1Mesa 22.1.0 Release Notes / 2022-05-18 2====================================== 3 4Mesa 22.1.0 is a new development release. People who are concerned 5with stability and reliability should stick with a previous release or 6wait for Mesa 22.1.1. 7 8Mesa 22.1.0 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 22.1.0 implements the Vulkan 1.2 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 df6270c1371eaa2aa6eb65b95cbbb2a98b14fa4b7ba0ed45e4ca2fd32df60477 mesa-22.1.0.tar.xz 25 26 27New features 28------------ 29 30- d3d12 GL4.2 31- GL_NV_pack_subimage 32- VK_EXT_depth_clip_control on lavapipe and RADV 33- Vulkan 1.3 support on lavapipe 34- VK_EXT_graphics_pipeline_library on lavapipe 35- VK_EXT_primitives_generated_query on lavapipe 36- VK_EXT_image_2d_view_of_3d on ANV and lavapipe 37- VK_KHR_swapchain_mutable_format on lavapipe 38- Intel DG2 support 39 40 41Bug fixes 42--------- 43 44- [radv] DCC causes artifacts in Senran Kagura Shinovi Versus 45- Radeon RX 5700XT crash when using prusa-slicer 46- turnip: gfxbench mh3.1 offscreen crash with ANGLE 47- Radv leaks a descriptor set layout 48- clover: Assertion \`NumContainedTys && "Attempting to get element type of opaque pointer"' failed 49- wsi/x11: Avoid using xcb_wait_for_special_event in FIFO modes causes memory leak in Apex Legends 50- freedreno: fragile location packing can break cull/clip distance 51- turnip: dEQP-VK.spirv_assembly.instruction.compute.float_controls.fp16.generated_args.* failures 52- Panfrost: wrong textures and lighting in "Thief" games in wine (not happening when using Software render) 53- [Bisected] r600: Big performance regression on some games 54- piglit ext_image_dma_buf_import.ext_image_dma_buf_import* regression 55- Huge amount of anon_inode:sync_file file descriptor created 56- Grid Autosport via proton/vulkan: glitches on the grass on amdgpu, fine on amdgpu-pro 57- NIR validation failed after nir_opt_if 58- intel: integer_mad_hi / integer_mad_sat / integer_mul_hi produce invalid results 59- Aperture Desk Job rendering artifacts Intel TGL 60- intel: integer_sub_sat produces stack corruption / OpISubBorrow compilation issue 61- 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 62- radv/aco: xfb bug 63- '../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' 64- Occasional flicker corruption in Rage 2, e.g. after loading, with ACO on RX 5700 XT 65- Vulkan::Calling vkWaitForFences Timeout 66- Intel (CHT) - Uplink text rendering bugged out in Mesa 22.0 67- gen9atom gpu hang on dEQP-VK.spirv_assembly.instruction.graphics.float16.arithmetic_1 68- bad memory managment on panfrost RK3399 - cannot alocate more ram - fury unleashed 69- radv,aco: Regression with global atomics with negative offsets 70- RADV: Quake 2 RTX v1.6.0 Corrupted Graphics 71- radv: nir validation error with invalid array access 72- [Intel, Iris] Broken rendering in Ryujinx on Tigerlake 73- intel: integer_mad_hi / integer_mad_sat / integer_mul_hi produce invalid results 74- llvmpipe: wrong color on big endian, with 16bits framebuffer. 75- llvmpipe: wrong color on big endian, with 16bits framebuffer. 76- radv: Poor performance in Ghostwire: Tokyo 77- Intel Iris Xe Geometry Flickering/Assets Disappearing 78- Rendering artifacts when playing Outer Wilds [Reproducible with latest ANV driver built from main] 79- ANV: EXT_depth_clip_control broken 80- VA-API accelerated videos are displayed all red w/ Encoder format conversion (EFC) support for AMD VCN2.0 devices patch 81- panfrost: BO importing crash on Android 82- [Intel,ANV] Feature request: a few formats 83- Vulkan wsi leaks vk_sync object on every wsi_AcquireNextImageKHR call 84- radv: Fog effects in Elden Ring broken 85- r300: Set PVS_LAST_VTX_SRC_INST and PVS_XYZW_VALID_INST correctly 86- [bisected] commit "vbo/dlist: use a single buffer object" results in heavy regression with virgl for glxgears 87- llvmpipe: [bisected] ec8104c6b2274 breaks virglrenderer's vtest_server 88- Textures colors distortion in "Black Geyser: Couriers of Darkness" with radeonsi 89- r300: Anmesia the dark descent corruption 90- ShaderStorageBlocksWriteAccess not set for spir-v shaders? 91- anv: EXT_color_write_enable severely broken 92- ICL gpu hang: dEQP-GLES31.functional.shaders.multisample_interpolation.interpolate_at_sample.centroid_qualified.multisample_rbo_16 93- null pointer deref crash in wsi_common_x11.c 94- Error compiling with LLVM-git/15 95- [bisected] panfrost(RK3399/T860): Emulationstation: broken, black or missing menus with v22.0.0 96- Plasma/KDE settings menus disappear on daily build 97- turnip: Unrecoverable hang in Age of Empires IV on A650/A660 98- anv regressions from commit 57445adc891 "anv: Re-enable CCS_E on TGL+" 99- anv: uninit compile warnings 100- VK_IMAGE_CREATE_EXTENDED_USAGE_BIT not accounted for in vkGetPhysicalDeviceImageFormatProperties 101- freedreno: crash in PUBG 102- [regression][bisected] MSVC: Build failure in libmesa_util when targeting x86 32-bit 103- vulkan: cmd queue codegen still needs work 104- Using VAAPI-AMDGPU, RGBA-surface converted to NV12-surface, and UV data is lost, only Y data 105- A crash in radeonsi driver 106- freedreno: deqp cts fails 107- turnip: autotune disabled under zink sometimes 108- radeonsi dEQP-GLES3.functional.buffer.map.write.explicit_flush.* flake crashes 109- [BDW-TGL] Piglit's ext_image_dma_buf_import-export-tex fails in the Intel CI 110- android: don't sync_wait() on in fence 111- va: incorrect frame_num logic causes corrupt h.264 encoding 112- Performance slowdown by ~20% in Unigine sanctuary with nir_to_tgsi 113- !15098 broke 4 piglit spec\@ext_external_objects tests 114- anv: Unable to import I915_FORMAT_MOD_Y_TILED surfaces produced by iHD vaapi driver 115- anv doesn't always resolve aux buffers with private bindings on transition to external queue 116- Square Artifacts Dragons Dogma 117- iris: Shared resources are allocated for scanout, but not configured for it 118- [Feature Request] turnip: Support of Gamescope 119- r300: shaders/closed/steam/massive-chalice/1707.shader_test reports absurd results 120- Redraw freeze after upgrade to Xwayland 21.1.3 121- [ANV] Rendering corruption in DOOM Eternal 122- radv: When using VS+PS only, primitive ID is mixed up after NGG culling 123- radv: Fails to build on Unix except Linux 124- [22.0.0_rc{1,2}] radeonsi: broken support for HD7850 (radeon 0000:0b:00.0: ring 0 stalled for more than...) 125- turnip: external memory import does not work 126- lavapipe: dEQP-VK.spirv_assembly.instruction.compute.float16.arithmetic_3.step fails 127- RADV: Texture seams in The Evil Within 2 (regression) 128- r300: some optimization oportunities when transforming SIN and COS inputs 129- microsoft/compiler: System variables that aren't VS input or PS output should have interpolation specified 130- ANV: Bad output from TransformFeedback . Regression from Mesa 21. Something to do with VB+XFB -> VB+XFB dependency? 131- undefined reference to \`intel_ds_end_dyn_render_pass' 132- radv: CullDistance fail 133- Changing the buffer modifer via Wayland dmabuf-feedback apparently does not reliably reset the buffer age to 0 134- r300: "deadcode" pass for removing unused values ignores "break" 135- turnip: dEQP-VK.pipeline.render_to_image.core.* failures on a618 136 137 138Changes 139------- 140 141Adam Jackson (20): 142 143- dri: Epoch how no-error context creation works 144- egl: Use the new no-error driver interface 145- glx: Use the new no-error driver interface 146- mesa: Enable GL_NV_pack_subimage 147- wsi/x11: xcb_wait_for_special_event failure is an error 148- dri: Fold away some unused indirection in __DriverAPIRec 149- dri: Implement __DRI_DRIVER_VTABLE 150- dri: Fill in the driver extensions for the legacy createNewScreen paths 151- dri: Remove the globalDriverAPI hacks 152- dri: Remove the megadriver compat stub 153- meson: Define a HAVE_XXXX macro for every gallium driver we build 154- kopper: Define the driver interface 155- egl: Learn about kopper 156- glx: Learn about kopper 157- gallium: Learn about kopper 158- glx/kopper: Wire up a way for SwapBuffers to fail 159- wsi/x11: Avoid using xcb_wait_for_special_event in FIFO modes 160- kopper: Grow a swap interval API 161- glx/kopper: Enable GLX_EXT_swap_control etc. 162- egl/kopper: Hook up eglSwapInterval 163 164Adrián Larumbe (1): 165 166- panfrost: fix segfault in pandecode 167 168Akihiko Odaki (2): 169 170- virgl/ci: Uprev virglrenderer 171- virgl: Check texture multisample compatibility 172 173Alejandro Piñeiro (4): 174 175- v3d/drm-shim: remove drm-shim driver 176- v3dv/pipeline: use new helper vk_shader_module_to_nir 177- docs: document v3d/v3dv envvars 178- broadcom: update language on V3D_DEBUG options 179 180Alex Xu (Hello71) (1): 181 182- r300/compiler/tests: print regoff_t as size_t 183 184Alexey Bozhenko (1): 185 186- spirv: fix OpBranchConditional when both branches are the same 187 188Alyssa Rosenzweig (345): 189 190- panfrost: Remove blend shader return value on v9 191- panfrost: Fix alignments on Valhall 192- panfrost: Add Valhall Plane Descriptor XML 193- panfrost: Add