper_xor on Mali-G31 671- panfrost/ci: Mark draw_buffers_indexed.* as flakes 672- panfrost: Use C11 static_assert for enums 673- nir: Add transform feedback system values 674- nir: Export nir_io_add_intrinsic_xfb_info 675- pan/bi: Add transform feedback lowering pass 676- pan/bi: Handle transform feedback intrinsics 677- panfrost: Don't allow vertex shaders to have side effects 678- panfrost: Wire up transfrom feedback sysvals 679- panfrost: Create transform feedback shaders 680- panfrost: Launch transform feedback shaders 681- pan/va: Do not insert NOPs into empty shaders 682- panfrost: Extract panfrost_get_index_buffer helper 683- panfrost: Don't calculate min/max indices on v9 684- pan/bi: Extract MUX to CSEL optimization 685- pan/va: Implement more lanes 686- pan/va: Lower MUX to CSEL where possible 687- pan/va: Add MUX lowering tests 688- pan/bi: Implement b2i with MUX 689- pan/va: Record which instructions are signed 690- pan/va: Try negating small constants when lowering 691- pan/va: Try widening small constants 692- pan/va: Unit test constant lowering pass 693- panfrost: Enable Mali-G57 694- docs: Document Mali-G57 conformance 695- panfrost: Fix inverted indirect dispatch condition 696- pan/bi: Implement load_push_constant 697- pan/bi: Don't reorder push with no_ubo_to_push 698- panfrost: Use push constants for indirect dispatch 699- panfrost: Use push constants for indirect draws 700- panvk: Use push constants for clear 701- panvk: Use push constants for copy shaders 702- panfrost: Deduplicate indirect dispatch structs 703- panfrost: Remove redundant first_tag access 704- panfrost: Remove is_64b assignments 705- panfrost: Add panfrost_query_l2_slices helper 706- pan/perf: Fix performance counters on G57 707- mesa,gallium: Make point coord origin a CAP 708- asahi,agx: Don't set PIPE_CAP_POINT_COORD_ORIGIN_UPPER_LEFT 709- panvk: Drop support for Midgard 710- panfrost: Allow pixels using discard to be killed 711- panfrost: Inline pan_fbd_has_zs_crc_ext 712- panfrost: Disable CRC at <16x16 tile sizes 713- panvk: Don't specialize clear shaders for RT 714- panvk: Remove unused pushmaps 715- panvk: Clear Z/S attachments without a shader 716- panvk: Fix stencil clears of combined ZS images 717- panvk: Simplify depth clear preload condition 718- panvk: Use common CmdBeginRenderPass 719- panfrost: Disable QUADS on Valhall 720- panfrost: Fix batch state changes on Valhall 721- panfrost: Add tristate_is_defined helper 722- panfrost: Implement provoking vertices on Valhall 723- pan/va: Move VA_NUM_GENERAL_SLOTS to common 724- pan/va: Simplify insert flow tests 725- pan/va: Workaround quirk of barrier handling 726- pan/va: Unit test barrier handling 727- pan/bi: Mark bi_postra_liveness_ins as MUST_CHECK 728- pan/bi: Constify bi_is_staging_src argument 729- pan/va: Add header guards to valhall_enums.h 730- pan/va: Move va_flow_is_wait_or_none to common 731- pan/va: Include BLEND for va_swap_12 732- pan/va: Use validate_register_pair for BLEND pack 733- pan/va: Mark last register reads 734- pan/va: Unit test va_mark_last 735- pan/bi: Fix LD_BUFFER.i16 definition 736- pan/bi: Align accesses with packed TLS 737- pan/va: Lower FADD_RSCALE.f32 to FMA_RSCALE.f32 738- pan/va: Implement sample positions FAU packing 739- pan/va: Rename RSCALE to LDEXP 740- pan/bi: Remove FRSCALE from IR 741- pan/bi: Model MKVEC.v2i8 742- pan/bi: Constant fold MKVEC.v2i8 743- pan/va: Pack MKVEC.v2i8 byte lanes 744- pan/va: Replace MKVEC.v4i8 with MKVEC.v2i8 745- pan/bi: Tune lower_vars_to_scratch 746- panfrost: Enable more FP16 caps on Valhall 747- panfrost: Enable CAP_INDIRECT_TEMP_ADDR on Valhall 748- panfrost: Bump ESSL_FEATURE_LEVEL on Valhall 749- v3d: Drop workaround for u_blitter bug 750- agx: Handle loop { if { loop { .. } } } 751- r600g: Remove streamout-based buffer copy path 752- u_blitter: Remove util_blitter_copy_buffer 753- lima,panfrost: Do not ralloc contexts against the screen 754- pan/va: Handle terminal barriers 755- ttn: Set nir->info.separate_shader 756- pan/bi: Fix unpack_32_2x16 definition 757- pan/va: Add missing <roundmode/> to V2F32_TO_V2F16 758- pan/bi: Implement f2f16{_rtz, _rtne} 759- panfrost: Separate core ID range from core count 760- panfrost: Move bifrost_lanes_per_warp to common 761- panfrost: Fix PIPE_COMPUTE_CAP_MAX_COMPUTE_UNITS 762- panfrost: Fix PIPE_COMPUTE_CAP_SUBGROUP_SIZE 763- panfrost/ci: Disable 0ad trace on T860 764- panfrost: Clear with a quad to avoid flushing 765- panfrost/ci: Disable T760 jobs 766- nir/opt_shrink_vectors: Round to supported vec size 767- panfrost: Revert provoking vertex assertion 768- panfrost: Only key points to point coord origin 769- pan/va: Dump unencodable instructions 770- pan/va: Use invalid_instruction in more places 771- pan/va: Print instructions with pack assert fails 772- pan/va: Handle BIFROST_MESA_DEBUG=nosb 773- panfrost: Make pan_select_max_tile_size O(1) 774- panfrost: Use implementation-specific tile size 775- panfrost: Protect pandecode by a mutex 776- pan/bi: Require ATEST coverage mask input in R60 777- pan/va: Stall after ATEST 778- panfrost: Fix shader_modifies_coverage on Valhall 779- panfrost: Add decoupled early-ZS helpers 780- panfrost: Unit test early-ZS helpers 781- panfrost: Add zsa->zs_always_passes flag 782- panfrost: Use early-ZS helpers 783- util: Enable list_assert in debugoptimized builds 784- util: Gate simple_mtx_assert_locked on !NDEBUG 785- agx: Port minifloat tests to GTest 786- pan/bi: Assert that we use the correct CLPER 787- pan/bi: Rename CLPER_V6.i32 to CLPER_OLD.i32 788- pan/decode: Fix overrun decoding planes 789- nir/lower_idiv: Be less creative about signs 790- agx: Fix packing of samplers in texture instrs 791- panfrost: Don't segfault on unknown models 792- pan/bi: Don't reorder image loads across stores 793- pan/bi: Don't allow ATEST to take a temporary 794- pan/mdg: Print 3 sources for CSEL 795- pan/bi: Fix dual texturing with uniforms 796- pan/bi: Fix out-of-bounds write in va_lower_split_64bit 797- pan/bi: Consider all dests in helper_block_update 798- agx: Fix float copyprop of neg(neg) case 799- panfrost: Respect buffer offset for OpenCL 800 801Andres Calderon Jaramillo (2): 802 803- nir: Account for YUV range. 804- mesa/st: Account for YUV color space and range. 805 806Andres Gomez (5): 807 808- radv/ci: update vkd3d-proton results for AMD's Kabini 809- ci: install a more recent version of Wine in the test-vk image 810- ci: install gfxinfo-mupuf in the test-vk image 811- ci: move b2c artifacts to its own directory 812- ci: add wrapper script for Valve's traces runner 813 814Anuj Phogat (2): 815 816- include/uapi: Update drm_fourcc.h from kernel 817- isl,iris: Add I915_FORMAT_MOD_4_TILED support for XeHP 818 819Arvind Yadav (7): 820 821- nir: Add a lowering pass for polygon and line smoothing 822- radeonsi: Add nir_lower_poly_line_smooth pass for polygon and line smoothing 823- nir: add load_point_coord_maybe_flipped intrinsics for point smoothing 824- nir: Add a lowering pass for point smoothing 825- ac/llvm : Adding Number of all interpolated inputs in ac_shader_abi 826- ac/llvm: Implement nir_intrinsic_load_point_coord_maybe_flipped opcodes 827- radeonsi: Enable nir_lower_point_smooth lowering pass for point smoothing 828 829Autumn on Tape (5): 830 831- gallivm: add subgroup shuffle support 832- lavapipe: enable subgroup shuffle operations 833- gallivm: use shufflevector for shuffles when index is constant data 834- gallivm: use VPERMPS (x86/AVX2) for 32-bit 8-element shuffles 835- lavapipe: enable subgroup quad operations 836 837Axel Davy (6): 838 839- frontend/nine: Skip invalid swvp calls 840- frontend/nine: Fix buffer tracking out of bounds 841- frontend/nine: Fix ATOC handling 842- frontend/nine: Fix cso restore bug 843- frontend/nine: Fix shader multi-use crash 844- frontend/nine: Fix ff position_t fallback when w = 0 845 846Bas Nieuwenhuizen (39): 847 848- vulkan/wsi/x11: Ensure we have the required number of images for acquire. 849- radv: always enable NGG on GFX11 850- radv: Mark all formats as DCC compatible with each other on gfx11. 851- radv: gfx11 register changes. 852- radv: Adjust for new SWIZZLE_ENABLE. 853- radv: Add gfx11 DCC fast clear support. 854- radv: Disable transform feedback for gfx11. 855- radv: update the initialization of SGPR0/1 registers for HS and GS on GFX11 856- radv: Deal with derefs from opaque types in function parameters. 857- radv: Add performance counter reg write. 858- radv: move ac_perfcounters to physical_device. 859- radv: Add pstate locking for perfcounters. 860- radv: Rename SPM functions. 861- radv: Make inhibiting clockgating not static. 862- radv: Add helper BO for perf counters. 863- radv: Add perf counter lock/unlock commandbuffers. 864- radv: Add a list of performance counters. 865- radv: Implement support for querying performance counters. 866- radv: Implement listing performance counters. 867- radv: Expose VK_KHR_performance_query. 868- radv: Add ability to override the build id for the cache. 869- radv: Add a 32bit memory type. 870- radv: Skip setting empty index buffers to avoid hang 871- radv: Expose function to write vertex descriptors for dgc. 872- radv: Always store stride in the vbo descriptor. 873- radv: Require 32bit memory for indirect buffers. 874- radv: Expose helper for base pa_su_sc_mode_cntl. 875- radv: Make radv_get_vgt_index_size non-static. 876- radv: Add helper to write scissors. 877- radv: Add flushing for DGC. 878- radv: Add DGC meta shader. 879- radv: Implement DGC generated command layout structure. 880- radv: Implement DGC cmdbuffer generation. 881- radv: Implement CmdExecuteGeneratedCommandsNV. 882- radv: Add stub for vkCmdBindPipelineShaderGroupNV. 883- radv: Expose VK_NV_device_generated_commands. 884- vulkan/wsi: Take max extent into consideration for modifier selection. 885- amd/common: Don't rely on DCN support checks with modifiers. 886- amd/common: Disable DCC retile modifiers on RDNA1 887 888Ben Skeggs (3): 889 890- nv50/ir: fix OUT.FINAL on ampere 891- nv50/ir: fix scope of memory ops on ampere 892- nv50/ir: fix ALD/AST where address comes from AL2P 893 894Benjamin Cheng (5): 895 896- radv: fix memory leak of descriptor set layout 897- radv: refactor image binding into struct 898- radv: consolidate descriptor set buffer counts 899- radv: implement disjoint memory for multiplane images 900- radv: expose disjoint image support 901 902Boris Brezillon (159): 903 904- dzn: Skip binding entries with zero descriptors 905- dzn: Make sure sampler heaps don't contain more than 2048 samplers 906- microsoft/compiler: Add a dunmmy SV_SampleIndex when needed 907- microsoft/spirv_to_dxil: Allow forcing per-sample shading 908- dzn: Add support for sampleRateShading 909- dzn: Pass the right type to CreateCommandList() in the reset path 910- dzn: refactor error-handling 911- ci/windows: Add a variable to globally disable jobs using windows runners 912- dzn: Use core helpers to fill physical device features/properties 913- dzn: Report actual device limits 914- d3d12: Move the D3D12ResourceState code to the gallium driver 915- ci/windows: Split the windows-related tests 916- dzn: Add support for EXT_vertex_attribute_divisor 917- dzn: Rename a bunch of dzn_descriptor_set helpers 918- dzn: Pass layout objects to various dzn_descriptor_set_ptr_xxx helpers 919- dzn: Split the write desc helpers in two halves 920- dzn: Add support for KHR_descriptor_update_template 921- microsoft/spirv_to_dxil: Add a draw_id field to vertex_runtime_data 922- dzn: Fix exec_buf_stride calculation 923- dzn: Simplify the triangle_fan_rewrite_index shader 924- dzn: Update the draw_id sysval when issuing indirect draws 925- dzn: Advertise KHR_shader_draw_parameters 926- dzn: Pass an image view to dzn_cmd_buffer_clear_attachment() 927- dzn: Set MAX_RTS to D3D12_SIMULTANEOUS_RENDER_TARGET_COUNT 928- dzn: Move all render-related states to a substruct 929- dzn: Inline dzn_cmd_buffer_clear_attachments() 930- dzn: Add support for dynamic rendering 931- dzn: Get rid of the render pass logic 932- dzn: Allow multiDrawIndirect and drawIndirectFirstInstance 933- dzn: Add KHR_draw_indirect_count support 934- dzn: Add missing VKAPI_{ATTR,CALL} specifiers to BeginCommandBuffer() 935- dzn: Pass the dzn_event pointer to _mesa_hash_table_insert() 936- dzn: Fix the STATIC_ASSERT() in dzn_meta_blits_get_context() 937- vulkan/wsi: Make wsi_win32_surface_get_capabilities() return the current extent 938- dzn: Fix dzn_cmd_buffer_collect_queries() 939- dzn: Use the vk_sync_binary abstraction 940- spirv: Fix windows build 941- dzn: Use the correct ICD path on Win32 942- microsoft/spirv_to_dxil: Fix push_constants type declaration 943- microsoft/spirv_to_dxil: Fix the push_constant UBO size calculation 944- dzn: Emit missing transition barriers for rendering attachments 945- ci/windows: Bump spirv-samples commit-ID 946- microsoft/spirv2dxil: Add a debug option 947- microsoft/compiler: Make sure all SV_Position components are written 948- microsoft/spirv_to_dxil: Extract NIR passes out of spirv_to_dxil() 949- microsoft/spirv_to_dxil: Constify the descriptor binding mapping fields 950- microsoft/compiler: Make input_mappings immune to driver_location holes 951- microsoft/spirv_to_dxil: Remove dead IO vars at the end of dxil_spirv_nir_passes() 952- microsoft/spirv_to_dxil: Add a linking helper 953- microsoft/compiler: Allow vulkan users to handle varyings linking 954- microsoft/spirv_to_dxil: vectorize IOs 955- dzn: Fix maxVertexInputAttributes limit 956- dzn: Handle varying linking explicitly 957- dzn: Fix unitialized var access in dzn_meta_compile_shader() 958- dzn: Fix vertex input handling 959- microsoft/spirv_to_dxil: Let spirv_to_dxil() handle the linking 960- microsoft/compiler: Drop the auto_link parameter passed to nir_to_dxil() 961- dzn: Switch to ID3D12Device2 962- dzn: Drop an unused argument passed to dzn_graphics_pipeline_translate_vi() 963- dzn: Use CreatePipelineState() 964- dzn: Enable depth-bounds testing 965- microsoft/compiler: Lower fquantize2f16 966- nir: Fix flat new_var assignment in create_new_io_vars() 967- microsoft/compiler: Pick a type that matches interpolation mode for structs 968- microsoft/compiler: Truncate function names when needed 969- vulkan/util: Fill VkPhysicalDeviceIDProperties::deviceNodeMask 970- spirv: Add a dst/src type to image deref loads/stores coming from OpAtomic{Load,Store} 971- microsoft/compiler: Images are no longer reprensented by uniform vars 972- microsoft/compiler: Set typed_uav_load_additional_formats when appropriate 973- microsoft/compiler: s/assert()/unreachable()/ in emit_intrinsic() 974- microsoft/compiler: Support load_sample_id_no_per_sample 975- dzn: Decorrelate external image info and external image properties 976- dzn: Return empty external properties 977- dzn: Add a dummy GetDeviceGroupPeerMemoryFeatures() 978- dzn: Add a dummy GetImageSparseMemoryRequirements2() 979- dzn: Add dummy {Create,Destroy}SamplerYcbcrConversion() implementations 980- dzn: Check that no export is requested in the memory allocation path 981- dzn: Lower alignment requirements when allocating buffers or single-sample images 982- dzn: Add a dzn_desc_type_has_sampler() helper 983- dzn: Implement GetDescriptorSetLayoutSupport() 984- dzn: Vulkan 1.1 wants maxPerSetDescriptors >= 1024 985- dzn: Prepare support for pipeline variants 986- dzn: Fix primitiveRestart support 987- dzn: Fix triangle-fan emulation 988- dzn: Support dynamic depth testing parameters properly 989- dzn: Support dynamic line width 990- ci/windows: Allow skipping deqp tests if we have to 991- ci/windows: Add dEQP-VK.dynamic_state.* to the test list 992- ci/windows: Skip dEQP-VK.api.command_buffers.record_many_draws* 993- d3d12: Keep access masks up-to-date when removing vars in d3d12_disable_multisampling() 994- microsoft/compiler: Move d3d12_fix_io_uint_type() to dxil_nir.c 995- microsoft/compiler: Fix Layer type 996- microsoft/compiler: Support gl_Layer 997- dzn: Disable rasterization if the last geometry stage doesn't write the position 998- dzn: Enable support for geometry shaders 999- ci/dzn: Move to a .toml description 1000- dzn: Fix the 2DMSArray case in dzn_image_get_dsv_desc() 1001- dzn: Fix CmdPushConstants() 1002- dzn: Clamp depthBiasConstantFactor when doing the float -> int conversion 1003- dzn: Transition resource to RENDER_TARGET/DEPTH_WRITE before clears 1004- microsoft/compiler: Fix emit_ubo_var() 1005- ci/dzn: Copy testlog.{css,xsl} to the result dir 1006- nir/serialize: Put dest last in packed_instr::tex 1007- dzn: Advertise anisotropic filtering support 1008- dzn: Advertise shaderImageGatherExtended support 1009- dzn: Enable dynamic indexing on all kind of descriptors 1010- dzn: Enable shader{Clip,Cull}Distance 1011- dzn: Enable the depthClamp feature 1012- nir/serialize: Support texop >= 16 1013- nir/serialize: Silence integer-overflow false positive 1014- vulkan: Add a vk_pipeline_hash_shader_stage() helper 1015- anv: Use vk_pipeline_hash_shader_stage() 1016- v3dv: Fix nir_shader leaks in v3dv_meta_{clear,copy}() 1017- v3dv: Use vk_pipeline_hash_shader_stage() 1018- tu: Use vk_pipeline_hash_shader_stage() 1019- radv: Kill unused fs_m local var in radv_create_shaders() 1020- radv: Use vk_pipeline_hash_shader_stage() 1021- dzn: Hook-up device-lost detection 1022- dzn: Fix indentation 1023- dzn: Initialize UUIDs 1024- dzn: Drop dzn_pipeline_cache.c 1025- dzn: Add missing D3D12_CACHED_PIPELINE_STATE to MAX_COMPUTE_PIPELINE_STATE_STREAM_SIZE 1026- dzn: Fix assertion in d3d12_pipeline_state_stream_new_desc() 1027- dzn: Use d3d12_compute_pipeline_state_stream_new_desc() in the compute path 1028- dzn: Pass the maximum stream size to d3d12_pipeline_state_stream_new_desc() 1029- dzn: Drop useless while(link_mask != 0) 1030- dzn: Drop unused allocator passed to dzn_graphics_pipeline_create() 1031- dzn: Let dzn_pipeline_init() initialize the root signature in the stream 1032- dzn: Fix potential nir_shader leak 1033- dzn: Use vk_to_mesa_shader_stage() 1034- dzn: Save a few indentation levels in graphics_pipeline_compile_shaders() 1035- dzn: Don't delegate binding translation to dxil_spirv_nir_passes() 1036- dzn: Hash pipeline layout stages 1037- dzn: Add a generic cached blob wrapper 1038- dzn: Drop unneeded goto statement in dzn_compute_pipeline_create() 1039- dzn: Move the compute shader compilation logic to a sub-function 1040- dzn: Cache NIR shaders 1041- dzn: Caculate a binding translation hash 1042- dzn: Cache DXIL shaders 1043- dzn: Cache pipeline info 1044- ci/dzn: Add pipeline cache tests to the include list 1045- dzn: Cache and pack transition barriers 1046- dzn: Don't call dzn_cmd_buffer_queue_transition_barriers() prematurely 1047- dzn: Rename dzn_cmd_buffer_queue_image_subres_range_transition() 1048- dzn: Introduce dzn_cmd_buffer_queue_image_range_state_transition() 1049- dzn: Allow shaders to read from depth/stencil attachments 1050- dzn: Fix copies of block-based formats 1051- dzn: Fix unused var warnings 1052- dzn: Fix 'statement with no effect' warning 1053- dzn: Fix printf() format in dxgi_get_factory() error message 1054- dzn: Fix {back,front}_wr_uses_ref assignment 1055- dzn: Fix -Wswitch warnings 1056- dzn: Use the vk_descriptor_set_layout base struct 1057- dzn: Use the vk_pipeline_layout base struct 1058- dzn: Mark transition barriers as executed when we execute barriers 1059- dzn: Query D3D12_FEATURE_D3D12_OPTIONS3 features 1060- dzn: Provide a helper to check if 2 formats are compatible 1061- dzn: Support native image copies when formats are compatible 1062- dzn: Take VK_IMAGE_CREATE_MUTABLE_FORMAT_BIT into account 1063 1064Boyuan Zhang (9): 1065 1066- radeonsi/vcn: add vcn 4.0 encode support 1067- radeonsi/vcn: add vcn 4.0 encode fw interface version 1068- radeonsi/vcn: add encode context ib for vcn 4.0 1069- radeonsi/vcn: add session init ib for vcn 4.0 1070- radeonsi/vcn: enable vcn 4.0 encode for gfx11 asic 1071- radeonsi/vcn: use calculated max hierarchy depth for hevc enc 1072- vl: add cabac init idc for h264 encode 1073- frontends/va: get entropy mode info for h264 encode 1074- radeonsi/vcn/enc: enable cabac entropy mode 1075 1076Brian Paul (101): 1077 1078- gallivm: assorted clean-ups in lp_bld_nir_aos.c 1079- gallivm: whitespace fixes in lp_bld_tgsi.h 1080- gallivm: asst. clean-ups in lp_bld_sample.h 1081- llvmpipe: asst. clean-ups in lp_linear_fastpath.c 1082- llvmpipe: asst. clean-ups in lp_setup.c 1083- llvmpipe: asst. clean-ups in lp_state_setup.c 1084- gallivm: asst. clean-ups in lp_bld_type.[ch] 1085- gallivm: cleanup lp_bld_arit.c 1086- gallivm: clean-up in lp_bld_nir.h 1087- gallivm: add const qualifier in lp_bld_nir_soa.c 1088- llvmpipe: asst. clean-up in lp_state_derived.c 1089- llvmpipe: asst. clean-ups in lp_rast_linear.c 1090- llvmpipe: fix comment typo in lp_linear_fastpath.c 1091- llvmpipe: asst. clean-ups in lp_rast.c 1092- llvmpipe: asst. clean-ups in lp_linear.c 1093- llvmpipe: asst. clean-ups in lp_scene.h 1094- llvmpipe: asst. clean-up and add comments in lp_state_fs.h 1095- gallivm: clean-up in lp_bld_sample.c 1096- llvmpipe: asst. clean-up in lp_setup_context.h 1097- gallivm: asst. clean-ups in lp_bld_sample_soa.c 1098- gallivm: clean-ups in lp_bld_tgsi.[ch] 1099- llvmpipe: asst. clean-ups in lp_state_fs.c 1100- llvmpipe: minor clean-ups in lp_jit.h 1101- llvmpipe: asst. clean-ups in lp_state_fs_linear_llvm.c 1102- llvmpipe: asst. clean-ups in lp_state_fs_linear.c 1103- llvmpipe: asst. clean-ups in lp_setup_tri.c, lp_setup_rect.c 1104- llvmpipe: whitespace clean-ups in lp_rast.h 1105- llvmpipe: minor whitespace/formatting cleanups in lp_state_setup.h 1106- llvmpipe: minor clean-ups in lp_state_fs_analysis.c 1107- llvmpipe: misc clean-up in lp_rast_tri.c 1108- llvmpipe: asst. clean-ups in lp_bld_blend_aos.c 1109- llvmpipe: asst. clean-ups in lp_rast_tri_tmp.h 1110- llvmpipe: asst. clean-ups in lp_rast_priv.h 1111- gallivm: minor clean-ups in lp_bld_sample_soa.c 1112- llvmpipe: misc clean-ups in lp_bld_interp.c 1113- gallivm: asst. clean-ups in lp_bld_nir.c 1114- llvmpipe: fix comment typo in lp_bld_interp.h 1115- llvmpipe: asst. clean-ups in lp_linear_sampler.c 1116- llvmpipe: asst. clean-ups in lp_tex_sample.c 1117- llvmpipe: asst. clean-ups in lp_scene.c 1118- llvmpipe: minor clean-up in lp_state_blend.c 1119- llvmpipe: use enum blend types instead of unsigned 1120- llvmpipe: s/unsigned/enum pipe_compare_func/ 1121- gallivm: asst. clean-ups in lp_bld_logic. 1122- llvmpipe: asst. clean-ups in lp_state_fs.c 1123- draw: s/unsigned/enum pipe_prim_type/ 1124- draw: asst. clean-up in draw_context.h 1125- draw: asst. clean-up in draw_gs.h 1126- draw: asst. clean-up in draw_pipe.c 1127- draw: asst. clean-up in draw_pipe_validate.c 1128- draw: asst. clean-up in draw_private.h 1129- draw: asst. clean-up in draw_pt.h 1130- draw: asst. clean-up in draw_pt.c 1131- draw: asst. clean-up in draw_pt_fetch_shader_emit.c 1132- draw: asst. clean-up in draw_pt_fetch_shader_pipeline.c 1133- draw: asst. clean-up in draw_pt_util.c 1134- draw: asst. clean-up in draw_pt_vsplit.c 1135- draw: asst. clean-up in draw_vbuf.h 1136- gallivm: use enum types instead of unsigned in lp_bld_sample.[ch] 1137- llvmpipe: add minor comments in lp_rast.h, lp_setup-rect.c 1138- llvmpipe: remove lp_rast_cmd_arg::state field 1139- gallivm: s/unsigned/enum pipe_swizzle/ 1140- llvmpipe: add simple assertion in generate_fragment() 1141- util/bitset: add BITSET_SIZE() 1142- lavapipe: rework code to compute textures_used, samplers_used 1143- llvmpipe: minor code re-org in lp_state_fs_analysis.c 1144- nir: add const qualifiers, move some decls in nir_to_tgsi_info.c 1145- llvmpipe: replace if/then with switch in llvmpipe_nir_fn_is_linear_compat() 1146- gallivm: s/0/LP_BLD_TEX_MODIFIER_NONE/ 1147- llvmpipe: replace GET_A0() macro w/ inline function 1148- llvmpipe: add missing tex_info->texture_unit assignment 1149- llvmpipe: fix invalid memory used in lp_fs_linear_run 1150- llvmpipe: fix texcoord analysis in llvmpipe_nir_fn_is_linear_compat() 1151- llvmpipe: replace LP_RAST_OP\_ #defines with enum type 1152- gallivm: increase LP_MAX_TGSI_SHADER_IMAGES from 16 to 32 1153- lavapipe: fix logicop, independent blend enable/disable 1154- llvmpipe: don't allow texture/resource swizzles on linear path 1155- llvmpipe: initialize a local var to fix compiler warning in release build 1156- lavapipe: fix incorrect sv[] array size 1157- llvmpipe: fix texcoord analysis regression 1158- llvmpipe: asst. clean-up in lp_bld_depth.c 1159- llvmpipe: minor white-space clean-up in lp_setup.h 1160- llvmpipe: asst. clean-ups in lp_setup_rect.c 1161- gallivm: asst. clean-ups in lp_bld_const.c 1162- llvmpipe: asst. clean-ups in lp_linear_interp.c 1163- llvmpipe: minor optimization in lp_linear_init_interp() 1164- llvmpipe: asst. clean-ups in lp_state_setup.c 1165- llvmpipe: remove unneeded casts 1166- llvmpipe: replace 'unsigned' with enum types 1167- gallivm: asst. clean-ups in lp_bld_swizzle.c 1168- gallivm: simplify shuffling in emit_store_reg() 1169- gallivm: fix incorrect memset() in SOA emit_load_const() 1170- gallivm: rename 'var' param in emit_store_reg(), emit_store_var(). 1171- llvmpipe: make opaque field a 1-bit field 1172- llvmpipe: minor whitespace clean-up in lp_state_fs.h 1173- llvmpipe: pass frontfacing param to jit_function() call 1174- llvmpipe: asst. clean-up in lp_rast_linear_fallback.c 1175- llvmpipe: remove unneeded casts in lp_rast_linear_fallback.c 1176- llvmpipe: consolidate code in lp_rast_linear_fallback.c 1177- llvmpipe: tighten up some code in lp_rast_linear_fallback.c 1178- llvmpipe: add some const qualifiers in lp_rast_linear_fallback.c 1179 1180Caio Oliveira (1): 1181 1182- intel/compiler: In XeHP prefer <1;1,0> regions before compacting 1183 1184Chad Versace (2): 1185 1186- venus: Refactor vn_fix_graphics_pipeline_create_info 1187- venus: Don't encode ignored pTessellationState 1188 1189Charlie Turner (4): 1190 1191- ci: fix shellcheck violations in the test-vk building script 1192- ci: add Linux build of apitrace 1193- ci: upgrade gfxreconstruct build to v0.9.10 1194- ci: upgrade to DXVK v1.10.1 1195 1196Charmaine Lee (20): 1197 1198- gallium/util: Increase the debug_flush map depth to 64 1199- draw: fix double free of NIR IR 1200- svga: fix shader IR type passed to draw create shader function 1201- svga/nir: enable PIPE_CAP_TGSI_TEXCOORD 1202- svga: enable PIPE_CAP_TGSI_TEXCOORD for vgpu10 and up 1203- svga/tgsi: add translation for DTRUNC 1204- svga/nir: enable lower double option 1205- svga: workaround for unexpected double swizzle 1206- svga: fix clip distance shadow copy writemask 1207- svga: add a helper function for common shader creation 1208- svga: create svga_shader_info 1209- svga: add svga_compile_shader helper function 1210- svga: add need_texcoord_semantic to tgsi_add_point_sprite & tgsi_add_aa_point 1211- svga: fix aa point 1212- svga: fix reference to NULL offset for atomic counter 1213- svga: fix shared memory index 1214- svga: fix mksstats build 1215- mesa/st: fix reference to nir->info after nir_to_tgsi 1216- svga: support TGSI_SEMANTIC_TEXCOORD in swtnl draw context 1217- svga: fix invalid component access of domain location 1218 1219Chia-I Wu (61): 1220 1221- turnip: fix drm modifier support with planar formats 1222- util: Keep quiet NaNs quiet when converting to half float. 1223- anv: advertise rectangularLines only for Gen10+ 1224- turnip: let modifier takes precedence over TU_DEBUG=noubwc 1225- freedreno/fdperf: make refresh rate configurable 1226- freedreno/fdperf: support dumping counters 1227- turnip: fix sampledImageIntegerSampleCounts 1228- turnip: emit VPC_SO_DISABLE in xfb begin/end 1229- ir3: fix mem_ctx for ir3_disasm_info::nir 1230- turnip: fix off-by-one in border color bitset 1231- turnip: fix tu6_pack_border_color for z24 1232- turnip: disable UBWC for SNORM formats 1233- turnip: if-checks autotune debug macros 1234- turnip: base pipeline draw costs on colorWriteMask 1235- turnip: include stencil test for drawcall costs 1236- turnip: estimate render pass costs 1237- turnip: consider render pass costs in autotune 1238- ci/turnip: add a link to VkPhysicalDeviceMemoryProperties failure 1239- turnip: force sample interpolations for sample shading 1240- turnip: advertise VK_EXT_border_color_swizzle 1241- freedreno/fdl: use XYZW swap for PIPE_FORMAT_X24S8_UINT 1242- vulkan/device_select: remove get_pdevice_proc_addr 1243- freedreno/drm-shim: add a660 1244- turnip: fix an assertion with drm-shim 1245- ir3: update ir3_const_state comment 1246- ir3: copy req_local_mem for MESA_SHADER_KERNEL 1247- ir3: fix output_loc size 1248- ir3: remove unused patch_vertices_in 1249- ir3: increment constoff right after it is assigned 1250- ir3: fix tess param allocation 1251- ir3: set UL flag before ir3_lower_subgroups 1252- turnip: remove shared_consts from tu_compiled_shaders 1253- turnip: shared_consts and push_consts are mutually exclusive 1254- turnip: add tu_common.h as the common header 1255- turnip: remove includes that are already in tu_common.h 1256- turnip: add tu_drm.h 1257- turnip: add tu_suballoc.h 1258- turnip: update tu_cs.h 1259- turnip: add tu_query.h 1260- turnip: add tu_image.h 1261- turnip: add tu_formats.h 1262- turnip: update tu_descriptor_set.h 1263- turnip: add tu_shader.h 1264- turnip: add tu_pipeline.h 1265- turnip: add tu_clear_blit.h 1266- turnip: add tu_dynamic_rendering.h 1267- turnip: add tu_lrz.h 1268- turnip: add tu_pass.h 1269- turnip: add tu_wsi.h 1270- turnip: update tu_autotune.h 1271- turnip: add tu_device.h 1272- turnip: add tu_cmd_buffer.h 1273- turnip: add tu_android.h 1274- turnip: update tu_util.h 1275- turnip: move away from tu_private.h 1276- turnip: remove tu_private.h 1277- turnip: remove headers from libtu_files 1278- turnip: use SPDX-License-Identifier 1279- turnip: fix a use-after-free in autotune 1280- turnip: lower the queue priority to 1 1281- ir3: fix predicate splitting in scheduler 1282 1283Christian Gmeiner (12): 1284 1285- isaspec: Extend split_bits(..) to accept a bitsize 1286- isaspec: Handle patterns bigger then 64 bit 1287- nir: Fix unused-variable compile warnings 1288- freedreno: Wait for async compile jobs to finish 1289- ir3: Create at least one thread for async shader compile 1290- gtest: Fix maybe-uninitialized compiler warning 1291- r600: Switch to unreachable(..) 1292- svga: Clear query before usage 1293- ci: Remove -Wno-error=maybe-uninitialized 1294- etnaviv: factor out shader screen related init/deint 1295- etnaviv: async shader compile 1296- etnaviv: Move nir_shader_compiler_options to compiler 1297 1298Chuansheng Liu (1): 1299 1300- iris,anv: correct the max thread number for DG2+ 1301 1302Clément Guérin (1): 1303 1304- venus: implement VK_EXT_physical_device_drm 1305 1306Connor Abbott (69): 1307 1308- tu: Delete unused tu_clear_blit GS handling 1309- ir3, freedreno, tu: Make ir3_shader_variant store stream_output 1310- ir3, fd, tu: Copy misc. info from ir3_shader to ir3_shader_variant 1311- ir3: Use ir3_shader_variant::type more often 1312- ir3: Allocate disasm_info under variant 1313- ir3: Add ir3_shader_variant::compiler 1314- tu: Keep original blit shaders separately 1315- ir3: Remove ir3_shader_variant::shader 1316- ir3: Add functions to serialize variants 1317- ir3: Support disabling the pipeline cache 1318- ir3: Add ir3_shader_create_variant() 1319- tu: Hash pipeline layout contents 1320- tu: Rewrite cache UUID based on radv 1321- tu: Include turnip debug flags in pipeline cache UUID 1322- tu: Implement pipeline caching with shared Vulkan cache 1323- tu: Zero-initialize compute driver key 1324- tu: Implement VK_EXT_pipeline_creation_feedback 1325- tu: Implement VK_EXT_pipeline_creation_cache_control 1326- tu: Convert to sync2 entrypoints 1327- tu: Expose VK_KHR_synchronization2 1328- tu: Implement VK_EXT_image_2d_view_of_3d 1329- tu: Enable shaderTessellationAndGeometryPointSize 1330- tu: Enable VK_KHR_swapchain_mutable_format 1331- tu: Fix partial-write LRZ determination 1332- tu: Implement extendedDynamicState2LogicOp 1333- tu: Implement VK_EXT_color_write_enable 1334- tu: Enable uniform texture buffers with NPOT formats 1335- freedreno/a6xx: Support more 3-component formats 1336- tu: Fix 3d GMEM store swizzle in texture descriptor 1337- tu: Add missing WFI to the 3d GMEM store path 1338- tu: Save/restore \*_BIN_CONTROL in 3d GMEM store path 1339- ir3/sched: Fix could_sched() determination 1340- tu: Fix linemode for tessellation with isolines 1341- ir3: Fix vectorizer condition for SSBOs 1342- freedreno: Document a650-specific CP_REG_WRITE flag 1343- freedreno/a6xx: Document border color/swizzle interaction 1344- freedreno/a6xx: Fix border color swizzling 1345- freedreno/a6xx: Use fdl format swizzle 1346- freedreno/fdl: Note border color constraints 1347- freedreno/a6xx: Fix border color with Z24UINT_S8UINT 1348- tu: Fix stencil border color with has_z24uint_s8uint 1349- mesa/st: Fix border color type for stencil sampling 1350- nv50/ir/ra: Fix copying compound for moves 1351- freedreno/a6xx: VPC_SO_NCOMP is actually VPC_SO_BUFFER_STRIDE 1352- tu: Use incoherent CCU write for buffer accesses 1353- tu: Support resolving D24S8 -> S8 1354- tu: Fix resolving d32s8 into s8 on fast path 1355- tu/lrz: Do not use framebuffer when inheriting LRZ 1356- tu: Treat CP_WAIT_FOR_ME as a cache invalidate 1357- tu: Rewrite tess factor emission 1358- tu: Move TU_DONT_CARE_AS_LOAD into attachment_set_ops() 1359- tu: Disable GMEM for multiview inside tu_render_pass_gmem_config 1360- tu: Move tu_render_pass definition up 1361- tu: Remove usage of RenderPassBeginInfo 1362- tu: Initial support for dynamic rendering 1363- tu: Remove has_subpass_predication 1364- tu: Update more state with secondaries 1365- tu: Split out some state into a separate struct 1366- util/trace: Fix u_trace_*_iterator with no tracepoints 1367- tu: Add support for suspending and resuming renderpasses 1368- tu: Use common vk_image struct 1369- tu: Use common vk_image_view base struct 1370- tu: Add debug option to use emulated renderpass support 1371- tu: Restore formatting of tu_clear_blit.c 1372- tu: Fix sysmem depth attachment clear flushing 1373- ir3/spill: Fix extracting from a vector at the end of a block 1374- tu/lrz: Fix multiple subpass case with secondaries 1375- tu/lrz: Fix multiple depth attachment case with secondaries 1376- tu: Don't preload variable-count descriptors 1377 1378Constantine Shablya (1): 1379 1380- anv: set image_read_without_format NIR option on Vulkan 1.3 1381 1382Corentin Noël (5): 1383 1384- mesa: Make sure to fallback to handling the original choose texture format 1385- virgl/ci: Update virglrenderer 1386- virgl/ci: Update virglrenderer 1387- virgl: Add support for passing the supported number of components from virgl 1388- virgl: Only progagate the uniform numbers if the numbers are actually right 1389 1390Cristian Ciocaltea (4): 1391 1392- virgl/ci: Rename virgl-lava-* jobs to virgl-iris-* 1393- virgl/ci: Update checksums for some virgl iris traces 1394- ci: Add CPU frequency adjustment capability 1395- ci: Limit Intel CPU scaling frequency for performance tests 1396 1397Daniel Schürmann (36): 1398 1399- nir/lower_alu_to_scalar: don't set the nir_builder cursor 1400- aco: remove 'max_waves' and use 'num_waves' to adjust for LDS and workgroup size 1401- aco: split num_waves adjustment into separate function 1402- aco: adjust num_waves for LDS before scheduling 1403- aco/optimizer: prevent any overflow between SGPR and const offset on MUBUF 1404- aco/ra: fix live-range splits of phi definitions 1405- aco: don't re-create vectors for load_barycentric_* intrinsics 1406- aco/ra: improve split_vector register assignment if the operand is not killed 1407- aco/optimizer: re-combine and copy-propagate p_create_vector(p_split_vector) 1408- radv: move two NIR passes out of tight optimization loop 1409- radv: call nir_lower_vars_to_ssa later 1410- aco: fix spilling of phis without temp operands 1411- nir/opt_vectorize: add callback for max vectorization width 1412- nir: introduce new nir_alu_alu_width() with nir_vectorize_cb callback 1413- nir/opt_vectorize: refactor src rewriting to avoid unnecessary mov instructions 1414- aco: correctly validate v_fma_mixhi_f16 register assignment 1415- radv: don't lower vectorized instructions to 32bit 1416- radv: use callback for nir_lower_to_scalar 1417- radv: vectorize nir_op_fdiv & nir_op_flrp & nir_op_ineg 1418- radv: vectorize nir_op_fabs 1419- radv: fix swizzles after nir_opt_algebraic_late 1420- radv: reorder some NIR passes 1421- nir/opt_shrink_vectors: fix re-using of components for vecN 1422- radv/shader_info: fix load_frag_coord and load_sample_pos read masks 1423- aco: fix packed 16bit fneg/fsat optimization 1424- aco: avoid unnecessary copies in emit_wqm() 1425- aco: Avoid live-range splits in Exact mode 1426- aco: fix assertion in insert_exec_mask 1427- spirv: handle phis decorated with RelaxedPrecision 1428- radv: Only create noop-fs for graphics pipelines 1429- radv: create RT traversal as separate shader 1430- radv/rt: use derefs for the traversal stack 1431- radv/rt: use stage ID as handle for general and closestHit shaders 1432- radv/rt: Don't load ClosestHit SBT on every hit, but only once after traversal 1433- radv/rt: fix nir_builder cursor in lower_rt_instructions() 1434- radv/rt: simplify lower_rt_instructions() 1435 1436Daniel Stone (10): 1437 1438- dzn/ci: Don't spam conformance warnings 1439- CI: Disable Windows jobs 1440- ci: Allow lima jobs to be globally disabled 1441- ci: Also disable Windows container builds when down 1442- ci/windows: Eliminate duplicate script definition 1443- ci/zink: Add glx-tfp to skips 1444- ci/lava: Sleep before, not after, API calls 1445- ci: freedreno farm is down 1446- CI: Disable Collabora lab 1447- Revert "CI: Disable Collabora lab" 1448 1449Danylo Piliaiev (53): 1450 1451- turnip: Try harder to keep LRZ valid and fix a few edge cases 1452- freedreno/a6xx: Add UNK fields to CP_REG_TEST and CP_COND_REG_EXEC 1453- turnip: Skip load/stores for tiles with no geometry 1454- turnip: Add debug option to print gmem load/store skip stats 1455- turnip: Fix tu_debug_flags values clashing 1456- tu: Implement VK_EXT_image_view_min_lod 1457- tu: Do not flush ccu in clear/blits during renderpass 1458- pps: Open writable renderer node in DrmDevice::create 1459- tu: Fix indices of drm_msm_gem_submit_cmd when filling them 1460- freedreno/a6xx: Add EARLYPREAMBLE flag to all a6xx_sp_xs_ctrl_reg0 1461- util/u_trace: Add json output 1462- docs/u_trace: document u_trace usage 1463- ir3: handle gl_Layer and gl_ViewportIndex when there is TES + GS 1464- ir3: Force late-z if FS has global store/atomic 1465- tu: Fix D32S8 resolves into D32 or S8 in sysmem mode 1466- spirv: Workaround for RelaxedPrecision on OpLogical* in 3DMark 1467- turnip: Allow nested CP_COND_REG_EXEC 1468- turnip: Fix out-of-bounds builder->shader_iova access 1469- tu/autotune: Prevent division by zero 1470- turnip: copy disasm strings for pipeline executables 1471- tu: Disable sample counting for 3d blits during occlusion query 1472- tu: Reset xfb_used at the end of a renderpass 1473- meson: Define _GNU_SOURCE for android host system 1474- meson/pps: Check if libdrm exists to compile pps 1475- meson/tu: Don't compile libdrm paths if KGSL is selected 1476- tu: Do not expose storage image/buffer features for PACK16 formats 1477- tu: Use hw binning or sysmem with QUERY_TYPE_PRIMITIVES_GENERATED 1478- tu: Don't count 3d blits in QUERY_TYPE_PRIMITIVES_GENERATED 1479- freedreno: Document rest of GRAS_LRZ_CNTL, clarify UNK_25 event 1480- freedreno: document GRAS_UNKNOWN_810A 1481- tu: Overhaul LRZ, implement on-GPU dir tracking and LRZ fast-clear 1482- ir3: Use NIR's info.writes_memory to detect when when to force late-z 1483- tu,freedreno: Refactored START/STOP events for pipeline stats 1484- tu: Fix prim gen query and pipeline stats query interaction 1485- zink: re-enable EXT_primitives_generated_query for Turnip 1486- util/u_trace: Fix iteration over config_control 1487- turnip: Refactor tracepoints generation to reduce duplication 1488- turnip: Add TU_GPU_TRACEPOINT envvar to toggle tracepoints 1489- freedreno: Add the rest of tracepoints with start/end to perfetto 1490- freedreno: Refactor tracepoints generation to reduce duplication 1491- freedreno: Add FD_GPU_TRACEPOINT envvar to toggle tracepoints 1492- docs/u_trace: Add documentation for \*_GPU_TRACEPOINT envvars 1493- turnip: Disable LRZ fast-clear for gen1 and gen2 1494- tu: Do not dereference descriptorSetLayout in push descriptors tmpl 1495- zink/turnip: Add comments to already triaged failures for tu baseline 1496- turnip: Simple breadcrumbs implementation to debug hangs 1497- tu: WFI after PC_TESSFACTOR_ADDR update 1498- freedreno: WFI after PC_TESSFACTOR_ADDR update 1499- ir3/ra: Always insert interval for precolored inputs 1500- ir3: Never remove GS_HEADER_IR3 sysval input 1501- tu: Disable LRZ write when alpha-to-coverage is enabled 1502- freedreno: Disable LRZ write when alpha-to-coverage is enabled 1503- ir3: Prevent reordering movmsk with kill 1504 1505Dave Airlie (82): 1506 1507- clc: add simple llvm initialise API 1508- clc: initialise one more llvm stage 1509- u_blitter/stencil: take dstbox x/y into accounts for dst fb width 1510- util/stencil: fix stencil fallback blit shader texture types. 1511- llvmpipe: add user memory resources to the debug global list. 1512- gallium/vl: wrap codec support checks in a common function. 1513- gallium/omx: add video codec supported hook for decode paths. 1514- meson: add a video codec support option 1515- gallium/vl: respect the video codecs configure in meson 1516- vulkan/wsi: keep allocate queue families in image, just don't fill them 1517- gallivm: move to new pass manager to handle coroutines change. 1518- radv: precalculate hs offchip parameters. 1519- radv: precalculate tess ring sizes/offsets. 1520- radv/ac: introduce a new common function for hs calcs. 1521- ac: port radeonsi tess factor calculations to common code. 1522- radeonsi: port tess ring calcs to the common helper. 1523- zink: avoid extra submits for empty wait cmdbufs. 1524- zink: don't finish barriers cmd buffer if not used. 1525- radv: cache the slab bo pointer in the pipeline. 1526- radv: only dirty line stipple when necessary. 1527- radv: optimised command buffer reset of vertex bindings. 1528- aco: move info pointer to a copy. 1529- aco: move to a minimal aco shader info struct. 1530- aco/info: reduce the gs ring info to what is needed. 1531- aco: remove radv specific streamout info 1532- aco: convert vs and so info over to aco structs. 1533- aco/info: add some more fields. 1534- aco/radv: convert to aco shader info at the radv level. 1535- aco: remove radv vs prolog key from aco internals. 1536- aco: drop unused radv include 1537- ac/radv: drop info pointer from the ac and radv shader structs 1538- aco/radv: drop radv_nir_compiler_options from aco. 1539- llvmpipe: align scratch size to 64-bit size. 1540- meson: add build-id to pipe libraries 1541- vl: fix codec checks to disable properly 1542- lavapipe: fix depth bias offset flag enables. 1543- llvmpipe: flush resources for kms swrast path. 1544- llvmpipe: handle timespec overflow on fence waits. 1545- nouveau: move codegen into a standalone library. 1546- nouveau: move codegen to a common higher level directory. 1547- nouveau/codegen: drop all ubytes from codegen. 1548- nouveau/codegen: drop gallium headers from the interface. 1549- llvmpipe: disable alpha test branch if multisample is enabled. 1550- egl/x11: split large put image requests to avoid server destroy 1551- egl/x11: add missing put_image cookie cleanups 1552- wsi/x11: add xcb_put_image support for larger transfers. 1553- glx/drisw: use xcb instead of X to query connection 1554- draw/cliptest: add support for guard band + full Z 1555- zink: workaround depth sampler border color when z24 is z32 1556- aco: refactor the radv binary builder out of the core aco fn. 1557- aco/radv: provide a callback from aco shader building to build binary 1558- aco/radv: provide a vs prolog callback from aco to radv. 1559- aco: drop radv_shader.h include 1560- radv: add a dynamic vertex format cache. 1561- lavapipe: drop unreachable pNext checks. 1562- lavapipe: enable variablePointers 1563- kms/dri: add mutex lock around map/unmap 1564- crocus: fail query begin if upload allocation fails. 1565- lavapipe: state latest conformance tests passed 1566- llvmpipe: fix aniso cube map arrays. 1567- crocus: fix leak in query code. 1568- llvmpipe: make last_fence a screen/rast object not a context one. 1569- llvmpipe: keep context list and use to track resource usage. 1570- gallivm: use progress from subgroup lowering. 1571- gallivm: fix casting around read invocation. 1572- llvmpipe: export ARB_shader_ballot 1573- lavapipe: expose some subgroup extensions 1574- zink: fixup subgroup vote/ballot enables. 1575- zink/llvmpipe/ci: skip broken CTS ballot test. 1576- util: add reallocarray wrapper for win32 1577- llvmpipe: Allocate scene tiles dynamically 1578- gallium: add a rasterizer state bit for unrestricted depth values. 1579- mesa/st: disable unrestricted fragment depth values for GL/GLES 1580- util/pack_color: clamp depth values outside range for unorm formats. 1581- llvmpipe/fs: handle unrestricted depth values. 1582- lavapipe: enable VK_EXT_depth_range_unrestricted. 1583- util/reallocarray: add errno.h include. 1584- lavapipe: enable VK_FORMAT_R5G5B5A1_UNORM_PACK16 1585- gallivm: fix printf hook for cached shaders. 1586- draw: don't touch info values that aren't valid. 1587- nir_to_tgsi_info: drop const_buffers_declared 1588- llvmpipe: finish rendering before flushing frontbuffer resources. 1589 1590David Heidelberg (58): 1591 1592- ci: skqp: patch skqp to report also GL results 1593- ci: skqp: upgrade to skqp based on Android CTS 11.0 (r7) 1594- ci: skqp: update URL 1595- ci: intel: Merge anv and iris into src/intel/ci 1596- ci/iris: Enable SKQP on Tiger Lake boards 1597- ci: traces: switch to Valve trimmed traces 1598- ci: traces: re-enable disabled traces which are now trimmed traces 1599- ci: traces: virgl: reenable previously crashing Godot trace 1600- ci: uprev piglit 2022-05-10 1601- ci/iris: skqp: add default GLES rendertests for TGL 1602- ci/iris: skqp: remove flaking atlastext for TGL 1603- ci/virgl: fix checksum for valve traces which run on iris 1604- ci/freedreno: enable ROR and Nheko traces 1605- ci/i915: add entries for RoR and Nheko traces 1606- ci/intel: add RoR and Nheko traces and reenable most of Valve traces 1607- ci/crocus: add RoR and Nheko traces 1608- ci/llvmpipe: add RoR and Nheko traces 1609- ci/radeonsi: add RoR and Nheko traces 1610- ci/virgl: traces: add RoR and Nheko traces 1611- ci/panfrost: add RoR and Nheko traces 1612- ci/iris: skqp: remove flaking atlastext for TGL (gl version) 1613- ci: traces: temporarily disable nheko trace 1614- ci/virgl: traces: temporarily disable nheko trace 1615- ci: uprev piglit 2022-06-09 1616- ci/etnaviv: sort fails.txt 1617- ci/etnaviv: add recent fails and remove 3 unexpected passes 1618- ci/etnaviv: disable tests which takes too long 1619- ci/etnaviv: add recent flakes for gc2000 1620- ci/etnaviv: gc2000: drop flakes from fails 1621- ci/freedreno: add Blender, Warzone2100, Freedoom and Unvanquished traces 1622- ci/freedreno: fix A530 glmark2@ideas:speed=10000 trace 1623- ci/llvmpipe: add Blender, Warzone2100, Freedoom and Unvanquished traces 1624- ci/iris: add Blender, Warzone2100, Freedoom and Unvanquished traces 1625- ci/virgl: add Blender, Warzone2100, Freedoom and Unvanquished traces 1626- ci/virgl: iris: disable Counter Strike 1.6 trace flaking 1627- ci/crocus: add Blender, Warzone2100, Freedoom and Unvanquished traces 1628- ci/panfrost: add Blender, Warzone2100, Freedoom and Unvanquished traces 1629- ci/iris: disable Blender Diorama trace (incompatible with perf. traces) 1630- ci/freedreno: update restricted traces to the brotli compressed version 1631- ci/freedreno: smuggle myself into restricted traces access list 1632- ci/freedreno: disable non-existing trace 1633- ci: Uprev apitrace to 790380e05854 1634- ci: traces: switch to brotli compressed traces 1635- ci/freedreno: add more restricted traces 1636- ci/freedreno: temporary disable AmnesiaTDD 1637- ci/freedreno: disable Stellaris trace 1638- ci/freedreno: disable SpecOps trace, each run flaky 1639- Revert "ci/freedreno: temporary disable AmnesiaTDD" 1640- ci/traces: piglit, be more verbose 1641- docs: briefly document Gallium Nine 1642- ci: add script for manage Mesa CI 1643- ci: Allow disabling the whole of the Igalia farm 1644- ci: move apitrace instalation outside of x86_test-vk 1645- ci: wine-apitrace bump to 11.1 1646- ci/freedreno: 3 pixel change in Raven restricted trace 1647- nine: char can be unsigned on non-x86 platforms 1648- mailmap: fix order of emails for Axel Davy 1649- ci: set reasonable amount of wget retries for each download 1650 1651Dawn Han (6): 1652 1653- Add iub features and properties passthrough 1654- Refactor the descriptor enums to be extensible 1655- Add iub type to vn_descriptor_type Verified the pNext VkWriteDescriptorSet is not NULL on iub type 1656- Add the iub binding count tracking 1657- Refactor and add template support for iub 1658- venus: enabled iub in physical device 1659 1660Denis Pauk (1): 1661 1662- panvk: Return VK_ERROR_INCOMPATIBLE_DRIVER for Midgard 1663 1664Dmitry Osipenko (1): 1665 1666- virgl: Fix unmapping of blob resources 1667 1668Dylan Baker (74): 1669 1670- VERSION: bump to 22.2-devel for next cycle 1671- docs: truncate new_features.txt 1672- docs: add release notes for 22.0.2 1673- docs: Add sha256sums for 22.0.2 1674- docs: update calendar and link releases notes for 22.0.2 1675- docs: update calendar for 22.1.0-rc1 1676- docs: update calendar for 22.1.0-rc2 1677- winsys/radeon: remove unused simple_list.h 1678- r300: remove unused simple_list.h 1679- gallium/rbug: replace simple_list.h with list.h 1680- gallium/driver_trace: remove simple_list.h 1681- util/list.h: add a function to move an item in a list 1682- gallium/util/u_cache.c: remove trailing whitespace 1683- gallium/util/u_cache: replace simple_list.h with list.h 1684- gallium/draw: remove trailing whitespace 1685- gallium/draw: use list.h instead of simple_list.h 1686- llvmpipe: cleanup trailing whitespace 1687- llvmpipe: replace uses of simple_list.h with list.h 1688- util: remove simple_list 1689- docs: Extend calendar entries for 22.0 by 1 releases. 1690- docs: update calendar and link releases notes for 22.0.3 1691- docs: Extend calendar entries for 22.1 by 1 releases. 1692- docs: update calendar for 22.1.0-rc3 1693- docs: update calendar for 22.1.0-rc4 1694- docs: add release notes for 22.0.3 1695- docs: Add sh256sum for mesa 20.0.3 1696- docs: update calendar and link releases notes for 22.1.0 1697- docs: add release notes for 22.1.0 1698- relnotes: Add sha256sum and fix minor formatting issues 1699- docs: Add calendar entries for 22.1 release. 1700- docs: add release notes for 22.0.4 1701- docs: add sha256sum to 22.0.4 notes 1702- docs: Extend calendar entries for 22.0 by 1 releases. 1703- docs: update calendar and link releases notes for 22.0.4 1704- docs: add release notes for 22.1.1 1705- docs: Add sh256sums for 22.1.1 1706- docs: update calendar and link releases notes for 22.1.1 1707- docs: add release notes for 22.0.5 1708- docs: update sha256 for 22.0.5 1709- docs: update calendar and link releases notes for 22.0.5 1710- docs: add release notes for 21.3.9 1711- docs Add sha256 sums for 21.3.9 1712- docs: link releases notes for 21.3.9 1713- docs: add release notes for 22.1.2 1714- docs: add sha256sum to 22.1.2 notes 1715- docs: update calendar and link releases notes for 22.1.2 1716- docs: drop 22.1.4 calendar entry. 1717- docs: Add calendar entries for 22.2 release candidates. 1718- docs: Extend calendar entries for 22.1 by 3 releases. 1719- docs: add release notes for 22.1.3 1720- docs: add sah256sum for mesa 22.1.3 1721- docs: update calendar and link releases notes for 22.1.3 1722- docs: Bump 22.2 branchpoint by two weeks 1723- anv: add gfx version 12.5 flushes to CCS path 1724- iris|anv: gfx version 12.5 data cache flush is not a workaround 1725- VERSION: bump 22.3.0-devel 1726- VERSION: bump for 22.2.0-rc1 1727- .pick_status.json: Update to 8e6bdb2ed31f27c28d258859ec48e8e67f51ff41 1728- .pick_status.json: Update to 0a0205f04552ab21971c431a5f8bb7486d664422 1729- .pick_status.json: Update to 1a3b086b06c6c1ff63a4abe3752ca0a7a854dc77 1730- .pick_status.json: Update to c67e60ae8fefaeadbe06bf0cbd8ad2ee267a6c0e 1731- .pick_status.json: Update to 70891edd9728e7dd73115171fdb03622ce6450ba 1732- VERSION: update to 22.2.0 1733- Revert "VERSION: update to 22.2.0" 1734- VERSION: bump to 22.2.0-rc2 1735- .pick_status.json: Update to a3bf0da1cbd4b10043c80bf44609a3024b5fcc36 1736- .pick_status.json: Update to 24b9ad7cd5ebc7cfa5d03cf0f243ea4841c971b9 1737- .pick_status.json: Update to 74fc367127ccf945f4c649dd6ddff955c802e36e 1738- .pick_status.json: Mark 11ab6087797f805cf158048915c67945613c9a72 as denominated 1739- VERSION: bump to 22.2.0-rc3 1740- .pick_status.json: Update to 0c6fbfca0c91ef012e8ab767a317c07f1f6dc5e6 1741- .pick_status.json: Update to 8eac45b27446cd9b9eaeb147af97fff1e09832cb 1742- .pick_status.json: Update to baf24dea943202b3a92cad0c9f9648597040955a 1743- .pick_status.json: Update to 3d4c36a3bcc51ed441b2667d92291bea30ef7449 1744 1745Eli Schwartz (1): 1746 1747- meson: add various generated header dependencies as order-only deps 1748 1749Ella Stanforth (1): 1750 1751- v3dv: Implement VK_KHR_performance_query 1752 1753Emil Velikov (1): 1754 1755- c11: reinstate the original license and authorship 1756 1757Emma Anholt (232): 1758 1759- nouveau/nir: Don't try to emit OP_FMA pre-nvc0. 1760- Revert "ci: Disable Google's lab" 1761- ci/virgl: Add piglit and GL4.3 testing to the virpipe path. 1762- nir: Add lowering for fround_even on r300. 1763- nir: Skip fround_even on already-integral values. 1764- nir: Avoid generating extra ftruncs for array handling. 1765- turnip: Add TU_DEBUG=layout for dumping image layouts. 1766- freedreno/ir3: Make sched nodes before adding deps. 1767- turnip: Add nir_opt_conditional_discard. 1768- freedreno/ir3: Call nir_opt_find_array_copies(). 1769- turnip: Ignore TOP/BOTTOM_OF_PIPE bits in subpass src/dst dep flags. 1770- turnip: Don't disable LRZ in subpasses after the first in the easy case. 1771- nir_to_tgsi: Add a workaround for virglrenderer TG4. 1772- nir_to_tgsi: Fix assertion failures handling 64-bit vec3/vec4 ssa undefs. 1773- nir_to_tgsi: Don't forget to split 64-bit store_per_vertex_output. 1774- nir_to_tgsi: Avoid swizzling from undefined channels in load_output. 1775- nir_to_tgsi: Allocate the primid sysval to num_inputs, not num_outputs. 1776- tgsi: Emit ureg HW_ATOMIC decls in range order. 1777- tgsi_to_nir: Emit load_ubo_vec4 instead of load_ubo on non-integer HW. 1778- nine: Disable optional use of TTN when MUL_ZERO_WINS is available. 1779- r600: Fix up some mis-indentation of blocks. 1780- ci/r600: Manual run updates. 1781- r600: Fix reading back from a temp array immediately after writing on RV770. 1782- r600/sb: Avoid causing an exception when getting the reciprocal of 0u. 1783- r600: Use nir-to-tgsi instead of TGSI when the NIR debug opt is disabled. 1784- nouveau/nv50: Print the number of loops in shader-db output. 1785- ci/nouveau: Add expectations files for GM206. 1786- ci/nouveau: Add nv92 xfails. 1787- nouveau/nir: Disable bitfield ops pre-nvc0. 1788- nouveau/nv50: Enable mesa/st alpha test lowering on nv50 with NIR. 1789- nouveau/nv50: Set the primid sysval flag if it's in the sysval list, too. 1790- nouveau/nir: Fix the inverted sense of usesSampleMaskIn. 1791- nouveau/nir: Fix edgeflag input detection. 1792- nouveau/nir: Set the input for vertex/instance ID like TGSI does. 1793- ci/dzn: Cut the runtime by running 1/3 of the enabled tests. 1794- nir_to_tgsi: Make vec_to_movs avoid unsupported coalescing for 64-bit. 1795- nir_to_tgsi: Lower FS input array indexing since we don't declare input arrays. 1796- tgsi: Extend array lengths when merging usage_masks. 1797- virgl: Switch to nir-to-tgsi by default. 1798- nir/lower_tex: Make the adding a 0 LOD to nir_op_tex in the VS optional. 1799- nir_to_tgsi: Drop the txl(lod=0)->tex nir_lower_tex() fixup. 1800- nouveau/nir: Move FS output stores to the end of the last block. 1801- ci/nouveau: Add MESA_GLES_VERSION_OVERRIDE=3.1 baseline state. 1802- nouveau/nir: Add support for pre-GF100 images and ssbos. 1803- nouveau/nir: Put the UBO offset indirect into the address reg. 1804- nouveau: Enable the NIR backend by default. 1805- util: Add some unit tests of the half-float conversions. 1806- nir: Eliminate out-of-bounds read/writes in local lowering. 1807- nir: Don't assert on tg4 offset range. 1808- svga: Add support for requesting NIR and translating to TGSI. 1809- ci/iris: Demote APL deqp to manual-only for now. 1810- ci/iris: Add a bunch of APL and KBL flakes recently. 1811- mesa/st: Always generate NIR from GLSL, and use nir_to_tgsi for TGSI drivers. 1812- mesa/st: Remove st_glsl_to_tgsi. 1813- mesa/st: Remove TGSI shader program variant handling. 1814- mesa/st: Drop the attrib remapping in st_prepare_vertex_program(). 1815- mesa/st: Remove TGSI-only shader lowering code. 1816- tgsi_scan: Remove unused input_array_last/output_array_last. 1817- glsl: Remove UBO reference lowering. 1818- mesa: Remove unused Mesa IR PROGRAM_* register files. 1819- gallium/tgsi: Remove tgsi_emulate. 1820- mesa/st: Remove now unused TGSI paths from disk cache. 1821- glsl: Retire the non-NIR GLSL linking paths. 1822- gallium: Remove now-unused shader caps. 1823- glsl: Remove the unused lower_if_to_cond_assign. 1824- glsl: Drop the dead MOD_TO_FLOOR path. 1825- mesa/st: Clean up a bit of st_prepare_vertex_program(). 1826- compiler/glsl: Remove the dead parts of build_program_resource_list(). 1827- glsl: Remove EmitNoLoops and the associated lower_jumps(lower_break=true) code. 1828- glsl: Remove unused lower_variable_index_to_cond_assign. 1829- svga: Set lower_bitops for vpu9 screen. 1830- glsl: Stop lowering ir_quadop_vector. 1831- nouveau: disable fences when running under drm-shim. 1832- Revert "ci: remove nouveau from shader-db runs" 1833- ci/iris: Cut the glk-deqp test coverage in half. 1834- glsl: Make all drivers take the GLSLOptimizeConservatively path. 1835- turnip: Add a TU_DEBUG=perf debug option. 1836- ci/crocus: Manual CI updates after CI was down for a bit. 1837- ci/crocus: Merge the piglit runs with the deqp runs. 1838- i915g/ci: Add depth-clear-precision-check xfails like everyone else. 1839- ci/crosvm: Simplify the CID setup. 1840- ci/crosvm: Rename VSOCK_TEMP_DIR -> VM_TEMP_DIR. 1841- ci/crosvm: Terminate the previous crosvm after a deqp-runner timeout. 1842- freedreno/ir3: Fix 16-bit bit_count. 1843- spirv: Use nir_vec_scalars() to simplify matrix transpose. 1844- spirv_to_nir: Cast RelaxedPrecision ALU op dests to mediump. 1845- turnip: Make RelaxedPrecision-decorated ALU ops 16-bit. 1846- freedreno/ir3: Add support for 16-bit nir_texop_lod. 1847- freedreno/ir3: Lower texture instructions used only for f2f16 to 16-bit. 1848- mesa/st: Only scalarize for doubles lowering if we're lowering doubles. 1849- glsl,nir: Move i/umulExtended lowering to NIR. 1850- panfrost: always print the bad ALU op if we're failing to translate. 1851- ci/turnip: Add missing xfails for a618 full VK run 1852- nir_lower_mediump: Drop assertion about not containing movs. 1853- freedreno/ir3: Enable load/store vectorization for SSBO access, too. 1854- freedreno/ir3: Make the types of tex coord padding match. 1855- freedreno/ir3: Fix validation of half-precision image store values. 1856- freedreno/ir3: Fold 16-bit conversions into image load/store src/dsts. 1857- i915g: Ignore the new SEPARABLE_PROGRAM property. 1858- ci/i915: Merge the piglit and deqp runs. 1859- ci/i915: Update manual piglit job expectations. 1860- turnip: Don't bother creating tile_load/store_cs for sysmem rendering. 1861- turnip: Refactor a bit of tu6_emit_tile_select(). 1862- turnip: Move binning decisions from FB usage time to FB creation time. 1863- turnip: Use fb->binning_possible to decide on conditional tile load/stores. 1864- ci/turnip: Add a little forced touch-testing of XFB with no binning requested. 1865- turnip: Refactor a bit of repeated code for subpass setup. 1866- turnip: Include 3d-based CmdClearAttachments() in binning visibility. 1867- turnip: Refactor a bit of subpass attachment processing. 1868- turnip: Allow load/store skipping in vkCmdClearAttachments(). 1869- turnip: Move tile loads back into the draw CS. 1870- nouveau/nv30: Make sure fsat is lowered in the VS. 1871- gallium+glsl: Remove EmitNoSat/PIPE_CAP_VERTEX_SHADER_SATURATE 1872- glsl: Remove stale lower_instructions comments. 1873- glsl: Move exp/log-to-exp2/log2 lowering to glsl-to-NIR. 1874- tgsi_exec: Fix inf/nan handling for divide by zero. 1875- glsl: Drop the div-to-mul-rcp lowering for floats. 1876- nir: Fix idiv lowering on !NativeIntegers when lower_fdiv is also set. 1877- glsl: Drop INT_DIV_TO_MUL_RCP lowering. 1878- ci/crocus: Drop g41's xfail for rgba_half_float_oes. 1879- ci/iris: Disable skqp until it can be stabilized. 1880- turnip: Reverse the order of walking pipes or tiles on odd rows. 1881- mesa/st: Abort the linking on driver link failure. 1882- .gitignore: Qualify the path for the ignored build directory. 1883- nir: Rename is_arb_asm to use_legacy_math_rules and document its meaning. 1884- gallium: Rename MUL_ZERO_WINS to LEGACY_MATH_RULES. 1885- nir_to_tgsi: Set LEGACY_MATH_RULES cap for use_legacy_math_rules shaders. 1886- r600/sfn: Implement 0*x=0 behavior for use_legacy_math_rules. 1887- nouveau/nir: Implement mul_zero_wins behavior for use_legacy_math_rules. 1888- tgsi_to_nir: Set use_legacy_math_rules for TGSI_PROPERTY_LEGACY_MATH_RULES. 1889- gallium: Rename PIPE_CAP_TGSI_LEGACY_MATH_RULES to drop "TGSI" 1890- Revert "nine: Disable optional use of TTN when MUL_ZERO_WINS is available." 1891- iris: Enable PIPE_CAP_LEGACY_MATH_RULES. 1892- crocus: Enable PIPE_CAP_LEGACY_MATH_RULES. 1893- ci/dzn: Skip a test that intermittently times out at a minute. 1894- ci/iris: Disable blender-demo-cube_diorama on APL. 1895- ci: Terminate capture-devcoredump before tarring up artifacts. 1896- turnip: Document some fields about resolves. 1897- ci/bare-metal: Apply autopep8 to our python scripts. 1898- ci/bare-metal: Add per-boot-stage timeouts for fastboot and poe. 1899- ci/bare-metal: Get rid of servo's serial feed threads. 1900- ci/bare-metal: Add handling of netboot firmwares for servo boards. 1901- ci/bare-metal: Add test phase timeouts to all boards. 1902- ci: Remove the unused .test-manual 1903- ci/vc4: Merge quick_shader in with deqp-gles 1904- ci/vc4: Turn on deqp-egl testing by default. 1905- vc4: Propagate txf_ms's dest_type to the lowered txf. 1906- freedreno/ir3: Disable image/ssbo 16-bit conversion folding pre-a6xx. 1907- freedreno/a5xx: Set the buffer bit appropriately in XS_CTRL_REG0. 1908- ci/freedreno: Turn a530 back on by default and update expectations. 1909- ci/bare-metal: Remove "stage: test" from .baremetal-test. 1910- ci/bare-metal: Consolidate needs declarations in .baremetal-test-\*. 1911- ci/bare-metal: Collapse artifacts wget by default. 1912- ci/traces: GC unused code for DXVK trace replay. 1913- ci/traces: Drop ZINK_USE_LAVAPIPE ICD override. 1914- ci/freedreno: Add vulkan+gl integration testing in piglit. 1915- ci/turnip: Test traces on turnip using zink. 1916- ci/turnip: Trim the a630 VK run a bit. 1917- ci: Use "!references" to manage scheduled pipeline rules. 1918- ci: Pull out farm disables to a reused rule. 1919- ci: Pull common zink frontend rules to a shared rule list. 1920- ci/freedreno: Use !references to clean up restricted traces rules. 1921- ci: Simplify vulkan rules using !references. 1922- ci: Split core GL from core VK-or-GL rules. 1923- ci/freedreno: Filter when we run test jobs for VK or GL changes. 1924- ci/zink: Reuse anv-rules. 1925- ci/zink: Simplify lavapipe rules setup, and clarify what the rules are for. 1926- ci: Remove the trailing "when: never"s from rules. 1927- ir3: Retire the cp postsched pass now that we do RA in SSA. 1928- ci: Uprev deqp to 1.3.3.0. 1929- ci/llvmpipe: Skip ssbo.layout.random.all_shared_buffer.36 1930- ci: Make the retry policy default for all jobs. 1931- ci: Rename .ci-run-policy rules to .build-rules. 1932- ci: Drop .build-rules from container jobs. 1933- ci: Remove .build-rules from core test job definitions. 1934- ci: Fix non-freedreno performance jobs running during Marge merges. 1935- ci: Remove GIT_STRATEGY from .test-manual-mr. 1936- mesa/st: Disable OES_texture_3D for PIPE_CAP_MAX_TEXTURE_3D_LEVELS==0. 1937- vc4: Disable OES_texture_3D being exposed. 1938- vc4: Move previous existing 3D xfails up to the group of 3d xfails. 1939- vc4: Add notes on the remaining dEQP failures. 1940- turnip: Account for additional_cs_reserve_size for both bin and render. 1941- turnip: Fix up per-stage additional size accounting. 1942- turnip: Fix the reservation for vertex inputs. 1943- turnip: Refactor vertex input setup a little. 1944- ir3: Make sure to pass the interp_mode through in our load_bary lowering. 1945- nir+ir3: Rename load_size_ir3 to load_center_rhw_ir3. 1946- freedreno: Rename the "SIZE" regs for interpolateAtOffset to "CENTERRHW" 1947- ir3: Clarify what's happening in the interpolateAtOffset() math. 1948- ir3: Use non-persp interpolation when appropriate for interpolateAtOffset. 1949- zink: Do the timestamp-to-ns math in a double to have better precision. 1950- mesa: Fix the error check for VertexAttrib*. 1951- vc4: Work around a HW bug with 2-vert line loops. 1952- vc4: Add links to test bug reports. 1953- kopper: Fix the return value of kopperSetSwapInterval(). 1954- kopper: Use the swap interval that was set at swapchain creation time. 1955- glx: Unify dri2/dri3 vblank_mode logic. 1956- kopper: Respect the vblank_mode env var. 1957- egl: Just include driconf.h instead of redefining its values. 1958- loader: Just include driconf.h instead of redefining its values. 1959- ci/freedreno: Mark an occasional flake pass that happens on a530. 1960- ir3: Fix the no-emitted-vertex condition emission in geom lowering. 1961- mesa/arbprog: Move the GLSLFragCoordIsSysVal handling to prog_to_nir. 1962- mesa/arbprog: Use nir_lower_io_to_temporaries. 1963- mesa/arbprog: Stop doing optimization in the ARB program IR. 1964- ci/crocus: Update portal 2 trace shas for the recent fix. 1965- ci/crocus: Disable the blender trace. 1966- ci/crocus: Drop xfails for the recent image external fix. 1967- freedreno/ir3: Enable core NIR's 16-bit ALU optimizations. 1968- intel/fs: Simplify brw_barycentric_mode() args. 1969- ci/bare-metal: Add timeouts to the shell commands called in fastboot. 1970- ci/freedreno: Switch a630 to manual/disabled for lab maintenance. 1971- glsl: Use the same NIR path for shared mem lowering as SPIRV does. 1972- glsl: Remove optimize_swizzles. 1973- ci/turnip: Add a bit of spilling-vs-ballot testing on a618. 1974- ci/freedreno: Add some more known flakes for a630 from our IRC logs. 1975- ci/turnip: Bump up the a630 full run timeout. 1976- ci/turnip: Add a couple of missing a630 fails. 1977- Revert "ci/freedreno: Switch a630 to manual/disabled for lab maintenance." 1978- nir: Split usub_sat lowering flag from uadd_sat. 1979- ci/crocus: Fix the commented YML for blender-demo-cube_diorama. 1980- gallivm: Refactor a bit of UBO/SSBO range checking. 1981- gallivm: Optimize single-invocation SSBO stores. 1982- ci/virgl: Make the VM-based virgl tests report to #virgl-ci, too. 1983- nir: Add an opt pass for phis after if choosing between true/false. 1984- glsl: Remove opt_conditional_discard(). 1985- nir: Use nir_foreach_phi_src consistently. 1986- ci/freedreno: Update a630 s8 resolve xfails. 1987- turnip: Split the tiling config into separate layouts based on CCU usage. 1988- turnip: Use the GMEM CCU space for attachments when the stores won't. 1989- zink: Make sure that we keep the existing ici pNext chain on inserts. 1990- spirv: Mark phis as mediump instead of directly lowering them to 16 bit. 1991 1992Enrico Galli (6): 1993 1994- microsoft/spirv_to_dxil: Add pass to lower dynamic accesses on ubo[1] 1995- microsoft/compiler: Fix for arrays of 1 element 1996- microsoft/compiler: Add support for unnamed ubos 1997- microsoft/compiler: Fix when using a shadow sampler more than once 1998- microsoft/spirv_to_dxil: Fix discard semantics 1999- microsoft/spirv_to_dxil: Fix discard to demote + return pass 2000 2001Eric Engestrom (36): 2002 2003- meson: replace deprecated dep.get_pkgconfig_variable(...) with dep.get_variable(pkgconfig : ...) 