remaining ZS/CRC XML 194- panfrost: Fix texel interleave flag on Valhall 195- nir,zink: Make lower_discard_if a common pass 196- agx: Call nir_lower_discard_if 197- asahi: Identify IOGPU_ATTACHMENT::size 198- asahi: Correctly set IOGPU_ATTACHMENT::size 199- agx: Add AUTO_LOD_BIAS mode 200- agx: Translate LOD modes more generically 201- agx: Implement nir_op_txb 202- agx: Handle texture array indices 203- agx: Don't kill helper threads in ld_var 204- asahi: Fix memory unsafety in delete_sampler_state 205- panfrost: Flesh out Buffer descriptor 206- panfrost: Strip % in GenXML names 207- panfrost: Flesh out tiler heap descriptor 208- panfrost: Update supported job types 209- panfrost: Remove some indexed formats on Valhall 210- panfrost: Remove unused layout enums 211- panfrost: Fix Depth Source enum 212- panfrost: Update primitive descriptor for Valhall 213- panfrost: Add more fields to Attribute Descriptor 214- panfrost: Clarify unknowns in z/stencil descriptor 215- panfrost: Simplify Valhall preload descriptor 216- panfrost: Flesh out the Shader Program Descriptor 217- panfrost: Add Valhall additions to the framebuffer 218- panfrost: Shuffle render target AFBC for Valhall 219- panfrost: Add Valhall fields to tiler descriptor 220- panfrost: Update Shader Environment descriptor 221- panfrost: Flesh out compute jobs 222- panfrost: Fix Malloc Vertex definition 223- nir: Set internal=true in nir_builder_init_simple_shader 224- panfrost,asahi,radv: Don't set internal=true manually 225- pan/mdg: Pull out skip_internal boolean 226- pan/mdg: Print optimized and scheduled shader 227- pan/mdg: Model zero/sign extension for 8/16-bit loads 228- pan/mdg: Handle 8/16-bit UBO loads 229- pan/mdg: Clarify some ISA unknowns 230- pan/mdg: Delete stray comment 231- pan/mdg: Delete dedicated fdot2 lowering 232- pan/mdg: Assert that we don't see unknown jumps 233- pan/mdg: Remove todo we'll probably never get to 234- panvk: Don't use UBOs for meta_clear 235- pan/va: Parse units from the XML 236- pan/va: Fix some units 237- pan/va: Make subgroup 4-bits 238- pan/va: Fix conservative branch handling 239- pan/va: Identify LEA_TEX_IMM table 240- pan/bi: Avoid \*FADD.v2f16 hazard in optimizer 241- pan/bi: Avoid \*FADD.v2f16 hazard in scheduler 242- pan/bi: Test avoiding \*FADD.v2f16 hazard in optimizer 243- pan/bi: Test avoiding FADD.v2f16 hazards in scheduler 244- nir: Check all sizes in nir_alu_instr_is_comparison 245- asahi: Streamline modifier selection 246- asahi: Don't redefine MIN2/MAX2 247- asahi: Identify Level field of render target descriptor 248- asahi: Respect mip level when rendering 249- asahi: Add 2D Array and 3D texture dimensions 250- asahi: Handle tiling of 2D arrays and 3D 251- asahi: Add some notes to XML about mipmapping 252- asahi: Dynamically configure tile size 253- asahi: Allow tiling of all bpps 254- asahi: Rename bpp to blocksize 255- asahi: Align allocations to effective tile size 256- asahi: Align linear texture's strides to 64 bytes 257- asahi: Handle page alignment of miptrees 258- asahi: Pass correct tile shift to tiling routines 259- asahi: Track mipmap state explicitly 260- asahi: Support 2D array and 3D textures 261- asahi: Add agx_map_texture_{cpu,gpu} helpers 262- asahi: Handle reloads of specific cube/mipfaces 263- asahi: Implement texturing with non-zero start level 264- asahi: Add AGX_PUSH_ARRAY_SIZE_MINUS_1 265- agx: Add typed move helper 266- agx: Add agx_size_align_16 helper 267- agx: Naturally align uniform pushes 268- agx: Round and clamp array indices 269- asahi: Allow GenXML to be used in C++ 270- asahi: Add LOD type 271- asahi: Add LOD clamp packing unit tests 272- asahi: Identify minimum/maximum LOD fields 273- asahi: Support LOD clamps 274- asahi: Wire in pure integer texture formats 275- pan/bi: Disambiguate IDVS variants in shader-db 276- pan/bi: Lower swizzles on CSEL.i32/MUX.i32 277- pan/bi: Lower swizzles on MUX.v2i16 278- pan/bi: Constant fold swizzles on constants 279- pan/bi: Optimize replication 280- pan/bi: Handle trivial i2i32 281- pan/bi: Handle vectorized u2f16/i2f16 282- pan/bi: Switch to lower_bool_to_bitsize 283- pan/bi: Revert "Fix load_const of 1-bit booleans" 284- pan/bi: Promote MUX to CSEL in the scheduler 285- asahi: Fix use-after-free in shader key 286- panvk: Use more reliable assert for UBO pushing 287- pan/bi: Specialize IDVS in NIR 288- pan/bi: Enable nir_opt_shrink_vectors 289- pan/bi: Clarify requirement for barriers 290- pan/bi: Cull DTSEL_IMM dests in post-RA DCE 291- pan/bi: Do not cull post-RA staging writes 292- pan/bi: Add bi_{start, exit}_block helpers 293- pan/bi: Use bi_exit_block 294- pan/bi: Clean up nits in liveness analysis 295- pan/bi: Add scoreboard state to IR 296- pan/bi: Print scoreboarding state 297- pan/bi: Implement basic scoreboarding pass 298- pan/bi: Add BIFROST_MESA_DEBUG=nosb option 299- pan/mdg: Fix partial execution mode names 300- panfrost: Use pan_shader_prepare_rsd in blitter 301- panfrost: Set defaults for deprecated DCD fields 302- panfrost: Annotate slow clears as such 303- panfrost: Decouple tiler job and DCD emit 304- panfrost: Inline pan_blit_emit_dcd 305- panfrost: Use txl instead of tex in the blitter 306- panfrost: Remove unrelated comment 307- panfrost: Remove pan_emit_fbd thunking 308- panfrost: Inline pan_emit_sfbd_tiler 309- ci: Disable windows-vs2019 310- panfrost: Fix FD resource_get_handle 311- panfrost: Simplify panfrost_resource_get_handle 312- iris,crocus,i915g: Don't stub flush_frontbuffer 313- pan/mdg: Fix overflow in intra-bundle interference 314- pan/bi: Reorder pushed uniforms to avoid moves 315- panfrost/ci: Move T720 flakes to skips 316- panfrost/ci: Move T860 flake to skip 317- panfrost: Remove Message Preload Descriptor from v6.xml 318- panfrost: Add an unpacked message preload struct 319- panfrost: Pack message preloads from compiler 320- pan/bi: Add bi_before_nonempty_block helper 321- pan/bi: Account for message preloading in shaderdb 322- pan/bi: Support message preloading 323- pan/bi: Unit test message preloading optimization 324- panfrost/ci: Update xfails list 325- panfrost: Fix set_sampler_views for big GL 326- panfrost: Handle NULL sampler views 327- panfrost: Handle NULL samplers 328- panfrost: Flush resources when shadowing 329- pan/va: Remove incorrect TEX test cases 330- pan/va: Add MUX.v2i16 and MUX.v4i8 opcodes 331- pan/va: Allow forcing enums for 1-bit modifiers 332- pan/va: Handle extended staging counts in assembler 333- pan/va: Don't use staging index as a sideband 334- pan/va: Fix definitions of TEX_SINGLE and TEX_FETCH 335- pan/va: Handle sr_write_count in the disassembler 336- pan/va: Add TEX_FETCH assembler case 337- panfrost: Extend SPD size 338- panfrost: Add Tile Render Order enum to fragment jobs 339- panfrost: Remove Invalidate Cache from Valhall job header 340- panfrost: Clarify contains descriptor? bit 341- panfrost: Make Divisor E an integer on v9 342- panfrost: Add an enum for Valhall resource tables 343- panfrost: Rename prepare_rsd->prepare_shader 344- panfrost: Add PAN_MESA_DEBUG=dump option 345- panfrost: Generalize some is_bifrost users 346- panfrost: Don't pack blend constants with blend shaders 347- panfrost: Allow uploading fragment SPDs 348- panfrost: Don't emit compression tags on Valhall 349- panfrost: Adapt estimate_texture_payload_size to Valhall 350- panfrost: Set texel_interleave on Valhall 351- panfrost: Unify barrier+helper handling 352- pan/bi: Mark NOP as having no destinations 353- pan/bi: Use a progress loop for constant folding 354- pan/bi: Allow CSE of preloaded registers 355- pan/bi: Support standalone Valhall disassembly 356- pan/bi: Wire Valhall disassembler into compiler 357- pan/bi: Add BI_SUBGROUP_SUBGROUP16 option 358- pan/bi: Trade off registers/threads on Valhall 359- pan/bi: Adapt bi_lower_branch for Valhall 360- pan/bi: Extract INSTRUCTION_CASE macro 361- pan/va: Add missing copyright notice 362- pan/va: Handle force_enum differing from name 363- pan/va: Add modifiers required for gathers 364- pan/va: Add TEX_DUAL instruction 365- pan/va: Add TEX_GATHER instruction 366- pan/va: Fix definitions of LD_VAR_BUF_IMM 367- pan/va: Fix LEA_BUF_IMM definition 368- pan/va: Remap "store segment" to "memory access" 369- pan/va: Add memory access modifier to LOADs 370- pan/bi: Model Valhall texture instructions 371- pan/bi: Extend BLEND to take a register format 372- pan/bi: Generalize I->table for Valhall 373- pan/bi: Add LD_VAR_BUF_IMM.f16/f32 instructions 374- pan/bi: Model LEA_BUF_IMM in the IR 375- pan/bi: Model pos/vary segments in STORE instructions 376- pan/bi: Model offset for LOAD/STORE 377- pan/bi: Model LD_BUFFER instructions 378- pan/bi: Add BRANCHZI instruction 379- pan/bi: Extend LD_TILE with a register format 380- pan/bi: Add arithmetic flag to RSHIFT ops 381- pan/bi: Run CSE after lowering FAU 382- panfrost: Push twice as many uniforms 383- mesa: Remove unused framebuffer validation 384- panfrost: Emulate GL_CLAMP on Bifrost 385- pan/bi: Handle non-2D arrays 386- panfrost: Handle txs of cube arrays 387- pan/va: Fix typo in BLEND text 388- pan/va: Add