2004- docs: show and link to the new macOS CI 2005- broadcom/rpi4-skips: drop duplicated lines 2006- mailmap: update my email to Igalia 2007- vulkan: add script to detect & replace identifiers that have been renamed 2008- vulkan/runtime: use updated tokens from vk.xml 2009- vulkan: use updated tokens from vk.xml 2010- nir: use updated tokens from vk.xml 2011- radv: use updated tokens from vk.xml 2012- v3dv: use updated tokens from vk.xml 2013- turnip: use updated tokens from vk.xml 2014- zink: use updated tokens from vk.xml 2015- lavapipe: use updated tokens from vk.xml 2016- pvr: use updated tokens from vk.xml 2017- anv: use updated tokens from vk.xml 2018- dzn: use updated tokens from vk.xml 2019- panvk: use updated tokens from vk.xml 2020- venus: use updated tokens from vk.xml 2021- zink: fix portability_subset usage after rename from EXTX to KHR 2022- zink: drop unused VkPhysicalDevicePortabilitySubsetPropertiesKHR 2023- zink: add missing guards around \`have_{ext}` 2024- docs/isl: fix typos & formatting 2025- vk/util: handful of pythonic cleanups 2026- vk/util: simplify extensions gen code 2027- util/list: rename LIST_ENTRY() to list_entry() 2028- nouveau: use existing list_first_entry() macro instead of re-writing it 2029- docs/features.txt: mark VK_ANDROID_native_buffer as supported on turnip 2030- docs/features.txt: mark VK_ANDROID_native_buffer as supported on v3dv 2031- docs/ci: fix Emma's name 2032- turnip: expose support for VK_EXT_acquire_drm_display 2033- ci: add \`--print-errorlogs` to \`meson test` 2034- bin/gen_release_notes.py: bump advertised vulkan version to 1.3 2035- vk/device-select-layer: fix .sType of VkPhysicalDeviceGroupProperties 2036- wsi/x11: fix memleak in wsi_x11_connection_create() 2037- meson: replace manual compiler flags with meson arguments 2038- broadcom: fix dependencies in static_library() calls 2039 2040Erico Nunes (11): 2041 2042- ci: put lima farm back online 2043- docs/ci: fix literalinclude of caching setup files 2044- lima/ppir: optimize branch comparison when possible 2045- lima/ppir: merge branch nodes in simple cases 2046- egl: Fix DETECT_OS macro usage 2047- ci: update linux kernel to v5.17 2048- ci: update docs for linux kernel uprevs 2049- lima/ci: update piglit ci expectations 2050- lima/ci: reenable lima-mali450-piglit-gpu:arm64 2051- lima/ci: Don't skip piglit opengl 1 tests 2052- ci: Revert "CI: Lima farm is offline" 2053 2054Erik Faye-Lund (157): 2055 2056- gallium: rename continue shader-cap 2057- gallium: rename dround shader-cap 2058- gallium: rename dfracexp/dldexp shader-cap 2059- gallium: rename ldexp shader-cap 2060- vulkan: explicitly cast object-type enum 2061- meson: turn on -Wno-unused-function project-wide 2062- nir: introduce and use nir_component_mask 2063- dzn: fixup indent 2064- dzn: remove all usage of ComPtr<T> 2065- dzn: always use ID3D12Device1 2066- dzn: always use ID3D12GraphicsCommandList1 2067- dzn: use ID3D10Blob instead of ID3DBlob 2068- dzn: pass IDXGIAdapter1 to d3d12_create_device 2069- dzn: port to d3d12 c-api 2070- dzn: add a bunch of missing struct-keywords 2071- dzn: drop auto usage 2072- dzn: drop using references 2073- dzn: no more reinterpret_cast 2074- dzn: use define instead of constexpr 2075- dzn: use c-style for-statement 2076- dzn: use c-style initialization 2077- dzn: c-style casts 2078- dzn: do not set unused default member initializer 2079- dzn: port code to plain c 2080- dzn: merge util sources 2081- dzn: remove needless defines 2082- dzn: remove some needless casts 2083- dzn: clean up pointer syntax 2084- dzn: add space before += operator 2085- dzn: drop needless enum-casts 2086- dzn: fixup indent 2087- dzn: remove stale cpp_args 2088- nir/lower_int64: do not try to clamp floats to int-range 2089- vulkan: do not depend on alignof(void) 2090- lavapipe: quiet non-conformant warning on ci 2091- panvk: quiet non-conformant warning on ci 2092- vulkan: use c_msvc_compat_args for shared code 2093- vulkan: drop empty vulkan_wsi_args 2094- microsoft/clc: fixup indentation 2095- microsoft/compiler: fixup indentation 2096- meson: remove unused defines 2097- meson: deprecate specifying osmesa-bits 2098- gallium/xlib: fix stale comment 2099- mesa: add missing error-path 2100- util: fix test on msvc 2101- lavapipe: tolerate NULL UBO and SSBO descriptors 2102- lavapipe: tolerate NULL image view descriptors 2103- lavapipe: tolerate NULL image-buffer view descriptors 2104- lavapipe: tolerate NULL sampler-buffer view descriptors 2105- lavapipe: add support for null-descriptors from EXT_robustness2 2106- zink: pass screen to extension getter macro 2107- zink: initialize dispatch-table before queue-init 2108- zink: call vk-functions through dispatch-table 2109- zink: pass screen to zink_create_instance 2110- zink: dynamically load a few functions 2111- zink: use run-time linking to loader 2112- dzn: reset correct list 2113- vulkan/wsi: unbreak win32-support 2114- d3d12: add missing cmath includes 2115- tests/graw: rename shaders from .sh to .txt 2116- dzn: add a debug-flag to wait for the debugger 2117- dzn: add debug option to redirect stdout/stderr 2118- editorconfig: remove scons-ism 2119- editorconfig: remove html-rule 2120- editorconfig: remove m4-rule 2121- editorconfig: remove pl-rule 2122- radeonsi: port amdgcn_glslc build to meson 2123- freedreno: remove stale makefile 2124- intel: remove stale makefile 2125- meson: modernize win_flex stdint.h logic 2126- c99_compat.h: remove inline and __func__ shims 2127- util: add missing c99_compat.h includes 2128- util: remove needless c99_compat.h includes 2129- egl: remove needless c99_compat.h includes 2130- gallium/util: remove needless c99_compat.h includes 2131- mapi: remove needless c99_compat.h includes 2132- gallium: remove needless c99_compat.h includes 2133- include: remove needless c99_compat.h includes 2134- intel/isl: remove needless c99_compat.h includes 2135- vulkan/device_select: remove needless c99_compat.h includes 2136- include: drop c99_math.h 2137- include: drop c11_compat.h 2138- pvr: do not use c_msvc_compat_args 2139- turnip: do not do STATIC_ASSERT on a variable 2140- freedreno: do not use variable in STATIC_ASSERT 2141- freedreno: do not do STATIC_ASSERT on variables 2142- iris: do not do STATIC_ASSERT on variables 2143- freedreno: degrade STATIC_ASSERT to assert 2144- v3dv: do not do STATIC_ASSERT on variables 2145- util: add IS_POT macro 2146- intel/compiler: use macro for power-of-two check 2147- broadcom/compiler: use macro for power-of-two check 2148- util: use static_assert directly 2149- util: implement STATIC_ASSERT using c++11 / c11 primitives 2150- meson: add back -Werror=vla to msvc compat flags 2151- ci/windows: add back build-error detection 2152- d3d12: move cubemap-lowering to common-code 2153- microsoft/compiler: mark image-functions as such 2154- microsoft/compiler: make sampler-lowering optional 2155- microsoft/spirv_to_dxil: lower cube-images to 2d arrays 2156- dzn: enable binding_model tests 2157- dzn: correct assert-condition 2158- dzn: handle stencil-attachment-optimal layout 2159- dzn: use a null-rtv to handle no-attachment 2160- dzn: enable two more exts 2161- util: sort includes 2162- llvmpipe: assume c99 2163- c99_compat.h: assume c99 support 2164- util: use c11 alignof instead of our own 2165- glcpp: remove outdated msvc hack 2166- vulkan/util: assume c11 2167- util: use c11 alignas instead of rolling our own 2168- gallium: remove stale comment 2169- gallium: reorganize includes 2170- llvmpipe: use c11 alignas instead of PIPE_ALIGN_VAR 2171- gallium: use c11 alignas instead of PIPE_ALIGN_VAR 2172- gallium: use c++11 alignas instead of PIPE_ALIGN_VAR 2173- gallium: remove repeated check 2174- gallium: remove unused macros 2175- gallium: clean up PIPE_ALIGN_STACK-definition 2176- docs/zink: correct requirement for tbos 2177- dzn: correct order of src_subres and dst_subres args 2178- docs: set language to english 2179- dzn: fill in driver name and info 2180- dzn: fill in float-control details 2181- dzn: fill in non-uniform-indexing props 2182- dzn: fill in bindless props 2183- dzn: fill in depth/stencil resolve props 2184- dzn: fill in minmax props 2185- dzn: fill misc props 2186- dzn: expose VK_KHR_driver_properties 2187- zink: remove pointless test 2188- zink: remove pointless assert 2189- zink: drop pointless comment 2190- llvmpipe: merge all rasterizer-bind functions 2191- llvmpipe: consolidate legacy_points and multisample state 2192- llvmpipe: clean up bresenham implementation 2193- docs/zink: correct name of extension 2194- docs/zink: document required VK_KHR_swapchain_mutable_format 2195- docs/zink: document required VK_EXT_border_color_swizzle 2196- zink: warn about missing image2DViewOf3D feature 2197- zink: fix EXT_color_write_enable check 2198- gallium/hud: do not use texture-rect for font 2199- llvmpipe: remove unused function 2200- zink: clean up extension emitting 2201- zink: set dynamic indexing caps 2202- zink: more accurately set {Sampled,Image}1D caps 2203- zink: more accurately set {Sampled,Image}Buffer caps 2204- zink: more accurately set ImageQuery cap 2205- docs: fixup link to virgl docs 2206- zink: add have_D32_SFLOAT_S8_UINT boolean 2207- zink: do not use VK_FORMAT_D32_SFLOAT_S8_UINT without checking 2208- zink: type_main -> type_void_func 2209- zink: add spirv_builder_function_call 2210- zink: wrap discard in a function 2211- zink: clamp miplodbias when creating sampler 2212- docs/zink: document rgtc requirement 2213 2214Fabrice Fontaine (1): 2215 2216- src/util/futex.h: fix build on 32-bit architectures using 64-bit time_t 2217 2218Feng Jiang (1): 2219 2220- virgl: do not share virgl_screen between different drm_files 2221 2222Filip Gawin (5): 2223 2224- r300: Print warning when stubbing derivatives 2225- r300: keep negation if w is an inline constant 2226- r300: don't check for unitialized reads when rewriting register 2227- r300: prefer old not native swizzle in constant folding 2228- r300: don't read from output transform_r300_vertex_SEQ/SNE 2229 2230Francisco Jerez (2): 2231 2232- intel/dev: Compute pixel pipe information based on geometry topology DRM query. 2233- intel/fs/xehp+: Emit scheduling fence for all NIR barriers on platforms with LSC. 2234 2235Frank Binns (3): 2236 2237- pvr: replace p_compiler.h boolean with stdbool 2238- pvr: shorten error to err in label names 2239- pvr: Add TI AM62 as a supported device. 2240 2241Georg Lehmann (74): 2242 2243- aco: Remove some old comments in aco_opcodes.py. 2244- aco: Reuse previous -1 result in find_msb to avoid using VOP3. 2245- nir/lower_mediump: Add a new pass to fold 16bit image load/store. 2246- aco: Implement image_store d16. 2247- aco: Implement image_load d16. 2248- radv: Use nir_fold_16bit_image_load_store_conversions. 2249- radv: Run copy_prop and dce after folding 16bit sampling/load/store. 2250- aco: Remove D16 zero components from image stores. 2251- meson: Tell glslang to be quiet. 2252- radeonsi: Use nir_fold_16bit_image_load_store_conversions. 2253- wsi/x11: Don't leak xcb_get_geometry_reply_t. 2254- nir/opt_algebraic: Fix mask in shift by constant combining. 2255- nir/opt_algebraic: Simple xor/ishr optimizations. 2256- radv: Don't retile read-only images. 2257- nir/opt_algebraic: Optimize Doom Eternal's word extract by LSB. 2258- aco: Implement uclz. 2259- aco: Implement usub_sat. 2260- aco: Implement isub_sat. 2261- ac/llvm: Implement usub_sat and isub_sat. 2262- ac/llvm: Implement uclz. 2263- nir: Add optional lowering for mul_32x16. 2264- radv: Lower mul_32x16. 2265- radv: Lower hadd/hradd. 2266- radv: Lower 8bit isub_sat/usub_sat. 2267- radv: Enable VK_INTEL_shader_integer_functions2. 2268- radv, aco: Packed usub_sat/isub_sat. 2269- zink: Use VK_USE_64_BIT_PTR_DEFINES to check for 64bit platforms. 2270- radv, aco: Lower txf offset in NIR. 2271- nir/lower_tex: Add lower_array_layer_round_even option. 2272- radv, aco: Round texture array layer in NIR. 2273- vulkan: Spec update to 1.3.217. 2274- radv: Implement VK_EXT_non_seamless_cube_map. 2275- aco: Print r128/a16 MIMG bits separately. 2276- aco: Remove r128_a16 MIMG builder option. 2277- anv: Implement VK_EXT_non_seamless_cube_map. 2278- radv: Remove redundant memset in radv_descriptor_set_create. 2279- radv: Check descriptor pool entry count before allocating a new set. 2280- radv: Don't check if we need to copy immutable samplers for non push templates. 2281- aco/ra: Static assert that changing instruction type to VOP2 is valid. 2282- aco/ra: Move mac encoding optimization to its own function. 2283- aco/ir: Pad SOP2 and SOPC to the same size as SOPK. 2284- aco/ra: Optimize some SOP2 instructions with literal to SOPK. 2285- aco/optimizer: Convert s_add_u32 with literals to s_add_i32 if carry is not used. 2286- aco/optimizer: Optimize SOPC with literal to SOPK. 2287- r600/sfn: Add missing std::array include. 2288- aco: Fix swapping sources in SOPC -> SOPK optimization. 2289- aco/assembler: Fix s_bitreplicate_b64_b32 on GFX9. 2290- nir_to_tgsi: Lower uadd_sat/usub_sat. 2291- zink: Lower uadd_sat/usub_sat. 2292- r600: Lower uadd_sat/usub_sat. 2293- ir3: Implement [iu]sub_sat. 2294- nir/lower_io: Fix array length of buffers larger than INT32_MAX. 2295- nir/opt_algebraic: Optimize [ui](add|sub)_sat with 0. 2296- radv,aco: Don't lower and vectorize 16bit iabs. 2297- radv: Remove ineg from lower_bit_size_callback callback. 2298- radv, aco: Don't lower 16bit isign. 2299- ir3: Only run 16bit tex NIR passes on a5xx+. 2300- ir3: Lower alu to scalar if nir_legalize_16bit_sampler_srcs made progress. 2301- nir: Rewrite and merge 16bit tex folding pass with 16bit image folding pass. 2302- nir/lower_mediump: Add an option to only fold if all tex sources can be folded. 2303- ir3: Stop using nir_legalize_16bit_sampler_srcs. 2304- aco: Add G16 opcodes. 2305- aco: Support 16bit sources for texture ops. 2306- radv: Fold 16bit tex sources. 2307- aco: Add a note about G16 without A16 to README-ISA.md. 2308- aco/ir: Generalize (un)ordered_swapped. 2309- aco/ir: Add integer get_cmp_info. 2310- aco/ir: Fix swapped nle. 2311- aco/ir: Add vcmpx opcode to get_cmp_info. 2312- aco/ir: Add v_cmp_class to get_cmp_info. 2313- aco/ir: Add swapped opcode for v_cmp_u/v_cmp_o. 2314- radeonsi: Stop using nir_legalize_16bit_sampler_srcs. 2315- nir/fold_16bit_tex_image: Default to only_fold_all. 2316- aco: Force tex operand to have the correct sub dword size before packing. 2317 2318Gert Wollny (85): 2319 2320- virgl: Extend integer write out output fix to all non-move integers ops 2321- r600/sfn: Fix store_shared_r600 write masks 2322- ntt: Add option to not optimized register allocation 2323- nir: Add pass to split 64 bit vec3 and vec4 variable access and phis 2324- ntt: Split 64 bit vec3 and vec4 local variables and phis 2325- virgl: Clear all indirect flags when writing to a temporary 2326- nir: Don't optimize to 64 bit fsub if the driver doesn't support it 2327- ntt: remove dead input variables before lowering FS IO 2328- r600/sb: Don't optimize float GT and GE 2329- r600: tune nir options 2330- r600: Add support for TGSI_OPCODE_ATOMIMIN and IMAX 2331- virgl: don't move input gl_SampleMaskIn to a temp 2332- r600: Allow eight bit, three channel formats for vertex buffers 2333- nir: honor lower_double options for ffloor and ffract 2334- r600/sb: Don't create three source ops with all kcache values 2335- r600/sb: Fall back to un-optimized shader if scheduling fails 2336- virgl: Don't let ntt optimize the register allocation 2337- tgsi: Don't bother setting the TGSI_PROPERTY_SEPARABLE_PROGRAM 2338- virgl: Submit the TGSI_PROPERTY_SEPARABLE_PROGRAM when the host understands it 2339- r600: Add number of ALU groups to statistics 2340- r600: Make sure that LDS instructions only use bank swizzle 012 2341- r600: Update nir options 2342- r600/sfn: rewrite NIR backend 2343- r600: enable sb by default also for NIR 2344- r600: Cleanup nir options 2345- r600/sfn: support nir_op_mulz and legazy math rules 2346- r600/sfn: Add missing copyright information 2347- r600/sfn: Don't use assert to test success of operation 2348- r600: limit loops when trying to merge alu groups 2349- r600/sfn: Never consider an op with register dest as dead 2350- r600/sfn: Add support for fdph 2351- r600/sfn: lower CLIPVERTEX to clip planes 2352- r600/sfn: emulate pmr::monotonic_buffer_resource if needed 2353- r600/sfn: clean up some commented out code 2354- r600/sfn: fix storing shader output array locations 2355- r600/sfn: lower undef to zero 2356- r600/sfn: count LDS queue pop reads separately in assembler 2357- r600/sfn: Fix the kcache failure handling 2358- r600/sfn: Fix emmision of LDS instructions 2359- r600/sfn: Fix used of local shared write mask 2360- r600/sfn: Fix streamout with non-zero component write 2361- r600: don't dump shader info to files on debug 2362- r600/sfn: make sure the memory pool is released after shader translation 2363- r600/sfn: lower the right shader variant 2364- r600: Delete possible old NIR variant when translating from TGSI 2365- r600: don't clone the nir shader, just use it 2366- r600ß/sfn: Don't fall through with VS outputs 2367- r600/dri-tweak: Disable ARB_shader_image_load_store for "Tomb Raider 2013" 2368- r600/sb: fall back to unoptimized shader if RA fails 2369- r600: Don't set TES_EVAL UCP dirty when it is not supported 2370- virgL: lower fneg and fabs 2371- virgl: honor host caps in shader cash sha evaluation 2372- r600/sfn: Factor out fragement shader class for EG 2373- r600/sfn: Support pre-EG style FS input 2374- r600/sfn: Prepare StreamOut instruction for pre EG opcodes 2375- r600/sfn: Handle slot differences between pre-EG and EG 2376- r600/sfn: Handle chipclass with kcache allocation 2377- r600/sfn: Handle buffer size query on pre-EG 2378- r600/sfn: Handle pre-EG GS input fetch 2379- r600/sfn: Handle primitive input handling for pre-EG 2380- r600/sfn: Handle non-compressed MS texel fetch 2381- r600/sfn: Set trans flag for flt_to_int on pre-EG 2382- r600/sfn: Handle pre-EG buffer fetch 2383- r600/sfn: Disable FP64 for pre-CEDAR 2384- r600/sfn: Advertise support for fmulz 2385- r600/sfn: Enable NIR for pre RG hardware 2386- r600/sfn: remove leftover debug output 2387- r600/sfn: Handle R600 class sin/cos 2388- r600: increase possible stack size in binary code 2389- r600/sfn: Schedule shift instruction on R600 in t-slot 2390- r600/sfn: Add GS thread fix just like the TGSI code path 2391- r600/sfn: Sort FS inputs to make interpolated values come first 2392- r600/sfn: Fix color outputs when color0 writes all 2393- r600: Force NOPs when loading AR on R600 class hardware 2394- r600/sfn: Handle R600 scratch read 2395- r600: Don't use SB with R600 style scratch reads 2396- r600/sfn: Handle color0 writes all on R700 like on EG 2397- r600/sfn: Don't tag mem-ring and stream instructions as exports 2398- r600/sfn: Don't schedule GDS instructions early 2399- r600/sfn: Don't scan the whole block for ready instructions 2400- r600/sfn: Use a heuristic to keep SSBO setup and store close 2401- r600: Fix reporting TGSI IR support 2402- r600/sfn: Use a low number for unused target register 2403- virgl: when reading back wait first, then do the transfer 2404- r600/sfn: override register ID when it doesn't matter 2405 2406Glenn Kennard (1): 2407 2408- nv30: Fix non-scissored clears after a scissor has been set 2409 2410Greg Depoire--Ferrer (5): 2411 2412- util/queue: add missing space to comment in util_queue_destroy 2413- zink: check for error when initializing util_queue 2414- lavapipe: check for error when initializing lvp_queue 2415- util/queue: handle thread cration failure in util_queue_adjust_num_threads 2416- zink: fix crash on dlclose when zink_internal_create_screen failed 2417 2418Guilherme Gallo (48): 2419 2420- ci/lava: Encapsulate job data in a class 2421- ci/lava: Filter log lines from LAVA return 2422- ci/lava: Return test-suite result as exit code 2423- ci/lava: Fix LAVA job validation 2424- ci/lava: Set lava-signal to kmsg 2425- ci/lava: Always validate the lava job 2426- ci/lava: Use lava-test-case to run custom scripts in LAVA 2427- ci/lava: Trap init-stage2.sh background processes 2428- ci/lava: Fix shebang in deqp-runner.sh 2429- ci/lava: Parse all test cases from 0_mesa suite 2430- ci/lava: Let LAVA job submitter run without JWT file 2431- ci/lava: Reduce LAVA boot phase timeout to 3 minutes 2432- ci/lava: Cancel the job if the script is interrupted 2433- ci/lava: Improve exception handling 2434- ci/lava: Retry when data fetching log RPC call is corrupted 2435- ci: Fix init-stage2 exit code 2436- ci: Update trace after CI fix 2437- ci: Fix tests expectations 2438- ci/lava: Make job submitter parse the job result 2439- ci/lava: Fix colored LAVA outputs 2440- ci/lava: Fix Gitlab Section markers 2441- ci/lava: Add support for more complex color codes 2442- ci/lava: Filter out undesired messages 2443- ci/lava: Improve result parsing regex 2444- ci/lava: Create Gitlab log sections handler 2445- ci/lava: Create LogFollower and move logging methods 2446- ci/lava: Follow job execution via LogFollower 2447- ci/lava: heartbeat: don't consider kernel message logs 2448- ci/lava: Wrap job info into a collapsed section 2449- ci/lava: Wrap job definition dump into a collapsed section 2450- ci/lava: Add an integration test for LAVA jobs 2451- ci/lava: Don't print LAVA debug messages 2452- ci/lava: Highlight kernel messages in bold 2453- ci/lava: Highlight job retrying message 2454- ci/lava: Stop printing after the result line 2455- ci/lava: Flexibilize section marker regexes 2456- ci/lava: Rename console color names 2457- ci/lava: Update license header 2458- ci/lava: Split lava_log into modules 2459- ci/lava: Detect R8152 issues preemptively and retry 2460- ci/lava: Make hung job status yellow 2461- ci/lava: Color red for fatal and yellow for warning 2462- ci/lava: Add \`slow` pytest marker 2463- ci/lava: Add canceled job status 2464- ci/lava: Add delay before calling lava-test-case 2465- ci/lava: Increase boot timeout 2466- ci/lava: Only parse result within testcase section 2467- ci/lava: Remove MINIO_RESULTS_UPLOAD if requested 2468 2469Hans-Kristian Arntzen (4): 2470 2471- radv: Fix RTPSO hashing of pGroups. 2472- vulkan: Update Vulkan XML and headers to 1.3.219. 2473- vulkan: Add common code for VK_EXT_shader_module_identifier. 2474- radv: Implement VK_EXT_shader_module_identifier. 2475 2476Hoe Hao Cheng (4): 2477 2478- zink/codegen: remember the fields in feats/props structs of extensions 2479- zink/codegen: do not include compilation structs with extension structs 2480- zink/codegen: do not automatically consider extensions promoted 2481- zink/codegen: use the updated identifiers 2482 2483Hyunjun Ko (15): 2484 2485- radv: fix assertion for the count of xfb output 2486- turnip: add an assertion for max descriptor set count. 2487- ir3: handle intrinsic_load_draw_id when scanning driver constants 2488- turnip: clamp to 1 for negative bottom-right of viewport. 2489- turnip: fix wrong size for lrz cs when lrz_track_quirk is enabled. 2490- freedreno,ir3: rename Z_CLAMP_ENABLE to Z_CLIP_DISABLE 2491- turnip: keep the depth_clip_disable state at the pipeline builder. 2492- zink: remove the workaround for depth_clip_enable on turnip 2493- turnip: Remove an unnecessary assert. 2494- turnip: expose VK_EXT_shader_module_identifier 2495- ir3: change maximum size of const files. 2496- ir3: handle shared consts. 2497- turnip: clean up unused parameters for user consts. 2498- turnip: Change to use shared consts for PushConstants 2499- tu: increase maxPushConstantsSize to 256. 2500 2501Iago Toral Quiroga (79): 2502 2503- v3dv: ignore no-op barriers 2504- v3dv: drop obsolete comment 2505- v3dv: don't emit final noop job if we don't have anything to signal 2506- v3dv: document Vulkan requirements for signaling operations 2507- v3dv: ignore barriers for image layout transitions from undefined layouts 2508- v3dv: drop default sampler states if not used 2509- v3dv: flag BCL barriers for all-graphics and all-commands stages 2510- v3dv: skip binning sync if binning shaders don't access external resources 2511- v3dv: use wait stage for wait semaphores to decide about binning syncs 2512- v3dv: check input attachment usage as sampled usage 2513- v3dv: implement VK_EXT_separate_stencil_usage 2514- v3dv: trivially implement VK_KHR_separate_depth_stencil_layouts 2515- docs/features: flag VK_KHR_timeline_semaphore as implemented 2516- v3dv: expose separateDepthStencilLayouts 2517- v3dv: implement vkGetPipelineExecutableInternalRepresentationsKHR 2518- v3dv: implement vkGetPipelineExecutablePropertiesKHR 2519- v3d/compiler: add more stats to prog_data 2520- v3dv: implement vkGetPipelineExecutableStatisticsKHR 2521- v3dv: expose VK_KHR_pipeline_executable_properties 2522- v3dv: don't leak variant QPU when pipeline compile fails 2523- v3dv: drop unsused struct 2524- v3dv: expose VK_KHR_shader_float_controls 2525- v3dv: expose VK_KHR_spirv_1_4 2526- v3dv: expose VK_KHR_vulkan_memory_model 2527- v3dv: fix format usage checks when extended usage flag is set 2528- broadcom/compiler: don't promote early fragment tests when writing sample mask 2529- v3dv: consume barriers at the right stages 2530- v3dv: only clear BCL barrier state if we don't have pending graphics barriers 2531- v3dv: use an explicit struct type to track barrier state 2532- v3dv: track sources of barriers 2533- v3dv: limit sync for barriers to hw queues selected by source mask 2534- v3dv: don't be too aggressive disabling early Z 2535- v3dv: use the global RCL EZ disable if we don't have any EZ draws in the job 2536- broadcom/compiler: disable flags optimization for loop conditions 2537- v3dv: merge pending secondary barrier state into primary command buffers 2538- v3dv: handle barriers at the end of a command buffer 2539- v3dv: fix leak 2540- broadcom/compiler: fix postponed TMU spills with multiple writes 2541- broadcom/compiler: don't predicate postponed spills 2542- v3dv: fix pool descriptor count for inline uniform buffers 2543- v3dv: fix merge jobs 2544- nir: fix documentation for uadd_carry and usub_borry opcodes 2545- v3dv: don't lower uadd_carry and usub_borrow 2546- v3dv: remove obsolete comment 2547- nir/serialize: fix missing divergence info after deserialization 2548- broadcom/compiler: don't try to hide TMU latency at QPU scheduling 2549- broadcom/compiler: don't over-estimate latency of TMU instructions 2550- v3dv: fix comment in texel buffer shader copy path 2551- v3dv: remove unused lowering for nir_intrinsic_load_layer_id 2552- v3dv: fix comment for point_sprite_mask filed in shader key 2553- v3dv: clean up get_internal_type_bpp_for_image_aspects 2554- v3d,v3dv: stop hardcoding various image limits 2555- v3d: stop using a smaller texture limit in OpenGL 2556- v3d,v3dv: stop copying and pasting the translate_swizzle helper 2557- v3dv: only save/restore push constant data for meta operations if needed 2558- v3dv: move push constant data to the command buffer state 2559- v3dv: limit upload of indirect push constant data 2560- v3dv: don't allocate MAX_PUSH_CONSTANTS_SIZE bytes for the push constants UBO 2561- v3dv: stop tracking push constant buffer references 2562- v3dv: allocate more push constant buffers if needed 2563- v3dv: add a dirty state for pending push constants UBO updates 2564- v3dv: explain why we clear certain state after a draw call 2565- v3dv: remove duplicate condition 2566- nir: add nir_address_format_2x32bit_global 2567- broadcom/compiler: implement 2x32 global intrinsics 2568- broadcom/compiler: track if a shader uses global intrinsics 2569- v3dv: implement VK_KHR_buffer_device_address 2570- v3dv: handle VK_STRUCTURE_TYPE_DRM_FORMAT_MODIFIER_PROPERTIES_LIST_2_EXT 2571- v3dv: remove assert that poolSizeCount must be larger than 0 2572- v3dv: fix offset reported by vkGetImageSubresourceLayout 2573- v3dv: remove unnecesary FIXME comment 2574- v3d,v3dv: add a common v3d_hw_prim_type helper 2575- v3dv: only check binning sync for semaphores for the first CL job 2576- v3dv: explicitly set sampler min/max filter properties to false 2577- v3dv: fill in Vulkan 1.2 missing device properties 2578- v3dv: enable missing Vulkan 1.2 features for supported extensions 2579- v3dv: enable missing mandatory Vulkan 1.2 features 2580- broadcom/ci: update list of slow tests 2581- v3dv: expose Vulkan 1.2 2582 2583Ian Romanick (26): 2584 2585- glsl: Fix mixed tabs and spaces in lower_mat_op_to_vec.cpp 2586- glsl: Add flag to disable part of do_vec_index_to_cond_assign 2587- nir: Use nir_vector_extract to generate code for ir_binop_vector_extract 2588- glsl: Delete lower_extracts code 2589- anv: Remove FS executables when applying the null FS optimization 2590- intel/fs: Add missing synchronization for WaW dependency 2591- nir: i32csel opcodes should compare with integer zero 2592- nir/algebraic: Fix NaN-unsafe fcsel patterns 2593- nir: Add and use algebraic property "is selection" 2594- nir/range_analysis: Teach range analysis about fdot opcodes 2595- intel/compiler: Rename vec4 state URB opcodes to have VEC4\_ prefix 2596- intel/eu: Validate some aspects of URB messages 2597- intel/compiler: Move logical-send lowering to a separate file 2598- intel/fs: Add _LOGICAL versions of URB messages 2599- intel/fs: Lower URB messages to SEND 2600- intel/fs: Remove non-_LOGICAL URB messages 2601- intel/vec4: Set lower_usub_sat 2602- spirv: Fix array length of buffers larger than INT32_MAX. 2603- intel/fs: Make logical URB write instructions more like other logical instructions 2604- intel/fs: Make logical URB read instructions more like other logical instructions 2605- intel/fs: Don't pass flags to lower_urb_read_logical_send or lower_urb_write_logical_send 2606- intel/fs: Eliminate "masked" and "per slot offset" URB messages 2607- intel/fs: Use canonical form for "work around" tags 2608- nir: spirv: Allow 32-bit version of nir_intrinsic_is_sparse_texels_resident 2609- radeonsi: r600: d3d12: st: Use NIR lowering for tg4 offset arrays instead of GLSL lowering 2610- glsl: Remove lower_offset_arrays pass 2611 2612Icecream95 (45): 2613 2614- clc: Use stringstream for printing spirv errors 2615- panfrost: Stop overallocating compressed textures 2616- panfrost: Enable NIR lowering of half float packing 2617- panfrost: Fix pack_32_2x16 implementation 2618- pan/bi: Use texture index instead of sampler for message preloading 2619- pan/va: Improvements to LEA_TEX instructions 2620- pan/va: Add more transcendental operations 2621- pan/va: Add SEG_ADD/SEG_SUB operation 2622- pan/va: Add absneg modifier to V2F32_TO_V2F16 2623- pan/va: Improve texture instructions 2624- nir/lower_tex: Copy more fields in lower_tex_to_txd and friends 2625- pan/mdg: Keep min_bound at 16 when alignment requires it 2626- pan/mdg: Use MAX2 to set min_alignment 2627- pan/mdg: Fix mask usage when filling before a spill 2628- pan/mdg: Fix disassembly of store instructions 2629- pan/mdg: Return the instruction from mir_insert_instruction_*_scheduled 2630- pan/mdg: Fix multiple spilt writes in the same bundle 2631- panfrost: Copy blend constant into variant even when reusing it 2632- panfrost: Increase the limit for blend shader variants 2633- panfrost: Move patched_s out of the pan_blitter_views struct 2634- nir: Add store_combined_output_pan BASE back 2635- pan/bi: Read base for combined stores 2636- pan/mdg: Read base for combined stores 2637- panfrost: Only write depth / stencil once if MRT is used 2638- pan/bi: Reverse linear constraint bits 2639- pan/bi: Add nodearray datastructure 2640- pan/bi: Use nodearrays for linear constraints 2641- panfrost: Constant stencil value tracking 2642- panfrost: Stop using sparse_array for batch BOs 2643- panfrost: Use a macro for checking for a shared bind type 2644- panfrost: Remove sync arguments from panfrost_batch_submit 2645- panfrost: Enable QUAD_STRIP and POLYGON on v6 2646- pan/va: Use the _safe iterator when adding blend shader calls 2647- pan/decode: fflush buffers after dumping and before aborts 2648- pan/decode: Use tag bits for resource entry count 2649- pan/decode: Change indent when decoding resources 2650- panfrost: Only emit images when they are present 2651- panfrost: Emit the correct number of attributes 2652- panfrost: Clear the GENERAL varying buffer field if unused 2653- panfrost: Correctly calculate prefetch suppresion varying index 2654- panfrost: Add a debug option for checking overflows on pool uploads 2655- panfrost: Don't unbind recently bound streamout targets 2656- panfrost: Allow NULL streamout targets 2657- nine: Make vdecl_index_map always signed 2658- nine: Only enable thread_submit by default on x86 2659 2660Igor Torrente (2): 2661 2662- venus: Add support to VK_KHR_maintenance4 extension 2663- venus: Use maintenance4 to get max_size_buffer 2664 2665Ikshwaku Chauhan (1): 2666 2667- Revert "radeon: add EFC support to only VCN2.0 devices" 2668 2669Illia Abernikhin (1): 2670 2671- vulkan/runtime: repair thread safety in method vk_queue_start_submit_thread() 2672 2673Indrajit Das (5): 2674 2675- radeonsi/gfx11: VRS changes 2676- radeonsi: save the fs constant buffer to the util blitter context 2677- gallium/u_blitter: clear color buffers using color from a constant buffer 2678- mesa/st: add nir shader to clear color buffers using constant value 2679- mesa/st: clear color buffers using color from a constant buffer 2680 2681Italo Nicola (3): 2682 2683- mesa/st: call pipe->link_shader even if ir loaded from cache when linking 2684- tgsi: Add SEPARABLE_PROGRAM property 2685- virgl: overpropagate precise flags 2686 2687Iván Briano (6): 2688 2689- vulkan/wsi/wayland: Fix double free on error condition 2690- anv: re-alloc push constants after secondary command buffers 2691- anv: do not get rid of empty/useless fragment shaders 2692- anv: assert inheritance_info is not NULL 2693- anv: emit scissors when the pipeline changes 2694- anv: pipelineStageCreationFeedbackCount is allowed to be 0 2695 2696James Zhu (5): 2697 2698- amd: update headers to support decode software ring 2699- radeonsi/gfx11: update codec support for gfx11 2700- radeonsi/vcn: add decode software ring support for gfx11 2701- radeonsi/vcn: add jpeg decode support for gfx11 2702- amd/common: some ASICs with gfx9 use compute rings for render 2703 2704Jan Beich (1): 2705 2706- vulkan/wsi: treat EBADF as missing DMA_BUF_IOCTL_{EXPORT,IMPORT}_SYNC_FILE 2707 2708Jan Palus (1): 2709 2710- pvr: use UINT64_C for 64-bit constant 2711 2712Jason Ekstrand (312): 2713 2714- clc: Only initialize LLVM once 2715- clc: Rework logging a bit 2716- clc: Declare LLVMContexts on the stack 2717- intel/compiler: Set lower_fisnormal 2718- nir,microsoft: Move scale_fdiv into a common NIR pass 2719- nir/opcodes: fisfinite32 should return bool32 2720- util: Remove util_cpu_detect 2721- util/blob: Add align helpers 2722- util/set: Respect found in search_or_add_pre_hashed 2723- vulkan: Add a common VkPipelineCache implementation 2724- vulkan/pipeline_cache: Add helpers for storing NIR in the cache 2725- anv: Rename a fail label in CreateDevice 2726- anv: Switch to the new common pipeline cache 2727- anv: Clean up pipeline cache helpers a bit 2728- nir: Lower all bit sizes of usub_borrow 2729- vulkan: Set signals[i].stageMask = ALL_COMMANDS for QueueSubmit2 wrapping 2730- vulkan: Use ALL_COMMANDS_BIT for waits/signals instead of ~0 2731- spirv: Handle Op*MulExtended for non-32-bit types 2732- vulkan/log: Allow but warn if called with a NULL object 2733- vulkan/log: Allow but warn for client-invisible objects 2734- nir/constant_folding: Break TXB folding into a helper function 2735- nir: Constant fold sampler/texture offsets 2736- panvk: Take buffer offsets into account in BindVertexBuffers 2737- panvk: Eliminate unused vertex attributes 2738- nir/deref: Add an alu-of-cast optimization 2739- panvk: Add address/range helpers for panvk_buffer 2740- vulkan/wsi: Close file descriptors in wsi_destroy_image 2741- vulkan/wsi: Only use a single fd per wsi_image 2742- vulkan/wsi: Hang on to file descriptors 2743- vulkan/wsi: Reset the image fence right before vkQueueSubmit 2744- vulkan/wsi: Set the right stage flags for semaphore waits 2745- vulkan/wsi: Add signal_fence/semaphore_for_image helpers 2746- nir/cf: Return a cursor from nir_cf_extract as well 2747- nir: Fix constant folding for non-32-bit ifind_msb and clz 2748- nir: Set image_buffers and msaa_images in lower_samplers_as_deref 2749- nir: Stop assuming shader_info::textures_used is 32-bit 2750- shader_info: Make images_used a bitset 2751- nir: Gather samplers_used separately from textures 2752- util/bitset: Support larger ranges in BITSET_TEST/CLEAR_RANGE 2753- shader_info: Bump the number of images and textures supported 2754- gallium/u_threaded_context: Use PIPE_MAX_SHADER_SAMPLER_VIEWS for sampler_buffers 2755- mesa/st: Clamp MaxImageUniforms to MAX_IMAGE_UNIFORMS 2756- gallium: Bump PIPE_MAX_SHADER_IMAGES to 64 2757- panvk: Call nir_lower_io_arrays_to_elements_no_indirects 2758- nir/lower_blend: Be more explicit about deref assumptions 2759- nir/lower_blend: Stop passing the whole options object around 2760- nir/lower_blend: Support SNORM and integer formats for logic ops 2761- nir/builder: Add a nir_trim_vector helper 2762- nir/builder: Add a nir_resize_vector helper 2763- midgard: Handle FB fetch from non-vec4 output variables. 2764- nir/lower_blend: Expand or shrink output variables as needed 2765- panvk: Advertise support for logicOp 2766- iris: Set BindingTableEntryCount for compute shaders 2767- nir: Mark negative re-distribution on fadd as imprecise 2768- clover: Set images/samplers_used when lowering images 2769- lavapipe: Set shader_info::samplers_used 2770- gallium: Set shader_info::samplers_used in pstipple_fs 2771- ttn: Set shader_info::samplers_used 2772- mesa/st: Set samplers_used in lower_tex_src_plane 2773- gallium/draw: Properly handle nr_samplers != nr_sampler_views in keys 2774- llvmpipe: Fill out samplers even if nr_samplers = 0 2775- lavapipe: Set images_used in lvp_lower_pipeline_layout 2776- ntt: Don't gather samplers_declared twice 2777- panvk: Move CreateDescriptorSetLayout to per-arch 2778- panvk: Break descriptor lowering into its own file 2779- panvk: Add a buffer to each descriptor set 2780- panfrost,panvk: Make fixed_sysval_ubo < 0 mean compiler-assigned 2781- panfrost: Add some sanity checking for sysvals 2782- bifrost,midgard: Allow providing a fixed sysval layout 2783- panvk: Get rid of the per-pipeline sysvals BO 2784- panvk: Use a flat sysvals struct 2785- panvk: Stop calling lower_uniforms_to_ubo 2786- panvk: Put the sysval and push const UBOs at fixed indices 2787- panvk: Interleave UBOs with multiple descriptor sets 2788- panvk: Put SSBO addresses in the descriptor buffer 2789- vulkan: Start collecting enabled features in vk_device 2790- panvk: Enable robustBufferAccess 2791- panvk: Advertise VK_KHR_variable_pointers 2792- panvk: Add an elems field to panvk_buffer_view 2793- pavnk: Pass bind layouts to texture and image descriptor helpers 2794- panvk: Implement texture/image queries 2795- FIXUP: Use 16-bit things for texture sizes 2796- panvk: Re-arrange descriptor set functions 2797- panvk: Rewrite the write portion of vkUpdateDescriptorSets 2798- panvk: Set immutable samplers properly up-front 2799- panvk: Implement descriptor copies properly 2800- panvk: Drop panvk_descriptor 2801- panvk: Round FillBuffer sizes down to a multiple of 4 2802- nir: Preserve metadata if remove_dead_derefs makes no progress 2803- nir: Add a var set version of lower_indirect_derefs 2804- mesa/st: Use nir_shader_instructions_pass for st_nir_lower_builtin 2805- mesa/st: Check deref modes in lower_builtin_instr 2806- mesa/st: Use lower_indirect_var_derefs in st_nir_lower_builtin 2807- nir: Put the builder first in lower_input_attachments helpers 2808- nir: Use nir_shader_instructions_pass in nir_lower_input_attachments 2809- bifrost: Run nir_lower_global_vars_to_local before nir_lower_vars_to_scratch 2810- panvk: So more nir_lower_tex before descriptor lowering 2811- panvk: Lower blending after lower_var_copies 2812- v3dv: Add a create_image_view helper for internal views 2813- vulkan: Only be clever about vk_image_view::view_format for normal views 2814- v3dv: Don't use color aspects for depth/stencil images 2815- v3dv: Drop the region temporary from blit_shader 2816- vulkan,v3dv: Add a driver_internal flag to vk_image_view_init/create 2817- radv: Only use PLANE_0 in meta when actually needed 2818- radv: Use vk_image as the base for radv_image 2819- radv: Use vk_image_view as the base for radv_image_view 2820- glsl: Drop this != NULL assertions 2821- vulkan/pipeline_cache: Implement deserialize for raw objects 2822- v3dv: Loosen an assert in copy_buffer_to_image_shader 2823- vulkan: Fall back to raw data objects when deserializing if ops == NULL 2824- lavapipe: Use the correct ICD path on Win32 2825- vulkan,anv,turnip: Add a common CmdBindVertexBuffers wrapper 2826- radv: Add a sqtt entrypoint for CmdBindVertexBuffers2 2827- radv: Use the common CmdBindVertexBuffers wrapper 2828- lavapipe: Use the common BindVertexBuffers wrapper 2829- nir/algebraic: Add two more pack/unpack rules 2830- intel/fs: Drop fs_visitor::emit_alpha_to_coverage_workaround() 2831- intel/fs: Copy color_outputs_valid into wm_prog_data 2832- anv: Drop alpha_to_coverage from the NULL FS optimization 2833- anv: Handle the null FS optimization after compiling shaders 2834- anv: Don't disable the fragment shader if XFB is enabled 2835- intel: Only set VectorMaskEnable when needed 2836- vulkan: Add a base struct for buffers 2837- anv: Use the base vk_buffer struct 2838- panvk: Use the vk_buffer base struct 2839- vulkan/runtime: Add min_lod to vk_image_view 2840- intel/isl: Add isl_view::min_lod_clamp for IVB+ 2841- anv: Implement VK_EXT_image_view_min_lod 2842- isl: Add some asserts about multisampled surfaces 2843- nir: Handle register sources in lower_phis_to_regs_block 2844- nir: Add a correctness note for nir_lower_phis_to_regs_block 2845- nir: Rename nir_gather_xfb_info to nir_shader_get_xfb_info 2846- nir: Add a nir_xfb_info to nir_shader 2847- anv: Use nir_shader_gather_xfb_info 2848- lavapipe: Use nir_shader_gather_xfb_info 2849- turnip: Use nir_gather_xfb_info 2850- radv: Use nir_gather_xfb_info 2851- nir: Drop nir_shader_get_xfb_info 2852- glsl/nir: Populate nir_shader::xfb_info after linking varyings 2853- glsl/nir: Stop leaking varyings_info 2854- glsl/nir: Stash the xfb_info in the nir_shader when linking XFB 2855- st,nir: Use nir_shader::xfb_info in nir_lower_io_passes 2856- nir/glsl: Use rzalloc for nir_xfb_info 2857- nir: xfb_buffer_info::stride is in bytes 2858- ttn: Populate the images/textures/samplers_used fields in shader_info 2859- vulkan,anv: Move the image offset/extent sanitize helpers to common code 2860- radv: Use the common image offset/extent sanitize helpers 2861- vulkan,radv: Move vk_format_get_plane_format to common code 2862- vulkan/format: Add a vk_format_get_aspect_format helper 2863- vulkan: Add a helper for image<->buffer copies 2864- anv: Use the common image<->buffer copy helper 2865- vulkan,radv: Steal some image offset/extent helpers from radv 2866- radv: Use vk_image_buffer_copy_layout 2867- panvk: Use vk_image_buffer_copy_layout 2868- lavapipe: Use vk_image_buffer_copy_layout 2869- nir/deref: Break out a helper for compare_deref_paths 2870- nir/deref: Use an index in compare_deref_paths 2871- nir/deref: Make compare_deref_paths take a stop callback 2872- nir/deref: Re-arrange variable checks in compare_deref_paths 2873- nir: Increase nir_variable_data::mode to 16 bits 2874- nir/vars_tests: Use nir_var_mem_global instead of ssbo 2875- nir/deref: Handle SSBO array bindings specially 2876- nir/deref: Handle RESTRICT for SSBO deref bindings 2877- compiler/types: Don't place members in the previous substruct's end padding 2878- anv/wsi: Stop resetting semaphores 2879- vulkan/wsi: Fix a signal_semaphore_with_memory check 2880- wsi: Always signal semaphores and fences in wsi_common_acquire_next_image 2881- panvk: Use the common AcquireNextImage implementation 2882- v3dv: Use the common AcquireNextImage implementation 2883- radv: Drop create_sync_for_memory 2884- lavapipe: Use the common AcquireNextImage implementation 2885- pvr: Use the common AcquireNextImage implementation 2886- turnip: Use the common AcquireNextImage implementation 2887- turnip: Use the common QueuePresent implementation 2888- turnip: Set supported_sync_types before wsi_init 2889- vulkan/wsi: Use a STACK_ARRAY for stage_flags 2890- vulkan/wsi: Reset fences earlier in wsi_common_queue_present 2891- vulkan/wsi: Re-arrange QueueSubmit code in wsi_common_queue_present 2892- vulkan/wsi: Set memory ownership after signaling fences and semaphores 2893- vulkan/wsi: Signal semaphores and fences from the dma-buf 2894- intel/compiler: Use NIR_PASS(_, ...) 2895- anv: Use NIR_PASS(_, ...) 2896- anv: Properly clamp attachment layer counts 2897- vulkan/wsi: Use HAVE_LIBDRM to detect DRM instead of !_WIN32 2898- vulkan: Add some border color helpers 2899- panvk: Use the new border color helpers 2900- lavapipe: Use the new border color helper 2901- turnip: Use the new border color helpers 2902- vulkan/nir: Make spirv_data const in vk_spirv_to_nir 2903- vulkan: Add a vk_pipeline_shader_stage_to_nir helper 2904- gallium: Add a u_default_clear_buffer helper 2905- iris: Use u_default_clear_buffer 2906- panfrost: Use u_default_clear_buffer 2907- nir/opt_memcpy: lower copies to/from tightly packed types 2908- nir: Add an options parameter to deref_instr_has_complex_use 2909- nir/opt_memcpy: Add another case for function_temp 2910- vulkan/render_pass: Add a better helper for render pass inheritance 2911- anv: Use CmdBeginRendering for resumes in BeginCommandBuffer when possible 2912- vulkan/render_pass: Use a special layout for self-dependencies 2913- vulkan/render_pass: Pass sample locations to barriers 2914- vulkan/render_pass: Allow for mixed sample counts 2915- vulkan/render_pass: Support VkAttachmentSampleCountInfoAMD 2916- vulkan: Fix a comment 2917- vulkan: Add a base struct for descriptor set layouts 2918- vulkan: Add a common implementation of pipeline layouts 2919- lavapipe: Use the vk_descriptor_set_layout base struct 2920- lavapipe: Use the vk_pipeline_layout base struct 2921- panvk: Use the vk_descriptor_set_layout base struct 2922- panvk: Use the vk_pipeline_layout base struct 2923- vulkan: Depend on vk_pipeline_layout in vk_cmd_enqueue 2924- anv: Replace an assert() with unreachable() 2925- util: Re-indent util_sign_extend, comment, and add asserts 2926- util: Use shifts in util_sign_extend 2927- util: Use util_sign_extend in fast_idiv_by_const 2928- mesa/bptc: Use util_sign_extend 2929- util,nir: Move mask_sign_extend from opt_load_store_vectorize to util 2930- isl: Use util_sign_extend 2931- intel/decoder: Use util_mask_sign_extend() 2932- panfrost,asahi: Use util_sign_extend for unpacking 2933- nir: Use util_mask_sign_extend when serializing constants 2934- vulkan/wsi: Pass the size to MapMemory in the SW path 2935- vulkan/wsi/wayland: Use a single memcpy in the SW path 2936- vulkan/wsi: Move select_memory_type to common and rework it a bit 2937- vulkan/wsi: Add a helper for creating CPU images 2938- vulkan/wsi/x11: Use wsi_configure_cpu_image 2939- vulkan/wsi/wayland: Use wsi_configure_cpu_image 2940- vulkan/wsi: Delete SW support from configure_native_image 2941- vulkan/wsi: Persistently map CPU images 2942- vulkan/wsi: Compute stride and size in configure_buffer_image 2943- vulkan/wsi: Align buffer image strides to optimalBufferCopyRowPitchAlignment 2944- vulkan/wsi: Support tiled CPU images 2945- vulkan/wsi/x11: Don't leak shm_reply if we don't have dri3 or present 2946- vulkan/wsi/x11: Only use MIT_SHM if the device supports EXT_external_memory_host 2947- vulkan/wsi/wayland: Use host pointer import when available 2948- dozen: Increase optimalBufferCopy*Alignment 2949- dozen: Stop asking for semaphore/fence signaling 2950- vulkan/wsi/wayland: Only memcpy if the swapchain is actually software 2951- vulkan/wsi: Fix structure chaining in wsi_create_buffer_image_mem 2952- shader_info: Move subgroup_size out of cs and make it an enum 2953- vulkan: Re-order pipeline hashing 2954- vulkan: Constify vk_spirv_version 2955- spirv,vulkan: Set shader_info::subgroup_size 2956- mesa,glsl,ttn: Set subgroup_size to UNIFORM 2957- anv: Drop unnecessary parameters to anv_pipeline_compile_cs 2958- anv: Put a VkPipelineShaderStageCreateInfo* in anv_pipeline_stage 2959- anv: Use vk_pipeline_shader_stage_to_nir 2960- vulkan: Add a vk_pipeline_shader_stage_is_null() helper 2961- anv: Use vk_pipeline_shader_stage_is_null() 2962- intel,anv,iris,crocus: Drop subgroup size from the shader key 2963- anv: Implement VK_EXT_shader_module_identifier 2964- anv: Init supported extensions before WSI 2965- vulkan/wsi/wayland: Add a BUFFER_TYPE flag 2966- vulkan/wsi: Add debug variables to force the SW and PRIME buffer blit paths 2967- intel/blorp: Set uses_sample_shading for MSAA blit shaders 2968- radv: Set uses_sample_shading for copy shaders 2969- intel/fs: Use shader_info::fs::uses_sample_shading 2970- nir: Add a pass for lowering shaders to single-sampled 2971- intel/fs: Use nir_lower_single_sampled 2972- glsl/nir: Set uses_sample_shading from glsl_to_nir 2973- spirv/nir: Set uses_sample_shading from spirv_to_nir 2974- nir/gather_info: Stop gathering uses_sample_shading 2975- glsl/nir: Set sample_shading if a FS output ever shows up as an rvalue 2976- intel/fs: Simplify persample_dispatch 2977- intel/fs,anv: Rework handling of coarse and sample shading 2978- panfrost: Simplify sample_shading 2979- anv: Stop compacting surface state tables 2980- vulkan: Detect pNext chain loops in vk_foreach_struct() 2981- anv: Don't use the wrong ARRAY_SIZE 2982- Revert "vulkan: Detect pNext chain loops in vk_foreach_struct()" 2983- vulkan: Add a vk_limits.h file for runtime limits 2984- vulkan: Add an enum for all dynamic graphics states 2985- vulkan: Add standard sample locations 2986- vulkan: Add data structures to store all graphics state 2987- util: Constify __bitset_test_range 2988- vulkan: Add a common data structure for dynamic states 2989- vulkan: Add a common implementation of vkCmdSet* 2990- vulkan: Copy the depth/stencil state optimization code from ANV 2991- Use vk_foreach_struct_const where needed 2992- vulkan: Detect pNext chain loops in vk_foreach_struct() (v2) 2993- anv: Add an instance multiplier to anv_pipeline 2994- anv: Handle vertex buffer sizes in anv_CmdBindVertexBuffers2 2995- anv: Stop recording sample locations per-sample-count 2996- anv: Convert to using vk_graphics_pipeline_state 2997- anv: Rework setting primitive topology 2998- anv: Stop merging DEPTH_STENCIL state 2999- anv: Switch to using common dynamic state tracking 3000- vulkan/graphics_state: Improve the depth/stencil optimization code 3001- anv: Do depth/stencil optimization for dynamic depth/stencil 3002- vulkan: Call lower_clip_cull_distance_arrays in vk_spirv_to_nir 3003- vulkan/nir: Don't remove dead XFB outputs 3004- vulkan: Call gather_xfb_info in vk_spirv_to_nir 3005- lavapipe: Use more Vulkan NIR heleprs 3006- vulkan: Allow up to 12 pointers in multialloc 3007- vulkan: Record shader stages in vk_graphics_pipeline_state 3008- vulkan: Multisample state isn't always included in fragment shader state 3009- vulkan: Depth/stencil isn't fragment output state but multisample is 3010- vulkan: Add a fully_dynamic_state_groups() helper 3011- vulkan: Fix pipeline libraries with dynamic-only VI or FSR state 3012- vulkan: Input assembly and depth/stencil can also be fully dynamic 3013- vulkan: Don't assert VkGraphicsPipelineLibraryCreateInfoEXT::Flags == 0 3014- zink: Mark depth/stencil visual tests as flakes 3015- lavapipe: Use shader_info::uses_sample_shading 3016- lavapipe: Use the vk_graphics_pipeline_state 3017- vulkan: Append subpass structures to VkRenderingInfo last 3018- vulkan: Include self-dep info in rendering continues 3019- intel/eu: Don't throw validation errors on float MOV_INDIRECT 3020- lima: Lower undefs to zero and run DCE after from_ssa 3021- nir/from_ssa: Ignore undef sources 3022- vulkan: Always populate vk_render_pass_state::render_pass 3023- vulkan: Improve the docs for vk_subpass_info 3024- radv: Use both aspects for depth/stencil blit destinations 3025- vulkan: Dirty VP_VIEWPORTS/SCISSORS when copying viewports/scissors 3026 3027Jason Volk (4): 3028 3029- r600: Fix userspace pointer support for evergreen compute. 3030- r600: Improve compute memory pool performance; reduce fragmentation. 3031- r600: Elide downloads for discarded and immutable compute memories. 3032- radeon: Support shared memory user pointers. 3033 3034Jeffrey Knockel (1): 3035 3036- egl: dereference XCB drawable pointers 3037 3038Jesse Natalie (158): 3039 3040- d3d12: Don't block DISPLAY based on format for non-Windows 3041- wgl: Implement WGL_ARB_create_context_robustness 3042- d3d12: Hook up robustness extensions 3043- d3d12: Plug context leaks 3044- d3d12: Plug screen leaks 3045- d3d12: Attempt screen reset during context create 3046- d3d12: Add a unit test for context reset recovery 3047- microsoft/compiler: Unload DXIL validator library \*after* calling Release() 3048- d3d12: Correctly key off of polygon stipple enable cap 3049- mesa: Ensure ARB programs end in a newline 3050- bin: Add a script for filtering/sorting refcount logs 3051- u_debug_refcnt: Don't loop for initial refcounts if the initial value is huge 3052- u_debug_symbol: Use correct sizeof for DbgHelp 3053- u_debug_symbol: Allow re-init of DbgHelp to work around already-initialized issues 3054- u_debug_stack: Lock around stack dumps to prevent interleaving 3055- u_primconvert: Refactor to remove recursion from util_primconvert_draw_vbo 3056- u_primconvert: Handle take_index_buffer_ownership 3057- d3d12: Use a pipe_reference in d3d12_bo 3058- d3d12: Fix leaks in map with do-not-wait 3059- d3d12: Fix a couple over-releases from incorrect take-ownership flag 3060- d3d12: Destroy empty residency bo set during early-return 3061- d3d12: Include windows.h on Windows before dxgicommon.h 3062- d3d12: Don't use VLAs 3063- d3d12: MSVC warning around operator precedence causing uint32_t==bool 3064- dzn: Add new parameter to vk_image_view_init call 3065- d3d12: Fix forward decl for sw_winsys 3066- microsoft/compiler: Fixup sampler derefs in tex instrs that don't \*need* samplers 3067- nir_lower_task_shader: Fix return from lower_task_intrin (bool, not void*) 3068- d3d12: Re-add missed bitmasks to shader key comparison 3069- CI/d3d12: Add skips and update baselines 3070- CI/windows: Install Vulkan SDK from LunarG directly isntead of Chocolatey 3071- CI/windows: Delete comment for installing vulkan-runtime 3072- CI/windows: Disable chocolatey progress printing to shrink log size 3073- CI/windows: Disable LLVM CMake developer warnings to shrink log size 3074- CI/windows: Move --fraction for deqp-runner to run command line 3075- dzn: Update fails/flakes 3076- CI: Re-enable Windows builds 3077- nir: Consider PNTC to be a varying 3078- wgl, d3d12: Handle front buffer writes for double-buffered MSAA surfaces backed by swapchains 3079- meson: Use a C99 STDC_VERSION for flex 3080- d3d12: Report PIPE_CAP_ACCELERATED as 0 on WARP 3081- d3d12: Properly set HS input control point count 3082- dzn: Init sync types before wsi 3083- mesa: Use C11 alignas for cross-platform SSE support 3084- CI: Lima farm is offline 3085- dzn: Skip another test that times out occasionally 3086- util/disk_cache: Implement disk_cache_get_function_identifier for Windows 3087- microsoft/clc: Enable tests that pass on server 2022 3088- microsoft/clc: Fix test double free in the case of compilation failure 3089- microsoft/clc: Remove dead image vars 3090- microsoft/clc: Add a unit test for unused image kernel args 3091- vulkan_entrypoints_gen: For Windows ARM64EC, prefix symbols with '#' 3092- dzn: Support Vulkan loader interface v5 3093- gallium/util: Move u_dl and u_pointer to src/util 3094- dzn: Annotate sometimes-unused variables. 3095- dzn: Define COBJMACROS before any includes 3096- dzn: Use os_time_sleep instead of Win32 Sleep 3097- dzn: Use u_dl to load D3D12 3098- dzn: Add an eventfd sync implementation 3099- dzn: Fix maybe-uninitialized warning 3100- dzn: Skip dxil validator and some Windows-only debug flags for Linux 3101- dzn: Use IUnknown instead of IDXGIAdapter1 as the stored adapter 3102- dzn: Don't hash adapter LUID as part of device UUID 3103- dzn: Use a custom adapter desc instead of DXGI adapter desc 3104- dzn: Move DXGI code to a separate file and only build it on Windows 3105- dzn: Add a DXCore enumeration path 3106- dzn: Add ABI helpers for D3D12 functions returning structs 3107- vulkan/wsi/win32: Use the new helpers and persistent map 3108- dzn: Fix winsys reporting 3109- dzn: Only support high/normal queue priorities 3110- dzn: Add for condition to break nested loop 3111- dzn: Re-order includes in dzn_nir to prevent TRUE redefinition 3112- dzn: Only add Werror flags that the compiler supports 3113- dzn: Disable Wunused-value for C++ files 3114- dzn: Fix comparison order for format enums 3115- dzn: Re-order dzn_physical_device_desc so = { 0 } doesn't initialize a struct 3116- dzn: Use fabs on floating point input 3117- CI: Add dzn to debian vulkan/clang/release builds 3118- mesa: Support D3D11/D3D12 memory imports 3119- mapi: Add more EXT_external_objects_win32 functions/enums 3120- gallium: Add a 'name' field to winsys_handle 3121- mesa: Implement ImportMemoryWin32NameEXT 3122- gallium: Add 'name' field to Win32 semaphore import 3123- driver_noop: Remove infinite recursion from create_fence_win32 3124- gallium: Add a new fence type with a pipe cap to indicate it can be imported 3125- mesa: Support importing D3D12 fences as timeline semaphores 3126- mesa: Implement ImportSemaphoreWin32NameEXT 3127- gallium, mesa: Support setting timeline semaphore values 3128- d3d12: Store the rest of the device IDs in the screen 3129- d3d12: Compute UUIDs required by external objects extension 3130- d3d12: Add pipe getters for Win32 and base external objects device matching 3131- d3d12: Support B4G4R4A4 format 3132- d3d12: Get adapter LUID after device creation 3133- d3d12: PIPE_BIND_SHARED doesn't mean linear and is always on opened resources 3134- d3d12: Support creating memory objects 3135- d3d12: Implement resource_from_memobj 3136- d3d12: Support opening resources and memobj by name 3137- d3d12: Support importing fences / timeline semaphores 3138- d3d12: Implement server signal/wait 3139- d3d12: Implement fence opening and value setting 3140- d3d12: Fix up resource import validation 3141- docs: Update features.txt and new_features.txt 3142- u_atomic: Implement p_atomic_xchg for Windows 3143- u_atomic: Fix MSVC p_atomic_add_return 3144- u_atomic: Add p_atomic_fetch_add which returns the old value 3145- zink: Use p_atomic_fetch_add 3146- simple_mtx: Replace GCC sync intrinsics with u_atomic ops 3147- Fix static glapi on Windows 3148- dzn: Missed ABI fixes for GetCustomHeapProperties 3149- dzn: Remove the cast when the SDK version is high enough 3150- d3d12: Add helpers to build with correct ABI for MinGW 3151- microsoft/clc: Add helpers to build with correct ABI for MinGW 3152- subprojects: Point DirectX-Headers at a specific tag 3153- CI: Update DirectX-Headers dependency for MinGW/Debian 3154- util: Add a Win32 futex impl 3155- d3d12: Support clip halfz 3156- d3d12: Add a blank d3d12_resource_state.h/cpp 3157- d3d12: Move desired resource state to new files 3158- d3d12: Move current resource state to new files 3159- d3d12: Give bos a unique identifier to be used for state tracking 3160- d3d12: Add a list of contexts alive for the current screen 3161- d3d12: Hold lock when removing resources from residency list 3162- d3d12: Notify contexts about deletion of bos 3163- d3d12: Add a context state tracking structure 3164- d3d12: Treat depth/stencil as planar for plane count helper 3165- d3d12: Create/free context state entries 3166- d3d12: Track a global resource state for non-simultaneous-access resources 3167- d3d12: Record a state fixup command list when necessary 3168- d3d12: Swap the remainder of state tracking to new method 3169- d3d12: Rename bind invalidate options to transition flags 3170- d3d12: Extract core barrier logic 3171- d3d12: Optimize transition_subresource_states that covers a whole resource 3172- d3d12: Add a transition flag indicating that state accumulation is needed 3173- d3d12: Remove state tracking implementation details from header 3174- nir_lower_io_to_scalar: Support arrayed (per-vertex) I/O 3175- microsoft/compiler: Add a max shader model option 3176- microsoft/compiler: Add a max validator version 3177- microsoft/compiler: Emit metadata based on long-lived signature data 3178- microsoft/compiler: Split signature processing into two parts 3179- microsoft/compiler: Ensure 4-component position writes via NIR 3180- microsoft/compiler: Lower I/O to scalar 3181- microsoft/compiler: Support DXIL validator version 1.5 3182- microsoft/compiler: Compute correct usage masks 3183- microsoft/compiler: Correctly compute dynamic indexing I/O masks 3184- microsoft/compiler: DXIL validator 1.6 counts UAVs differently 3185- microsoft/compiler: Set the barycentrics flag for attribute_at_vertex 3186- microsoft/compiler: DXIL validator 1.6 uses a new PSV struct version 3187- microsoft/compiler: DXIL validator 1.6 uses a new PSV resource struct 3188- microsoft/compiler: Support DXIL validator 1.6 3189- microsoft/compiler: Support DXIL validator 1.7 3190- microsoft/compiler: Add a getter for a validator version from DXIL.dll 3191- microsoft/compiler: Blacklist DXIL validator 1.6 from 20348 SDK 3192- d3d12: Retrieve validator version 3193- dzn: Retrieve validator version 3194- egl/wgl: Delete unused variables/code 3195- egl/wgl: Fix some awkward sizeof formatting 3196- microsoft/compiler: Discard shouldn't be marked readnone 3197- microsoft/compiler: Fix PSV struct when numthreads is 0 3198 3199Jiang Feng (1): 3200 3201- virgl: Set res->maybe_busy to true when creating resources 3202 3203Jonathan Gray (2): 3204 3205- intel/dev: sync ADL-S pci ids with linux 3206- intel/dev: add RPL-S pci ids from drm-intel-next 3207 3208Jonathan Marek (2): 3209 3210- freedreno/registers: add a7xx registers for drm/msm kernel driver 3211- freedreno/registers: add missing varset="chip" for new enum values 3212 3213Jonathan Weinstein (1): 3214 3215- lavapipe: support instance rate zero for VK_EXT_vertex_attribute_divisor 3216 3217Jordan Justen (48): 3218 3219- intel/dev: Add device info for RPL-P 3220- vulkan/wsi/x11: Use atomic_int with sent_image_count 3221- drm-uapi/i915_drm.h: Update from drm-next (2022-04-28) 3222- intel/dev: Read hwconfig from i915 3223- intel_dev_info: Add --hwconfig command line parameter 3224- iris: Fix assertion meant to only target the clear-color stride 3225- nir/divergence: handle more \*_intel intrinsics 3226- intel/dev: Add INTEL_PLATFORM_DG2_G12 3227- intel/dev: Add DG2 G12 PCI IDs 3228- intel/dev: Enable first set of DG2 PCI IDs 3229- intel: Fix build of mi_builder_tests by including c99_compat.h 3230- intel: Build mi_builder_test whenever build-tests is set 3231- drm-uapi/i915_drm.h: Update from drm-next (2022-05-06) 3232- intel/gem: Add support for I915_ENGINE_CLASS_COMPUTE 3233- anv: Add support for I915_ENGINE_CLASS_COMPUTE in init_device_state() 3234- anv: Move STATE_BASE_ADDRESS programming into init_common_queue_state() 3235- anv, iris: Add support for I915_ENGINE_CLASS_COMPUTE 3236- anv, iris: Enable compute engine with INTEL_COMPUTE_CLASS=1 3237- intel/dev: Add devinfo::mem to store i915 regions information 3238- anv: Use memory info from devinfo 3239- iris/bufmgr: Use memory info from devinfo 3240- intel/tools: Print memory info in intel_dev_info 3241- vulkan/wsi: Disable dma-buf sync file if ENOSYS is returned 3242- intel/compiler: Don't create vec4 reg-set for gen8+ 3243- intel/dev: Add vram.unmappable.size region info 3244- intel/tools: Print unmappable region info in intel_dev_info 3245- intel/dev: Add intel_vram_all_mappable() 3246- iris/resource: Assert that DG2 CCS buffers don't also try to set BO_ALLOC_SMEM 3247- iris/resource: Avoid mapping when not needed in iris_resource_init_aux_buf() 3248- iris/bufmgr: Add all_vram_mappable which is currently always true 3249- iris/bufmgr: Set mmap_mode to IRIS_MMAP_NONE for lmem in small-BAR mode 3250- iris: Set clear_color_unknown if the bo is not mappable 3251- anv/allocator: Add assert and TODO comment for future small BAR uapi 3252- iris/bufmgr: Add assert and TODO comment for future small BAR uapi 3253- drm-uapi/i915_drm.h: Update from drm-next (2022-07-22) 3254- iris/bufmgr: Add I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS for vram mappable buffers 3255- intel/dev: Use i915 region probed_cpu_visible_size when non-zero 3256- intel/dev: Determine the amount of free vram using small BAR uapi 3257- intel/dev: Enable remaining DG2 and ATS-M device IDs 3258- anv: Fix PHYSICAL_DEVICE_MEMORY_BUDGET_PROPERTIES with large BAR 3259- intel/dev: Fill in system memory info when using INTEL_DEVID_OVERRIDE 3260- iris: Drop extra file-descriptor dup in iris_drm_screen_create() 3261- intel/pci_ids: Update ADL-S strings 3262- intel/pci_ids: Add 0x468b ADL-S PCI-id 3263- intel/pci_ids: Update ATS-M device names 3264- intel/pci_ids: Update DG2 device names 3265- intel/pci_ids: Add dg2 0x5698 pci-id 3266- intel/pci_ids: Drop non-upstream dg2 pci-ids 3267 3268Jose Maria Casanova Crespo (1): 3269 3270- v3dv/ci: increase fraction to 10 on v3dv ci jobs. 3271 3272Josh Billingsley (1): 3273 3274- driconf: add SD Gundam G Generation Cross Rays 3275 3276Joshua Ashton (1): 3277 3278- aco: Use movk for AddressHi bits in vertex prolog 3279 3280José Fonseca (6): 3281 3282- trace: Fix framebuffer state serialization. 3283- lavapipe: Prevent mapping buffers beyond their size. 3284- trace: Dump NIR. 3285- trace: Parse character data more efficiently. 3286- trace: Bring state dump up to speed. 3287- trace: Allow to control nir dumping via an environment variable. 3288 3289Juan A. Suarez Romero (37): 3290 3291- ci/v3dv: remove fixed test 3292- v3d: report the correct unsupported blit format 3293- v3d: do not tile 1D textures 3294- v3d: use surface format defined on pipe_blit 3295- gallium: add hook on getting canonical format 3296- v3d: define our own canonical supported formats 3297- v3d: enable GL_ARB_copy_image extension 3298- gallium: move get_canonical_format hook to pipe_screen 3299- v3dv: duplicate key on hashtable insert 3300- v3dv: store device_id on device init 3301- v3d/simulator: add support for AMD cards 3302- mesa: unref syncobj after wait_sync 3303- v3d: enable early Z/S clears 3304- v3d: disable early-Z on odd frame dimensions 3305- v3d: allow TFU blitting for single layer textures 3306- v3d: add hook to get on-disk shader cache 3307- v3d: add builtin support for white/black clamp-to-border 3308- v3dv: check BO mapping result 3309- docs: update extensions for V3D and VC4 3310- ci: bump VK-GL-CTS to 1.3.2.0 3311- v3d: fix blending for mixed RT formats 3312- v3d: expose GL_ARB_draw_buffers_blend 3313- v3d/ci: Add traces 3314- v3d: use function to initialize refcount 3315- v3d: save only required states in blitter 3316- v3d: release all color buffers on context destroy 3317- v3d: add ARB_polygon_offset_clamp extension support 3318- v3dv/ci: Update expected results 3319- v3d/ci: Add flake test 3320- v3d/v3dv/ci: update expected results 3321- v3d/ci: (partially) revert expected list 3322- v3d/ci: update expected list 3323- v3dv/ci: Update expected list 3324- v3d/vc4/ci: Fix typo in skipped list 3325- v3d/ci: enable ARB_texture_cube_map piglit tests 3326- vc4: properly restore vc4 debug option 3327- vc4: store tex sampler in proper register 3328 3329Kai Wasserbäch (1): 3330 3331- fix(gallivm): Replace LLVMConstF* with LLVMBuild* methods. 3332 3333Karmjit Mahil (26): 3334 3335- pvr: Fix clang-format errors caused by vk outarrays. 3336- pvr: Add PVR_WORKGROUP_DIMENSIONS define. 3337- pvr: Add initial implementation of vkCmdDispatch(). 3338- pvr: Add compute update shared kernel in vkCmdDispatch(). 3339- pvr: Add compute update kernel in vkCmdDispatch(). 3340- pvr: Update pvrsrvkm to fw 1.17 . 3341- pvr: Add csb enum helper funciton header. 3342- pvr: Write immutable descriptor words in vkAllocateDescriptorSets(). 3343- pvr: Fix incorrect samples to cr_isp_aa_mode conversion. 3344- pvr: Change compute hard coding infrastructure. 3345- pvr: Remove unused output_selects from vertex shader state. 3346- pvr: Add graphics pipeline hard coding infrastructure. 3347- pvr: Enable 33.15.11.3 (AXE-1-16M). 3348- pvr: Extend hard coding infra to allow per stage per device coding. 3349- pvr: Fix physical device limits. 3350- pvr: Handle vdm degen_cull_enable. 3351- pvr: Fix off by 1 error in buffer_id for ubo pds program. 3352- pvr: Rename loop iterator variable. 3353- pvr: Update pvrsrv build version for fixed size fw. 3354- pvr: Rename uniform pds program to descriptor pds program. 3355- pvr: Add static consts support in descriptor pds program. 3356- pvr: Add static consts to hard coding infra. 3357- pvr: Add initial vkCmdPipelineBarrier skeleton. 3358- pvr: Add IDF/WDF program for compute pipeline barrier. 3359- pvr: Add compute pipeline barrier handling. 3360- pvr: Fix cdm shared reg usage reported to fw. 3361 3362Karol Herbst (20): 3363 3364- nir: add 16 and 64 bit fisnormal lowering 3365- vtn: clamp SpvOpImageQuerySize dest to 32 bit 3366- nv50/nir: align tlsspace to 0x10 3367- llvmpipe: PIPE_COMPUTE_CAP_GRID_DIMENSION is uint64_t 3368- llvmpipe/fence: make the fence id counter atomic 3369- iris/cs: take buffer offsets into account for CL 3370- nir: add a nir_remove_non_entrypoints helper 3371- ntt: make use of new samplers_used field 3372- clc: parse localSize and localSizeHint 3373- clc: drop parsingComplete check 3374- clc: speed up compilation by not relying on opencl-c.h 3375- clc: fix compiler features_macro CTS Test 3376- clc: disable opaque pointers until they are supported 3377- nv50/ir: recoginse Ampere 3378- nir/gather_info: allow to run it before inlining 3379- nir/gce: pin call instructions 3380- nouveau: use the contexts pushbuf and client where possible 3381- nouveau/mm: make code thread safe 3382- nv50/ir: fix OP_UNION resolving when used for vector values 3383- nv50: properly flush the TSC cache on 3D 3384 3385Ken Johnson (2): 3386 3387- d3d12: Reduce cost of shader key comparisons 3388- mesa/main: Correct alloca() use without #include "c99_alloca.h" 3389 3390Kenneth Graunke (41): 3391 3392- anv: Fix INTEL_DEBUG=bat on XeHP 3393- intel/decoder: Fix binding table pointer decoding with large offsets 3394- iris: Add FLUSH_HDC to PIPE_CONTROL_CACHE_FLUSH_BITS 3395- Revert "st/mesa: Transcode ASTC to BC7 (BPTC) where possible" 3396- intel: Drop Wa_1409226450 (stall before instruction cache invalidation) 3397- intel: Fix analysis invalidation in eliminate_find_live_channel 3398- intel/compiler: Move spill/fill tracking to the register allocator 3399- mesa: Split tx_compress_dxtn into per-format functions 3400- mesa: Avoid temp images in _mesa_texstore_rgb_dxt1 for GL_RGBA source 3401- mesa: Require temp images in _mesa_texstore_*_dxt* with SkipImages 3402- mesa: Avoid temp images in _mesa_texstore_*_dxt* for stride = 0 3403- isl: Add an isl_format_supports_typed_atomics() helper. 3404- intel/compiler: Fix uncompaction of signed word immediates on Tigerlake 3405- intel: Set a more useful fake devinfo->gtt_size in no-hw mode 3406- intel/compiler: Fix A64 header construction with a uniform address 3407- iris: Update comment about 2GB dynamic state range 3408- iris: Use PIPE_* defines rather than ones from main/config.h 3409- crocus: Use PIPE_* defines rather than ones from main/config.h 3410- intel/compiler: Stop including src/mesa/main/config.h 3411- intel/compiler: Fix brw_gfx_ver_enum.h to be a proper header file 3412- intel/compiler: Split 3DPRIM_* defines out to a separate header. 