start property to source 389- pan/va: Handle 64-bit sources in message instrs 390- pan/va: Fix BLEND instruction 391- pan/va: Rewrite FAU handling in dis/assembler 392- pan/va: Handle uniforms from page 1 393- pan/va: Rename imm_mode -> fau_page 394- pan/va: Use 64-bit special FAU for pages 1 and 3 395- pan/va: Remove immediate modes from XML/asm 396- pan/va: Use boring names for FAU special pages 1/3 397- pan/va: Use XML for special FAU page 0 398- pan/decode: Handle blend arrays on Valhall 399- pan/decode: Unify tiler job handling 400- panfrost: Correct ASTC decode mode XML 401- panfrost: Fix primitive restart with 32-bit indices 402- panfrost: Fix definition of DCD on v9 403- panfrost: Refactor XML to permit non-IDVS jobs 404- panfrost: Add Tiler Job to v9 XML 405- pan/va: Correct definition of ZS_EMIT 406- pan/va: Model LEA_TEX_IMM more accurately 407- pan/va: Add LEA_ATTR_IMM instruction 408- pan/va: Add missing .auto32 register format 409- pan/va: Align error messages in disassembler tests 410- pan/va: Fix ST_CVT definitions 411- pan/bi: Add helpers to get vertex/instance ID 412- pan/bi: Use vertex/instance ID helpers 413- pan/bi: Print Valhall-specific FAU indices 414- pan/bi: Don't analyze helper reqs in !frag shaders 415- pan/bi: Add Valhall-specific zero builder 416- pan/bi: Model Valhall action on bi_instr 417- pan/bi: Emit arch-specific code for bi_dontcare 418- pan/bi: Use bi_dontcare for ZS_EMIT 419- asahi: Identify IOGPU Clear Z/S structure 420- asahi: Identify aux framebuffer data structure 421- asahi: Identify IOGPU Internal Pipelines structure 422- asahi: Add stencil buffer attachment type 423- asahi: Identify IOGPU_MISC data structure 424- asahi: Add size field to slices 425- asahi: Add separate_stencil, internal_format fields 426- asahi: Generate IOGPU attachments dynamically 427- asahi: Wire in u_transfer_helper 428- asahi: Handle flushes of depth-only rendering 429- asahi: Don't clobber clear colours 430- asahi: Port driver to macOS 12.x ABI 431- panfrost: Process scissor state earlier 432- pan/va: Allow forcing staging flags to read-write 433- pan/va: Allow omitting staging registers 434- pan/va: Add atomic instructions 435- pan/bi: Use consistent modifier lists in packing 436- pan/bi: Gate late DCE/CSE on "optimize" 437- pan/bi: Rename PATOM_C to ATOM 438- pan/bi: Add ATOM_RETURN pseudo-instruction 439- pan/bi: Model Valhall-style A(CMP)XCHG 440- pan/bi: Allow branch_offset on BLEND 441- pan/bi: Check return addresses in blend shaders 442- pan/bi: Augment ST_TILE with register format 443- pan/bi: Model LD_VAR_BUF instructions 444- pan/bi: Rename I->action to I->flow 445- pan/va: Add ST_TILE instruction 446- pan/va: Add LD_VAR_BUF instructions 447- pan/va: Add Bifrost-style LD_VAR instructions 448- pan/va: Unify flow control 449- pan/va: Permit encoding more flags 450- pan/va: Build opcode info structures 451- pan/va: Generate header containing enums 452- pan/va: Add helpers for swapping bitwise sources 453- pan/va: Add packing routines 454- pan/va: Optimize add with imm to ADD_IMM 455- pan/va: Add unit tests for ADD_IMM optimizations 456- pan/va: Add FAU validation 457- pan/va: Validate FAU before packing 458- pan/va: Add constant lowering pass 459- pan/va: Add instruction selection lowering pass 460- pan/va: Lower branch offsets 461- pan/va: Test instruction selection lowerings 462- pan/va: Implement the cycle model 463- pan/va: Add shader-db support 464- pan/va: Add packing unit tests 465- pan/va: Lower BLEND to call blend shaders 466- pan/bi: Add .shadow modifier to TEX_GATHER 467- pan/bi: Fix write_mask size 468- pan/bi: Call Valhall backend passes on v9 469- pan/bi: Use nir_tex_instr_has_implicit_derivative 470- pan/bi: Split out load/store to thread storage 471- pan/bi: Use ID accessors for LEA_ATTR 472- pan/bi: Preload r60/r61 for MSAA + blend shader 473- panfrost: Disable AFBC on Valhall 474- panfrost: Handle Valhall IDVS in job_uses_tiling 475- panfrost: Restrict Z/S formats for Valhall 476- panfrost: Add Valhall compressed formats 477- pan/bi: Model Valhall image loads 478- pan/va: Add indirect LEA_{ATTR, TEX} 479- pan/va: Pack LEA_TEX_IMM 480- pan/va: Model image load instructions 481- pan/va: Don't truncate slots 482- pan/va: Add flow control lowering pass 483- pan/va: Allow small constants in register pairs 484- panfrost: Add a table for images 485- pan/bi: Mark LD_TILE as w=format 486- pan/bi: Fix spilling on Valhall 487- pan/bi: Waits before tilebuffer access on Valhall 488- pan/bi: Specialize BLEND emit for Valhall 489- pan/bi: Emit Valhall texture instructions 490- pan/bi: Handle Valhall texturing in helper analysis 491- pan/bi: Track whether the malloc IDVS flow is used 492- pan/bi: Emit Valhall-style varying loads 493- pan/bi: Emit Valhall-style varying stores 494- pan/bi: Set table for Valhall LD_ATTR 495- pan/bi: Force psiz to mediump 496- pan/bi: Lower gl_PointSize to FP16 on Valhall 497- pan/bi: Make psiz variants 498- pan/bi: Generate LD_BUFFER on Valhall 499- pan/bi: Avoid masked writes for now 500- pan/bi: Report whether workgroups can be merged 501- pan/bi: Don't lower vertex_id for malloc IDVS 502- pan/bi: Consider flow control in DCE 503- pan/va: Add LD_TILE.v3.f16 packing test 504- panfrost: Handle Valhall texturing 505- panfrost: Handle stencil texturing on Valhall 506- panfrost: Control tiler memory usage 507- panfrost: Hide parts of pan_encoder.h for Valhall 508- panfrost: Add panfrost_make_resource_table helper 509- panfrost: Add shader_stage helper 510- panfrost: Adapt pan_shader.h for Valhall 511- panfrost: Hide AFBC on Valhall 512- panfrost: Move assign_vertex_buffer to pan_helpers 513- panfrost: Make alpha=0 NOP / 1 store Bifrost only 514- panfrost: Disable PIPE_CAP_PRIMITIVE_RESTART on v9 515- panfrost: Add helpers to set batch masks 516- panfrost: Split out image access tracking 517- panfrost: Don't check alpha test in fs_required on Bifrost+ 518- panfrost: Don't set a default for blend count 519- panfrost: Adapt panfrost_rasterizer for v9 520- panfrost: Add a pool to sampler_view 521- panfrost: Specialize vertex state for Valhall 522- panfrost: Split out panfrost_get_blend_shaders 523- panfrost: Split out allow_fpk helper 524- panfrost: Add Valhall fields to panfrost_batch 525- panfrost: Add valhall_has_blend_shader field 526- panfrost: Hide some Bifrost-specific functions 527- panfrost: Adapt viewport/scissor to Valhall 528- panfrost: Add helpers to emit Valhall data structures 529- panfrost: Use track_image_access on Bifrost 530- pan/bi: Don't use funny round modes in tests 531- pan/bi: Mark some opcodes as default round-to-zero 532- pan/bi: Use should_skip in bi_builder generation 533- pan/bi: Imply round mode most of the time 534- nir: Don't set writes_memory for reading XFB 535 536Andrey Konovalov (1): 537 538- ir3: set local_size for shaders of MESA_SHADER_KERNEL type 539 540Andrii Pauk (1): 541 542- venus: Allow usage of virtio-mmio based device 543 544Andrii Simiklit (1): 545 546- glsl: add member's location layout qualifier rules for \`arrayed` in/out blocks 547 548Anuj Phogat (3): 549 550- anv, iris: Add Wa_16011411144 for DG2 551- include/uapi: Update drm_fourcc.h from kernel 552- isl,iris: Add I915_FORMAT_MOD_4_TILED support for XeHP 553 554Autumn on Tape (4): 555 556- gallivm: add subgroup shuffle support 557- lavapipe: enable subgroup shuffle operations 558- gallivm: use shufflevector for shuffles when index is constant data 559- gallivm: use VPERMPS (x86/AVX2) for 32-bit 8-element shuffles 560 561Bas Nieuwenhuizen (9): 562 563- radv: Add submit locking with trace bo. 564- radv: Use larger arena sizes. 565- radv: Fix preamble argument order. 566- radv: Only wait on CS/PS to finish if we wait on a semaphore. 567- vk: Update xml and headers to 1.3.207. 568- radv: Expose VK_VALVE_descriptor_set_host_mapping for vkd3d only. 569- radv: Fix vk_queue_to_radv for radv_image_queue_family_mask. 570- radv: Add more BVH vertex formats. 571- vulkan/wsi/x11: Ensure we have the required number of images for acquire. 572 573Benjamin Cheng (3): 574 575- vulkan/queue: Destroy wait temps if they are skipped 576- anv: drop from_wsi bit from anv_image 577- radv: fix memory leak of descriptor set layout 578 579Boris Brezillon (81): 580 581- pan/midg: Add intra-bundle interferences 582- pan/midg: Remove spurious printf() in print_vector_constants() 583- pan/midg: Prefix scalar immediates with '#' instead of '<' 584- pan/midg: Fix swizzling on 8-bit sources 585- pan/midg: Fix 64-bit swizzle printer 586- pan/midg: Fix the upper/lower limit on 8bit vectors 587- pan/midg: Fix swizzle packing on 64bit instructions with src-expansion + dst-shrinking 588- pan/midg: Add a pass to lower non-logbase2 global/shared loads 589- pan/midg: Support 8/16 bit load/store 590- vulkan/wsi: Use ALIGN_POT() instead of open-coding it 591- vulkan/wsi: Don't open-code vk_format_get_blocksize() 592- panvk: Add support for push constants 593- panvk: Add support for storage buffers 594- panvk: Support creation of compute pipelines 595- panvk: Add support for storage/uniform buffers with dynamic offsets 596- panvk: Move dummy attribute buffer emission out of emit_{attribute,varying}_bufs 597- panvk: Add support for storage image 598- panvk: Implement vkCmdDispatch() 599- vulkan/cmd_queue: Constify vk_cmd_queue.alloc 600- vulkan/cmd_queue: Remove duplicate entries in MANUAL_COMMANDS 601- vulkan/cmd_queue: Properly deconstify array of pointers 602- lavapipe: Re-use auto-generated vk_cmd_enqueue entrypoints 603- panvk: Add a dummy sampler for NIR tex operations that don't take one 604- panvk: Add support for texel buffers 605- panvk: No-op zero-vertex draws 606- panvk: Fix per-instance attribute handling 607- panvk: Implement indexed rendering 608- vulkan/cmd_queue: Track allocation errors in vk_cmd_queue 609- vulkan/cmd_queue: Fix the allocation scope 610- panvk: Refcount the descriptor set and pipeline layouts 611- aco: Fix an MSVC warning 612- amd: Fix ac_gpu_info.c compilation on windows 613- radv: Don't use VK_OUTARRAY_MAKE()/vk_outarray_append() 614- lavapipe: Don't use VK_OUTARRAY_MAKE()/vk_outarray_append() 615- Revert "ci: Disable windows-vs2019" 616- vulkan/runtime: Add vk_cmd_queue.h to idep_vulkan_runtime_headers 617- v3dv: Stop using VK_OUTARRAY_MAKE() 618- turnip: Stop using VK_OUTARRAY_MAKE() 619- pvr: Stop using VK_OUTARRAY_MAKE() 620- anv: Stop using VK_OUTARRAY_MAKE() 621- panvk: Stop using VK_OUTARRAY_MAKE() 622- venus: Stop using VK_OUTARRAY_MAKE() 623- vulkan/device_select: Stop using VK_OUTARRAY_MAKE() 624- vulkan/wsi: Stop using VK_OUTARRAY_MAKE() 625- vulkan/util: Get rid of VK_OUTARRAY_MAKE() 626- vulkan/image: Make MSVC C++ compiler happy 627- vulkan/util: Make STACK_ARRAY() work for arrays of pointers 628- dzn: Compile-test the driver 629- dzn: Add Missing return type to dzn_translate_sampler_filter() 630- dzn: Remove the dzn_cmd_exec_functions file 631- dzn: Make a bunch of functions private 632- dzn: Fix alpha blend factor translation 633- dzn: Properly support static blend constants 634- dzn: Pass a NULL ralloc context to dxil_create_validator() 635- dzn: Fix pipeline creation when rasterization is disabled 636- dzn: Fix dzn_translate_viewport() when height < 0 637- dzn: Don't crash when EndCommandBuffer() returns an error 638- dzn: Align the default case in dzn_image_view_prepare_dsv_desc() 639- dzn: Replace C++ references by pointers 640- dzn: Fix dzn_image_get_rtv_desc() for 3D views 641- dzn: Support 2Darray views on 3D images for color attachments 642- dzn: Fix 3D <-> 2D image copies 643- dzn: Return a valid imageFormatProperties.maxMipLevels 644- dzn: Make sure the properties are all zero when the format is not supported 645- dzn: Set bufferFeatures to zero on depth/stencil formats 646- dzn: 3D array images don't exist 647- dzn: Get rid of dzn_GetPhysicalDeviceFeatures() 648- dzn: Get rid of dzn_GetPhysicalDeviceProperties() 649- dzn: Force sampleCounts to 1 for bgra4 images 650- dzn: Check image view usage instead of image usage when creating an image view 651- dzn: Drop extra blank line in dzn_CmdCopyImage2() 652- vulkan/util: Make STACK_ARRAY() C++-friendly 653- dzn: Lower partial copy of multisample resources to blits 654- dzn: Fix 2D <-> 3D blits 655- dzn: Support independent depth/stencil access 656- dzn: Fix loop condition in dzn_descriptor_set_copy() 657- dzn: Pass the right type to CreateCommandList() in the reset path 658- dzn: Add missing VKAPI_{ATTR,CALL} specifiers to BeginCommandBuffer() 659- dzn: Pass the dzn_event pointer to _mesa_hash_table_insert() 660- dzn: Fix the STATIC_ASSERT() in dzn_meta_blits_get_context() 661- ci/windows: Add a variable to globally disable jobs using windows runners 662 663Brian Paul (1): 664 665- vulkan/wsi/x11: add null pointer check for the has_dri3_v1_2 test 666 667Caio Oliveira (7): 668 669- anv: Enable requiredSubgroupSize for Task/Mesh 670- intel/compiler: Lower Task/Mesh I/O before SIMD specific lowering 671- intel/compiler: Use pass helper in brw_nir_adjust_offset_for_arrayed_indices 672- intel/fs: Initialize the sample mask in flags register when using demote 673- intel/fs: Fix IsHelperInvocation for the case no discard/demote are used 674- intel/compiler: Use nir_var_mem_task_payload 675- intel/compiler: Inline TUE map computation into TUE Input lowering 676 677Chad Versace (1): 678 679- intel/tools: Fix build without drivers 680 681Charles Baker (2): 682 683- zink: Fix MSVC RTC in zink_get_framebuffer_imageless() 684- mesa: align constant/uniform uploads to driver expected alignment 685 686Charlie Turner (8): 687 688- ci, valve: Add support scripts for the Valve bare-metal farm. 689- amd, ci: Remove unused runners. 690- amd, ci: Drop log level in SPIRV -> NIR code generator. 691- amd, ci: Categorize the sections of the CI file. 692- ci, valve: Add the dEQP runners for Valve CI 693- ci, valve: Show real kernel addresses in KFENCE reports. 694- ci, valve: Bump the trigger container 695- ci, radv: Update flake expectations 696 697Charmaine Lee (1): 698 699- mesa: fix misaligned pointer returned by dlist_alloc 700 701Chia-I Wu (20): 702 703- zink: set dma-buf bit for shared resources 704- zink: always chain wsi_image_create_info for scanout images 705- zink: set needs_mesa_flush_wsi for venus 706- venus: fix two VN_TRACE_SCOPE's in the same scope 707- venus: use 64KB alignment for suballocations 708- venus: cache VkFormatProperties 709- venus: trace vn_ring_wait_space 710- venus: abort when stuck 711- venus: fix properties of unsupported external fences/semaphores 712- venus: update venus-protocol headers 713- venus: add VK_EXT_shader_demote_to_helper_invocation 714- venus: add VK_EXT_conservative_rasterization 715- venus: add VK_EXT_depth_clip_enable 716- venus: add VK_EXT_robustness2 717- venus: add VK_EXT_shader_stencil_export 718- venus: add VK_EXT_vertex_attribute_divisor 719- venus: add vn_extension_get_spec_version 720- venus: update venus-protocol headers 721- util: Keep quiet NaNs quiet when converting to half float. 722- anv: advertise rectangularLines only for Gen10+ 723 724Christian Gmeiner (3): 725 726- etnaviv: drop TGSI based backend compiler 727- Revert "nir: make tgsi_varying_semantic_to_slot(..) public" 728- nir: Use const for nir_shader_get_entrypoint(..) 729 730Connor Abbott (57): 731 732- ir3/spill: Fix simplify_phi_nodes with multiple loop nesting 733- nir/serialize: Don't access blob->data directly 734- util/blob: Clarify rules on blob::data 735- spirv: Rewrite determinant calculation 736- ir3: Use CAN_REORDER instead of NON_WRITEABLE 737- freedreno: Replace A6XX_IBO with A6XX_TEX_CONST 738- freedreno/fdl: Set swizzle on storage descriptor 739- ir3: Don't always set bindless_tex with readonly images 740- ir3/nir: Fix 1d array readonly images 741- tu: Call nir_opt_access 742- ir3: Use isam for bindless images 743- ir3/cp_postsched: Support multiple destinations 744- ir3/dce: Support multiple destinations 745- ir3/sched: Support multiple destinations 746- ir3/ra: Fix tied destination handling with multiple destinations 747- ir3/ra: Fix ra_foreach_dst_n 748- ir3/ra: Sanitize parallel copy flags better 749- ir3/ra: Add proper support for multiple destinations 750- ir3/ra: Add IR3_REG_EARLY_CLOBBER 751- ir3/spill: Mark reload destination as early-clobber 752- util/bitset: Fix off-by-one in __bitset_set_range 753- ir3: Track physical edges when inserting (ss) for shared regs 754- ir3: Add support for subgroup arithmetic 755- tu: Expose subgroup arithmetic 756- tu: Rewrite dynamic descriptor handling 757- tu: Handle UBO/SSBO descriptors with different sizes 758- tu: Add an extra storage descriptor for isam 759- ir3/parser: Don't use right recursion 760- ir3/lower_spill: Fix corner case with oob offsets 761- ir3: Actually use wrmask in emit_sam 762- ir3: Use isam for bindless readonly ssbo loads 763- nir: Add preamble functions 764- nir: Add a "deep" instruction clone 765- nir: Add a preamble optimization pass 766- ir3: Fix scan.macro valid flags 767- ir3: Don't count reserved user consts in ubo_state::size 768- ir3: Implement basic shader preamble intrinsics 769- ir3: Better assemble/disassemble stc 770- ir3: Plumb through store_uniform_ir3 intrinsic 771- ir3/legalize: Handle inserting (ei) with preamble 772- ir3: Support prefetching with preambles 773- ir3: Insert frag coord code after preamble 774- ir3: Don't include preamble instructions in stats 775- ir3: Add preamble optimization pass 776- ir3: Implement and document ldc.k 777- ir3: Refactor ir3_compiler_create() to take an options struct 778- ir3, turnip: Use ldc.k to push UBOs 779- tu: Enable UniformBufferUpdateAfterBind 780- tu: Trivially implement VK_EXT_texel_buffer_alignment 781- tu: Actually expose VK_EXT_texel_buffer_alignment 782- tu: Correctly handle VK_IMAGE_CREATE_EXTENDED_USAGE_BIT 783- freedreno/ci: Fix skip comment 784- tu: Implement GetDevice*MemoryRequirements() 785- tu: Fill out maxBufferSize 786- tu: Remove tu_pipeline::layout 787- tu: Expose VK_KHR_maintenance4 788- freedreno/a6xx: Fix SP_DS_CTRL_REG0 definition 789 790Corentin Noël (6): 791 792- ci: Uprev virglrenderer and crosvm 793- virgl: Update virgl_protocol and use the provided constants 