3413- intel/tools: Stop malloc'ing device info in i965_disasm 3414- intel/compiler: Move opcode_desc handling to a separate header 3415- intel/compiler: Introduce a new brw_isa_info structure 3416- intel/compiler: Remove use of thread_local for opcode tables 3417- intel/compiler: Convert brw_eu.cpp back to brw_eu.c 3418- intel/compiler: Handle split-sends in EOT high-register pinning case 3419- intel/fs: Opportunistically split SEND message payloads 3420- intel/compiler: Avoid copy propagating large registers into EOT messages 3421- intel/compiler: Don't set SBID on EOT send messages 3422- mesa: Mark render-to-texture as unsafe if there's no pipe_resource 3423- intel/eu: Handle compaction when inserting validation errors 3424- intel/eu: Fix XeHP register region validation for hstride == 0 3425- intel/eu: Clarify spec citations for XeHP region restrictions 3426- intel/eu: Mark header present in URB memory fences on XeHP 3427- intel/compiler: Lower FIND_[LAST\_]LIVE_CHANNEL in IR on Gfx8+ 3428- intel/compiler: Delete unused Gfx8+ code in brw_find_live_channel() 3429- iris: Fix PIPE_CAP_UMA 3430- iris: Use linear for exported resources if we can't convey tiling 3431- iris: Delete unused iris_screen::aperture_bytes field 3432- crocus: Fix memory leaks on iris_resource_create failure paths 3433 3434Konrad Dybcio (1): 3435 3436- freedreno: Enable A619 3437 3438Konstantin Seurer (106): 3439 3440- radv: Fix ray queries with !15854 3441- radv: Do not discard hits with t=tmax 3442- radv: Enable rt primitive culling for spirv2nir 3443- ci: Always install glslangValidator for build jobs 3444- radv: Use VAs in radv_meta_buffer.c 3445- radv: Make fill_buffer_shader non-static 3446- radv: Format radv_acceleration_structure.c 3447- radv: Add the fuchsia radix sort 3448- radv: Build accaleration structures using LBVH 3449- radv: Clean up the accel-struct build shaders 3450- radv: Fix lowering ignore_ray_intersection 3451- vulkan: Add a shader module clone helper 3452- radv: Copy shader modules to avoid use after free 3453- radv/radix_sort: Add missing entry points 3454- radv/radix_sort: Make variable names consistent 3455- radv: Use the entire morton code as sort key 3456- ac/nir: Do not set cursor in try_extract_additions 3457- radv: Move RADV_RT_STAGE_BITS to radv_private.h 3458- radv: Treat rt stages like compute stages 3459- radv: Use RADV_RT_STAGE_BITS more often 3460- nir: Add a ray launch size addr intrinsic 3461- radv: Allow indirect dispatches without bo 3462- radv: Indirect unaligned dispatches 3463- radv,aco: Use ray_launch_size_addr 3464- radv: Implement vkCmdTraceRaysIndirectKHR 3465- radv: Advertise indirect ray tracing 3466- radv: Allow radv_fill_buffer to work with VAs only 3467- radv: Use radv_fill_buffer for accel struct builds 3468- Revert "radv: Make fill_buffer_shader non-static" 3469- radv: Add and use radv_bvh_node_type 3470- radv: Replace nir_vector_extract with nir_channel 3471- radv/ci: Skip slow/crashing ray tracing tests 3472- radv: Enable ray queries by default 3473- radv: Add fail/flake lists for navi22 3474- radv: Fix handling of primitiveOffset 3475- radv: Ignore transformOffset if transformData is 0 3476- vulkan: Use the correct enum in vk_buffer 3477- radv: Use vk_buffer 3478- radv: Add and use radv_cp_dma_wait_for_stages 3479- radv: Some acceleration structure cleanups 3480- radv: Require an alignment of 64 for accel structs 3481- radv: Fix the accel struct size calculation 3482- vulkan: Spec update to 1.3.216 3483- spirv: Add plumbing for ray_cull_mask 3484- radv: ray cull mask 3485- radv: Add accel struct queries for maintenance1 3486- radv: ACCESS_2_SHADER_BINDING_TABLE_READ_BIT_KHR 3487- radv: PIPELINE_STAGE_2_ACCELERATION_STRUCTURE_COPY 3488- nir: Handle ray_launch_size_addr in opt_preamble 3489- radv: vkCmdTraceRaysIndirect2KHR 3490- radv: Advertise VK_KHR_ray_tracing_maintenance1 3491- nir,spirv: Preserve inbounds access information 3492- nir/nir_lower_io: Optimize 32-bit inbounds access 3493- radv: Move some rt intrinsics to the top 3494- radv: Introduce an enum for the any hit status 3495- radv: Fix the t comparison for triangles 3496- radv: Refactor lower_rt_derefs 3497- nir: Add a nir_test_mask helper 3498- radv: Use nir_test_mask instead of i2b(iand) 3499- nir: Use nir_test_mask instead of i2b(iand) 3500- d3d12: Use nir_test_mask instead of i2b(iand) 3501- intel: Use nir_test_mask instead of i2b(iand) 3502- microsoft: Use nir_test_mask instead of i2b(iand) 3503- dozen: Use nir_test_mask instead of i2b(iand) 3504- radv: Use two bools for ahit_status 3505- radv: Skip pipeline_no_null_shaders_flag tests 3506- radv: Fix acceleration structure size queries 3507- radv: Use a global address for ray_launch_size 3508- radv: Use a global address for sbt_base 3509- radv: Only set rt stack size for dynamic stacks 3510- vulkan: Common vk_format_get_component_bits 3511- nir: Add a common gen_rect_vertices implementation 3512- radv: Use nir_gen_rect_vertices 3513- v3dv: Use nir_gen_rect_vertices 3514- radeonsi: Set vertex_id_zero_based 3515- aco: Remove dead nir_intrinsic_load_vertex_id case 3516- ac/llvm: Remove load_vertex_id handling 3517- util/format: Use an explicit length for the descs 3518- util/format: Assert that formats are valid 3519- radv: Remove format desc null checks 3520- radv: Use desc->format 3521- v3dv: Remove format desc null asserts 3522- turnip: Remove format desc null assert 3523- gallivm: Remove format desc null checks 3524- tgsi: Remove format desc null check 3525- util: Remove format desc null check 3526- agx: Remove format desc null check 3527- etnaviv: Remove format desc null check 3528- llvmpipe: Remove format desc null checks 3529- panfrost: Remove format desc null check 3530- r300: Remove format desc null check 3531- r600: Remove format desc null checks 3532- radeonsi: Remove format desc null checks 3533- softpipe: Remove format desc null check 3534- virgl: Remove format desc null checks 3535- pvr: Remove format desc null check 3536- util/format: Remove format desc null checks 3537- gallium/tests: Remove format desc null checks 3538- radv: Add meta saving/restoring for predicating 3539- radv: Use RADV_META_SUSPEND_PREDICATING 3540- radv: Simplify the meta init fail path 3541- radv: Fix out of bounds access in RTPSO creation 3542- vulkan: Fix descriptor set layout allocation scope 3543- vulkan: Handle descriptor set layout alloc fails 3544- radv: Use vk_descriptor_set_layout 3545- radv: Fix stack size calculation with stage ids 3546 3547Kristian Høgsberg (1): 3548 3549- util/u_debug: Use 'initialized' instead of 'first' 3550 3551Kyle Russell (1): 3552 3553- etnaviv: don't halve max_instructions for GC400 3554 3555Leandro Ribeiro (1): 3556 3557- Revert "egl/wayland: deprecate drm_handle_format() and drm_handle_capabilities()" 3558 3559Lionel Landwerlin (120): 3560 3561- anv: use shadow surface for stencil input attachment on gfx7 3562- intel: fix URB programming for GT1s 3563- intel/fs: require memory fence commit bit on Gfx9 3564- anv: improve INTEL_DEBUG for submit 3565- anv/intel: add a new debug flag for stalling after every draw/dispatch 3566- anv: implement DEBUG_SYNC 3567- anv: allow getting the address of the beginning of the batch 3568- intel/fs: tidy up lower of ray queries 3569- ci/iris: mark another test as flaky 3570- anv: implement VK_EXT_primitives_generated_query 3571- nir/divergence: handle load_global_block_intel 3572- anv: fix INTEL_DEBUG=sync 3573- intel: fixup number of threads per EU on XeHP 3574- anv: remove unused enum 3575- anv: fix acceleration structure descriptor template writes 3576- anv: skip acceleration structure in binding table emission 3577- anv: reemit 3DSTATE_STREAMOUT after memcpy 3578- anv: fix clflush usage on utrace copy batch 3579- anv: limit clflush usage 3580- vulkan/runtime: 0-out pipeline cache object 3581- iris: use new kernel uAPI to compute video memory 3582- anv: fix invalid utrace memcpy l3 config on gfx < 11 3583- anv: reset all dynamic state after secondary execution 3584- anv: add missing logic op set in pipeline dyn state 3585- Revert "anv: fix dynamic state emission" 3586- anv: rework sample location 3587- anv: never emit 3DSTATE_CPS in the pipeline batch 3588- anv: don't emit 3DSTATE_STREAMOUT in pipeline batch 3589- anv: don't emit 3DSTATE_WM in pipeline batch 3590- anv: don't emit 3DSTATE_BLEND_STATE_POINTERS in pipeline batch 3591- anv: don't emit 3DSTATE_DEPTH_BOUNDS in pipeline batch 3592- anv: don't emit 3DSTATE_VF_TOPOLOGY in pipeline batch 3593- anv: remove static_state_mask 3594- docs: explain state emission in Anv 3595- intel: fix stall debug option 3596- nir/cf: return cursor after insertion of cf_list 3597- nir/lower_shader_calls: don't use nop instructions as cursors 3598- nir/lower_shader_calls: don't insert code after break/continue 3599- nir/lower_shader_calls: put inserted instructions into a dummy block 3600- anv: fix primitives generated queries values 3601- intel/perf: add support for new opcodes in code generation 3602- intel/perf: store a copy of devinfo 3603- intel/perf: add support for dualsubslice count variable 3604- intel/perf: add support for GtSlice/GtSliceXDualsubsliceY variables 3605- intel/perf: stop overriding oa_format 3606- intel/perf: add OA A counter type 3607- intel/perf: add support new variable counting the number of EUs in slice0-3 3608- intel/perf: add max vfuncs 3609- intel/perf: add new layout for Gfx12.5 products 3610- intel/perf: disable sseu setting on Gfx12.5+ 3611- intel/perf: deal with OA reports timestamp values on DG2 3612- intel/ds: fix compilation 3613- u_trace/anv/iris: drop cs argument for recording traces 3614- intel/disasm: add missing handling of <1;1,0> 3615- util/debug: add an enable parsing helper 3616- util/perf: add options to enable/disable tracepoints 3617- anv: move internal RT shaders to the internal cache 3618- nir/divergence: handle load_scratch_base_ptr 3619- nir/divergence: handle load_ray_num_dss_rt_stacks_intel 3620- anv: don't expose EXT_border_color_swizzle on gfx7 3621- intel/validator: validate dst/src types against devinfo support 3622- anv: fix index buffer emission 3623- vulkan/wsi: fix crash with debug names on swapchain 3624- intel/dev: add a helper to update memory info 3625- intel/dev: add a fallback when memory regions are not available 3626- anv: limit RT writes to number of color outputs 3627- anv: disable perf queries on non RCS engines 3628- anv: make apply_pipeline_layout/compute_push_layout visible to NIR debug 3629- anv: silence border color swizzle debug message 3630- clc: add new feature options for intel_clc 3631- intel/clc: enable fp16 & subgroups for GRL 3632- anv: disable non uniform indexing of UBOs 3633- anv: remove device features handling now in runtime 3634- anv: use the right helper to invalidate memory 3635- intel/fs: make sure memory writes have landed for thread dispatch 3636- intel/nir: temporarly disable opt_uniform_atomics for RT/CL 3637- anv/utrace: use a bo pool for utrace buffers 3638- intel/fs: ray query fix for global address 3639- nir/serialize: restore ray query variables 3640- isl: add new helper for format component compatibility 3641- anv: track if images can be fast clear with non-zero color 3642- anv: move gfx12+ CCS_E clear color restriction 3643- anv: deal with isl format swizzles for buffer views 3644- anv: validate image view lowered storage formats for storage 3645- anv: add a comment about handling buffer view swizzles on gfx7 3646- intel: protect against empty invalidate ranges 3647- anv: skip flush/invalidate faster 3648- spirv: switch to uint64 for rayquery internal type 3649- anv: ensure tile flush before streamout writes 3650- intel/fs: Set NonPerspectiveBarycentricEnable when the interpolator needs it. 3651- anv: remove local computation of dynamic states 3652- anv: simplify dynamic buffer count in pipeline layout 3653- anv: break up anv_pipeline_compile_graphics() 3654- anv: split graphics nir loading 3655- anv: rename internal function for consistency 3656- anv: move CreateGraphicsPipelines to common code 3657- anv: move CreateComputePipelines to common code 3658- anv: move CreateRayTracingPipelines to common code 3659- anv: reorder & document fields of anv_graphics_pipeline 3660- anv: name non dynamic state fields correctly 3661- anv: fix primitive topology dynamic state emission on gfx7 3662- anv: move restart index to gfx state 3663- anv: remove pipeline setting dynamic state 3664- intel/dev: deal with i915 unallocated_size on smem 3665- anv: allocate RT scratch in local memory 3666- anv: make use of the new smallbar uAPI 3667- anv: decode init batch with INTEL_DEBUG=bat 3668- intel/dev: move verification function to a header 3669- intel/tools/drm-shim: fixup eu_stride for topology 3670- intel/tools/stub: fixup parsing of --platform= 3671- intel/dev: add a test verifying that device override works 3672- intel/dev: remove INTEL_DEVID_OVERRIDE 3673- anv: fixup PIPE_CONTROL restriction on gfx8 3674- anv: don't return incorrect error code for vkCreateDescriptorPool 3675- anv: fix GetPipelineExecutableStatistics for ray tracing pipelines 3676- anv: fixup assertions on lowered storage formats 3677- anv: fix assert in memory budget code when extension is not supported 3678- intel/fs: fix load_scratch intrinsic 3679- intel/fs: fixup scratch load/store handling on Gfx12.5+ 3680- intel/fs: fixup SEND validation check on overlapping src0/src1 3681 3682Liviu Prodea (7): 3683 3684- dzn: Fix incompatible pointer type error affecting MSYS2 MINGW32 Suggested-by: Yonggang Luo <luoyonggang@gmail.com> Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6807 3685- ci/MSVC: LLD is not really used anywhere 3686- ci/MinGW: LLD is not really used anywhere 3687- d3d12/dzn/spirv2dxil: Require version library 3688- Microsoft clc: strip lib prefix Otherwise OpenCLon12 ICD can't load it 3689- meson: Microsoft / maybe Intel CLC need the all-targets workaround just like clover 3690- meson: Only draw with llvm depends on native directly Tests, softpipe or AMD drivers don't depend on it directly 3691 3692Louis-Francis Ratté-Boulianne (4): 3693 3694- d3d12: Move d3d12_nir_lower_vs_vertex_conversion() to a common place 3695- microsoft/compiler: Add support for more scaled formats 3696- dzn: Allow some non-native formats to be used as vertex inputs 3697- dzn: Enable dEQP-VK.pipeline.vertex_input.* test cases 3698 3699Lucas Stach (37): 3700 3701- etnaviv: update headers from rnndb 3702- etnaviv: set VIVS_GL_VERTEX_ELEMENT_CONFIG depending on prim type 3703- etnaviv: don't supertile textures if supertiling is disabled via debug option 3704- etnaviv: fill all minor GPU features from the kernel 3705- etnaviv: update headers from rnndb 3706- etnaviv: correct bits per tile and clear value for HALTI5 3707- etnaviv: use new PE pipe address states on >= HALTI0 3708- etnaviv: use feature flag to determine which RS states to use 3709- etnaviv: use feature bit for one const src per instuction limitation 3710- etnaviv: clean up tiling setup in etna_compile_rs_state 3711- etnaviv: add support for big tile RS states 3712- etnaviv: properly set additional DEC400 compression states 3713- etnaviv: use feature bit to check for big tile support 3714- etnaviv: add tile size helper 3715- etnaviv: reinstate accidentially commented assert 3716- etnaviv: improve single buffer setup 3717- etnaviv: add linear PE support 3718- etnaviv: switch depth mode when linear PE is used 3719- etnaviv: expose real map buffer alignment 3720- etnaviv: allow mapped buffers during execution 3721- etnaviv: drm: always use hash to track BO index 3722- etnaviv: drm: don't cache mmap offset 3723- etnaviv: drm: make etna_bo_map thread safe 3724- etnaviv: drm: rename etna_drm_table_lock 3725- etnaviv: add function to get resource status 3726- etnaviv: rework resource status tracking (again) 3727- etnaviv: drop scanout assumption for resources allocated via modifier interface 3728- kmsro: add 'imx-lcdif' driver support 3729- etnaviv: move dummy BOs to screen 3730- etnaviv: add texture descriptor suballocator 3731- etnaviv: tex_desc: make error handling more consistent 3732- etnaviv: tex_desc: remove descriptor patch TODO comment 3733- etnaviv: fix use after free in async shader compile 3734- etnaviv: remove etna_align_up() function 3735- etnaviv: move checking for MC2.0 for TS into screen init 3736- etnaviv: use linear PE rendering only on properly aligned surfaces 3737- etnaviv: add debug option to disable linear PE feature 3738 3739M Henning (7): 3740 3741- nouveau/nir: Allow up to 6 nested joins 3742- nouveau: Emit cctl to flush L1 cache for atomics 3743- nouveau: Skip cctl for atomic counters in tgsi 3744- nouveau/nir: Split fewer 64-bit loads 3745- ttn: Make ttn_optimize_nir match gl_nir_opts 3746- nv50: Fix tlsSpace alignment 3747- nouveau: Drop C++03 compat code 3748 3749Manas Chaudhary (2): 3750 3751- panvk: Add PANVK_DEBUG_DUMP dump mappings 3752- panvk: Allow PAN_BO_INVISIBLE to be mapped with dump flag 3753 3754Manuel Stoeckl (4): 3755 3756- dri_interface: add XBGR16161616 format 3757- gbm: add GBM_FORMAT_[AX]BGR16161616 3758- dri: fix incorrect fourcc for R16G16B16A16_UNORM format 3759- dri: add [ax]bgr16161616 to format lookup tables 3760 3761Marcin Ślusarz (40): 3762 3763- anv: invalidate all metadata in anv_nir_lower_ubo_loads 3764- intel/compiler: invalidate all metadata in brw_nir_lower_intersection_shader 3765- intel/compiler: invalidate metadata in brw_nir_initialize_mue 3766- compiler: add VARYING_SLOT_CULL_PRIMITIVE 3767- intel/compiler: handle VARYING_SLOT_CULL_PRIMITIVE in mesh 3768- intel/compiler: handle gl_Viewport and gl_Layer in FS URB setup 3769- anv: update task/mesh distribution with the recommended values 3770- intel/compiler: print name of the unhandled intrinsic 3771- anv: disable streamout before emitting mesh shading state 3772- intel/compiler: implement primitive shading rate for mesh 3773- anv: handle primitive shading rate for mesh 3774- intel/tools: add support for GEM_CREATE_EXT in intel_dump_gpu 3775- intel/tools: clean up mmio regs definitions 3776- intel/tools: drop wrappers around mmio regs macros 3777- intel/tools/aubinator: add support for 2 "new" subopcodes 3778- intel/tools/aubinator: list all platforms in help message 3779- intel/tools: add macros for gfx12+ variant of VCSUNIT0 3780- nir: add load_mesh_inline_data_intel intrinsic to nir_divergence_analysis 3781- nir: add load_task_payload intrinsic to nir_divergence_analysis 3782- intel/tools: fix 32-bit build 3783- anv: remove redundant code calculating dynamic states mask 3784- anv: remove invalid copy/pasted comment 3785- anv: mask out not applicable state flags when setting up mesh pipeline 3786- intel/dev: fixup max_cs_workgroup_threads after hwconfig is read 3787- nir/opt_load_store_vectorize: handle task payloads 3788- intel/compiler: vectorize task payload loads/stores 3789- nir/lower_task_shader: insert barrier before/after shared memory read/write 3790- intel/compiler: assert that base is 0 for [load|store]_shared intrins 3791- intel: fix typos found by codespell 3792- intel/common: allocate space for at least one task urb 3793- intel/compiler: adjust task payload offsets as late as possible 3794- anv: disable injection of primitive shading rate for mesh 3795- intel/compiler: print shaders after nir_remove_unused_varyings 3796- nir: convert unused mesh outputs to shared memory 3797- intel/compiler: emit URB fences for TASK/MESH 3798- intel/compiler: there are 4 types of fences on gfx >= 12.5 3799- intel/compiler: insert URB fence before task/mesh termination 3800- intel/compiler: use NIR_PASS more 3801- intel/compiler: fix mesh urb write regression 3802- anv: disable task redistribution 3803 3804Marek Olšák (311): 3805 3806- gallium: move drivers/radeon/* into drivers/radeonsi/ 3807- gallium: move radeon_winsys.h into gallium/include/winsys/ 3808- gallium/winsys: move {amdgpu,radeon_drm}_public.h contents into radeon_winsys.h 3809- gallium/ddebug: implement pipe_vertex_state callbacks 3810- nir: fix an uninitialized variable valgrind warning in nir_group_loads 3811- ac/surface: fix an addrlib race condition on gfx9 3812- radeonsi: set more state bits for disabled binning 3813- radeonsi: don't compile monolithic PS for dual source blending 3814- radeonsi: add an SQTT workaround for chips with disabled RBs 3815- radeonsi: set PA_SC_EDGERULE according to bottom_edge_rule state 3816- radeonsi: initialize line stipple config registers 3817- radeonsi: fix ALPHA_IS_ON_MSB determination 3818- radeonsi: handle an out-of-bounds sampler/image index better 3819- radeonsi: don't treat skipped GS copy shader compilation as successful 3820- radeonsi: dump shaders after si_shader_binary_upload to fix printed LDS stat 3821- radeonsi: ignore AMD_DEBUG=nodcc if the texture is imported 3822- radeonsi: run nir_io_add_const_offset_to_base for TES/TCS as late optimizations 3823- radeonsi: scalarize IO instructions 3824- radeonsi: add pipe_stream_output_info into si_shader_context 3825- radeonsi: use nir_shader_get_entrypoint in si_nir_scan_shader 3826- radeonsi: apply key.ge.opt.kill_{outputs,pointsize,clipdistance} in NIR 3827- radeonsi: gather pipe_stream_output_info from NIR intrinsics 3828- radeonsi: move most "info" fields from si_shader_selector into si_shader_info 3829- radeonsi: move si_shader_info::stage into si_shader_selector 3830- ac: add ac_nir_optimize_outputs, a NIR version of ac_optimize_vs_outputs 3831- radeonsi: assign param export indices before compilation 3832- ac/llvm: remove LLVM pass ac_optimize_vs_outputs 3833- amd: document chips 3834- ac/surface/tests: generalize and extend gfx10 tests 3835- ac: add more non-shadowed registers to the lists 3836- ac/llvm: add AC_WAIT_EXP for ac_build_waitcnt 3837- ac/llvm: remove immoffset parameter from ac_build_tbuffer_store 3838- ac/llvm: remove immoffset parameter from ac_build_tbuffer_load 3839- ac/llvm: remove inst_offset parameter from ac_build_buffer_store_dword 3840- ac/llvm: use soffset correctly in ac_build_buffer_load 3841- ac/llvm: remove inst_offset parameter from ac_build_buffer_load 3842- ac/llvm: remove immoffset parameter from ac_build_tbuffer_load_byte/short 3843- ac/llvm: remove unused code from ac_llvm_build.c/h 3844- ac/llvm: don't use tbuffer_store as a fallback for swizzled stores 3845- ac/llvm: rename tbuffer -> buffer where tbuffer instructions are not generated 3846- ac,radeonsi: rework how scratch_waves is used and move it to ac_gpu_info.c 3847- ac/gpu_info: remove old and unused fields from radeon_info 3848- ac/llvm: remove unused LLVM helpers 3849- ac/llvm: don't create the target machine if the LLVM processor is unsupported 3850- radeonsi: remove an obsolete comment about LLVM 14 3851- radeonsi: fail to create pipe_screen if LLVM doesn't support the GPU 3852- winsys/amdgpu: fix a mutex deadlock when we fail to create pipe_screen 3853- radeonsi/ci: add polaris11 baseline 3854- util: add pipe formats R10G10B10X2_SINT, B10G10R10X2_SINT, G16R16_SINT 3855- util: change util_format_snorm8_to_sint8 -> snorm_to_sint for radeonsi 3856- ac/gpu_info: rework how num_se is derived 3857- ac: remove really_needs_scratch, parameter from ac_parse_shader_binary_config 3858- ac/llvm: set the correct cache policy for sparse buffer loads 3859- ac/llvm: don't set ac_dlc in visit_intrinsic 3860- ac: support GR channel order in ac_choose_spi_color_formats 3861- ac,radeonsi: report SCALED formats as unsupported by samplers and color buffers 3862- radeonsi: always print the lower-case GPU name for radeonsi-run-tests.py 3863- radeonsi: cleanup local includes 3864- radeonsi: align the tessellation ring address to 2MB 3865- radeonsi: remove unused si_cs_shader_state::uses_scratch 3866- radeonsi: simplify setting COMPUTE_PGM_RSRC2.SCRATCH_EN 3867- radeonsi: code style fix 3868- radeonsi: add si_can_use_compute_blit 3869- radeonsi: don't report R64_*INT as a sampler format because it doesn't work 3870- radeonsi: don't allow broken compute image copies into A8R8_UNORM 3871- radeonsi: add color buffer support for 2_10_10_10 pipe formats 3872- radeonsi: add color buffer support for R9G9B9E5_FLOAT on gfx10.3 3873- radeonsi: implement compute_copy_image between 1D_ARRAY and other texture types 3874- radeonsi: cosmetic changes in si_compute_copy_image 3875- radeonsi: assert format support in si_compute_copy_image 3876- radeonsi: don't expose formats that CB supports, but not shaders 3877- radeonsi: rename si_test_blit -> si_test_image_copy_region 3878- radeonsi/test_image_copy: remove whole image copies 3879- radeonsi/test_image_copy: remove linear and power-of-two size testing 3880- radeonsi/test_image_copy: remove special SDMA codepaths 3881- radeonsi/test_image_copy: simplify texture size generation 3882- radeonsi/test_image_copy: separate generating a random pipe_resource 3883- radeonsi/test_image_copy: separate printing pipe_resource into a function 3884- radeonsi/test_image_copy: test all interesting texture targets 3885- radeonsi/test_image_copy: test copying a block-compressed format 3886- radeonsi/test_image_copy: test mipmaps 3887- radeonsi/test_image_copy: test all formats 3888- radeonsi/ci: update sienna_cichlid results 3889- radeonsi: handle compressed formats in si_compute_copy_image 3890- radeonsi: improve the subsampled copy_image calculation 3891- radeonsi: remove compressed and subsampled gfx copy from resource_copy_region 3892- radeonsi: remove si_create_sampler_view_custom and related code 3893- radeonsi: remove si_create_surface_custom 3894- mesa: consider the sample count when choosing a texture format 3895- frontend/dri: allow swapped BGR->RGB channel order for MSAA color buffers 3896- amd/registers: hardcode GC base offsets in the json generator 3897- amd/registers: add gfx11 to the json generator 3898- amd/registers: add gfx11.json 3899- amd/registers: add gfx11-rsrc.json 3900- amd: enable gfx11 in header generator, fix drivers with renamed gfx6-10 defs 3901- amd: add gfx11 to packet definitions 3902- amd: import gfx11 addrlib 3903- amd: add Mesa-only addrlib changes 3904- ac: don't align VGPRs to 8 or 16 for gfx11 3905- ac: implement ac_get_tbuffer_format for gfx11 3906- ac: scratch buffer register changes for gfx11 3907- ac: implement register shadowing for gfx11 3908- ac/llvm: export mrt0 instead of null on gfx11 3909- ac/llvm: don't set DLC on gfx11 because it means something else there 3910- ac/llvm: update pknorm and waitcnt for gfx11 3911- ac/llvm: don't set GLC for stores on gfx11 3912- ac/gpu_info: set cu_mask correctly for gfx11 3913- ac/surface: gfx11 changes 3914- ac/surface: define gfx11 modifiers 3915- ac/surface: add gfx11 support to modifiers tests 3916- radeonsi/gfx11: increase the hw screen offset alignment 3917- radeonsi/gfx11: always allow DCC stores 3918- radeonsi/gfx11: expect packed threadID VGPRs 3919- radeonsi/gfx11: enable NGG-only draw paths 3920- radeonsi/gfx11: enable arbitrary DCC format reinterpretation 3921- radeonsi/gfx11: use the new TCS WaveID SGPR to compute vs_rel_patch_id 3922- radeonsi/gfx11: program inst_pref_size for graphics 3923- radeonsi/gfx11: buffer descriptor changes 3924- radeonsi/gfx11: add CB deltas 3925- ac,radeonsi/gfx11: remove FMASK loads 3926- ac,radeonsi/gfx11: set SWIZZLE_ENABLE correctly 3927- radeonsi/gfx11: don't set non-existent SPI_SHADER_USER_DATA_VS_x 3928- radeonsi/gfx11: don't set COMPR for exports, use 0x3 channel mask instead 3929- radeonsi/gfx11: implement attributes through memory 3930- radeonsi/gfx11: scattered register deltas 3931- radeonsi/gfx11: don't set non-existent CP_COHER_START_DELAY 3932- radeonsi/gfx11: emit SQ_NON_EVENT for tessellation at the end of IBs 3933- radeonsi/gfx11: don't use FLUSH_AND_INV_DB_META 3934- radeonsi/gfx11: TF_RING_SIZE changed to a per-SE size 3935- radeonsi/gfx11: limit MSAA color buffers to the RGBA channel order 3936- radeonsi/gfx11: update the initialization of SGPR0/1 registers for HS and GS 3937- radeonsi/gfx11: change LDS allocation granularity for PS 3938- radeonsi/gfx11: don't count the non-existent scratch_byte_offset SGPR 3939- radeonsi/gfx11: resolve MSAA using u_blitter 3940- radeonsi/gfx11: mark streamout as unimplemented for now 3941- radeonsi/gfx11: limit CP DMA to max 32KB sizes 3942- radeonsi: inline si_cp_dma_prefetch in si_draw_vbo for lower overhead 3943- radeonsi/gfx11: add a workaround for CB perf counters 3944- amd: add chip identification for gfx1100-1103 3945- radeonsi: remove tautologies when setting CU_EN 3946- amd: rename ring_type --> amd_ip_type and match the kernel enum values 3947- amd: rename RING_* enums to AMD_IP_* 3948- amd: replace num_rings[type] with ip[type].num_queues 3949- ac/gpu_info: simplify HW IP querying 3950- ac/gpu_info: remove dead GDS query code 3951- ac/gpu_info: print all IP versions reported by the kernel 3952- ac/gpu_info: fix incorrect IP versions reported by the kernel 3953- radeonsi: report 10_10_10_2 scaled formats as unsupported on gfx6-9 as well 3954- radeonsi: report more missing/broken texture formats as unsupported on gfx6-9 3955- ac/llvm: implement memory_barrier_tcs_patch properly to fix LLVM 15 3956- radeonsi: wait before s_barrier in TCS epilog to fix LLVM 15 3957- ac/llvm: inline emit_membar 3958- ac/llvm: replace LLVMBuildFence with ac_build_waitcnt(VSTORE) 3959- ac/llvm: add gl_shader_stage parameter into ac_build_s_barrier 3960- ac/llvm: move the gfx6 optimization for TCS barriers into ac_build_s_barrier 3961- ac/llvm: fix the remaining s_barriers for LLVM 15 3962- radeonsi: add si_get_tcs_epilog_key 3963- ac/llvm: skip s_barrier if tess patches don't cross a wave boundary 3964- radeonsi: inline si_nir_build_llvm 3965- ac/llvm: remove ac_shader_abi::emit_outputs 3966- radeonsi: cleanups getting rid of the ambigous "epilogue" word 3967- radeonsi: get rid of the ambiguous "prologue" word 3968- radeonsi: try to group stage-specific code in si_llvm_translate_nir 3969- ac: replace 5 ac_shader_abi::load_* callbacks with 1 intrinsic_load callback 3970- amd: change chip_class naming to "enum amd_gfx_level gfx_level" 3971- radeonsi: rename RADEON_FLAG_UNCACHED -> RADEON_FLAG_GL2_BYPASS 3972- winsys/amdgpu-radeon: rework how we describe heaps 3973- winsys/amdgpu: handle RADEON_FLAG_ENCRYPTED as a heap modifier 3974- radeonsi: use PIPE_RESOURCE_FLAG_UNMAPPABLE and DRIVER_INTERNAL more 3975- winsys/amdgpu-radeon: clean up code related to heaps 3976- include: update amdgpu_drm.h 3977- radeonsi: use the new flag AMDGPU_GEM_CREATE_DISCARDABLE 3978- radeonsi: use AMDGPU_VM_PAGE_NOALLOC to disable MALL (infinity cache) 3979- radeonsi: fix a crash when failing to create a context 3980- radeonsi: print an error when failing to create a context 3981- radeonsi/gfx11: set BIG_PAGE for the attribute ring 3982- ac/gpu_info: silence a valgrind warning due to amdgpu_query_hw_ip_info 3983- radeonsi: remove GFX9_MERGED_NUM_USER_SGPR definition 3984- ac/gfx11: fix the scratch buffer 3985- radeonsi/gfx11: fix alpha-to-coverage with stencil or samplemask export 3986- radeonsi/gfx11: don't insert shader code for GS_PIPELINE_STATS_EMU 3987- radeonsi/gfx11: fix the value of VGT_GS_OUT_PRIM_TYPE at the beginning of IBs 3988- Revert "radeonsi/gfx11: limit MSAA color buffers to the RGBA channel order" 3989- radeonsi/gfx11: fix VM faults due to the attribute ring 3990- radeonsi/gfx11: optimize attribute stores 3991- radeonsi: merge all preamble states into one 3992- radeonsi: move CS preamble emission into the winsys 3993- winsys/amdgpu: initialize IB_PREAMBLE in advance 3994- radeonsi: fix resource_copy_region with ETC formats (e.g. for Stoney) 3995- Revert "frontend/dri: allow swapped BGR->RGB channel order for MSAA color buffers" 3996- Revert "mesa: consider the sample count when choosing a texture format" 3997- amd: rename fishes to Navi21, Navi22, Navi23, Navi24, and Rembrandt 3998- r300,r600,svga: save the FS constant buffer for u_blitter to fix a regression 3999- st/mesa: remove unused GENERIC input and output from the clear VS 4000- ac/gpu_info: always retile DCC on gfx10 and newer chips 4001- mesa: clamp UINT values greater than INT_MAX for glGetInteger 4002- mesa: change GL_MAX_COMPUTE_WORK_GROUP_COUNT to UINT 4003- mesa: change GL_MAX_UNIFORM_BLOCK_SIZE to UINT 4004- mesa: change GL_MAX_TEXTURE_BUFFER_SIZE to UINT 4005- mesa: change GL_MAX_SHADER_STORAGE_BLOCK_SIZE to UINT 4006- st/mesa: limit MaxComputeWorkGroupCount to INT_MAX - 1 due to failing tests 4007- st/mesa: round down MaxUniformBlockSize to a multiple of 4 for piglit 4008- gallium: rename PIPE_SHADER_CAP_MAX_CONST_BUFFER_SIZE to \*_BUFFER0_* 4009- gallium: add PIPE_CAP_MAX_CONSTANT_BUFFER_SIZE_UINT for UBOs 4010- gallium: rename PIPE_CAP_MAX_TEXTURE_BUFFER_SIZE->MAX_TEXEL_BUFFER_ELEMENTS_UINT 4011- gallium: rename PIPE_CAP_MAX_SHADER_BUFFER_SIZE -> \*_UINT 4012- radeonsi: increase the max compute LDS size to 64KB for gfx7+ 4013- radeonsi: report correct maximum compute grid sizes 4014- radeonsi: clamp against MAX_TEXEL_BUFFER_ELEMENTS correctly 4015- radeonsi: change max TBO/SSBO sizes again and rework max alloc size 4016- radeonsi: compute PIPE_CAP_MAX_TEXEL_BUFFER_ELEMENTS_UINT correctly 4017- radeonsi: set the max UBO size same as the max SSBO size 4018- winsys/amdgpu: use AMDGPU_IB_FLAG_PREAMBLE for the CS preamble on gfx10+ 4019- radeonsi: print LDS size in bytes 4020- radeonsi: don't use info.gs.invocations if it's not GS 4021- radeonsi: fix uninitialized wait_mem_scratch_tmz 4022- radeonsi: remove unused dword from wait_mem_scratch 4023- radeonsi: fix polygon stippling without color and Z outputs (v2) 4024- radeonsi: restructure PS no-export fixups 4025- radeonsi: determine DB_SHADER_CONTROL in si_shader_ps 4026- radeonsi: set INTERPOLATE_COMP_Z to 0 to work around an EQAA bug 4027- radeonsi: add BREAK_BATCH at the beginning of IBs 4028- radeonsi: simplify how pipeline statistic offsets are computed 4029- radeonsi: rework how VS_STATE_BITS are set for VS, TES, and GS 4030- radeonsi: rename and regroup VS_STATE definitions 4031- radeonsi: move GS_STATE bits to the end to make space at the beginning 4032- radeonsi: rework how vs_state_bits is set and unpacked 4033- radeonsi: fix an NGG streamout hang with monolithic shaders 4034- radeonsi: fix a crash in gfx10_sh_query_get_result_resource 4035- radeonsi: unconditionally enable the streamout overflow query with NGG 4036- radeonsi: inline gfx10_emit_streamout_begin/end 4037- radeonsi: fix NGG streamout hang by allocating GDS in the right place 4038- radeonsi: remove streamout code from shaders if no streamout buffers are bound 4039- winsys/amdgpu: add a kernel GDS management workaround retrying on -ENOMEM 4040- radeonsi: allocate GDS only once per process 4041- radeonsi: allocate only 1 GDS OA counter for gfx10 NGG streamout 4042- ac,radeonsi: don't export null from PS if it has no effect on gfx10+ 4043- ac/gpu_info: clamp gart_size_kb and vram_size_kb to fix buggy kernel driver 4044- gallium/u_threaded: fix buffer busyness tracking 4045- ac/llvm: fix tcs_wave_id unpacking on gfx11 4046- amd/gfx11: add PixelWaitSync packet fields 4047- radeonsi/gfx11: compile monolithic PS if it writes memory 4048- radeonsi/gfx11: allocate more space for pipeline statistics 4049- radeonsi/gfx11: synchronize correctly before setting SPI_ATTRIBUTE_RING_* 4050- radeonsi/gfx11: don't use memory for waiting for cache flushes 4051- radeonsi/gfx11: don't allocate unused wait_mem_scratch 4052- radeonsi/gfx11: drop the ES vertex count requirement 4053- radeonsi/gfx11: add missing register shadowing code 4054- radeonsi/gfx11: rework GDS streamout code to single-lane and enable streamout 4055- amd: update addrlib - trivial changes 4056- ac/surface: expose all 64K_R_X and 256K_R_X modifiers on gfx11 4057- radeonsi: fix random PS wave size 4058- util: remove our redefinition of assert 4059- Change all debug_assert calls to assert 4060- util: remove now unused debug_assert 4061- ttn: set dest_type for TXQ 4062- gallium: add pipe_blit_info::dst_sample, update u_blitter to write that sample 4063- ac/gpu_info: rename info fields to num_cu, memory_bus_width, memory_freq_mhz 4064- ac/gpu_info: remove tabs 4065- ac/gpu_info: rework and extend device info to be more informative 4066- ac/gpu_info: use hw_ip::ip_discovery_version to set IP versions 4067- ac/gpu_info: use the kernel-reported GFX IP version to set gfx_level 4068- ac/gpu_info: remove vram_size and gtt_size in favor of \*_kb variants 4069- ac/gpu_info: remove redundant uvd_decode 4070- ac/gpu_info: remove redundant jpeg_decode 4071- ac/gpu_info: remove redundant uvd_encode 4072- ac/gpu_info: remove redundant vce_encode 4073- ac/gpu_info: remove redundant vcn_encode 4074- amd: require amdgpu DRM 3.2.0 from April 2016 4075- gallium/radeon: require radeon DRM 2.45.0 from April 2016 4076- radeonsi: remove workarounds for radeon DRM < 2.45.0 4077- ac/gpu_info: use drm_amdgpu_device_info instead of amdgpu_gpu_info 4078- ac/gpu_info: remove amdgpu_gpu_info parameter from ac_query_gpu_info 4079- ac/gpu_info: remove unused has_unaligned_shader_loads 4080- gallium/radeon: require radeon DRM 2.50.0 (kernel 4.12) from July 2017 4081- ac/gpu_info: require amdgpu DRM 3.15.0 (kernel 4.12) from July 2017 4082- util/u_threaded: fix a CPU storage crash due to an invalid codepath taken 4083- glthread: don't ignore glPushAttrib/glPopAttrib when tracking GL_CULL_FACE 4084- glthread: don't sync on IsEnabled(GL_DEPTH_TEST) by tracking it in glthread 4085- util/format: add util_format_rgbx_to_rgba helper 4086- amd/addrlib: fix 3D texture allocation failures on gfx11 4087- ac/surface: don't forbid 256KB swizzle modes on smaller gfx11 chips 4088- radeonsi: unify VGT_TESS_DISTRIBUTION programming 4089- radeonsi: move SI_MAX_VRAM_MAP_SIZE to si_debug_options.h 4090- radeonsi: call pipe->blit instead of util_blitter_blit after MSAA resolving 4091- radeonsi: make si_launch_grid_internal static 4092- radeonsi: add common helper si_launch_grid_internal_images that is more robust 4093- radeonsi: remove compute-based DCC decompression because it's broken 4094- radeonsi: don't do image stores with RGBX, L, LA, I, and SRGB formats 4095- radeonsi: follow shader_info.float_controls_execution_mode (mostly) 4096- radeonsi: add need_fmask_expand parameter into si_decompress_subresource 4097- radeonsi: check for 16-bit hw support instead of relying on options.fp16 4098- radeonsi: move compute-related code from si_blit.c to si_compute_blit.c 4099- radeonsi: fold async_copy into the preceding conditional in si_blit 4100- radeonsi: cosmetic changes around do_hardware_msaa_resolve 4101- radeonsi/gfx11: use correct VGT_TESS_DISTRIBUTION settings 4102- radeonsi/gfx11: set SAMPLE_MASK_TRACKER_WATERMARK = 15 and clean up 4103- radeonsi/gfx11: use better PRIM_GRP_SIZE_GFX11 setting 4104- radeonsi/gfx11: skip code in si_update_shaders that has no effect 4105- radeonsi/gfx11: rename si_calc_inst_pref_size -> si_get_shader_prefetch_size 4106- radeonsi/gfx11: enable shader prefetch except for initial chip revisions 4107- radeonsi/gfx11: use a better workaround for the export conflict bug 4108- radeonsi: make various blit functions non-static 4109- radeonsi: allow texture_map to upload only 1 sample for MSAA instead of all 4110- radeonsi: add a randomized blit test 4111- radeonsi: move the no-AA small prim precision cull constant into an SGPR 4112- radeonsi: move small prim precision computation out of si_emit_cull_state 4113- radeonsi: don't assume that TC_ACTION_ENA invalidates L1 cache on gfx9 4114- glthread: unbind framebuffers in glDeleteFramebuffers 4115- glthread: call _mesa_glthread_DeleteBuffers unconditionally 4116- ac/surface: disallow 256KB swizzle modes on gfx11 APUs 4117 4118Marek Vasut (3): 4119 4120- dri: Check buffer height and avoid overflow 4121- etnaviv: Rename etna_emit_tex() args 4122- etnaviv: Implement TXD 4123 4124Mark Collins (1): 4125 4126- tu: Clamp priority in DRM submitqueue creation 4127 4128Mark Janes (11): 4129 4130- intel: parse intel_measure environment without side effects 4131- intel: relax assertion in INTEL_MEASURE 4132- anv: fix performance bug in INTEL_MEASURE 4133- intel: provide a callback to clean up after intel_measure_gather 4134- iris: provide a callback to INTEL_MEASURE to clean up snapshots 4135- util: define helpful macros for compiler diagnostic features 4136- iris: pad all structures used in a shader key 4137- iris: reorder to minimize padding 4138- intel/compiler: pad all data structures used by shader cache keys 4139- intel/compiler: reorder shader cache keys to minimize padding 4140- intel/dev: drop warning for unhandled hwconfig keys 4141 4142Martin Roukala (né Peres) (27): 4143 4144- ci/b2c: fix the generation of the IMAGE_UNDER_TEST variable 4145- ci/b2c: bump the kernel to 5.17.1 4146- ci/b2c: update b2c to enable monitoring support 4147- radv/ci: add a Van Gogh dEQP runner 4148- ci: add ACO_DEBUG to the list of variables to pass down for testing 4149- ci/radv: remove amdgpu.gpu_recovery=0 from the kernel cmdline 4150- radv/ci: reboot the DUTs when hitting a ring timeout 4151- radv/ci: reduce the execution timeout 4152- radv/ci: re-enable fossils testing 4153- radv/ci: enable the experimental support for mesh shaders 4154- radv/ci: update the list of flakes 4155- radv/ci: allow polaris10 to retry more times 4156- radv/ci: bump the timeout of vangogh/renoir 4157- radv/ci: update the vkcts expectations for the CI machines 4158- radv/ci: remove an expected fail from all families 4159- radv/ci: try to guess failure expectations for Fiji/Raven/Navi14/Navi22 4160- ci/valve: do not extract the install tarball on the runner 4161- ci/valve: uprev boot2container 4162- Revert "ci/valve: uprev boot2container" 4163- radv/ci: test vkd3d on kabini 4164- ci/radv: move a test from the fail list to the flake list 4165- ci/radv: add a missing flake 4166- ci/radv: enable vkcts testing on kabini 4167- ci/x86: update to llvm 13 4168- radv/ci: test the llvm backend on navi21 4169- radv/ci: allow specifying a warning pattern 4170- radv/ci: do not reboot on soft-recovered hangs, just warn 4171 4172Matt Coster (20): 4173 4174- pvr: csbgen: Formatting pass (PEP-8 plus other minor changes) 4175- pvr: csbgen: Add a missing @staticmethod decorator 4176- pvr: csbgen: Make some loops more pythonic 4177- pvr: csbgen: Cleanup imports 4178- pvr: csbgen: Remove "dim" functionality from Group 4179- pvr: csbgen: Remove unused function parameters 4180- pvr: csbgen: Add typing information 4181- pvr: csbgen: Isolate "main" function with __name__ check 4182- pvr: csbgen: Replace map() and filter() with list comprehension 4183- pvr: csbgen: Fix "local variable may be undefined" lints 4184- pvr: csbgen: Add \*_unpack() functions for all generated struct types 4185- pvr: debug: Implement PVR_DEBUG variable 4186- pvr: csbgen: Make all generated enums unambiguous 4187- pvr: Add a const qualifier to inline pvr_csb_*() helpers 4188- pvr: Remove pvr_cmd_struct() macro 4189- pvr: Make pvr_cmd_pack() macro clearly internal 4190- pvr: Split out unioned structs from struct pvr_sub_cmd 4191- pvr: Implicitly assert that the correct sub-command type is present 4192- pvr: Split pvr_dev_addr_t into a separate header 4193- pvr: Add helper macros for creating pvr_dev_addr_t instances 4194 4195Matt Turner (3): 4196 4197- mesa: Drop dead #include "sparc/sparc.h" 4198- freedreno/ir3: Move the texture array coord fixup to nir 4199- freedreno/ir3: Pass 16-bit sampler coordinates when possible. 