794- nir_to_tgsi: Require the block index to always be populated 795- nir_to_tgsi: Handle blocks defined as arrays of arrays 796- ci: Only apply patches with the build-skqp prefix 797- virgl/ci: Uprev virglrenderer and crosvm 798 799Cristian Ciocaltea (29): 800 801- ci: Ensure Mesa Shader Cache resides on tmpfs 802- ci: Add socat utility 803- ci: Enable kernel virtio transport for Virtual Sockets 804- virgl/ci: Setup virtio-vsock based IPC 805- ci: Increase limit of concurrent crosvm instances per runner 806- ci: Improve interrupt signal handling in crosvm-runner.sh 807- ci/zink: Report flake test 808- radeonsi/ci: Mark a bunch of flaky tests on stoney 809- ci: Avoid altering EXTRA_CARGO_ARGS environment variable 810- ci: Convert generate-env.sh to a POSIX compliant script 811- ci: Set CI_JOB_JWT_FILE to a fixed path outside /tmp 812- ci: Build crosvm for LAVA runners 813- ci: Add crosvm runtime dependencies for LAVA 814- ci: Enable KVM_AMD and KVM_INTEL kernel modules 815- ci: Load KVM kernel module for LAVA runners 816- ci: Make kernel image available in LAVA for KVM use cases 817- ci: Use script relative paths in crosvm-runner 818- ci: Add PIGLIT_REPLAY_LOOP_TIMES to generate-env.sh 819- ci: Remove obsolete CROSVM_TEST_SCRIPT env var 820- ci: Provide consistent results location in LAVA 821- ci: Allow specifying any shell command via HWCI_TEST_SCRIPT 822- ci: Dynamically adjust LIBGL_ALWAYS_SOFTWARE for crosvm 823- virgl/ci: Add jobs for running trace tests on LAVA 824- ci: Make bash available in LAVA rootfs 825- Revert "ci: Convert generate-env.sh to a POSIX compliant script" 826- virgl/ci: Add support for dEQP GL vtest-ing 827- ci: Add Intel GPU frequency utility 828- ci: Provide intel-gpu-freq.sh in LAVA and bare-metal rootfs 829- ci: Lock Intel GPU frequency for performance tests 830 831Daniel Schürmann (39): 832 833- aco: emit nir_intrinsic_discard() as p_discard_if() 834- aco: remove block_kind_discard 835- aco: make Preserve_WQM independent from block_kind_uses_discard_if 836- aco: merge block_kind_uses_[demote|discard_if] 837- aco: optimize discard_if when WQM is not needed afterwards 838- radv: remove exports without color attachment or writemask 839- nir: split nir_opt_shrink_stores from nir_opt_shrink_vectors 840- nir/opt_shrink_vectors: Remove shrinking of store intrinsics data source 841- radv: move nir_opt_shrink_stores from radv_optimize_nir() 842- aco/insert_exec_mask: stay in WQM while helper lanes are still needed 843- aco: don't propagate WQM for p_as_uniform 844- aco: don't emit WQM for bool_to_scalar_condition 845- aco/insert_exec_mask: remove Preserve_WQM flag 846- aco/insert_exec_mask: remove some unnecessary WQM loop handling code 847- aco/insert_exec_mask: remove ever_again_needs and Exact_Branch 848- aco/insert_exec_mask: refactor and simplify get_block_needs() 849- aco/insert_exec_mask: refactor and remove some unnecessary WQM handling code 850- aco: relax condition to remove branches in case of few instructions 851- aco/ra: don't immediately assign a register for p_branch 852- nir/opt_shrink_vectors: shrink load_const properly 853- nir/opt_shrink_vectors: remove duplicate components from vecN 854- nir/opt_shrink_vectors: update docstring 855- aco/ra: count constant moves in get_reg_create_vector() 856- aco/ra: special-case get_reg_for_create_vector_copy() 857- aco/ra: refactor find_vars() to return a vector 858- aco/ra: refactor collect_vars() to return a sorted vector 859- nir: rename nir_src_is_dynamically_uniform to nir_src_is_always_uniform 860- aco/optimizer: fix call to can_use_opsel() in apply_insert() 861- aco: remove 'high' parameter from can_use_opsel() 862- aco: use branch definition as scratch register for SSA lowering 863- aco/ra: fix stride check on subdword parallelcopies for create_vector 864- aco/optimizer: check recursively if we can eliminate s_and exec 865- aco/ra: only use VCC if program->needs_vcc == true 866- aco/ra: create VCC-affinities during RA 867- aco/ra: omit VCC affinity on VOPC_SDWA for GFX9+ 868- aco: make program->needs_vcc independent of VCC hints 869- aco: remove occurences of VCC hint 870- aco: remove register hints entirely 871- aco/ra: fix live-range splits of phi definitions 872 873Daniel Stone (11): 874 875- egl/wayland: Reset buffer age when destroying buffers 876- egl/wayland: Don't replace existing backbuffer in get_buffers 877- ci: Disable Windows for now 878- Revert "ci: Disable Windows for now" 879- CI: Disable panfrost-t760 880- CI: Disable Panfrost T720 jobs 881- Revert "CI: Disable panfrost-t760" 882- Revert "CI: Disable Panfrost T720 jobs" 883- ci: Add new Panfrost G52 skip 884- CI: Disable Windows jobs 885- ci: Also disable Windows container builds when down 886 887Danylo Piliaiev (37): 888 889- turnip: Add TU_GMEM envvar to test different gmem sizes 890- turnip: Do not use hw binning if tiles per pipe are over the limit 891- turnip/doc: Update turnip extension list 892- turnip: Use the shared helpers to expose 1.3 core extensions/limits 893- turnip: Expose VK_EXT_image_robustness 894- turnip: Implement VK_KHR_zero_initialize_workgroup_memory 895- turnip: Expose VK_KHR_shader_non_semantic_info 896- turnip: Unconditionaly remove descriptor set from pool's list on free 897- tu: Implement VK_AMD_buffer_marker to support Graphics Flight Recorder 898- turnip: Depth/stencil formats should not expose any bufferFeatures 899- freedreno/pps: Expose same counters as blob 900- ir3: Limit the maximum imm offset in nir_opt_offset for shared vars 901- turnip: Add a refcount mechanism to BOs 902- turnip: Use LATE_Z when there might be depth/stencil feedback loop 903- turnip: Merge LRZ and DEPTH_PLANE draw states 904- turnip: Implement VK_ARM_rasterization_order_attachment_access 905- turnip: Always use GMEM for feedback loops in autotuner 906- turnip: Set drmFormatModifierTilingFeatures 907- turnip: Implement VK_EXT_physical_device_drm 908- turnip: Add "rast_order" debug option to force rast order access 909- tu: Refactor VS DECODE/DEST to be emitted in two pkt4 910- turnip: Make autotuner work with reusable command buffers 911- tu: Implement VK_EXT_depth_clip_control 912- turnip: Force linear mode for non-ubwc R8G8 formats 913- turnip: Disallow non-linear tiling when casting R8G8 to other fmts 914- turnip: Fix the lack of WFM before indirect draws 915- turnip: enable has_ccu_flush_bug workaround for a660 916- turnip: Use correct type for OUTARRAY in FormatProperties2 917- turnip: Correctly store separate stencil in gmem store 918- turnip: Ignore aspectMask for D32S8 framebuffer attachment 919- turnip: Add "unaligned_store" debug option to better test gmem stores 920- turnip: Fix subpassLoad from CUBE input attachments 921- ci/freedreno: Add fractional test of forced unaligned gmem store 922- turnip: Implement VK_EXT_primitives_generated_query 923- tu: Do not flush ccu in clear/blits during renderpass 924- pps: Open writable renderer node in DrmDevice::create 925- tu: Fix indices of drm_msm_gem_submit_cmd when filling them 926 927Dave Airlie (85): 928 929- llvmpipe: optimise triangle setup a bit. 930- llvmpipe: inline retry_triangle_ccw 931- llvmpipe/setup: remove opaque from setup triangle 932- llvmpipe: refactor lp_rast_shader_inputs. 933- llvmpipe: just move opaque alpha lookup closer to use. 934- llvmpipe/triangle: don't store area in fixed_position. 935- crocus: find correct relocation target for the bo. 936- lavapipe: fix sampler + sampler view leaks. 937- ci/lavapipe: update lvp asan results after leak fixes. 938- gallivm: add coroutine attribute that llvm requires. 939- gallivm/st/lvp: add flags arg to get_query_result_resource api. 940- gallivm: fix missing cast in 4-bit blending paths. 941- llvmpipe: fix linear rast samples check. 942- lavapipe: reference gallium fences correctly. 943- crocus: fix leak on gen4/5 stencil fallback blit path. 944- drisw: fence drawing to the swap/copy buffers. 945- lavapipe: fix pipeline statistic query results with availability. 946- lavapipe: handle endless fence timeout properly. 947- lavapipe: execute a finish in pipeline barrier and event waiting. 948- lavapipe: don't flush on transfer operations. 949- lavapipe: handle non-timeline semaphores wait/signal. 950- llvmpipe: convert texture barrier to a finish. 951- llvmpipe/flush: always finish whether for cpu/gpu access. 952- llvmpipe/scene: move to slab allocated objects for scenes. 953- llvmpipe: base the scene queue size of the max number of scenes. 954- llvmpipe: handle dynamically creating scenes when needed 955- llvmpipe: size initial allocation and free scenes 956- llvmpipe: add writeable resource tracking to the scene. 957- llvmpipe: pass ssbo write mask down into setup. 958- llvmpipe: add ssbo to resources reference by scenes. 959- llvmpipe: add images to the scene resource tracker. 960- llvmpipe: check framebuffer resources for all scenes for references. 961- gallium: add partial bit to the query flags. 962- llvmpipe/query: add support for partial query waits. 963- lavapipe: pass partial results flags through. 964- llvmpipe: add support for fence_server_sync. 965- llvmpipe: allow vertex processing and fragment processing in parallel 966- ci/lavapipe: fixup results after proper reference counting. 967- draw/so: don't use pre clip pos if we have a tes either. 968- vulkan/wsi: handle queue families properly for non-concurrent sharing mode. 969- lavapipe: always set read/write on ssbo/images. 970- llvmpipe/linear: fix disk caching. 971- gallivm/nir: split load_const out into backend helper. 972- gallivm/llvmpipe: add support for NIR to the linear/aos paths. 973- crocus: change the line width workaround for gfx4/5 974- gallivm/nir: extract a valid texture index according to exec_mask. 975- zink: workaround depth texture mode alpha. 976- lavapipe: remove broken workaround for zink depth texturing. 977- crocus: don't map scanout buffers as write-back 978- radv: abstract queue family away from queue family index. 979- intel: add some missing debug recompile info. 980- crocus: force ignore_sample_mask_out on gen4/5 for precompile 981- radv: try and fix internal transfer queue mapping 982- radv/winsys: complete ring/ip translations. 983- radv/winsys: add nop packets for uvd and vcn dec. 984- radv/winsys: add a ring level detection for ib bo usage. 985- radv/winsys: add support for queues without user fences. 986- llvmpipe/fs: add missing depth_clamp key printing 987- vulkan: update vk video headers for new vulkan headers. 988- clover/nir: respect lower to scalar options. 989- lavapipe: add EXT_texel_buffer_alignment support. 990- gallivm/sample: detect if rho is inf or nan and flush to zero. 991- zink: update resource layout in copy_scanout 992- util/format: add new z24/s8 packing helper to pack z32/s8. 993- u_transfer: refactor out code to check interleave/deinterleave path. 994- u_transfer_helper: add a new option for handling z24 stored in z32 995- lavapipe: add loop unrolling. 996- Reinstate: llvmpipe: allow vertex processing and fragment processing in parallel 997- zink/query: consolidate xfb_buffers into one array. 998- zink/query: collapse the xfb_query_pool array into the normal one. 999- zink: refactor out number of vk queries per gallium query helper 1000- zink/query: use a single query pool for XFB queries. 1001- zink/query: refactor get_query_result to map upfront. 1002- zink/query: only reset the range of queries in use. 1003- zink/query: rewrite the query handling code to pass validation. 1004- llvmpipe: fix nr_sampler_view in key creation. 1005- radv: use flush vgt streamout like PAL does. 1006- zink: fix tessellation shader key matching. 1007- zink/query: refactor out vk queries and allow sharing them 1008- draw: handle tess eval shader when getting num outputs 1009- u_blitter/stencil: take dstbox x/y into accounts for dst fb width 1010- util/stencil: fix stencil fallback blit shader texture types. 1011- vulkan/wsi: keep allocate queue families in image, just don't fill them 1012- llvmpipe: add user memory resources to the debug global list. 1013- meson: add build-id to pipe libraries 1014 1015Dmitry Baryshkov (2): 1016 1017- freedreno/registers: add new register for 7nm DSI PHY v4.3 (sm8450) 1018- freedreno/regs: remove 5nm DSI PHY regs 1019 1020Dylan Baker (56): 1021 1022- VERSION: bump version for 22.0 release 1023- docs: reset new_features.txt 1024- docs: update calendar for 22.0.0-rc1 1025- meson: add support for \`meson devenv` with vulkan 1026- meson: add LIBGL_DRIVERS_PATH to the devenv 1027- meson: add radv to meson devenv 1028- docs: update calendar for 22.0.0-rc2 1029- docs: update calendar and link releases notes for 22.0.0 1030- docs: Add calendar entries for 22.0 release. 1031- docs: add release notes for 22.0.0 1032- util/list.h: Add docstrings for list_add and list_addtail 1033- mesa/main: replace use of simple_list with util/list 1034- gallium/opencl: set OCL_ICD_FILENAMES with devenv 1035- docs: Add calendar entries for 22.1 release candidates. 1036- docs: add release notes for 22.0.1 1037- docs: add sah256 sum for mesa 22.0.1 1038- docs: update calendar and link releases notes for 22.0.1 1039- VESRION: bump for 22.1.0-rc1 release 1040- .pick_status.json: Update to e509598470c000498c3e1328e012142d9047e292 1041- VERSION: bump for 22.1.0-rc2 1042- .pick_status.json: Update to 9f44a264623461c98368185b023d99446676e039 1043- .pick_status.json: Update to fbece25a451bb7915891851ee5c72724974ae5e2 1044- .pick_status.json: Update to a6a4bf0f1eae36cb68d5c67653ac013fe0fbde8a 1045- .pick_status.json: Update to f329f67243d671965d73bd2243cffc4e1e68c4a3 1046- VERSION: bump for 22.1.0-rc3 1047- .pick_status.json: Update to f17d3a125126dd7e0f1c2072545976cec1eeb1c8 1048- .pick_status.json: Update to 1dc697942af7254798f83583f5788e1abf136d27 1049- .pick_status.json: Update to 2505afc430c6b151e7e4243503953ba0b0a2c842 1050- .pick_status.json: Mark c025cb9ee9d79ebfb66a577556e04deecfe012ed as backported 1051- .pick_status.json: Update to 575068a1656ab4303647ade1491da7d711d36db7 1052- VERSION: bump for 22.1.0-rc4 release 1053- .pick_status.json: Update to 7f91e8fad94dd34f83c6a124dbbe5d210be7715f 1054- .pick_status.json: Update to 17c98393f9f3cb0801b73b2d4c62442859417368 1055- .pick_status.json: Update to 14b1ed1ce105d42652f70e2fd13c90fc4f2e7ffc 1056- .pick_status.json: Mark 6317f88b044501354a052064478d5b43dfe41809 as backported 1057- .pick_status.json: Mark 5ff3fa5912778adb8117fa26bfe4786b583e741b as backported 1058- .pick_status.json: Mark 9a412c10b7a96adf71c9a2ca44a0abca75de1c49 as backported 1059- .pick_status.json: Mark 0e49ef5c9f2ec34567613226ad498edca28bce88 as backported 1060- .pick_status.json: Mark ae369e9f6d4f2c826f1f2a748c32a14f9d5f1f54 as backported 1061- .pick_status.json: Mark 8b28d1751c76b0ba72dae1f6b916cb7f2bbc1c0c as backported 1062- .pick_status.json: Mark 8daf8ff730b777bb512a237a36783d781cbf9de9 as backported 1063- .pick_status.json: Mark 8c1d9c7b744b2e5b40fd42cfd51256b16deea6a8 as denominated 1064- VERSION: bump for 22.1.0-rc5 1065- .pick_status.json: Update to 57293dee2b11ba7e52052edc4d0437f08db19144 1066- .pick_status.json: Mark 5a3aee78cbb70918b413cdd40dffcae7c9e97d8c as denominated 1067- .pick_status.json: Update to 5c90eb1c53f46e86717c6bf4d5253dd23c4dac1f 1068- .pick_status.json: Update to 07eba9a15a06ceda3469892822e8b539effc6788 1069- .pick_status.json: Update to 29a8f1f03bed317370e07288fd9d4b85a24562e6 1070- .pick_status.json: Mark 07efe6f129956d0cce0fb167d08b5f8af25d0c95 as backported 1071- .pick_status.json: Mark 1c17502ab38a238efe5d9aa4bc0ff07c967adb80 as backported 1072- .pick_status.json: Mark b6eec12327b65038dd56742dbe272c21c63a3844 as backported 1073- .pick_status.json: Mark 7057a363591279c3f65a3c4af41687e6f8142649 as backported 1074- .pick_status.json: Mark 9924fecee6412f1cad65d7d7347681856e3a525c as backported 1075- .pick_status.json: Mark 7aee9f12b9c1f17ae2662c8043c8679ce698b731 as backported 1076- .pick_status.json: Mark 5e4009fe6c91c34a7cdf8c16aa5aa53c3ff41222 as backported 1077- .pick_status.json: Mark b18448e7acfc765ee95505c7cd5e2ce6e6fc36cc as backported 1078 1079Emma Anholt (145): 1080 1081- r300: Fix missing \\n in an error message. 1082- r300: Set up shadow sampler lowering in precompiles. 1083- r300: Simplify DCE by assuming all output writes are used. 1084- ci/freereno: Reduce run-by-default a630-vk coverage. 1085- ci/turnip: Extend the full-vk-run job timeouts. 1086- ci/freedreno: Reduce concurrency for a618 vk_full. 1087- ci/i915: Update rendering hash for plot3d trace. 1088- llvmpipe: Disable an assertion that may not be quite right. 1089- ci: Uprev vulkan-cts to 1.2.8.0 1090- ci/softpipe,llvmpipe: Disable Xvfb server reset on piglit runs. 1091- ci/crocus: Add recent flakes from #intel-ci 1092- ci/iris: Add skips and flakes notes for recent #intel-ci logs. 1093- ci/freedreno: Add another unsizedArrayLength flake. 1094- ci/panfrost: Add a flake a few of us have run into in the last couple days. 1095- ci/broadcom: Remove unused v3dv xfails file. 1096- ci: Bump VK-GL-CTS to 1.3.1.0. 1097- r300: Request that nir-to-tgsi avoid generating TGSI_OPCODE_CMP. 1098- r300: Demote a compiler assert(0) to a compile failure. 1099- r300: Throw a compile error instead of an assert in r300 swizzle rewrites. 1100- nir: Split the flag for lowering of fabs and fneg to source modifiers. 1101- nir_to_tgsi: Add a flag for lowering fabs, and use it in r300/i915. 1102- i915g: Report the temps usage 1103- tgsi: Refactor out a tgsi_util_get_src_usage_mask(). 1104- nir_to_tgsi: Track our TGSI insns in blocks before emitting tokens. 1105- nir_to_tgsi: Replace the NIR SSA liveness with TGSI reg-level liveness. 1106- nir: Delete the per-instr SSA liveness impl. 1107- ci/r300: Drop xfails that were fixed with the VK-GL-CTS 1.3.1.0 uprev. 1108- ci/lvp: Add a flake that's shown up a couple of times since VKCTS 1.3.1. 1109- ci/freedreno: Try to detect a wedged MMU that's happened recently. 1110- tgsi_translate: Make the procType public when translating. 