4200 4201Matti Hamalainen (7): 4202 4203- gallium/tools: fixes to option handling 4204- gallium/tools: add option for ignoring junk calls in trace dumper 4205- gallium/tools: reimplement tracediff completely in Python 4206- pytracediff: add per-line difference highlighting for blocks 4207- pytracediff: make -M ("method only") option print arguments for differing calls 4208- pytracediff: change how 'junk' calls are handled 4209- pytracediff: implement pager ('less') invocation internally 4210 4211Max Kellermann (2): 4212 4213- gallium/u_threaded: add missing reference counts for draw_multi slots 4214- gallium/u_threaded: fix offset calculation for draw_multi slots 4215 4216Michael Olbrich (1): 4217 4218- crocus: export GEM handle with RDWR access rights 4219 4220Michael Skorokhodov (2): 4221 4222- glsl: Fix ir_quadop_vector validation 4223- anv: Update line range 4224 4225Michel Zou (7): 4226 4227- gallium: fix unused symbols warnings 4228- meson: call run_command with check=true 4229- vulkan/wsi: fix missing unistd include 4230- zink: fix pointer size conversion warning 4231- lavapipe: allow null handleTypes 4232- meson: drop dladdr check on win32 4233- util/xmlconfig: fix -Wmaybe-uninitialized warning 4234 4235Mihai Preda (28): 4236 4237- gallium: add opaque pointers shim for LLVM < 8.0 4238- gallium: use LLVM opaque pointers in draw_llvm.c 4239- gallium/llvmpipe: use LLVM opaque pointers in lp_bld_interp.c 4240- gallium/llvmpipe: use LLVM opaque pointers in lp_bld_depth.c 4241- gallivm: use LLVM opaque pointers in lp_bld_arit.c 4242- gallivm: use LLVM opaque pointers in lp_bld_assert.c 4243- gallivm: use LLVM opaque pointers in lp_bld_format_aos.c 4244- gallivm: use LLVM opaque pointers in lp_bld_conv.c 4245- gallivm: use LLVM opaque pointers in lp_bld_coro.c 4246- gallivm: use LLVM opaque pointers in lp_bld_printf.c 4247- gallivm: use LLVM opaque pointers in lp_bld_flow.c 4248- gallivm: use LLVM opaque pointers in lp_bld_intr.c 4249- gallivm: use LLVM opaque pointers in lp_bld_format_s3tc.c 4250- gallivm: use LLVM opaque pointers in lp_bld_gather.c 4251- gallivm: use LLVM opaque pointers in lp_bld_nir_soa.c 4252- gallivm: use LLVM opaque pointers in lp_bld_sample.c 4253- gallivm: use LLVM opaque pointers in lp_bld_sample_soa.c 4254- gallivm: use LLVM opaque pointers in lp_bld_struct.c 4255- gallivm: use LLVM opaque pointers in lp_bld_tgsi_soa.c 4256- gallivm: use LLVM opaque pointers in lp_bld_tgsi_aos.c 4257- gallivm: LLVM opaque pointers small changes 4258- gallium: refactor a channel loop in draw_llvm.c 4259- Revert "gallivm: use LLVM opaque pointers in lp_bld_tgsi_soa.c" 4260- gallium: LLVM-15 contexts use non-opaque pointers 4261- gallivm: fix a few llvm non-opaque pointers 4262- gallivm: refactor a bit the cache access in view of LLVM opaque pointers 4263- gallivm: add lp_build_struct_get() variants that take the LLVM type 4264- gallivm: push LLVM version guard into assert 4265 4266Mike Blumenkrantz (634): 4267 4268- nir/lower_tex: fix rect queries with lower_rect set 4269- mesa/st: set normalized coords for RECT samplers if rects are unsupported 4270- zink: unset PIPE_CAP_TEXRECT 4271- zink: radv baseline fix 4272- zink: nv ci update 4273- zink: clamp max shader images to 32 4274- zink: remove tcs patch slot map 4275- zink: disable EXT_extended_dynamic_state2 on AMDPRO 4276- zink: add baseline for amdpro 4277- zink: reject resource creation if format features don't match attachment 4278- zink: use mixed zs renderpass for depth read/write 4279- zink: unconditionally set line width on rasterizer state change 4280- zink: support restart with PIPE_PRIM_LINES_ADJACENCY if ext is available 4281- zink: fix extended restart prim types without dynamic state2 4282- zink: do not create fences at all if timeline semaphores are supported 4283- zink: require timeline semaphores 4284- zink: remove batch lock 4285- zink: rename zink_query::batch_id 4286- zink: remove driver-based max_fences throttling 4287- zink: hook up VK_KHR_create_renderpass2 4288- zink: only use VK_DEPENDENCY_BY_REGION_BIT if sync2 is available 4289- zink: require renderpass2 4290- zink: require KHR_imageless_framebuffer 4291- zink: delete all non-imageless framebuffer code 4292- zink: remove framebuffer indirection 4293- zink: rename imageless framebuffer functions 4294- zink: remove the worst part of basic framebuffer support 4295- zink: remove ZINK_NO_TIMELINES 4296- zink: make a kopper debug print into an error 4297- zink: rework zink_kopper_update() assert 4298- zink: always bind gfx pipeline at the top of draw 4299- zink: store min required stride values on the vertex state 4300- zink: selectively disable dynamic vertex stride 4301- nir/fold_16bit_sampler_conversions: add a mask for supported sampler dims 4302- lavapipe: run nir_fold_16bit_sampler_conversions 4303- zink: set VK_QUERY_RESULT_WAIT_BIT when copying to qbo 4304- zink: NV_linear_color_attachment 4305- zink: nv ci update 4306- zink: fix synchronization when drawing from streamout 4307- zink: fix xfb counter buffer barriers 4308- zink: remove xfb_barrier flag 4309- zink: handle device-local unsynchronized maps 4310- util/draw: fix map size of indirect buffer in util_draw_indirect_read 4311- util/draw: handle draw_count=0 when reading indirect parameters 4312- util/draw: fix indirect draw count readback 4313- gallium: add PIPE_CAP_MULTI_DRAW_INDIRECT_PARTIAL_STRIDE 4314- zink: unset PIPE_CAP_MULTI_DRAW_INDIRECT_PARTIAL_STRIDE 4315- zink: move the kopper present fence to the displaytarget object 4316- wgl: support GL 4.6 4317- zink: make unsupported blit formats an error 4318- zink: make update_framebuffer_state() public 4319- zink: rework clear_depth_stencil hook 4320- zink: add a stencil blit fallback path 4321- zink: fix tcs control barriers for use without vk memory model 4322- zink: fix semantics mask for compute control barriers 4323- zink: hook up VK_KHR_vulkan_memory_model 4324- zink: use vulkan memory model shader semantics for tcs barriers 4325- zink: always use vk mem model for shaders if available 4326- zink: fix memory_barrier intrinsic 4327- zink: fix stride=0 for dynamic state vertex stride checking 4328- zink: add synchronization for buffer clears 4329- mesa/st: clamp GL_RENDERBUFFER to GL_TEXTURE_2D for sparse queries 4330- aux/trace: make get_sparse_texture_virtual_page_size useful 4331- glsl/nir: set new_style_shadow for sparse tex ops as necessary 4332- zink: fix group memory barrier emission 4333- vulkan: bump layer api versions to current vk header version 4334- vk/cache: unbreak vk_pipeline_cache_create with flags 4335- kopper: add DISPLAY_TARGET bind for depth buffer 4336- zink: fix/improve swapchain surface info updating 4337- zink: fix up swapchain depth buffer geometry during fb update 4338- drisw: remove dead code 4339- kopper: always fetch and store drawable info 4340- kopper: move drawable geometry updating up in function 4341- kopper: store whether screen has dmabuf support 4342- kopper: copy a bunch of code for texture_from_pixmap 4343- zink: add supported present modes to kopper displaytarget 4344- zink: add a util function for finding swapchain resource 4345- zink: flag swapchains when updating fails 4346- zink: handle dead swapchains in acquire 4347- zink: unset image layout when killing swapchain 4348- zink: rework swapchain object check for acquires 4349- zink: fix return for zink_kopper_acquire 4350- zink: handle swapchain acquire failures more directly 4351- zink: flag swapchain resources as swapchains 4352- zink/kopper: add a mechanism for checking swapchain status 4353- zink: radv ci updates 4354- ac: remove gallivm header include 4355- kopper: add some error logging for pixmap->texture failure 4356- kopper: check for modifiers to use modifier functionality 4357- zink/kopper: add a function for determining if running on software 4358- kopper: check whether zink is using sw 4359- kopper: rename a confusing variable 4360- kopper: implement texture-from-pixmap for software (non-dmabuf) 4361- zink: fix dmabuf plane returns 4362- zink: export fd info for all 2d images 4363- zink: ignore KMS handle types 4364- zink: remove drm_fd 4365- llvmpipe: add a ci flake 4366- llvmpipe: disable statistic queries during draw if queries are disabled 4367- llvmpipe: remove misleading query comment 4368- llvmpipe: disable compute statistics queries if queries are disabled 4369- wgl: always set alpha on kopper windows 4370- zink: hook up VK_EXT_primitives_generated_query 4371- zink: pass query object to get_num_queries() 4372- zink: pass query object to get_num_query_pools() 4373- zink: pass query object to get_num_results() 4374- zink: pass screen param to convert_query_type() 4375- zink: add and use a function to detected emulated primgen queries 4376- zink: add a flag to zink_query to trigger rasterizer discard workaround 4377- zink: use VK_EXT_primitives_generated_query when available 4378- wgl: don't auto-load zink before software drivers 4379- lavapipe: lower quad_broadcast intrinsics 4380- zink: add a ci flake 4381- kopper: invalidate drawables when resizing textures in place 4382- kopper: pass the current context to dri_flush 4383- util/blitter: fix sampler restore with 0 saved samplers 4384- radv: emit fewer framebuffer registers 4385- zink: restore conditional ordering for query begin/end 4386- nir/gather_info: flag fbfetch on subpass image loads 4387- llvmpipe: never infer early zs tests when fbfetch is active 4388- zink: split renderpasses for TextureBarrierNV() usage 4389- zink: scale depth bias by factor of 2 4390- zink: pass sparse backing page offset to binding function 4391- zink: pass sparse bind bo offset through for texture binds 4392- zink: set all usage flags when querying sparse features 4393- zink: fix multisample conditional in sparse image query 4394- zink: bump number of image binds that can be batched to 50 4395- zink: semi-handle 1D sparse texture rewrites for drivers that don't support them 4396- zink: decouple descriptor templates from layouts 4397- zink: flatten push descriptor template into normal template array 4398- zink: implement indirect buffer indexing 4399- zink: delete nir_lower_dynamic_bo_access 4400- lavapipe: enqueue pipeline destruction 4401- zink: fix null buffer/surface formats 4402- zink: add more image usage for null surfaces 4403- zink: use descriptor surfaces for notemplates ref updating 4404- lavapipe: handle null samplerviews 4405- gallivm: fix ssbo oob reads 4406- lavapipe: fix CmdPushDescriptorSetWithTemplateKHR with refcounting 4407- st/draw_feedback: set constant buffer stride 4408- gallivm/draw: fix oob ubo reads 4409- llvmpipe: always set ssbo data pointers for draw 4410- gallivm: fix oob txf swizzling 4411- lavapipe: set robustBufferAccess2 and robustImageAccess2 4412- lavapipe: ci updates 4413- docs: EXT_robustness2 for lavapipe 4414- zink: ci updates 4415- kopper: remove unused zink include 4416- zink: remove unused headers/struct members/linkage 4417- zink: outdent some code 4418- zink: add a flag to indicate whether a resource is exportable 4419- zink: break out resource bind-adding into separate function for reuse 4420- zink: introduce copy contexts 4421- zink: create a copy context for the screen on init 4422- Revert "zink: export fd info for all 2d images" 4423- zink: always check for fb rebinds when starting renderpass 4424- zink: unset pipe_resource::next pointers when creating resource copies 4425- zink: add a bind flag to indicate a resource is exported as a dmabuf 4426- zink: fix conditional for modifier usage 4427- zink: force memory exports if dmabuf bind is specified 4428- zink: add a LINEAR drm modifier if rebinding to add dmabuf export 4429- zink: rebind resources for export as needed 4430- zink: create images with modifiers any time there is an import handle 4431- zink: fix up sparse texture sampling for shadow samplers 4432- zink: clamp out min_lod operands for explicit lod ops 4433- zink: fix sparse binding for arrayed textures 4434- zink: set sparse flag in cubemap lowering 4435- zink: remove misleading sparse comment 4436- zink: fix sparse texture depth calcs for arrayed textures 4437- zink: remove implicit query resets 4438- zink: remove refs from shader images 4439- zink: reuse local res pointer in set_shader_images 4440- zink: simplify dumb update flagging in set_shader_images 4441- zink: read shader image r/w usage from incoming data struct 4442- zink: copy incoming shader image struct after doing updates 4443- zink: stop leaking shader image surfaces 4444- zink: fix 3d shader image miplevel calc for depth 4445- zink: directly reuse surface ivci when rebinding 4446- gallium/noop: implement pipe_screen::create_fence_win32 4447- aux/trace: screen::create_fence_win32 4448- zink: add win32 guards for device extensions with win32 in name 4449- lavapipe: ignore depth/stencil states if attachment is not present 4450- lavapipe: fix pipeline handling for dynamic render zs state with pipeline library 4451- spirv: fix barrier scope assert 4452- zink: manually validate shaders in debug builds 4453- zink: fix init with MESA_SHADER_CACHE_DISABLE enabled 4454- zink: use copy context to eliminate dependency on EXT_calibrated_timestamps 4455- zink: move get_timestamp to screen 4456- zink: add tu baseline for a630 4457- zink: add an anv-tgl ci job 4458- zink: never create a sw context unless sw is requested 4459- zink: only reject non-blittable resource creation if it has no other binds 4460- zink: add separate error message for push descriptor set alloc fail 4461- zink: add extra validation for resource creation 4462- zink: remove input attachment usage when pruning surface usage 4463- zink: add extended usage and retry if resource creation fails 4464- zink: fix surface/bufferview cache comparisons 4465- zink: force render target usage for sampler-only resources 4466- zink: clamp 1D_ARRAY sparse textures to 2D as needed 4467- zink: handle PIPE_BUFFER sparse texture queries 4468- zink: fix non-dynamic vertex stride update flagging 4469- zink: add c++ guards for zink_screen.h 4470- zink: add ZINK_DEBUG=sync 4471- zink: update radv baseline 4472- zink: rename a fb surface variables 4473- zink: always use zink_batch_rp to init renderpass 4474- zink: use zink_batch_no_rp to end renderpasses 4475- zink: move renderpass handling to zink_render_pass.c 4476- zink: use global image rebind counter for dmabuf export 4477- zink: remove fb rebind check during renderpass begin 4478- zink: use local variable for getting rt barrier info 4479- zink: break out sample location updating to separate function 4480- zink: break up swapchain renderpass update to util function 4481- zink: remove renderpass state flag for swapchain 4482- zink: pass rt attrib to zink_render_pass_attachment_get_barrier_info() 4483- zink: fix indentation 4484- zink: break out color rt attrib analysis into util function 4485- zink: break out zs rt attrib analysis into util function 4486- zink: track whether there's any transient attachments bound 4487- zink: lift some renderpass mechanics up a level in the api 4488- zink: split out scissor flagging from framebuffer updating 4489- zink: use dynamic rendering (most of the time) 4490- turnip: fix assert for max xfb outputs 4491- zink: only warn once for SRGB fb without KHR_swapchain_mutable_format 4492- zink: clamp dynamic render layerCount to 1 4493- zink: create all descriptor ref templates at once 4494- zink: add screen-based indexing to descriptor sets 4495- zink: use descriptor indices in compiler 4496- zink: use descriptor set index indirection in program init 4497- zink: use descriptor set index indirection in cache state updates 4498- zink: flag descriptor sets as changed before updating hashes 4499- zink: add a "compact" descriptor mode 4500- zink: use XOR for descriptor hash accumulation 4501- zink: radv ci updates 4502- zink: use a zink_render_pass_pipeline_state bit for fbfetch 4503- zink: add renderpass bits for color/depth r/w 4504- zink: update radv piglit fails 4505- vtn: assert that vector shuffle indices are in-bounds 4506- vtn: assert that composite members have the same bit size as the result 4507- zink: require draw params ext/cap in ntv if drawid is used 4508- zink: emit spirv cap for subgroup vote ops 4509- zink: use correct scope for subgroup vote ops 4510- zink: fix subgroup vote detection 4511- zink: lower subgroup width to 1 for unsupported subgroup vote stages 4512- zink: remove first_frame stalling 4513- zink: drop wideLines requirement 4514- zink: drop largePoints requirement 4515- st/pbo_compute: do pbo readback directly to the buffer object if it exists 4516- st/pbo_compute: use different calc for non-3d compute buffer sizing 4517- st/pbo_compute: make compute download conditional in shader slightly more readable 4518- st/pbo_compute: fix z coords for compute pbos 4519- zink: only infer modifiers if winsys handle has a stride 4520- zink: drop requirement for 6 descriptorsets to use bindless 4521- zink: store the real (non-compact) binding usage for programs 4522- zink: fix change flagging for compact descriptor cache 4523- zink: flag all cached descriptors as needing update on program change 4524- zink: add a #define for descriptor compaction 4525- zink: make swapchain kill message more descriptive 4526- st/texture: use base teximage for compute blits 4527- st/texture: allow compute blits for (some) non-finalized textures 4528- lavapipe: fix dynamic patchControlPoints 4529- mesa: throw a log warning any time a fallback texture is used 4530- zink: fix radv piglit fails 4531- zink: force queries to start/end out of renderpass 4532- zink: tu ci updates 4533- zink: reorder pipeline dynamic state1 substruct 4534- zink: improve packing for pipeline dynamic state1 info 4535- zink: break out dynamic state1 pipeline info to struct 4536- zink: use memcmp for comparing dynamic state1 info 4537- zink: use dynamic cull mode 4538- zink: break out pipeline dynamic state2 into struct 4539- zink: use dynamic patch vertices if available 4540- zink: expand patch_vertices pipeline key bitsize 4541- zink: fix framebuffer attachment usage asserts for dummy attachments 4542- aux/trace: fix sampler view dumping 4543- aux/trace: implement pipe_screen::is_compute_copy_faster 4544- zink: disable compute pbos on turnip 4545- egl/wayland: manually swap backbuffer when using zink 4546- egl/wayland: skip buffer creation on zink 4547- kopper: use get_drawable_info path for non-x11 drawables 4548- ci: disable unit tests 4549- zink: fix dmabuf plane layout struct scoping 4550- zink: represent plane offsets using offset from plane 0 vs size of plane 4551- zink: rename a variable 4552- zink: handle aux plane imports 4553- Revert "zink: remove drm_fd" 4554- zink: add back kms handling 4555- mesa: improve relocation problem message 4556- mesa/st: bump param reservation to 20 4557- format_utils: properly parenthesize macro params 4558- d3d12: skip time-elapsed piglit tests in ci 4559- virgl: add some ci flakes 4560- mesa/st: bump param reservation to 28 4561- mesa: add statevar for atomic counter offsets 4562- mesa: conditionally set constants dirty for atomic counter binds 4563- st/glsl_to_nir: call st_set_prog_affected_state_flags() as late as possible 4564- mesa: handle atomic counter lowering for drivers with big ssbo offset aligns 4565- zink: remove ANV depth clip control workaround 4566- zink: disable EXT_primitives_generated_query on turnip 4567- zink: track invalidation for image resources 4568- zink: flag renderpass for change if image resource changes valid state 4569- zink: rename renderpass attrib value 4570- zink: split out a dynamic render ternary 4571- zink: hook up surface invalidation to LOAD_OP_DONT_CARE 4572- zink: invalidate blit dsts if fully covered 4573- zink: remove buffer valid range tracking from blit 4574- mesa/streaming-memcpy: add memcpy fallback 4575- streaming-memcpy: move to src/util/ and compile unconditionally 4576- mesa/st: use util_streaming_load_memcpy for compute pbo copying 4577- zink: break out of zs mixed layout update loop when work is done 4578- zink: only update layout when doing mixed zs attachment renderpass check 4579- aux/trace: add pipe_context::link_shader 4580- aux/tc: handle link_shader 4581- zink: simplify some compiler bo code 4582- zink: fix buffer descriptor hashing 4583- zink: always create descriptor templates 4584- zink: flag optimization pass after inlining uniforms 4585- zink: break out nir alu typing in ntv 4586- zink: run nir_lower_alu_to_scalar in optimizer loop for 64bit lowering 4587- zink: always scalarize pack/unpack alu ops 4588- zink: don't lower pack/unpack intrinsics 4589- zink: run more int64 passes during optimization loop if int64 isn't available 4590- zink: fix ntv partial stores 4591- zink: flag all shader i/o variables as !is_xfb 4592- zink: fix variable locations in manual xfb emission 4593- zink: if a variable is flagged as being xfb, treat it as manual xfb output 4594- zink: improve/expand manual xfb emission 4595- zink: implement nir_op_unpack_64_2x32 in ntv 4596- zink: rewrite 64bit shader i/o as 32bit 4597- zink: scalarize when rewriting explicit 64bit io 4598- zink: avoid adding more 64bit alus during bo rewriting 4599- zink: fix 32bit bo rewriting 4600- zink: update tu a630 baseline 4601- zink: add new flakes 4602- zink: fix feature requirement 4603- zink: add feature req for border colors 4604- zink: more requirement docs 4605- zink: rewrite atomic ssbo intrinsics as atomic derefs 4606- zink: run copy_prop_vars during optimization 4607- zink: warn on missing customBorderColorWithoutFormat 4608- zink: more lavapipe glcts fails 4609- zink: cap driver inlining using ssa allocation limit 4610- zink: use separate pointer for swapchain presents 4611- zink: keep a count of async presents occuring for a given swapchain 4612- zink: defer old swapchain destruction 4613- zink: add implicit sync workaround for non-mesa drivers 4614- zink: unset res->swapchain upon killing a swapchain 4615- zink: fix up KILL to a more sensible log message 4616- lavapipe: VK_EXT_border_color_swizzle 4617- anv: VK_EXT_border_color_swizzle 4618- zink: fix dynamic stride conditional in pipeline creation 4619- zink: fix generated tcs update path 4620- zink: use the patch_vertices value from the shader key, not the context 4621- zink: unset generated tcs pointer from tes on deletion 4622- zink: fix generated tcs deletion 4623- zink: only add necessary binds during dmabuf export 4624- zink: avoid uninit values in renderpass state 4625- zink: don't print VK_EXT_shader_atomic_float warning for nir_intrinsic_image_deref_atomic_exchange 4626- zink: hook up VK_EXT_non_seamless_cube_map 4627- zink: use VK_EXT_non_seamless_cube_map when available 4628- zink: rename nonseamless struct members to emulate_nonseamless 4629- zink: skip some emulated cube code if the extension is present 4630- zink: fix cubemap lowering for array samplers 4631- zink: fix cubemap lowering bit size 4632- zink: clamp dynamic renderArea to framebuffer size 4633- zink: check for pending clears to determine write status of zs attachments 4634- zink: add a function for getting the minimum framebuffer layers 4635- zink: clamp dynamic fb layers better 4636- zink: clamp renderpass layers better 4637- zink: anv ci updates 4638- zink: use uint64_t for batch_id 4639- zink: remove thread_queue 4640- zink: use a dynarray for semaphore wait flags 4641- zink: add batch api for adding a wait semaphore 4642- zink: add semaphore handling for sparse binds 4643- zink: detect sparse-only binding queue 4644- zink: always bind on the sparse queue 4645- zink: only enable vtn debugging if validation is active 4646- zink: tu ci baseline update 4647- zink: radv baseline updates 4648- drisw: probe kms if fd is passed 4649- kmsro: add error message on drm ioctl failure 4650- zink: rename a variable 4651- zink: unify actual map calls in buffer mapping 4652- zink: use only the extents for buffer unmap flushes 4653- zink: fix buffer transfer flushes with offsets 4654- zink: store and reuse memory heap flags for buffer placement 4655- zink: move BAR allocation demotion up the stack 4656- zink: be a little selective about BAR fallback memory type 4657- zink: hook up VK_EXT_border_color_swizzle 4658- zink: use VK_EXT_border_color_swizzle 4659- zink: print runtime warning on missing EXT_border_color_swizzle 4660- mesa/st: add PIPE_CAP_DITHERING 4661- zink: unset PIPE_CAP_DITHERING 4662- mesa/st: increment num_uniforms for hw accel select shaders 4663- zink: move draw-time barrier generation down a little 4664- zink: track vertex buffer bind counts on resources 4665- zink: rework buffer barrier generation 4666- zink: track image binds 4667- zink: add a #define for vk shader bits 4668- zink: collect gfx stages for all bindings during barrier generation 4669- zink: don't short-circuit gfx stage finding during barrier generation 4670- zink: relax zink_resource_buffer_needs_barrier checks 4671- zink: ci updates 4672- glx/drisw: store the flush extension to the screen 4673- glx/drisw: invalidate drawables upon binding context if flush extension exists 4674- zink: fix dual_src_blend driconf workaround 4675- nir/lower_point_size: apply point size clamping 4676- mesa: break out PointSizeIsOne setting to util function 4677- mesa: rename PointSizeIsOne -> PointSizeIsSet 4678- mesa: skip pointsize exports if pointsize attenuation is enabled 4679- mesa: enforce pointsize exports if pointsize is being clamped 4680- mesa: explicitly disallow multiple pointsize exports from generating 4681- zink: ci updates 4682- zink: add flag to indicate if a resource is a dmabuf 4683- zink: flag dmabufs for foreign queue transition on flush_resource call 4684- lavapipe: zero out blend info if blend isn't enabled 4685- lavapipe: copy more pNexts for pipeline creation 4686- lavapipe: fix renderpass info handling during pipeline creation 4687- lavapipe: add a pipeline library assert 4688- lavapipe: skip post-copy pNext checking during pipeline creation for composites 4689- zink: use the bigger of the variable type and interface type for bo sizing 4690- zink: track ssbo bind counts 4691- zink: track overall samplerview bind counts 4692- zink: fix image bind counting 4693- zink: track gfx/compute descriptor barrier info 4694- zink: use tracked barrier info for generated barriers 4695- zink: remove another zink/tu fail 4696- zink: fix-ish depth clipping without VK_EXT_depth_clip_enable 4697- zink: add a turnip driver workaround for EXT_depth_clip_enable 4698- zink: remove swizzle from fbfetch lowering 4699- lavapipe: always set point_tri_clip 4700- zink: fix kopper_acquire return value 4701- zink: track whether current swapchain has data 4702- zink: move to AoS for storing swapchain struct data 4703- zink: move 'acquired' flag onto swapchain struct 4704- zink: use better determination for handling swapchain acquire submits 4705- zink: remove broken kopper assert 4706- zink: remove unnecessary code in kopper readback acquire 4707- zink: improve no-op acquire detection 4708- zink: flag batch as having work during frontbuffer flush 4709- zink: ensure pending present flushes are handled during frontbuffer flush 4710- zink: delete zink_resource_object::acquire 4711- zink: no-op pixmap frontbuffer flush calls 4712- zink: acquire swapchain images on image map 4713- zink: use kopper_displaytarget type directly 4714- zink: delete zink_resource_object::dt_has_data 4715- zink: set PIPE_RESOURCE_FLAG_DONT_MAP_DIRECTLY for device-local resources 4716- lavapipe: break out part of shader optimization into function 4717- lavapipe: streamline xfb shader setup 4718- nir/types: fix glsl_matrix_type_is_row_major() assert 4719- zink: store context flags 4720- zink: don't increment screen->num_contexts for copy context 4721- zink: disable turnip traces temporarily 4722- lavapipe: don't remove xfb outputs 4723- zink: enforce viewport depth clamping 4724- zink: size uniform_0 ubo accurately 4725- zink: further improve bo sizing 4726- zink: force unsized array in bo creation using the last block member type 4727- zink: move/refactor bo_vars struct creation in compiler 4728- zink: handle bitsizes in get_bo_vars() analysis 4729- zink: don't run nir_lower_uniforms_to_ubo repeatedly 4730- zink: pass zink_shader to optimize_nir during final optimize pass 4731- zink: add an optimizer pass to enforce bo bounds 4732- zink: use std430 packing by default 4733- zink: use a std430 interface for bo types 4734- mesa: remove incomplete texture warning 4735- mesa/st: make get_sampler_view_format() public 4736- mesa/st: add PIPE_QUIRK_TEXTURE_BORDER_COLOR_SWIZZLE_FREEDRENO 4737- zink: init driver workarounds earlier in screen creation 4738- zink: disable customBorderColorWithoutFormat on turnip 4739- zink: don't warn for missing customBorderColorWithoutFormat on turnip 4740- zink: support PIPE_QUIRK_TEXTURE_BORDER_COLOR_SWIZZLE_FREEDRENO 4741- zink: flush pending clears for fb texture barriers 4742- mesa: fix SignalSemaphoreEXT behavior 4743- lavapipe: don't overwrite entire VkFormatProperties3 struct 4744- lavapipe: don't crash on null xfb buffer pointer 4745- zink: reject swizzled format blits 4746- lavapipe: support inlined shader spirv for compute 4747- zink: allow multiple tex components for depth tg4 4748- zink: always use 32bit sample ops 4749- zink: add a compiler pass to match up tex op dest types 4750- zink: use sampler_bind_count to simplify some code 4751- zink: break out samplerview layout reset code 4752- zink: always update sampler descriptor layouts on fb surface unbind 4753- zink: always use 32bit floats for so output types 4754- zink: handle bare matrix types in xfb emission 4755- zink: split xfb block emission from array/matrix handling 4756- zink: fix xfb array inlining 4757- zink: use right glsl length getter for ntv partial stores 4758- zink: bitcast extracted streamout components to uint before creating uvec 4759- zink: add a compiler pass to split xfb block outputs 4760- zink: fix xfb emit check in compiler 4761- mesa: move is_wrap_gl_clamp() to samplerobj.h and deduplicate 4762- mesa: track which sampler wrap params use GL_CLAMP 4763- mesa: update GL_CLAMP emulation when binding/unbinding textures 4764- zink: remove bogus range tracking from texture clear 4765- zink: fix transient attachment rp assert 4766- zink: remove format check from clear texture 4767- zink: fix coverage check for texture clears 4768- zink: stop using u_blitter for texture clears 4769- zink: improve zink_clear_depth_stencil check for current attachment 4770- zink: massively simplify zink_clear_depth_stencil 4771- zink: remove non-renderpass clear path from zink_clear_texture 4772- zink: remove u_blitter usage from zink_clear_render_target 4773- zink: always use storeOp=STORE for depth renderpass 4774- zink: track a bitmask of fb attachments with mismatched layer counts 4775- zink: add explicit (awful) handling for fb layer mismatch clears 4776- zink: clamp color clear values based on format 4777- zink: lift clearing on fb state change up a level 4778- zink: remove out-of-renderpass clears 4779- zink: delete zink_fb_clear_util_unpack_clear_color 4780- zink: delete srgb tracking for clears 4781- zink: simplify zink_framebuffer_clear_data union 4782- zink: inject a 0,0,0,1 clear for RGBX formats 4783- zink: break out tc/trace context unwrapping 4784- zink: nv baseline update 4785- zink: tu a630 baseline update 4786- mesa/st: add implicit zeroing of clipdistance array 4787- zink: move descriptor_mode to be a global 4788- zink: delete ZINK_DESCRIPTOR_MODE_NOFALLBACK 4789- zink: add an explicit ZINK_DESCRIPTOR_MODE_CACHED 4790- zink: default to lazy/templated descriptors 4791- zink: handle max_vertices=0 in geometry shader 4792- zink: handle null samplerview in get_imageview_for_binding() 4793- zink: fix crashing on bo allocation failure 4794- zink: improve oom error message 4795- zink: call bind_last_vertex_stage() when binding vs 4796- zink: fix viewport count tracking 4797- zink: use gfx_barrier for synchronization when binding buffer descriptors 4798- vulkan/wsi: fix multiple acquires for sw without mit-shm 4799- llvmpipe: ignore multisample state for alpha_to_one 4800- zink: don't try to start a renderpass before CmdClearAttachments 4801- zink: don't flag renderpass change on depth write change as often 4802- zink: flag renderpass change if renderpass clears change 4803- zink: do dynamic render implicit clear updates on rp_changed 4804- zink: unflag ctx->rp_changed after dynamic render update 4805- zink: remove redundant renderpass update check for clears 4806- zink: add an assert to verify that deferred clears aren't added in renderpass 4807- zink: split out rp_changed to be more granular for dynamic render 4808- zink: break out renderpass attachment load op calc 4809- zink: break out zs renderpass layout calc 4810- zink: break out color rt layout calc 4811- zink: reuse rt layout functions to refactor zink_render_pass_attachment_get_barrier_info() 4812- zink: add a bits union for zink_render_pass_state comparisons 4813- zink: improve granularity of renderpass switching 4814- zink: allow no-op renderpass updates in zink_batch_rp() 4815- zink: move layout-setting for dynamic render attachments after no-op case 4816- zink: handle null samplerview/image rebinds more gracefully 4817- zink: scale number of shader cache-get threads 4818- zink: invoke descriptor_program_deinit for programs on context destroy 4819- vulkan/wsi: return VK_SUBOPTIMAL_KHR for sw/x11 on window resize 4820- zink: add env var to abort on device-lost if no reset callback is set 4821- util/vbuf: handle multidraws 4822- anv: disable dev.i915.perf_stream_paranoid=0 warning 4823- lavapipe: propagate xfb info for pipeline library rasterization stages 4824- zink: remove deqp fails for lavapipe 4825- zink: only flag fbfetch as rp update if fbfetch wasn't already configured 4826- zink: assert !rp_changed after renderpass has started 4827- zink: avoid crashing when bindless texture updates aren't used 4828- lavapipe: EXT_non_seamless_cube_map 4829- lavapipe: accurately set image/ssbo access based on shader usage 4830- lavapipe: implement extreme uniform inlining 4831- lavapipe: remove some skips 4832- zink: add a ci flake 4833- lavapipe: move vertex input dynamic states into vertex input block 4834- lavapipe: make binding/layout helper functions static inline 4835- lavapipe: use correct offset calculation for flagging ssbo/image write access 4836- lavapipe: use 64bit masks for shader access 4837- lavapipe: set image access on pipeline bind, not descriptor bind 4838- llvmpipe: remove weird clamping for format support with sample counts 4839- llvmpipe: allow resource_copy_region to perform replicate operations 4840- lavapipe: fix stencil clears 4841- llvmpipe: add a shortcut for blitter resolves 4842- vulkan: hook up VK_EXT_multisampled_render_to_single_sampled 4843- lavapipe: VK_EXT_multisampled_render_to_single_sampled 4844- lavapipe: advertise VK_EXT_multisampled_render_to_single_sampled 4845- util/blitter: respect PIPE_TEXTURE_RECT 4846- lavapipe: fix shader access stage conversion from mesa -> pipe 4847- zink: stop ORing unordered barrier access 4848- zink: convert ZINK_DEBUG to enum 4849- zink: add doc for ZINK_DEBUG=compact 4850- zink: add ZINK_DEBUG=noreorder 4851- zink: zink_resource::unordered_barrier -> unordered_exec 4852- zink: explicitly unset unordered_exec when resource can't be unordered 4853- zink: rework unordered_exec handling 4854- zink: add a second resource param to get_cmdbuf() for unordered promotion 4855- zink: promote image/buffer copy operations to unordered 4856- zink: move buffer/image functions down in file 4857- zink: expand unordered_exec 4858- zink: make get_cmdbuf() public 4859- zink: enable unordered blit/resolve ops 4860- virgl: more ci flakes 4861- d3dadapter: fix sw pipe loading 4862- nine: set view_mask=0 4863- zink: fix gfx program cache pruning with generated tcs 4864- zink: don't flag lazy push constant set dirty on batch change 4865- zink: combine loops for lazy descriptor program deinit 4866- zink: handle mutable swapchain images with dmabuf 4867- zink: store VkFormatFeatureFlags on creation 4868- zink: use modifier feature flags during surface creation when necessary 4869- zink: init cache_put program fence on program creation 4870- zink: force flush clears on fb change if fb geometry changes 4871- zink: force a new framebuffer for clear_depth_stencil if the clear region is big 4872- zink: drop mode_changed check from linewidth/depthbias draw updates 4873- zink: handle unscaled depth bias from nine 4874- zink: handle !half_pixel_center 4875- zink: fix return for PIPE_CAP_DEPTH_CLIP_DISABLE 4876- nine: check return on resource_get_handle 4877- zink: demote dmabuf tiling to linear if modifiers aren't supported 4878- zink: use modifier_aspect to check for modifier plane in zink_resource_get_param 4879- zink: don't add modifiers if EXT_image_drm_format_modifier isn't present 4880- zink: add all format modifiers when adding for dmabuf export 4881- zink: don't fixup sparse texops 4882- vulkan: Update the XML and headers to 1.3.224 4883- radv: fix return type for meta resolve shaders 4884- nir/validate: clamp unsized tex dests to 32bit 4885- mesa: fix blending when using luminance/intensity emulation 4886- mesa: require render target bind for A/L/I in format selection 4887- vk/render_pass: don't deref null resolve attachments 4888- zink: handle nir_intrinsic_sparse_residency_code_and mechanics 4889- zink: ignore nir_texop_lod for tex dest matching 4890- tu: fix invalid free on alloc failure 4891- zink: don't call util_queue_fence_init in zink_screen_get_pipeline_cache() 4892- zink: don't emit entrypoints for function temp variables 4893- zink: check the variable mode before taking samplemask path in ntv 4894- zink: explicitly use unsigned types for bit shifts 4895- zink: only add srgb mutable for images with modifiers 4896- zink: don't emit illegal interpolation 4897- llvmpipe: don't assume pipe_context is always available in flush_frontbuffer 4898- zink: flag all assigned output slots as mapped 4899- zink: handle split acquire/present 4900- st_pbo/compute: fix 1D_ARRAY offsets 4901- st_pbo/compute: fix 1D coord dimension by pre-trimming vectors 4902 4903Mike Lothian (1): 4904 4905- util: Fix invalid usage of alignas in u_cpu_detect.c 4906 4907Mykhailo Skorokhodov (3): 4908 4909- anv: Use sampleLocationsEnable for sample locations 4910- iris: Move Wa_1806527549 and enable by default 4911- anv: Move Wa_1806527549 and enable by default 4912 4913Nagappa Koppad, Basanagouda (1): 4914 4915- iris:Duplicate DRM fd internally instead of reuse. 