1111- virgl: Work around old virglrenderer's BARRIER counting bug. 1112- virgl: Add a workaround for virglrenderer output writemask bugs. 1113- virgl: Apply TGSI transforms to compute shaders, too. 1114- virgl: Add workarounds for virglrenderer input/sv signedness bugs. 1115- virgl: Move tex immediate operands to a temp to avoid virglrenderer bug. 1116- virgl: Move double operands to a temp to avoid double-swizzling bugs. 1117- nir_to_tgsi: Don't vectorize 64-bit instructions, to keep virgl happy. 1118- nir_to_tgsi: Add support for FBFETCH. 1119- r300: Delete the loop unrolling. 1120- i915g: Initialize the rest of the "from_nir" temporary VS struct. 1121- draw: Don't look at .nir if !IR_NIR. 1122- nv30/40: Switch to using NIR-to-TGSI by default. 1123- nir: Add some notes about const/uniform array access rules in GL. 1124- ci/freedreno: Move a 60s timeout test to skips instead of flakes. 1125- ci/freedreno: Cut down pre-merge a630 VK coverage. 1126- ci/freedreno: Add a known spilling hangcheck flake. 1127- turnip: Request no implicit sync when we have no implicit-sync WSI BOs. 1128- ci: Stash the ldd and ccache stats output under collapsed sections. 1129- ci/llvmpipe: Move most of testing to shared 64-core runners at Google. 1130- ci/lavapipe: Test 1/3 of lavapipe on the shared 64-core google runners. 1131- ci/softpipe: Move most of testing to shared 64-core runners at Google. 1132- ci/virgl: Drop the bvec4_from_mat4x2_vs xfail. 1133- ci/lvp: Update the asan fails list. 1134- ci/zink: Move testing to shared 64-core runners at Google. 1135- ci/zink: Add testing of dEQP GLES3.1/3.2. 1136- nir: Allow the _replicates opcodes to have num_components != 4. 1137- freedreno: Use the resource size rather than BO size for VFD_FETCH[].SIZE. 1138- freedreno: Fix start_slot handling in set_vertex_buffers. 1139- freedreno: Improve robustness behavior for VBs with offset > size. 1140- ir3: Don't assert on not finding the VS output for an FS input. 1141- ci/freedreno: Consolidate some information about an a630 flake. 1142- nir: Introduce a nir_vec_scalars() helper using nir_ssa_scalar. 1143- nir: Add a helper for setting up a nir_ssa_scalar struct. 1144- nir: Switch to using nir_vec_scalars() for things that used nir_channel(). 1145- intel/perf: Move some static blocks of C code out of the python script. 1146- ci/bare-metal: Drop the BM_POE_USERNAME/PASSWORD env var checks. 1147- ci/bare-metal: Increase maximum retry count for POE boots. 1148- ci: Stop xz-compressing firmware for ramdisks. 1149- ci/nouveau: Add nouveau support to the rootfs. 1150- ci/nouveau: Add a manual run for the Jetson Nano (GM20B). 1151- docs/ci: Update some bare-metal CI docs. 1152- docs/ci: Add docs for using a POE switch to control boards, like nouveau. 1153- turnip: Enable VK_EXT_display_control using the common code. 1154- turnip: use vk_shader_module_to_nir(). 1155- ci: Drop skips of spv-stable-pillars-volatile-nontemporal-store 1156- ci/lvp: Stop skipping spv-stable-maze-flatten-copy-composite 1157- ci/freedreno: Drop the skips of spirv_ids_abuse in pre-merge. 1158- ci/freedreno: Remove some xfails for tests that now skip. 1159- turnip: Use the DRM or KGSL GPU reset status ioctls to report device loss. 1160- ci/turnip: Drop alpha_to-coverage flake note on a618. 1161- vulkan: Make sure we've loaded our connectors when querying plane props. 1162- ci/turnip: Increase the hangcheck timer to 2 seconds. 1163- ci/traces: Make sure we have no pre-existing traces-db before starting. 1164- nouveau/nir: Enable nir_opt_move/sink. 1165- r600: Drop nr_ps_max_color_exports 1166- r600: Update the PS state before checking for cb_misc update. 1167- r600: Update the PS state when MSAA-ness changes, too. 1168- r600: Add shader precompile and shader-db support. 1169- ci/r600: Check in some expectation files for rv770 and Turks. 1170- r600: Add shader-compiler debug knobs to the shader cache key. 1171- r600: Drop unused sbcl debug option. 1172- r600: Drop unused debug options from the fork off of radeonsi. 1173- r600: Fix ordering of SSBO loads versus texturing. 1174- r600: Add a workaround and explanation for shadowcubearray TG4. 1175- r600: Disable SB in the presence of indirection on temp arrays. 1176- r600: Disable SB when INTERP_SAMPLE is used. 1177- r600: Implement memoryBarrier() in the non-SFN path. 1178- glsl: Delete unused EmitNoPow path. 1179- glsl: Delete dont_lower_swz path of lower_quadop_vector. 1180- nir_to_tgsi: Add support for nir_intrinsic_image_samples. 1181- turnip: Add support for VK_KHR_format_feature_flags2. 1182- freedreno/a6xx: Set the color_swap field for storage descriptors. 1183- turnip: Disable tiling on 1D images. 1184- turnip: Allow image access on swapped formats. 1185- util/log: Don't print an extra \\n if the format string had one. 1186- spirv: Silence "Decoration not allowed on struct members: SpvDecorationRestrict" 1187- ci/deqp: Add gitlab-ci sections to deqp-runner.sh. 1188- ci/deqp: Move the set +e just before the deqp-runner invocation. 1189- ci/lava: Simplify passthrough of the request to upload results/ to minio. 1190- util/log: Add support for logging once. 1191- gallivm/nir: Don't do uniform-and-broadcast access on inactive invocations. 1192- gallivm/nir: Pull some repeated exec_mask computation out of loops. 1193- gallivm/nir: Refactor out some repeated logic for SSBO/shared access. 1194- gallivm/nir: Refactor out some repeated code to generate 0 values. 1195- gallivm/nir: Add a short circuit uniform-offset mode for load_global. 1196- gallivm/nir: Add a short circuit uniform-offset mode for load_ssbo/load_shared. 1197- ci/turnip: Drop xfails for create_list_modifiers. 1198- ci: Uprev deqp-runner and piglit. 1199- freedreno: Fix the cpu-prep wait to be "infinite". 1200- Revert "ci/freedreno: Reduce concurrency when replaying traces on a630" 1201- virgl: Disable nir_op_ffloor to avoid sending DFLR to virglrenderer. 1202- nir_to_tgsi: Fix emitting the sample number for non-array MSAA image access. 1203- r600: Stop using ArrayID to look up atomic counters. 1204- r600: Add a helper function for rat_index_mode, with documentation and assert. 1205- nir_to_tgsi: Extract const components of atomic counter offsets into Index. 1206- nir_to_tgsi: Fix the address reg mapping for images and SSBOs to match G-T-T. 1207- tgsi/transform: Make tgsi_transform_shader() manage token allocation. 1208- tgsi/transform: Drop a stale comment. 1209- ci/iris: Disable pixmark-piano trace testing. 1210- ci/crocus: Disable pixmark-piano trace testing. 1211- st/glsl-to-tgsi: Fix handling of csel(bool, vec, vec). 1212- ci/softpipe: Mark some flakes that have appeared across a few MRs. 1213- ci/zink: Mark a new GLX flake that hit an innocent MR. 1214- turnip: Track refcounts on BOs in kgsl as well. 1215- turnip: Stop allocating unused pvtmem space in the pipeline CS. 1216- turnip: Sub-allocate pipelines out of a device-global BO pool. 1217- turnip: Skip telling the kernel the BO list when we don't need any. 1218- turnip: Reduce the pipeline's CS allocation a bit. 1219- turnip: Get autotune off of ralloc destructors. 1220- turnip: Move autotune buffers to suballoc. 1221- nir_to_tgsi: Do the required cleanup for nir_opt_find_array_copies(). 1222- Revert "ci: Disable Google's lab" 1223- nir: Add lowering for fround_even on r300. 1224- nouveau/nir: Fix the inverted sense of usesSampleMaskIn. 1225- nouveau/nir: Move FS output stores to the end of the last block. 1226 1227Enrico Galli (1): 1228 1229- microsoft/spirv_to_dxil: Add missing ralloc_free 1230 1231Eric Engestrom (7): 1232 1233- docs: add release notes for 21.3.6 1234- docs: update calendar and link releases notes for 21.3.6 1235- docs: add release notes for 21.3.7 1236- docs: update calendar and link releases notes for 21.3.7 1237- Revert "glx: Fix build errors with --enable-mangling (v2)" 1238- docs: add release notes for 21.3.8 1239- docs: update calendar and link releases notes for 21.3.8 1240 1241Erico Nunes (14): 1242 1243- lima/gpir: clean up override-init warnings 1244- lima/ppir: clean up override-init warnings 1245- lima/ppir: avoid ppir_codegen_outmod implicit conversion 1246- lima/ppir: initialize spill_costs array in regalloc 1247- lima: fix warning of garbage value access 1248- lima: add some checks for potential null pointer dereference 1249- lima: remove an unneeded lima_job_get assignment 1250- lima/gpir: avoid invalid write in regalloc 1251- lima/ppir: initialize slots array for dummy/undef 1252- lima/ppir: refactor bitcopy to use unsigned char 1253- lima/ci: update deqp results 1254- lima/ci: enable CI again 1255- lima/ci: enable piglit in lima CI 1256- lima: fix vector const src referenced multiple times 1257 1258Erik Faye-Lund (81): 1259 1260- vulkan/util: Add extern "C" to allow inclusion from c++ 1261- vulkan/util: Add explicit casts to make c++ happy 1262- vulkan/util: simplify multialloc init 1263- zink: do not copy colors through floats 1264- nir/spirv: guard macros in case of redefinition 1265- vulkan/wsi: pass win32-swapchain directly 1266- vulkan/wsi: untangle buffer-images from prime 1267- vulkan/wsi: add transition to/from transfer-src state 1268- vulkan/wsi: use buffer-image code-path on Windows 1269- docs: add license to the redirects script 1270- docs: remove incorrect drivers from extension 