4916 4917Nanley Chery (11): 4918 4919- iris: Return a 64B stride for clear color plane 4920- intel/isl: Add a score for I915_FORMAT_MOD_4_TILED 4921- isl,iris: Add DG2 CCS modifier support for XeHP 4922- mesa: Simplify UNORM8 conversions for sRGB ASTC textures 4923- mesa: Follow ASTC decode mode extension for RGBA8 output 4924- iris: Don't leak surface states for compressed resources 4925- iris: Don't leak compressed resources in iris_create_surface 4926- iris: Use fill_surface_states for compressed resources 4927- anv: Make the D16 reg mode single-sampled 4928- iris: Make the D16 reg mode single-sampled 4929- iris: Dedent enum iris_depth_reg_mode 4930 4931Nicolas Caramelli (1): 4932 4933- egl: Fix EGL_EXT_platform_xcb name string to match the registry 4934 4935Panagiotis Apostolou (1): 4936 4937- util: Don't block SIGSEGV for new threads 4938 4939Patrick Lerda (1): 4940 4941- panfrost: Fix unwanted valgrind message related to restart_index 4942 4943Paulo Zanoni (4): 4944 4945- iris: fix race condition during busy tracking 4946- intel/compiler: compute int64_options based on devinfo->has_64bit_int 4947- intel/compiler: split handling of 64 bit floats and ints 4948- intel/compiler: call ordered_unit() only once at update_inst_scoreboard() 4949 4950Pavel Asyutchenko (7): 4951 4952- gallium: add PIPE_CAP_FBFETCH_ZS and expose extension 4953- glsl: add language support for GL_ARM_shader_framebuffer_fetch_depth_stencil 4954- nir_to_tgsi: Don't count ZS fbfetch vars as outputs 4955- llvmpipe: enable per-sample shading when FB fetch is used 4956- llvmpipe: simplify early/late zs tests selection 4957- llvmpipe: implement FB fetch for depth/stencil 4958- llvmpipe: enable PIPE_CAP_FBFETCH_ZS 4959 4960Pavel Ondračka (13): 4961 4962- r300: set PIPE_BIND_CONSTANT_BUFFER for const_uploader 4963- r300: skip draws instead of using a dummy vertex shader 4964- r300: fix vertex shader control flow in loops 4965- r300: deduplicate common NIR options 4966- r300: be less agresive with copy propagate in loops 4967- r300: don't try to use inline constants instead of constant swizzles 4968- r300: Update list of RV515 dEQP failures and add some flakes 4969- r300: merge simple movs with constant swizzles together 4970- r300: expose PIPE_CAP_LEGACY_MATH_RULES 4971- r300: only run merge_movs pass on R500 4972- r300: Keep rc_rename_regs() from overflowing RC_REGISTER_MAX_INDEX 4973- r300: disable PIPE_CAP_PREFER_IMM_ARRAYS_AS_CONSTBUF 4974- r300: fix variables detection for paired ALU and TEX instructions in different branches 4975 4976Pierre-Eric Pelloux-Prayer (81): 4977 4978- radeonsi: make si_pc_emit_shaders non static 4979- radeonsi: add helpers to deal with spm counters 4980- radeonsi: add spm counters setup code 4981- radeonsi/sqtt: setup spm bo 4982- radeonsi/sqtt: enable spm counters on gfx10+ 4983- radeonsi: make sure profile_peak is used before using sqtt 4984- ac/spm: setup write broadcasting correctly 4985- ac/surface: adjust gfx9.pitch[*] based on surf->blk_w 4986- radeonsi: remove commutative_blend_add option 4987- radeonsi: drop assume_no_z_fights option 4988- drirc: remove i965 entry 4989- gallium: add PIPE_CAP_ALLOW_DRAW_OUT_OF_ORDER 4990- radeonsi: enable PIPE_CAP_ALLOW_DRAW_OUT_OF_ORDER 4991- radeonsi: fix VS kill_outputs handling 4992- radeonsi: don't always strip NIR debug infos 4993- radeonsi: generate unique shader name in si_get_nir_shader 4994- amd: update gfx10_format_table.py for gfx11 4995- radeonsi/gfx11: add assert in legacy vs path 4996- radeonsi/gfx11: register changes 4997- radeonsi/gfx11: image descriptor changes 4998- radeonsi/tests: allow empty line and comments in csv files 4999- radeonsi/tests: add a heuristic to pick the baseline 5000- radeonsi/tests: reformat with black 5001- u_threaded: clear non-async debug callback correctly 5002- radeonsi: flush VGT streamout like PAL 5003- radeonsi: implement pipeline stats workaround 5004- gallium: add a union to access queries counters 5005- gallium: always return the correct pipeline stats index 5006- gallium: deduplicate target -> stats index code 5007- radeonsi: deduplicate query offsets 5008- radeonsi: store the pipeline stats index 5009- radeonsi: implement GL_GEOMETRY_SHADER_PRIMITIVES_EMITTED_ARB in shaders 5010- radeonsi/test: update sienna_cichlid results 5011- radeonsi/query: use the qbo correct size 5012- radeonsi/ngg: reuse the pipeline stats buffer when using atomics 5013- radeonsi: fix gs_invocation query with NGG 5014- radeonsi: don't use wave32 for GE on gfx10 if culling is used 5015- amd: fix ac_build_mbcnt_add in wave32 mode 5016- util/u_queue: rework UTIL_QUEUE_INIT_SCALE_THREADS to scale faster 5017- radeonsi: remove unused params from si_shader_select_with_key 5018- radeonsi: scale the number of shader compiler threads 5019- radeonsi: store shader variants in an array 5020- glx: set max values for pbuffer width / height 5021- radeonsi: fix glTexBuffer max size handling 5022- radeonsi: wait for PS idle in si_set_framebuffer_state 5023- radeonsi/tests: add a --slow option 5024- radeonsi/tests: use a smaller tests-per-group value 5025- radeonsi: don't use sel->nir in si_check_blend_dst_sampler_noop 5026- ac: export LLVM c++ ResetCommandLineParser method 5027- ac: reset LLVM command line parser 5028- ac/llvm: add use_waterfall_for_divergent_tex_samplers option 5029- radeonsi: enable use_waterfall_for_divergent_tex_samplers 5030- st/mesa: use mutex in st_get_texture_sampler_view_from_stobj 5031- gallium: update winsys_handle documentation 5032- gallium: add size attribute to winsys_handle 5033- va/surface: set the correct size in vaExportSurfaceHandle 5034- radeonsi: set size in si_texture_get_handle 5035- tradeonsi: fix preamble state producing incorrect packets 5036- radeonsi: prevent recursion in si_decompress_dcc 5037- radeonsi: add helper to use si_screen::aux_context 5038- radeonsi: use helpers to access si_screen::aux_context 5039- Revert "winsys/amdgpu: use AMDGPU_IB_FLAG_PREAMBLE for the CS preamble on gfx10+" 5040- dri: store internal_format when known 5041- mesa/st: don't guess the internal format if it's known 5042- util/blob: use memcpy in read functions 5043- mesa: skip draws with invalid indices offset 5044- vulkan/wsi: define pWaitDstStageMask in the blit submission 5045- ac: use LLVMContextSetOpaquePointers if available 5046- radeonsi: use LLVMBuildLoad2 where possible 5047- ac: add per output is_16bit flag to ac_shader_abi 5048- ac/llvm: handle opaque pointers in visit_store_output 5049- ac/llvm: use LLVMBuildLoad2 in visit_load 5050- radeonsi: use LLVMBuildLoad2 in llvm PS 5051- radeonsi: use LLVMBuildLoad2 for inter-stage outputs loads 5052- radeonsi: check last_dirty_buf_counter and dirty_tex_counter 5053- amdgpu/bo: update uses_secure_bos when importing buffers 5054- nir: add a nir_opt_if_options enum 5055- radeonsi: prevent u_blitter recursion in si_update_ps_colorbuf0_slot 5056- radeonsi: use nir_opt_large_constants earlier 5057- mesa: avoid reading back textures from VRAM 5058- radeonsi: invalidate L2 when using dcc stores 5059 5060Qiang Yu (63): 5061 5062- radeonsi: lower nir_intrinsic_sparse_residency_code_and 5063- mesa/program: fix nir output reg overflow 5064- mesa/vbo: remove unused vbo_context->binding 5065- nir/builder: add load/store array variable helper functions 5066- mesa: add hardware accelerated select constant 5067- mesa: add _mesa_bufferobj_get_subdata 5068- mesa: refine name stack code to prepare for hw select 5069- mesa: add hw select name stack code path 5070- mesa/vbo: enclose none-vertex functions with HW_SELECT_MODE 5071- mapi: add api setup header for hw select mode 5072- mesa: set CurrentServerDispatch too when glBegin/End 5073- mesa: add HWSelectModeBeginEnd dispatch table 5074- mesa: pass select result buffer offset as attribute/varying 5075- mesa/st: implement hardware accelerated GL_SELECT 5076- zink: reset PIPE_CAP_ACCELERATED when cpu soft rendering 5077- virgl: return -1 for PIPE_CAP_ACCELERATED 5078- mesa: enable HardwareAcceleratedSelect 5079- ac/llvm: get back nir_intrinsic_load_tess_rel_patch_id_amd 5080- nir: add nir_intrinsic_load_lshs_vertex_stride_amd 5081- ac/nir: use nir_intrinsic_load_lshs_vertex_stride_amd 5082- radeonsi: implement load_lshs_vertex_stride abi 5083- radeonsi: add tcs_vgpr_only_inputs parameter to si_get_nir_shader 5084- ac/nir: skip gl_Layer/gl_ViewportIndex write for LS 5085- radeonsi: replace llvm ls/hs interface lds ops with nir lowered ones 5086- gallium: disable hardware select for crocus 5087- ac/llvm: conditionally check wave id in gs sendmsg 5088- radv: no need to do gs_alloc_req for newer chips in ngg vs/tes 5089- radeonsi: deserialize nir binary in si_check_blend_dst_sampler_noop 5090- radeonsi: use si_shader as parameter in si_get_nir_shader 5091- radeonsi: support multi stage shader state creation in nir shaderlib 5092- radeonsi: add si_create_passthrough_tcs 5093- radeonsi: replace llvm based fixed tcs with nir 5094- radeonsi: implement nir_intrinsic_load_tcs_num_patches_amd 5095- ac/nir: remove unused parameter in tes input lower 5096- ac/nir: add nir_intrinsic_load_hs_out_patch_data_offset_amd 5097- ac/nir: use nir_intrinsic_load_hs_out_patch_data_offset_amd in tess lower 5098- ac/nir: add pass_tessfactors_by_reg param to hs output lower 5099- ac/nir: add wave_size parameter to ac_nir_lower_hs_outputs_to_mem 5100- radeonsi: preload tess offchip ring for tcs 5101- radeonsi: implement nir_intrinsic_load_ring_tess_offchip(_offset)_amd 5102- radeonsi: implement nir_intrinsic_load_tess_rel_patch_id_amd for both tcs and tes 5103- ac/llvm: handle write mask for nir_intrinsic_store_buffer_amd 5104- ac/nir: add no_input_lds_space param to hs output lower 5105- radeonsi: replace llvm tcs output with nir lower pass 5106- radeonsi: enable PIPE_CAP_GLSL_TESS_LEVELS_AS_INPUTS 5107- radeonsi: set uses_vmem_load_other for nir_intrinsic_load_buffer_amd 5108- radeonsi: replace llvm tes input load with nir lowering 5109- ac/llvm: fix code format alignment in visit_load_local_invocation_index 5110- radeonsi: implement two esgs ring nir intrinsic 5111- ac/nir: skip gl_ViewportIndex and gl_Layer write in ES 5112- radeonsi: set lds for gs/es to handle nir shared memory load/store 5113- ac/nir: remove unused param num_reserved_es_outputs from gs input lower 5114- ac/nir: change es output lower param to esgs_itemsize 5115- radeonsi: replace llvm es output with nir lowering 5116- ac/nir: add triangle_strip_adjacency_fix to gs input lower 5117- ac/llvm: get back nir_intrinsic_load_gs_vertex_offset_amd 5118- radeonsi: replace llvm gs input handle with nir lowering 5119- ac/nir/ngg: fix nogs culling scratch size 5120- ac/nir/ngg: add a barrier before prim id export 5121- ac/nir/ngg: Decouple primitive ID store and primitive export. 5122- nir/lower_gs_intrinsics: fix primitive count for points 5123- radeonsi: fix tcs_out_lds_offsets arg alignment 5124- winsys/amdgpu: fix non-page-aligned sparse buffer creation 5125 5126Rajnesh Kanwal (29): 5127 5128- pvr: Remove double error reporting in error path. 5129- pvr: Change phase_rate_change type to bool from uint32_t. 5130- pvr: Setup common nop shader. 5131- pvr: Add services winsys transfer context support. 5132- pvr: Add support to create transfer context and setup required shaders. 5133- pvr: Add vk_sync support and remove service winsys syncobjs interface. 5134- pvr: Fix argument comments in render job submission. 5135- pvr: Pass device info struct in pds generator. 5136- pvr: Add support for R32G32B32A32_UINT, R32_UINT and R8_UINT formats. 5137- pvr: Allow signal_sync pointer to be NULL in job submission. 5138- pvr: Remove vk_format_is_pure_integer and use common helper. 5139- pvr: Add transfer 3d parameter heap support. 5140- pvr: Fix ISP_RENDER_MODE_TYPE enum in rogue_cr.xml 5141- pvr: Restrict argument to dev_info where possible. 5142- pvr: Fix UTIL_ARCH_BIG_ENDIAN check in vk_format.h. 5143- pvr: Change ALIGN macro to ALIGN_ATTR. 5144- pvr: Add services winsys transfer cmd submit interface. 5145- pvr: Move transfer logic to pvr_job_transfer.[ch] 5146- pvr: Remove clang-format off comment from vk_format.h 5147- pvr: Implement vkCmdDraw API. 5148- pvr: Init and store default sampler in pvr_device. 5149- pvr: Use common vk_buffer implementation. 5150- pvr: Implement vkCreateBufferView and vkDestroyBufferView. 5151- pvr: Move binding related checks in common code. 5152- pvr: Add support for sampler descriptor. 5153- pvr: Add support for combined image and sampler descriptor. 5154- pvr: Add support for image descriptor. 5155- pvr: Add support for storage and uniform texel descriptor. 5156- pvr: Add support for input attachment descriptor. 5157 5158Renato Pereyra (4): 5159 5160- venus: add support for vk_xwayland_wait_ready 5161- venus: Add driconf option to enable implicit fencing 5162- Revert "wsi/x11: Don't leak xcb_get_geometry_reply_t." 5163- Revert "wsi/x11: Avoid using xcb_wait_for_special_event in FIFO modes" 5164 5165Rhys Perry (94): 5166 5167- aco: fix disassembly of SMEM with both SGPR and constant offset 5168- ac/nir: properly handle large global access constant offsets 5169- nir/gather_info: fix system_value_read for rt/mesh system values 5170- nir/builder: add nir_{ine,ibfe,ubfe}_imm() helper 5171- radv: remove unnecessary align_mul/offset initialization 5172- radv: initialize meta shader options earlier 5173- radv: use radv_meta_build_nir_vs_generate_vertices() more 5174- radv: use nir_op_imm helpers 5175- radv: use inot(a) instead of ine(a, true)/ieq(a, false) 5176- ac/nir: fix 64-bit NGG GS output stores 5177- radv: always use nir_lower_io_lower_64bit_to_32 5178- nir/copy_prop_vars: fix non-vector shader call payloads 5179- radv: fix clearing of TRUNC_COORD with tg4 and immutable samplers 5180- radv: consider radix sort shaders to be internal/meta 5181- aco/ra: fix vgpr_limit 5182- aco: ensure that definitions fixed to operands have matching regclasses 5183- radv: add radv_has_shader_buffer_float_minmax 5184- radv: allow LBVH on GFX6+ 5185- radv: remove subtractions in address calculations 5186- radv: use SMEM for a few load_global 5187- aco: split load_sbt_amd result 5188- aco: skip s_barrier if TCS patches are within subgroup 5189- ac/nir: skip s_barrier if TCS patches are within subgroup 5190- aco: fix cmpswap global atomic definition on GFX6 5191- nir: allow NIR_PASS(_, ) 5192- nir: print file when validation fails 5193- nir: call nir_metadata_preserve in nir_io_add_const_offset_to_base 5194- nir: call nir_metadata_preserve in nir_vectorize_tess_levels 5195- nir: call nir_metadata_preserve in nir_lower_memory_model 5196- radv: call nir_metadata_preserve in various lowering passes 5197- radv: add missing NIR_PASS() and switch from NIR_PASS_V() 5198- radv: validate shaders after linking passes 5199- aco: fix p_constaddr with a non-zero offset 5200- aco/tests: add test for p_constaddr with a non-zero offset 5201- aco: only add/subtract low bits of program addresses 5202- aco/ra: fix usage of invalidated iterator 5203- aco: use p_parallelcopy for uniform reduction with zero source 5204- aco: clarify a portion of do_pack_2x16 5205- aco/tests: disable regalloc.subdword_alloc.reuse_16bit_operands on GFX11 5206- aco/tests: fix gfx11 variants printed as gfx12 5207- aco: disable sdwa on gfx11 5208- aco: use v_perm_b32 to copy 0xff00/0x00ff/0xff/0x00 5209- aco/tests: update for GFX11's removal of SDWA 5210- radv,aco: disable shader cache when ACO debug options are used 5211- aco: fix SMEM load_global_amd with non-zero offset 5212- aco: fix SMEM load_global with VGPR address and non-zero offset 5213- aco: consider fma with multiplication by power-of-two unfused 5214- ac: use ResetAllOptionOccurrences instead of ResetCommandLineParser 5215- gitlab: ask for useful titles 5216- gitlab: ask that reporters don't include long logs in descriptions 5217- nir/deref: stop assuming coherent accesses of different SSBOs may alias 5218- aco: update SMEM offset workaround for LLVM 15 5219- aco: don't skip VS->TCS barrier if TCS output vertices doesn't match input 5220- aco: cleanup force-waitcnt output 5221- aco/ra: update register file when updating phi definition 5222- radv/ci: re-enable vega10 fossils testing 5223- aco: add and use constantValue16() 5224- aco: don't accept med3 opcodes in get_minmax_info() 5225- aco: include _e64 variants of 16-bit min/max in minmax optimizations 5226- aco: ignore precise flag when optimizing integer clamps 5227- aco: fix single-alignbyte do_pack_2x16() path with fp inline constants 5228- aco: fix propagate_constants_vop3p with integer vop3p and 16-bit constants 5229- aco: don't use opsel to fold constants into dot accumulation sources 5230- aco: fix redirect combine in propagate_constants_vop3p() with negatives 5231- aco: try sign-extending or shifting constants in propagate_constants_vop3p 5232- aco: don't use 32-bit fp inline constants for fp16 vop3p literals 5233- aco/tests: add vop3p constant combine tests 5234- nir: rename fsin_r600/fcos_r600 to fsin_amd/fcos_amd 5235- nir/algebraic: optimize bcsel(c, fsin/cos_amd(a), fsin/cos_amd(b)) 5236- nir: allow 16-bit fsin_amd/fcos_amd 5237- radv,aco,ac/llvm: use nir_op_f{sin,cos}_amd 5238- aco: make flat access latency match mtbuf/mubuf/mimg 5239- aco: include flat-like in vmem clause statistics 5240- aco: make FLAT_instruction::offset signed 5241- aco: improve support for scratch_* instructions 5242- aco: combine additions and constants into scratch load/store 5243- aco: handle subtractions in parse_base_offset 5244- aco: refactor VGPR spill/reload lowering 5245- aco: avoid WAW hazard with BVH MIMG and other VMEM 5246- aco: include scratch/global in VMEM WAW optimization 5247- aco: treat flat-like as vmem in some scheduling heuristics 5248- aco: initialize scratch base registers on GFX9-GFX10.3 5249- aco: use scratch_* for scratch load/store on GFX9+ 5250- aco: use scratch_* for VGPR spill/reload on GFX9+ 5251- radv: fix vbo_bound_mask indexing 5252- radv: don't update misaligned_mask in CmdBindVertexBuffers2 5253- aco: remove unnecessary exp_cnt increments 5254- aco: fix LdsBranchVmemWARHazard with 2+ branch chains 5255- aco: fix hash statistic 5256- radv: remove claimed support for sRGB vertex buffer formats 5257- radv: fix 16-bit support in radv_lower_vs_input 5258- aco: fix 16-bit VS inputs 5259- aco: don't expand vec3 VS input load to vec4 on GFX6 5260- aco: add SCC clobber in build_cube_select 5261 5262Ricardo Garcia (1): 5263 5264- vulkan: allow null descriptor set layouts in pipeline layouts 5265 5266Riteo (1): 5267 5268- vulkan/device_select_wayland: fix a memory leak with DRM device handling 5269 5270Rob Clark (64): 5271 5272- freedreno: Misc indent fixes 5273- freedreno/drm/virtio: Appease valgrind 5274- freedreno/drm: Drop FD_PP_PGTABLE 5275- freedreno/drm: Close bo handle after bo->destroy() 5276- freedreno/drm: Move bo common init 5277- freedreno/drm/virtio: Support ring_idx 5278- freedreno/drm/virtio: Use userspace IOVA allocation 5279- freedreno/drm/virtio: Drop blocking in host 5280- freedreno/drm/virtio: Fix SHAREABLE+MAPPABLE 5281- freedreno/drm/virtio: Protocol updates 5282- freedreno/drm/virtio: Pass guest handles to execbuf 5283- freedreno/drm/virtio: Async ccmd batching 5284- freedreno/drm/virtio: Split up large uploads 5285- freedreno/a6xx: A couple more FD_BO_NOMAP 5286- freedreno: Split out helper for staging blit 5287- freedreno: Use staging transfer if mmap fails 5288- freedreno/drm: Fix bos_on_stack calculation 5289- drm-shim: Add GET_UNIQUE support 5290- freedreno/drm-shim: Update to latest uapi version 5291- freedreno/drm-shim: Robustify error handling 5292- drm-shim: Better mmap offsets 5293- drm-shim: Add error handling for bo_init() 5294- freedreno/drm-shim: Better iova handling 5295- drm-shim: Cleanup on device file close 5296- freedreno/drm: Use DEBUG_GET_ONCE_OPTION() 5297- freedreno/drm/virtio: Don't try to mmap imported bo's 5298- freedreno/devices: Add another SKU 5299- Revert "ci: freedreno farm is down" 5300- freedreno: Add pkt4 assert 5301- freedreno/registers: Small cleanup 5302- freedreno/a6xx: Split VFD_FETCH[] if needed 5303- freedreno/a6xx: Initialize VFD_FETCH[n].SIZE to zero 5304- freedreno/a6xx: assert valid vertex_flags reg 5305- turnip: assert valid vertex_flag reg 5306- freedreno/ir3: Don't lower_gs multiple times 5307- freedreno/ir3: Add copy_vars() helper 5308- freedreno/ir3: Copy vars if needed on EndPrimitive() 5309- freedreno/ir3: Add more tess varying slots 5310- freedreno/a6xx: Fix indentation 5311- freedreno/a6xx: Fix VS const packet size 5312- freedreno/ir3: Fix GS clip-plane lowering 5313- freedreno: rename ir3_emit_driver_params() 5314- freedreno/a6xx: Handle driver-params in GS/DS 5315- freedreno/a6xx: Do clip-plane lowering in backend 5316- freedreno/gmem: Reverse order of alternative tile rows 5317- util: Fix c++ enum casting pickiness 5318- libsync: Fix compiler warning 5319- freedreno/a6xx: Remove fd6_format.[ch] 5320- freedreno/a6xx: Fix enum tag 5321- freedreno: Extract common helper macros 5322- freedreno/ir3: Remove unneeded forward declaration 5323- freedreno/autotune: Make 'offset' macro "private" 5324- freedreno/drm: Combine upper and lower 32b of OR val 5325- freedreno/a6xx: De-open-code CACHE_FLUSH enum 5326- freedreno: Use enum for primtypes table 5327- freedreno/ci: Update unit test reference decodes 5328- freedreno/registers: Whitespace fix for gen_header.py 5329- freedreno/registers: Allow varset to be specified on enum 5330- freedreno/registers: Move varset to <enum> 5331- freedreno/registers/a6xx: Some reg64 conversion 5332- freedreno/drm: Fix potential bo cache vs export crash 5333- freedreno/gmem: Fix col0 calc 5334- llvmpipe: Add some missing locking 5335- freedreno: We really don't need aligned vbo's 5336 5337Roman Stratiienko (3): 5338 5339- Android.mk: Fix c11-related build failures 5340- v3dv: Enable sync_fd importing/exporting on Android 5341- v3dv: Limit API version to v1.0 for Android 5342 5343Ruijing Dong (6): 5344 5345- radeonsi/vcn: Add support of array_mode for gfx11 5346- radeonsi/vcn: update av1 decoding to support vcn4 5347- radeonsi/vcn: prepare for unified queue in vcn4 5348- radeonsi/vcn: support unified queue in vcn4 5349- frontends/va: do texture_map when needed 5350- frontends/va: WA for ffmpeg 10bit encoding crash 5351 5352Ryan Neph (5): 5353 5354- venus: update venus-protocol with VK_EXT_image_view_min_lod 5355- venus: enable VK_EXT_image_view_min_lod 5356- venus: support VK_KHR_copy_commands2 5357- venus: refactor VkCommandBufferBeginInfo fixups to function 5358- venus: add support for VK_KHR_dynamic_rendering 5359 5360Sagar Ghuge (1): 5361 5362- anv: Disable storage image compression for possible atomic ops 5363 5364Sami Kyöstilä (2): 5365 5366- subprojects: Roll Perfetto to v27.1 5367- util: Shut down Perfetto before driver unload 5368 5369Samuel Pitoiset (213): 5370 5371- radv: exclude PRIMITIVE_{COUNT,INDICES} from the per-vertex output mask 5372- radv: use shader_info::writes_memory 5373- radv: use shader_info::gs::active_stream_mask 5374- radv: use shader_info::inputs_read/outputs_written for FS IO 5375- radv: use shader_info::outputs_written/per_primitive_outputs for VS outputs 5376- radv: use shader_info::system_values_read 5377- radv: fix enabling adjust_frag_coord_z and apply per-pipeline 5378- radv: add support for independent descriptor set layouts 5379- radv: add support for VkShaderModuleCreateInfo as pNext 5380- radv: fix initializing pipeline_key::topology for GFX9 and older 5381- radv: mark all active stages earlier in the pipeline creation path 5382- radv: use the hardware primitive topology everywhere 5383- radv: initialize the vertex input interface state in only one place 5384- radv: do not support UNIFORM_TEXEL_BUFFER with SRGB 5385- radv: only apply enable_mrt_output_nan_fixup for 32-bit float MRTs 5386- radv/ci: update the flakes list for GFX9 chips 5387- aco: fix load_barycentric_at_{sample,offset} on GFX6-7 5388- nir: fix marking XFB varyings as always active IO 5389- nir: mark XFB varyings as unmoveable to prevent them to be remapped 5390- radv: clarify why STAGE_2_CLEAR_BIT needs to wait for CP DMA to be idle 5391- radv: remove redundant VK_PIPELINE_STAGE_2_TRANSFER_BIT for CP DMA idle 5392- radv: add an SQTT workaround for chips with disabled RBs 5393- radv: fix handling divisor == 0 with dynamic vertex input state 5394- radv: allow to disable sinking of load inputs for FS via drirc 5395- radv: enable radv_disable_sinking_load_input_fs for Grid Autosport 5396- radv: use correct push constants range for internal operations 5397- radv/ci: update list of expected failures for Bonaire (GFX7) 5398- radv/ci: remove empty flakes lists for GFX9 5399- radv: fix the number of generated primitive queries with NGG GS vs legacy 5400- radv/ci: stop skipping dEQP-VK.synchronization.* on Bonaire 5401- radv: use 3D views for 3D internal operations on GFX6-8 5402- radv: re-emit dynamic line stipple state if the primitive topology changed 5403- radv/radix: handle intentional allocation failures properly 5404- ac,radeonsi: add has_sqtt_auto_flush_mode_bug 5405- radv/sqtt: fix configuring AUTO_FLUSH_MODE on GFX10.3 5406- radv: only init acceleration structure if RT is enabled 5407- radv: disable instance packing to fix pipeline query stats 5408- radv: remove useless check against nir_texop_samples_identical 5409- radv/winsys: allow to use RADV_FORCE_FAMILY=gfx1100 5410- radv: disable DCC for Senra Kagura Shinovi Versus 5411- aco: recognize GFX11 in few places 5412- aco: do not align VGPRS to 8 or 16 on GFX11 5413- radv,aco: add support for packed threadID VGPRs on GFX11 5414- radv,aco: use the new TCS WaveID SGPR to compute vs_rel_patch_id on GFX11 5415- aco: export MRT0 instead of NULL on GFX11 5416- aco: do not set DLC for loads on GFX11 5417- aco: do not set GLC stores on GFX11 5418- aco: update LDS allocation granularity for PS on GFX11 5419- aco: update waitcnt on GFX11 5420- aco: do not set RESOURCE_LEVEL for buffer descriptors on GFX11 5421- aco: do not set COMPR for exports but use 0x3 channel mask on GFX11 5422- radv,aco: export alpha-to-coverage via MRTZ on GFX11 5423- aco: only retrieve the scratch offset when it's declared 5424- radv,aco: do not implicitly export the primitive ID for mesh shaders 5425- aco: remove unreachable code about viewport index/layer and mesh shaders 5426- radv,aco: add a workaround for binding 2D views of a 3D image on GFX9 5427- radv: implement VK_EXT_image_2d_view_of_3d 5428- radv: advertise VK_EXT_image_2d_view_of_3d 5429- radv: do not declare the scratch offset on GFX11 5430- aco: do not emit the primitive ID twice for NGG VS or TES with GS 5431- radv: do not align VGPRS to 8 or 16 on GFX11 5432- radv: use the new format table on GFX11 5433- radv: update scratch buffer registers on GFX11 5434- radv: update sampler registers on GFX11 5435- radv: update buffer descriptor registers on GFX11 5436- radv: update image descriptor registers on GFX11 5437- radv: use PIXEL_PIPE_STATE_DUMP event instead of ZPASS_DONE on GFX11 5438- radv: configure DB_RENDER_CONTROL on GFX11 5439- radv: do not emit non-existent CP_COHER_START_DELAY on GFX11 5440- radv: do not emit FLUSH_AND_INV_DB_META on GFX11 5441- radv: update TF_RING_size to a per-SE size on GFX11 5442- radv: update VRS registers on GFX11 5443- radv: apply a workaround for CB perf counters on GFX11 5444- radv: limit CP DMA to max 32KB sizes on GFX11 5445- radv: more register changes on GFX11 5446- radv: do not lower loading TESS/ESGS rings using the ABI for LLVM 5447- aco: use ac_is_llvm_processor_supported() for checking LLVM asm support 5448- radv: rename shader compile functions to spirv_to_nir/nir_to_asm 5449- radv: do not try to dump the NIR of the trap handler shader 5450- radv: add new pipeline helpers for NIR->ASM compilation 5451- radv: init states from pTessellationState at only one place 5452- radv: init states from pViewportState at only one place 5453- radv: init states from pRasterizationState at only one place 5454- aco: remove unnecessary intrinsics that are lowered at the ABI level 5455- radv: determine if shaders use NGG before linking 5456- radv: export implicit primitive ID in NIR for legacy VS or TES 5457- radv: constify radv_pipeline in more radv_pipeline_generate_XXX() helpers 5458- radv: rename radv_pipeline_generate_XXX() to radv_pipeline_emit_XXX() 5459- radv: remove couple of useless pCreateInfo pointers in the emit path 5460- radv: remove unused radv_get_conservative_raster_mode() helper 5461- radv: move gs_table_depth to the physical_device 5462- radv: cleanup using device/physical_device during pipeline creation 5463- radv: split radv_pipeline into radv_{graphics,compute,library}pipeline 5464- radv: introduce radv_graphics_pipeline_info and split existing info structs 5465- radv/ci: fix fails list for NAVI21 5466- radv: move HS info and task_num_entries to the physical device 5467- radv: rename radv_cmd_state::pipeline to graphics_pipeline 5468- radv: unify radv_pipeline_has_XXX() helpers 5469- radv: move streamout_shader to radv_graphics_pipeline 5470- radv: move active_stages to radv_graphics_pipeline 5471- radv: rename VERT_GRP_SIZE to VERTS_PER_SUBGRP on GFX11 5472- radv: do not emit more non-existent registers on GFX11 5473- radv: update framebuffer registers on GFX11 5474- radv: update VRS rates on GFX11 5475- radv: use the fragment resolve path by default on GFX11 5476- radv: report adjusted LDS size for fragment shaders on GFX11 5477- radv: fix configuring COLOR_INVALID on GFX11 5478- radv: fix VK_BLEND_FACTOR_CONSTANT_COLOR translation on GFX11 5479- radv: configure DB_Z_INFO.NUM_SAMPLES correctly on GFX11 5480- radv: disable VK_AMD_shader_fragment_mask on GFX11 5481- radv: disable attachementFragmentShadingRate on GFX11 5482- radv: use pipeline->slab_bo in more places 5483- radv: fix writing buffer markers with non-zero memory offset 5484- radv: copy viewport/scissor when initializing radv_viewport_info 5485- radv: init states from pMultisampleState at only one place 5486- radv: init states from pDepthStencilState at only one place 5487- radv: init states from VkPipelineRenderingCreateInfo at only one place 5488- radv: use AMD values for computing blend related state 5489- radv: init states from pColorBlendState at only one place 5490- radv: init states from VkAttachmentSampleCountInfo at only one place 5491- radv: init states from VkPipelineDiscardRectangleState at only one place 5492- radv: init states from VkPipelineFragmentShadingRateState at only one place 5493- radv: ignore DYNAMIC_STENCIL_OP if stencil test isn't enabled 5494- radv: reduce radv_tessellation_info::patch_control_points to 8-bit 5495- radv: reduce radv_input_assembly_info::primitive_topology to 8-bit 5496- radv: reduce radv_rasterization_info::polygon_mode to 8-bit 5497- radv: reduce radv_discard_rectangle_info::count to 8-bit 5498- radv: fix lowering GS intrinsics if NGG is disabled per pipeline 5499- radv/ci: update list of failures for Pitcairn 5500- radv: enable radv_zero_vram for Hammerting 5501- radv/winsys: add get_chip_name() to the null winsys 5502- radv: report the marketing name as part of the device name 5503- radv: rename ngg_gs_state to ngg_query_state 5504- radv: track if primitives generated query features are enabled 5505- radv: add primitives_generated_query to the graphics pipeline key 5506- radv: declare the NGG query argument for primitives generated query 5507- ac/nir/ngg: count the number of generated primitives for VS and TES 5508- radv: flush the NGG query state when the argument is declared 5509- radv: add few helpers related to streamout 5510- radv: implement VK_EXT_primitives_generated_query 5511- radv: advertise VK_EXT_primitives_generated_query 5512- radv/ci: rename deqp-XXX jobs to vkcts-XXX 5513- zink/ci: skip two KHR-GL46 tests that timeout most of the time with RADV 5514- zink/ci: update list of expected failures with RADV 5515- radv: remove redundant check when importing vertex input info 5516- radv: update radv_is_vrs_enabled() to use radv_graphics_pipeline_info 5517- radv: update the check to determine if dynamic discard rectangle is needed 5518- radv: update the check to determine if dynamic sample location is needed 5519- radv: update the check to determine if rasterization is enabled 5520- radv: remove remaining unused pCreateInfo pointers 5521- ci: bump the hang-detection tool in the test image for Vulkan 5522- ci: fix passing down MESA_LOADER_DRIVER_OVERRIDE for zink-anv-tgl 5523- zink/ci: change the surface type used for dEQP 5524- ci/valve: split .b2c-test into .b2c-test-vk and .b2c-test-gl 5525- radv/ci: add GLES/GLCTS testing with Zink on NAVI10 5526- zink/ci: update list of failures again 5527- radv/ci: refactor test machines and dEQP rules to reduce copy&paste config 5528- ci: update vkd3d-proton builder/runner 5529- ci: print a message when vk3d-proton returns successfully 5530- ci: add VKD3D_PROTON_RESULTS to the list of variables to pass down for testing 5531- radv/ci: add vkd3d-proton testing 5532- radv: advertise VK_EXT_border_color_swizzle on GFX10+ 5533- ci: uprev vkd3d-proton to 39d07dea2cef34bfb3ed39741f026bc637e3eec4 5534- radv/amdgpu: make sure to reset the number of BO when there is no ranges 5535- radv/amdgpu: do not add "global" BO to the virtual BO list 5536- aco: fix validation of SOP1 instructions without definitions 5537- radv,aco: rename radv_prolog_binary/radv_shader_prolog to shader_part 5538- radv/llvm: always emit a null export even if the FS doesn't discard 5539- radv: disable small primitive culling for user sample locations 5540- radv: use RADEON_FLAG_VA_UNCACHED for the trace BO 5541- radv: fix command line for dumping waves with UMR 5542- radv: dump UMR waves before UMR rings 5543- radv: do not emit SQTT user data packets on the SDMA queue 5544- radv/ci: add CI lists for LLVM on NAVI21 5545- radv: use LOAD_CONTEXT_REG to load the opaque buffer size on GFX10+ 5546- radv: do not abort if SPM isn't supported for the current GPU 5547- radv: remove old workaround for HTILE layers with F1 2021 5548- zink/ci: update list of failures for piglit with RADV 5549- radv: disable DCC for Melty Blood Actress Again Current Code 5550- radv: always enable VK_EXT_debug_utils 5551- zink/ci: update list of failures 5552- radv/ci: update list of failures against CTS 1.3.3.0 5553- aco: fix load_barycentric_at_sample without MSAA 5554- radv: fix dumping VS prologs assembly 5555- radv/ci: enable fossils testing for GFX1100 5556- radv: remove the radv_report_apu_as_dgpu workaround for Red Dead Redemption 2 5557- radv: fix wide points/lines by configuring the guardband correctly 5558- zink/ci: update list of CTS flakes for RADV 5559- ci: uprev vkd3d-proton to 5b73139f182d86cd58a757e4b5f0d4cfad96d319 5560- radv,aco: track if a fragment shader needs an epilog 5561- radv,aco: introduce {radv,aco}_ps_epilog_key 5562- radv: declare a new user SGPR arg in FS for the epilog PC 5563- radv: add a function that declares PS epilog shader arguments 5564- aco: add new pseudo instruction p_jump_to_epilog 5565- aco: emit p_jump_to_epilog if the main fragment shader has an epilog 5566- aco: do not abort if the FS doesn't export anything but has an epilog 5567- aco: prevent adding DONE/VM to the last export if the FS has an epilog 5568- aco: add aco_postprocess_shader() helper 5569- radv,aco: rename radv_aco_build_prolog to radv_aco_build_shader_part 5570- aco: refactor export_fs_mrt_color() for PS epilogs preparation 5571- aco: add support for compiling PS epilogs 5572- radv: implement PS epilogs 5573- radv: disable viewport depth clamping only when necessary 5574- aco: requires Exact for p_jump_to_epilog 5575- radv: only force 1x sample for Bresenham lines when pipeline draws lines 5576- vulkan: add support for VK_IMAGE_LAYOUT_ATTACHMENT_FEEDBACK_LOOP_OPTIMAL_EXT 5577- radv: implement VK_EXT_attachment_feedback_loop_layout 5578- radv: ignore out-of-order rasterization if stencil write mask is dynamic 5579- radv: fix gathering XFB info if there is dead outputs 5580- radv: fix cleaning the meta query state if an error occured 5581- radv: re-emit viewports if negative one to one or depth clamp mode changed 5582- radv: fix bogus assertion with RADV_FORCE_VRS 5583- radv: fix pipelineStageCreationFeedbackCount when it's 0 5584 5585Sarah Walker (2): 5586 5587- pvr: Update for firmware 1.17@6256262 5588- pvr: Move BRN 44079, 48492 and 66011 code into pvrsrvkm specific directory 5589 5590Sathishkumar S (2): 5591 5592- radeon/vcn: skip create and destroy message for jpeg 5593- radeon/vcn: engage all available jpeg engines 5594 5595Sebastian Keller (1): 5596 5597- egl/wayland: Don't try to access modifiers u_vector as dynarray 5598 5599Sergi Blanch Torne (1): 5600 5601- ci/crosvm: clean crosvm socket 5602 5603Sidney Just (19): 5604 5605- wgl: add a flag to determine if running on zink 5606- wgl: add zink to the list of auto-loaded drivers 5607- zink: support VK_KHR_win32_surface 5608- kopper: add win32 loader interface 5609- zink: support win32 wsi 5610- wgl: support kopper 5611- gallium: support for implementing EXT_external_objects_win32 5612- mapi: added EXT_external_objects_win32 definitions 5613- gallium/noop: implement win32 device LUID and node mask getters 5614- gallium/trace: implement win32 device LUID and node mask getters 5615- mesa: implement GL_DEVICE_LUID_EXT and GL_DEVICE_NODE_MASK_EXT getters 5616- mesa: implement import win32 memory and semaphore 5617- mesa: pipe cap for EXT_memory_object_win32 and EXT_semaphore_win32 5618- zink: codegen for Win32 external object extensions 5619- zink: implement win32 fence import 5620- zink: implement win32 memory handle import 5621- zink: implement device LUID and node mask 5622- zink: enable pipe cap for win32 external memory and fences 5623- docs: updated zink features and release notes 5624 5625Sil Vilerino (39): 5626 5627- gallium/vl: Add software winsys and offscreen winsys 5628- util/u_format: Drop assert that has valid/well-defined behavior 5629- util/vl_vlc: Support compiling in C++ 5630- st_vdpau: Pass format when opening resource from handle in st_vdpau_resource_from_description 5631- d3d12: Add .clang_format file 5632- gallium/va/radeonsi: Using private as a parameter name conflicts with C++ keywords 5633- gallium va: VaDeriveImage support stride/offset being different for NV12 planes 5634- gallium: Add values to pipe_video_cap for multi-slice and multi-reference encode 5635- gallium va: Handle new VA attributes with new pipe video caps 5636- gallium: Add multiple slice support to pipe_h264_enc_picture_desc 5637- gallium va: Add support for multiple slices encoding 5638- gallium radeon/r600/omx/va: Adds support for multiple reference encoding 5639- gallium: Add MinLumaBiPredSize8x8 to pipe_h264_sps, pic_init_qs_minus26 to pipe_h264_pps 5640- gallium omx: Fill out MinLumaBiPredSize8x8 and pic_init_qs_minus26 5641- gallium vdpau: Fill out level_idc and MinLumaBiPredSize8x8 5642- gallium va: VaHandlePictureParameterBufferH264 fill out new pipe params MinLumaBiPredSize8x8, pic_init_qs_minus26, chroma_format_idc, bit_depth_chroma, bit_depth_chroma_minus8 5643- d3d12: Add d3d12_promote_to_permanent_residency 5644- ci: Update x86 debian build to pick up tag v1.602.0-r1 from DirectX-Headers 5645- d3d12: Add video decode implementation of pipe_video_codec 5646- d3d12: Add video encode implementation of pipe_video_codec 5647- d3d12: Add create_video_codec and create_video_buffer entrypoints 5648- d3d12: Add util video functions to d3d12_format 5649- d3d12: Improve planar resource support to handle video requirements 5650- d3d12: Add support for d3d12 video in d3d12_screen 5651- d3d12: Add GLSL singleton refcounting to screen for offscreen non-GL rendering 5652- meson: Support d3d12 as a video-supporting driver 5653- winsys: Do not use Display type when X11 is not present in build 5654- meson: Update DirectX-Headers dependency to 1.602.0 version with fallback to the wrap 5655- meson: Add build option for gallium-d3d12-video feature 5656- d3d12: Align output buffer offset access to D3D12_FEATURE_DATA_VIDEO_ENCODER_RESOURCE_REQUIREMENTS.CompressedBitstreamBufferAccessAlignment 5657- d3d12: Fix debug_printf format strings for cross platform builds 5658- d3d12: Video - Remove unused spCopyQueues from enc/dec objects 5659- pipe/video: Add PIPE_VIDEO_ENTRYPOINT_PROCESSING 5660- vl: Replace usage of entrypoint UNKNOWN with PROCESSING for VP 5661- va: Replace usage of entrypoint UNKNOWN with PROCESSING for VP 5662- gallium/video: Add video post processing interface 5663- va: Add support for VPP rotation, flip, alpha blend, crop, scaling 5664- d3d12: Add pipe_video_codec::process_frame implementation 5665- d3d12: Enable VPP rotation, flip, alpha blend, crop, scaling via pipe_video_codec::process_frame 5666 5667SoroushIMG (4): 5668 5669- zink: Fix BO size when it's not aligned to 16 bytes 5670- zink: Fix spirv stream 0 vertex emit for multistream shaders 5671- Zink: Fix clear being missed when using emulated draws in zink_blit 5672- zink: Fix incorrect emission of SPIR-V shift ops 5673 5674SureshGuttula (4): 5675 5676- radeonsi: Set display_remote for non-refernced frames 5677- radeonsi/vcn : update enc->dpb ref_use for index 0 5678- Revert "radeonsi: Set display_remote for non-refernced frames" 5679- Revert "radeon: hardcode uvd/vce encoder not_referenced value to false" 5680 5681Sviatoslav Peleshko (7): 5682 5683- mesa: flush bitmap caches when changing scissors or window rects state 5684- anv: workaround apps that assume full subgroups without specifying it 5685- intel/blorp/gen6: Set BLEND_STATEChange only if emitting the blend state 5686- anv: Dirty all dynamic state bits when creating command buffer state 5687- intel/blorp: Dirty depth bounds dynamic state bits after blorp 5688- intel/nullhw: Use correct macro to fix build regression 5689- iris: Always initialize shader compilation queue ready fence 5690 5691TSnake41 (1): 5692 5693- zink: print result code string on vulkan failure 5694 5695Tapani Pälli (5): 5696 5697- isl: disable mcs (and mcs+ccs) for color msaa on DG2 5698- anv: use anv_cmd_dirty_mask_t type for dynamic state 5699- intel/fs: setup SEND message descriptor from nir scope 5700- iris: implement Wa_14015264727 for DG2 5701- anv: implement Wa_14015264727 for DG2 5702 5703Tatsuyuki Ishi (9): 5704 5705- radv: Fix redundant subpass barriers due to erroneous comparison 5706- radv/ci: skip dEQP-VK.fragment_operations.transient_attachment_bit 5707- radv/ci: Move transient_attachment_bit from fail to skip list 5708- radv/ci: skip image.sample_texture.*_compressed_format* 5709- radv/ci: Move sample_texture.*_compressed_format_* to faillist for gfx<=9 5710- radv: Fix vkCmdCopyQueryResults -> vkCmdResetPool hazard. 5711- amd: Revert gfx10 addrlib changes 5712- radv: Only set pstate for the first hw_ctx. 5713- radv: Implement radv_flush_before_query_copy to workaround UE Vulkan bugs. 5714 5715Thomas Debesse (2): 5716 5717- gallium/clover: LLVM setLangDefaults moved from clangFrontend to clangBasic 5718- gallium/clover: pass -no-opaque-pointers to Clang 5719 5720Tiago Koji Castro Shibata (2): 5721 5722- d3d12: fallback to integrated adapter instead of arbitrary one 5723- d3d12: add more formats to supported conversions 5724 5725Timothy Arceri (71): 5726 5727- glsl/st: vectorise interfaces of SSO shader programs 5728- nir: fix setting varying from uniform as flat 5729- nir: fix sorting before assigning varying driver locations 5730- nir: always set the exact_trip_count_unknown loop terminator property 5731- nir: remove unreachable loop terminators 5732- svga: remove duplicate nir compile options 5733- nir/glsl: add glsl_record_compare() wrapper 5734- nir/glsl: wrapper contains_{double,interger}() 5735- nir/glsl: wrapper field_index() 5736- nir/glsl: wrap component_slots_aligned() 5737- glsl/mesa: move parse_program_resource_name() to common linker_util code 5738- glsl: add new build program resource helpers 5739- nir: add variable data fields required for NIR glsl varying linking 5740- nir: abort io info gathering if location is not set or is a temp value 5741- nir: skip lowering io to scalar for must_be_shader_input 5742- glsl: implement lower_packed_varyings() as a NIR pass 5743- glsl: add a NIR based varying linker 5744- glsl: implement opt_dead_builtin_varyings() as a NIR pass 5745- glsl: implement lower_xfb_varying() as a NIR pass 5746- glsl: enable the use of the nir based varying linker 5747- glsl: remove now unused GLSL IR varying linker code 5748- glsl: simplify finding cursor in varying packing code 5749- nir: add support for forced sampler indirect loop unrolling 5750- gallium/drivers: set force_indirect_unrolling_sampler for all required drivers 5751- glsl: move validation of sampler indirects to the nir linker 5752- i915g: use a valid setting for force_indirect_unrolling 5753- softpipe: switch to NIR loop unrolling 5754- asahi: switch to NIR loop unrolling 5755- d3d12: switch to NIR loop unrolling 5756- etnaviv: switch to NIR loop unrolling 5757- i915g: switch to NIR loop unrolling 5758- r300: disable GLSL IR loop unrolling 5759- r600: switch to NIR loop unrolling 5760- vc4: disable GLSL IR loop unrolling 5761- nir/i915g/r300/nv30: skip marking varyings as flat in some drivers 5762- etnaviv: assert if etna_shader_io_file reg overrun 5763- etnaviv: vectorise io 5764- nir/gcm: fix pushing instructions into if blocks 5765- ci: uprev piglit 2022-05-31 5766- nir: add nir based version of the lower_const_arrays_to_uniforms pass 5767- glsl/nir: skip adding hidden uniforms to the remap tables 5768- glsl/nir: allow the nir linker to remove dead uniforms we created 5769- glsl: move common link time optimisation calls to linker code 5770- glsl: move gl_nir_link_opts() call out of the st code 5771- glsl: switch to NIR based implementation of lower_const_arrays_to_uniforms() 5772- glsl: remove now unused lower_const_arrays_to_uniforms() 5773- gallivm: disable GLSL IR loop unrolling in LLVMPIPE 5774- freedreno/ir3: tidy up duplication of common nir options 5775- freedreno: switch to NIR loop unrolling 5776- lima: lower all undefs to zero in vs 5777- lima: fixup nir indirect unroll options to match gallium CAP 5778- lima: switch to NIR loop unrolling 5779- nouveau/nv30: disable GLSL IR loop unrolling 5780- nouveau/nv50: disable GLSL IR loop unrolling 5781- nouveau/nvc0: disable GLSL IR loop unrolling 5782- svga: disable GLSL IR loop unrolling 5783- gallium: remove PIPE_SHADER_CAP_MAX_UNROLL_ITERATIONS_HINT CAP 5784- glsl: remove the now unused GLSL IR loop unrolling code 5785- glsl: drop extra optimise swizzles call 5786- glsl: tidy up link_varyings_and_uniforms() 5787- glsl: remove never true do_dead_code() parameter 5788- glsl: inline do_common_optimization() call 5789- util: add dri config option to disable GL_MAP_UNSYNCHRONIZED_BIT 5790- util: use force_gl_map_buffer_synchronized workaround with RAGE 5791- glsl: merge lower_buffer_access with lower_shared_reference 5792- glsl: drop non-nir path for atan in builtin functions 5793- st/glsl: fix broken vertex attrib mapping 5794- nir/loop_unroll: clean up after complex_unroll_single_terminator() 5795- glsl: correctly track cross slot component packing 5796- Revert "nir: Preserve offsets in lower_io_to_scalar_early" 5797- glsl: dont lower precision for textureGatherOffsets 5798 5799Timur Kristóf (119): 5800 5801- radv: Minor formatting fix in radv_device.c file. 5802- radv: Move queue submit sparse bindings to a separate function. 5803- radv: Move empty queue submit code path to a separate function. 5804- radv: Move normal (non-empty) queue submit to a separate function. 5805- radv: Move up early exit for transfer queues in radv_get_preamble_cs. 5806- radv: Simplify some boolean code in radv_get_preamble_cs. 5807- radv: Initialize BO pointers when creating preambles. 5808- radv: Refactor cache flush code for the initial preambles. 5809- radv: Don't create continue preamble when it's not needed. 5810- radv: Rename fill_geom_tess_rings to radv_fill_shader_rings. 5811- radv: Don't use pointers to pointers when updating the preambles. 5812- radv: Only emit what is appropriate to the queue family for preambles. 5813- ac/nir: Add ac_nir_load_arg helper for shader arguments. 5814- radv: Move radv_nggc_settings enum out of radv_cmd_buffer. 5815- nir: Add upper bound for AMD shader arg intrinsics. 5816- radv: Lower ABI in NIR for tess/ESGS/NGG shader arguments. 5817- ac/nir: Remove now-superfluous ac_nir_lower_tess_to_const. 5818- aco: Remove now-superfluous intrinsics. 5819- ac/llvm: Remove now-superfluous intrinsics and ABI callbacks. 5820- radv/amdgpu: Group queue submit info into a structure. 5821- radv/amdgpu: Initialize semaphore info with a designated initializer. 5822- radv/amdgpu: Pass new queue submit info structure to internal function. 5823- radv/amdgpu: Pass correct struct type instead of repeating the cast. 5824- radv/amdgpu: Use scheduled dependency for submitting to multiple queues. 5825- ac/nir: Add I/O lowering for task and mesh shaders. 5826- ac: Add task ring entry shader argument. 5827- radv: Add radv_pipeline_has_task helper. 5828- radv: Set user data register for task shaders. 5829- radv: Set wave size for task shaders. 5830- radv: Fill task shader info. 5831- radv: Add task shader arguments. 5832- radv: Add task ring entry argument for mesh shaders. 5833- radv: Implement task shader intrinsics in the ABI. 5834- radv: Enable nir_opt_offsets for task shaders. 5835- radv: Use I/O lowering for task and mesh shaders. 5836- radv: Postprocess task shader configuration. 5837- radv: Allow linking task shaders. 5838- radv: Lower shared and task_payload variables in task/mesh shaders. 5839- radv: Print task shader stage name before disasm. 5840- radv: Fix task shader info. 5841- radv: Fix loading task shader ring buffer addresses. 5842- aco: Fix scratch with task shaders. 5843- radv: Introduce ring info structure for queues, refactor preamble generation. 5844- radv: Use a ws variable during preamble creation. 5845- radv: Pass radv_device pointer to some functions. 5846- radv: Refactor queue state to separate structure. 5847- radv: Clarify emitting graphics shader pointers. 5848- nir: Add explicit task payload atomic intrinsics. 5849- radv: Implement conditional rendering for async compute queue. 5850- ac: Remove trailing white space. 5851- aco: Remove trailing whitespace. 5852- radv: Remove trailing whitespace. 5853- radv: Disable predication for supass clear and image clears. 5854- nir: Add new launch_mesh_workgroups intrinsic. 5855- nir: Add common task shader lowering to make the backend's job easier. 5856- radv: Add mesh and task stage names to pipeline executable properties. 5857- nir: Keep track of cross-invocation mesh shader output access. 5858- radv, ac/nir: Fix multiview layer export for mesh shaders. 5859- ac/nir/taskmesh: Use task shader lowering from common NIR code. 5860- ac/nir/taskmesh: Use 3 dimensional workgroup ID. 5861- ac/nir: Add remappability to tess and ESGS I/O lowering passes. 5862- ac: Add RDNA2 task+mesh shader draw packet opcodes. 5863- ac: Add task shader ring information. 5864- radv: Implement task shader draw and payload rings. 5865- nir: Introduce new intrinsics for AMD specific mesh shader task ring. 5866- ac/nir/ngg: Clean up mesh shader output LDS layout. 5867- ac/nir/ngg: Use mesh shader scratch ring when outputs don't fit LDS. 5868- ac/nir/ngg: Use variables for outputs without cross-invocation access. 5869- radv: Implement mesh shader scratch ring. 5870- radv: Enable NGG wave ID for mesh shader scratch ring. 5871- ac/nir/ngg: Add mesh shader workgroup index. 5872- radv: Lower mesh shader 3D workgroup ID to 1D index. 5873- ac/nir/taskmesh: Preserve workgroup ID Y and Z when applying firstTask. 5874- radv: Always use 3D block ID and grid size in task shaders. 5875- gallium/u_blitter: Fix depth. 5876- zink: Enable the VK_EXT_depth_clip_enable extension. 5877- zink: Always enable depth clamping, make depth clipping independent. 5878- nir/lower_task_shader: don't use base index for shared memory intrinsics 5879- ac/nir/ngg: Refactor LDS instructions in NGG GS vertex emit and export. 5880- ac/nir/ngg: Ignore driver location for mesh shader outputs. 5881- radv: Don't assign driver locations to mesh shader outputs. 5882- radv: Add CULL_PRIMITIVE to special output mask. 5883- radv: Use NIR optimization to move discards to the top. 5884- radv: Only initialize DGC state when DGC is enabled. 5885- radv: Remove trailing whitespace introduced by DGC commits. 5886- radv: Refactor predication for compute queues. 5887- radv: Refactor some CP DMA functions to work with radeon_cmdbuf. 5888- radv: Refactor radv_emit_userdata_address to work with radeon_cmdbuf. 5889- radv: Refactor radv_emit_descriptor_pointers to work with radeon_cmdbuf. 5890- radv: Refactor radv_emit_inline_push_consts to work with radeon_cmdbuf. 5891- radv: Refactor view index emit to use a per-stage function. 5892- radv: Move inline push constants to a new function. 5893- radv: Minor cleanup of radv_queue_submit_normal. 5894- radv: Allow reusing pipeline compute state emit functions. 5895- radv: Add dispatch_initiator_task field to radv_device. 5896- radv: Create internal cmdbuf when a graphics pipeline needs compute. 5897- radv: Flush descriptors and push constants for task shaders. 5898- radv: Introduce radv_before_taskmesh_draw. 5899- radv: Implement mesh shading draw calls with task shaders. 5900- radv: Synchronization for task shaders. 5901- radv: Support task shaders in secondary cmd buffers. 5902- radv: Workaround MEC taskmesh dispatch hang when count buffer has zero. 5903- radv: Submit internal compute cmdbuf. 5904- radv: Copy BO list to ACE internal CS. 5905- radv: Enable task shader feature for NV_mesh_shader. 5906- ac/nir/ngg: Move primitive ID workgroup barrier to proper place. 5907- aco: Remove hack for primitive ID export. 5908- ac/nir/ngg: Create output variable for primitive ID export. 5909- ac/llvm: Implement load_num_subgroups for NGG shaders. 5910- ac/llvm: Add LLVM bug workaround to ac_build_mbcnt_add. 5911- ac/llvm: Use gs_prim_id for NGG VS. 5912- ac/nir/ngg: Copy comment about LDS layout for NGG GS. 5913- ac/nir/ngg: Remember proper bit sizes of GS output variables. 5914- ac/llvm: Implement GDS atomic add NIR intrinsic. 5915- radv/llvm: Remove incorrect hardcoded workgroup size from NGG GS. 5916- radv: Properly set LDS size for LLVM NGG shaders. 5917- radv: Use NIR-based NGG lowering with LLVM. 5918- aco: Fix p_init_scratch for task shaders. 5919- nir/gather_info: Clear cross-invocation output mask. 5920 5921Tomeu Vizoso (4): 5922 5923- lima/ci: Disable lima-mali450-piglit-gpu:arm64 5924- panvk/ci: Disable CI for a while 5925- ci: Disable jobs to the Collabora lab 5926- Revert "ci: Disable jobs to the Collabora lab" 5927 5928Vadym Shovkoplias (4): 5929 5930- anv: Fix geometry flickering issue when compute and 3D passes are combined 5931- driconf: Add a limit_trig_input_range option 5932- drirc: Set limit_trig_input_range option for the Wolfenstein Youngblood 5933- drirc: Set limit_trig_input_range option for glmark2 5934 5935Vasily Khoruzhick (7): 5936 5937- lima: Implement texture_barrier 5938- lima: enable PIPE_CAP_BUFFER_MAP_PERSISTENT_COHERENT 5939- lima: Lower nir_op_idiv with other scalar math operations. 5940- u_transfer_helper: flush temporary transfer first for MSAA 5941- lima: implement lima-specific blitter 5942- lima: wire up MSAA 4x support 5943- lima: fix buffer overallocation for index, vertex and constant buffers 5944 5945Victor Hermann Chiletto (1): 5946 5947- radv: always check entry count in descriptor pool when allocating 5948 5949Viktoriia Palianytsia (1): 5950 5951- anv: Add workaround for sample mask with multisampling 5952 5953Vinson Lee (14): 5954 5955- panfrost: Fix memory leaks on unit test failures. 5956- zink: Fix memory leak on error path. 5957- radeonsi: Move NULL check before dereference. 5958- dri2: Fix memory leak. 5959- d3d12: Initialize d3d12_video_encoder_bitstream member m_uiOffset. 5960- amd: Initialize Gfx11Lib members in constructor. 5961- d3d12: Remove unnecessary NULL check. 5962- clc: Fix build with llvm-15. 5963- microsoft/spirv_to_dxil: Fix missing-prototypes errors. 5964- radv: Remove unnecessary null check. 5965- microsoft/compiler: Fix assert. 5966- tu: Check dereferenced value of rop_reads_dst. 5967- aco: Initialize spill_ctx members in constructor. 5968- r600/sfn: Initialize TestShaderFromString member m_instr_factory. 5969 5970Vlad Zahorodnii (2): 5971 5972- winsys/amdgpu-radeon: Allow specifying context priority 5973- radeonsi: Add support for EGL_IMG_context_priority 5974 5975Yevhenii Kolesnikov (1): 5976 5977- nir: Remove single-source phis before opt_if_loop_last_continue 5978 5979Yiwei Zhang (51): 5980 5981- venus: override aspectMask for internal tiling modifier 5982- venus: use linear modifier for legacy common wsi path 5983- venus: fix view format for ahb image 5984- venus: refactor to add struct vn_env 5985- venus: add env perf options and introduce no_async_set_alloc 5986- venus: add VN_PERF option no_async_buffer_create 5987- venus: add VN_PERF option no_async_queue_submit 5988- venus: refactor vn_cmd_submit 5989- venus: flush when batched draw calls reach a threshold 5990- venus: disable VK_EXT_extended_dynamic_state2 5991- venus: refactor android gralloc pieces 5992- venus: cache front_rendering_usage bit at gralloc init 5993- venus: update vn_GetSwapchainGrallocUsage2ANDROID for shared present 5994- venus: enable ANB shared presentable image prop 5995- venus: add VN_RELAX_BASE_SLEEP_US env var for tuning cpu utils 5996- venus: log performance envs upon instance creation with debug init 5997- venus: extend buffer cache to cover layering usage 5998- vulkan: remove the VkPhysicalDeviceMemoryProperties workaround 5999- radv: remove VkPhysicalDeviceMemoryProperties2 workaround 6000- venus: use common entry point for VkPhysicalDeviceMemoryProperties 6001- venus: resolve AHARDWAREBUFFER_FORMAT_Y8Cb8Cr8_420 to external format 6002- venus: fix a bug on buffer cache init failure path 6003- venus: fix the queue init failure path 6004- venus: fix a missing mtx_destroy in vn_device_init 6005- venus: adjust device resources init order 6006- venus: let device track queue families used for device creation 6007- venus: add vn_feedback_pool backed by coherent buffer (part 1) 6008- venus: add vn_feedback_pool backed by coherent buffer (part 2) 6009- venus: add feedback check/set/reset helpers 6010- venus: add event feedback 6011- venus: add NO_FENCE_FEEDBACK perf option and disable sparse resource 6012- venus: create per queue family command pool for feedback cmd 6013- venus: renderer to store allow_vk_wait_syncs capset 6014- venus: add fence feedback 6015- venus: resolve YV12 AHB to VK_FORMAT_G8_B8_R8_3PLANE_420_UNORM 6016- venus: use narrow range to match up with mesa EGL 6017- venus: swizzle the chroma channels for YVU420 to match the VkFormat 6018- venus: GetPhysicalDeviceSparseImageFormatProperties to match sparse support 6019- venus: suballocate more for layering 6020- venus: add more tracepoints for perf analysis 6021- venus: filter out VK_EXT_physical_device_drm on the driver side 6022- venus: check dynamic state for raster enablement 6023- venus: reenable VK_EXT_extended_dynamic_state2 6024- Revert "venus: suballocate more for layering" 6025- anv: enable VK_FORMAT_G8_B8_R8_3PLANE_420_UNORM for modifier support 6026- venus: fix external memory ext filtering 6027- venus: avoid feedback for external fence 6028- venus: allow no external memory handle when renderers lacks support 6029- zink: fix zink_create_fence_fd to properly import 6030- zink: fix in-fence lifecycle 6031- venus: ignore pInheritanceInfo if not secondary command buffer 6032 6033Yogesh Mohan Marimuthu (12): 6034 6035- radeonsi/gfx11: make flat_scratch changes for graphics 6036- radeonsi/gfx11: make flat_scratch changes for compute 6037- radeonsi/gfx11: interp changes for 32bit 6038- radeonsi/gfx11: interp changes for 16bit 6039- radeonsi/gfx11: instruction cache line size is 128 bytes 6040- radeonsi/gfx11: program inst_pref_size for compute 6041- radeonsi/gfx11: program db render control register 6042- radeonsi/gfx11: export alpha through mrtz for alpha-to-coverage if mrtz is there 6043- ac,radeonsi/gfx11: swizzle MRT0/1 for dual source blending 6044- vulkan/wsi: fix extra free if buffer_blit_queue 6045- radeonsi: remove tabs from code 6046- radeonsi: no need to call si_pm4_clear_state() in si_pm4_free_state() 6047 6048Yogesh Mohanmarimuthu (1): 6049 6050- radeonsi/gfx11: use PIXEL_PIPE_STATE_DUMP event instead of ZPASS_DONE 6051 6052Yonggang Luo (119): 6053 6054- zink: Remove redundant framebuffer_mtx from zink_screen.h 6055- win32: Fixes 32 bits visual studio module definition files by add script gen_vs_module_defs.py 6056- win32: Do not use BUILD_GL32, we use def file to export win32 dll symbols. 6057- misc: Replace \`#ifdef\\t__cplusplus` with \`#ifdef\\s\\s__cplusplus` 6058- misc: replace \`#ifdef\\s\\s__cplusplus` with \`#ifdef\\s__cplusplus` 6059- zink: Trim spaces in zink.rst 6060- zink: Improve zink.rst with links 6061- st: Improve _mesa_error about GL_OUT_OF_MEMORY in st_cb_texture.c 6062- llvmpipe: Do not use _Atomic keyword that doesn't support by MSVC 6063- radv: fixes msvc compiling error 6064- radv: Fixes compiling error with msvc 6065- radv: Add losing member hw_ip_version_major and hw_ip_version_minor for win32 6066- radv: Add macros for win32 that accessed 6067- nir: Fixes MSVC compiling error about unused variable \`_` 6068- egl/wgl: On win32, there is no support for EGL_EXT_device and EGL_EXT_platform_device 6069- util: Fixes test_util_get_process_exec_path on windows host with msys2/mingw 6070- util: Fixes typo in test_util_get_process_exec_path 6071- ci/vs2019: Clear CI_COMMIT_MESSAGE and CI_COMMIT_DESCRIPTION for please meson 6072- radv: Remove usage of \`cnd_monotonic.h` 6073- d3d12: Fixes d3d12 compiling errors in \`D3D12ResourceState.h` for mingw 6074- d3d12: Move shared code that include d3d12 headers into d3d12_common.h 6075- d3d12: Fixes compiling error by include \`<unknwn.h>` 6076- d3d12: \`#include <windows.h>` instead \`#include <Windows.h>` for building under linux with mingw 6077- d3d12/wgl/test: Fixes wgl_tests.cpp for d3d12 with mingw 6078- dzn: Fixes compiling error when build with msys2/mingw 6079- dzn: Fixes compiling error by include \`<unknwn.h>` 6080- microsoft/clc: Fixes compiling of microsoft clc with mingw 6081- docs: Fixes dispatch.rst with __THREAD_INITIAL_EXEC 6082- meson: Fixes timespec_get detect code 6083- c11/threads: Remove the detecting code for \`HAVE_TIMESPEC_GET` in threads_win32.h 6084- c11: Implement c11/time.h with c11/impl/time.c 6085- c11: Implement timespec_get on win32 properly when not available 6086- util: Use timespec_get directly, it's always present 6087- util: Remove usage of designated initializers in timespec_test.cpp 6088- util: always enable timespec_test.cpp 6089- util: Remove unused function timespec_passed 6090- util: Add unittest for timespec_get 6091- ci/vs2019: There is no meaning to set $env:ErrorActionPreference 6092- ci/vs2019: switch to powershell 7 6093- ci/vs2019: Update editorconfig for ps1 files 6094- ci/vs2019: Give each option its own line in mesa_build.ps1 6095- ci/vs2019: Rename to vsInstallPath 6096- ci/vs2019: Do not install subprojects 6097- ci/vs2019: Ignore cert install output, it's too long 6098- ci/vs2019: Ignore error when removing directory recursively 6099- ci/vs2019: Disable progress when install python 6100- ci/vs2019: Disable progess bar for pip3 install 6101- ci/vs2019: Use shared VULKAN_SDK_VERSION for install and build vulkan related packages 6102- ci/vs2019: Add a dedicated script for entering VS dev shell 6103- ci/vs2019: Split call to C:\\vs_buildtools.exe into multiple lines 6104- ci/vs2019: Show the total vs2019 install time 6105- ci/vs2019: Reduce the vs2019 build tools components to be installed 6106- ci/vs2019: ignore _build and _install directory by write * into .gitignore 6107- ci/vs2019: Install all build artifacts into C:\\mesa-deps 6108- ci/vs2019: Split choco install output 6109- ci/vs2019: Split install of vs2019 and choco into separate Docker image 6110- ci/vs2019: Remove comment for windows_test_vs2019 that not true 6111- ci/vs2019: Convert mesa_build.ps1 to using PowerShell 7 && operator 6112- microsoft/compiler: Include stddef.h before using size_t 6113- microsoft/compiler: Fix error from double extern 6114- dxil: Fixes compiling dxcapi.h with mingw. 6115- dxil: Sync dxil_spirv_shader_stage with gl_shader_stage 6116- dxil: Fixes warning about comparing different enum value and enum literal 6117- microsoft/spirv_to_dxil: Delete unused local variable 6118- microsoft/clc: Delete unused local variables 6119- microsoft/clc: Add missing void to no-parameter function signature 6120- microsoft/spirv_to_dxil: Fixes maybe-uninitialized compiling error in dxil_spirv_nir.c 6121- ci: Building spirv-to-dxil on both mingw and debian/x64 6122- intel: using C++11 keyword thread_local 6123- glx: using C++11 keyword thread_local 6124- util: Define WIN32_LEAN_AND_MEAN before include of \`windows.h` in u_thread.h 6125- c11: Move the implementation of threads.h into c source code 6126- c11: Fixes memory leak of pack in thrd_create on win32 6127- c11: Implement thread_local in c11/threads.h 6128- util/c11: Update function u_thread_create to be c11 conformance 6129- meson: Using get_argument_syntax as the \`--compiler_id` option for gen_vs_module_defs.py 6130- meson: Use cc.get_argument_syntax instead cc.get_id when possible. 6131- meson/rtti: Using get_argument_syntax to check msvc compatible compiler 6132- meson/clover: Getting clover can be built with no-rtti llvm in mingw 6133- ci/vs2019: Upgrade Windows 10 SDK version to 20348 6134- docs: Update windows build requirement with Visual Studio 6135- d3d12: Use static_cast instead of dynamic_cast in d3d12_video_enc_h264.cpp 6136- d3d12: Convert #include <Windows.h> to #include <windows.h> for mingw on linux 6137- d3d12: Fixes compiling error in d3d12/wgl/d3d12_wgl_framebuffer.cpp with gcc 6138- microsoft/clc: Fixes narrowing error in clc_compiler_test.cpp with mingw/gcc 6139- microsoft/clc: Disable clc_compiler_test on non-windows platform 6140- dzn: Fixes incompatible pointer type error 6141- ci/x86_build: Getting pushd popd be paired, avoid using cd 6142- ci: Prepare the container for building all mesa components with mingw under linux 6143- ci: Trigger the new mingw/linux dockers to be build 6144- d3d12: Turn d3d12_format.h to include d3d12_common.h 6145- ci: Building all mesa functional with mingw on debian 6146- meson: Enable wgl tests on mingw 6147- CODEOWNERS: Update c11 code owners 6148- CODEOWNERS: evelikov renamed to xexaxo 6149- gallium: Move -DHAVE_PIPE_LOADER_DRI and -DHAVE_PIPE_LOADER_KMS to be pre_args 6150- gallium: Do not call to pipe_loader_sw_probe_kms when not present 6151- gallium: Rename macros HAVE_PIPE_LOADER* 6152- gallium: Disable dri2 interface on OSX 6153- glx: __glX_tls_Context should use __THREAD_INITIAL_EXEC 6154- util: thread_local doesn't support to be exported on OSX 6155- glx: Convert tab to space in applegl_glx.c 6156- dri: Trim trailing space in dri/dri_util.* 6157- ci: Building osmesa on OSX 6158- util: Disable tests/sparse_array_test.cpp on MacOS as it's too slow 6159- mapi: Trim trailing spaces in stub.c and u_current.c 6160- meson: Remove usage of use_elf_tls 6161- egl: Remove usage of USE_ELF_TLS macro 6162- mapi: Remove usage of USE_ELF_TLS 6163- loader: Remove usage of USE_ELF_TLS in loader.c 6164- glx: Remove usage of USE_ELF_TLS 6165- util: Remove usage of USE_ELF_TLS in u_thread.h 6166- meson: Remove pre_args += '-DUSE_ELF_TLS' 6167- glx: Remove usage of pipe_tsd_set and pipe_tsd_get 6168- auxiliary: Remove pipe_tsd 6169- d3d12: Fixes compile error with mingw/gcc-x64 when static linkage to runtime library 6170- microsoft/clc: Fixes compiling errors with clang/mingw64 in clc/clc_compiler_test.cpp 6171- util: Fixes invalid assumption that return non null by function util_format_fetch_rgba_func 6172- c11: #include <threads.h> when the os/platform provide it 6173 6174Yurii Kolesnykov (2): 6175 6176- Remove Travis CI config 6177- Init macOS GitHub Action 6178 6179Zack Rusin (4): 6180 6181- svga: Don't try to build x86/x64 assembly on different arch's 6182- svga: finish readbacks before mapping resources 6183- svga: Use direct maps when GB objects are present 6184- svga: Add support for SVGAv3 6185 6186Zhang, Jianxun (1): 6187 6188- iris: Wa_14016820455 for GFX_VERx10 == 12.5 6189 6190illiliti (1): 6191 6192- Use proper types for meson objects 6193 6194jheaff1 (1): 6195 6196- build(glx): Fix build by adding missing deps 6197 6198naveen (1): 6199 6200- github/ci: Set permissions for GitHub actions 6201 6202newbluemoon (1): 6203 6204- nine: replace ulimit with sysconf call 6205 6206prashanth (1): 6207 6208- vulkan/device_select: force default device if environment variable set 6209 6210sjfricke (2): 6211 6212- isl: fix bug where sb.MOCS is not being set 6213- anv: fix assert to build with shader cache disabled 6214 6215xperia64 (1): 6216 6217- wgl: Add driver_zink as a dependency of the wgl frontend 6218 6219İlhan Atahan (1): 6220 6221- Add Adreno 616 and 620 to use turnip on these GPU's . 6222