1271- docs: import virgl docs 1272- docs: remove a few repeated words 1273- docs: add missing get 1274- docs: qemu -> QEMU 1275- docs: Virgl -> VirGL 1276- docs: master -> main 1277- docs: link to gitlab instead of cgit 1278- docs: update virgl description 1279- docs: mark virgl gles2 renderer as done 1280- docs: update irc channel 1281- docs: fix a broken link 1282- docs: match build-flags markup with meson docs 1283- docs: fixup zink gl 4.3 requirements 1284- docs: improve language in zink article 1285- Revert "ci: downgrade sphinx to v3.x" 1286- gallium: rename pack half-float cap 1287- gallium: rename vs instance id cap 1288- gallium: rename layer-viewport caps 1289- gallium: rename pixel-coord caps 1290- gallium: rename fine derivative cap 1291- gallium: rename sysval caps 1292- gallium: rename array-components cap 1293- gallium: rename read-outputs cap 1294- gallium: rename texture query samples cap 1295- gallium: rename group-vote cap 1296- gallium: rename clock cap 1297- gallium: rename ballot cap 1298- gallium: rename window-space position cap 1299- gallium: rename image atomic float-add cap 1300- gallium: rename image atomic inc-wrap cap 1301- pvr: use zloadformat instead of zstoreformat 1302- pvr: fixup typos when allocating object 1303- microsoft/compiler: ralloc incoming phi-values 1304- microsoft/compiler: remove phi-value limit 1305- pvr: use a helper to translate compare-ops 1306- pvr: use a helper to translate stencil-ops 1307- pvr: zero-initialize variable 1308- docs: fixup breakage in release-calendar 1309- docs: add a minimal docs page for radv 1310- pvr: do not use fallthrough for unreachable code 1311- pvr: do not use fallthrough for unreachable code 1312- microsoft: Initial vulkan-on-12 driver 1313- vbo/dlist: do not try to pad an empty draw 1314- aco: do not use designated initializers 1315- nir/tests: do not use designated initializers in c++ code 1316- dozen: require c++20 for designated initializers 1317- ci: do not specify c_std and cpp_std for windows-build 1318- d3d12: fix return-code without dxcompiler.dll 1319- microsoft/compiler: add common dxil-validator API 1320- d3d12: use dxil_validator 1321- microsoft/spirv_to_dxil: use dxil_validator 1322- dzn: use dxil_validator 1323- clc/tests: use dxil_validator 1324- dzn: remove needless using 1325- dzn: remove unused struct 1326- dzn: drop needless includes 1327- dzn: add D3D12_IGNORE_SDK_LAYERS define 1328- dzn: fixup indent 1329- dzn: remove unused variable 1330- dzn: drop unused header 1331- dzn: drop incorrect return statement 1332- dzn: drop unused include 1333- dzn: add missing space 1334- wgl: rename force-msaa env-var 1335- util: limit error-dialogs to win32 1336- wgl: do not disable error-dialogs by default 1337- vulkan: explicitly cast object-type enum 1338- meson: remove unused defines 1339- gallium/xlib: fix stale comment 1340- mesa: add missing error-path 1341 1342Ernst Sjöstrand (1): 1343 1344- intel/compiler: Fix non-trivial designated initializer 1345 1346Felix DeGrood (4): 1347 1348- anv/measure: Fix INTEL_MEASURE for ANV 1349- anv: add indirect draw to INTEL_MEASURE 1350- intel: increase INTEL_MEASURE batch/buffer sizes 1351- intel: change INTEL_MEASURE output to microseconds 1352 1353Filip Gawin (2): 1354 1355- r300: fix swizzle handling in transformation of abs 1356- r300: Print warning when stubbing derivatives 1357 1358Francisco Jerez (5): 1359 1360- iris: Demote all callers of iris_flush_and_dirty_for_history() to iris_dirty_for_history(). 1361- iris: Remove remaining history flushes. 1362- iris: Replace unconditional QBO flush with iris_dirty_for_history(). 1363- intel/perf: Fix OA report accumulation on Gfx12+. 1364- intel/dev: Compute pixel pipe information based on geometry topology DRM query. 1365 1366Frank Binns (2): 1367 1368- pvr: Add a Vulkan driver for Imagination Technologies PowerVR Rogue GPUs 1369- pvr: fix clang unused function warning 1370 1371Ganesh Belgur Ramachandra (1): 1372 1373- radeonsi: NIR equivalent of si_create_clear_buffer_rmw_cs() 1374 1375Georg Lehmann (20): 1376 1377- nir/lower_mediump: Treat u2u16 like i2i16. 1378- radv, aco: Add u_foreach_bit to .clang-format. 1379- nir: Fix source type for fragment_fetch_amd. 1380- radv: Don't hash ycbcr sampler base object. 1381- nir/fold_16bit_sampler_conversions: Don't fold dest upcasts. 1382- nir/fold_16bit_sampler_conversions: Fix dest type mismatches. 1383- nir/fold_16bit_sampler_conversions: Fix src type mismatches. 1384- nir/legalize_16bit_sampler_srcs: Don't guess source type. 1385- radv: Add a vkCmdBuildAccelerationStructuresIndirectKHR stub. 1386- radv: Add more RT pipeline stubs. 1387- aco: Remove 0 data components from image stores. 1388- nir: Move lower_usub_sat64 to nir_lower_int64_options. 1389- nir: Add an option to lower 64bit iadd_sat. 1390- aco: Implement 64bit uadd_sat. 1391- aco: Implement scalar iadd_sat. 1392- radv: Lower 64bit iadd_sat. 1393- radv, aco: Packed iadd_sat/uadd_sat. 1394- radv: Enable global bo list if 1.2 features are used. 1395- wsi/x11: Don't leak xcb_get_geometry_reply_t. 1396- nir/opt_algebraic: Fix mask in shift by constant combining. 1397 1398Gert Wollny (24): 1399 1400- virgl: Enable PIPE_CAP_TGSI_TEXCOORD when the host supports it 1401- virgl: Fix texture transfers by using a staging resource 1402- virgl: Add a few more formats to the format table 1403- Revert "llvmpipe: allow vertex processing and fragment processing in parallel" 1404- Revert "virgl: Enable PIPE_CAP_TGSI_TEXCOORD when the host supports it" 1405- virgl: re-enable PIPE_CAP_TGSI_TEXCOORD with new host versions 1406- virgl: Don't support QUADS natively 1407- r600: Only emit the NOP group triggered by dest.rel after a full group 1408- virgl: Propagate precice flag through moves 1409- virgl: Always make some extra temps available for transformations 1410- virgl: Add an extra mov for int outputs from constant and immediate inputs 1411- r600: Don't limit scheduling of PARAM_SRC values 1412- r600: ignore dest sel for non-write targets when counting registers 1413- r600: don't reschedule INTERP_LOAD_P0 1414- r600: Force last instruction of group when starting a new CF 1415- r600: Set the last bit if an alu group is split by kcache allocation 1416- r600: make r600_load_ar available to driver code 1417- ntt: translate nir_intrinsic_shader_clock 1418- r600: Assign shader type when creating a new CS state 1419- virgl: Apply integer op fix only for ALU ops and clear modifiers 1420- virgl: Fix relocating the re-writing the transformation code 1421- r600/sfn: Fix store_shared_r600 write masks 1422- virgl: don't move input gl_SampleMaskIn to a temp 1423- r600: Allow eight bit, three channel formats for vertex buffers 1424 1425Guilherme Gallo (5): 1426 1427- ci: Make LAVA jobs fail CI job when retry is exhausted 1428- ci: Install pytest and freezegun plugin 1429- ci: Add unit tests for lava_job_submitter 1430- ci: skqp: Build skqp from android-cts-10.0_r11 tag with Clang 1431- ci: skqp: Add Vulkan support for a630_skqp job 1432 1433Gurchetan Singh (1): 1434 1435- zink: emulate some more memory 1436 1437Hans-Kristian Arntzen (1): 1438 1439- radv: Implement VK_VALVE_descriptor_set_host_mapping. 1440 1441Henry Goffin (1): 1442 1443- frontends/va: ignore incoming frame_num from VA picture parameters 1444 1445Iago Toral Quiroga (50): 1446 1447- broadcom/compiler: fix offset alignment for ldunifa when skipping 1448- broadcom/compiler: allow ldunifa with read-only SSBOs 1449- broadcom/compiler: choose compile strategy with lowest spilling 1450- broadcom/compiler: define max number of tmu spills for compile strategies 1451- broadcom/compiler: fix end of TMU sequence check 1452- broadcom/compiler: reset spill/fill counts after lowering thread count. 1453- broadcom/compiler: do not rebuild the interference graph after each spill 1454- broadcom/compiler: drop spill_count and add spilling boolean 1455- broadcom/compiler: document that spill_base is used for spills and scratch 1456- broadcom/compiler: only patch temps that existed before the current spill 1457- broadcom/compiler: fix register class patching for postponed spills 1458- nir/nir_opt_move: allow to move uniform loads 1459- nir: add a nir_instr_def_is_register helper 1460- nir/nir_opt_move: handle non-SSA defs 1461- broadcom/compiler: move uniforms right before their first use after scheduling 1462- broadcom/compiler: increase cost of TMU spills to 10 1463- broadcom/compiler: disallow TMU spills if max tmu spills is 0 1464- broadcom/compiler: move constants before their first user 1465- broadcom/compiler: sink uniform loads 1466- broadcom/compiler: don't sort nodes for register allocation 1467- broadcom/compiler: simplify node/temp translation during register allocation 1468- broadcom/compiler: stop moving UBO loads before NIR scheduling 1469- nir/schedule: fix handling of generic memory barrier 1470- nir/schedule: handle nir_intrinsic_group_memory_barrier 1471- nir/schedule: use larger delay for non-filtered memory reads 1472- nir/schedule: allow drivers to decide about instruction latency 1473- broadcom/compiler: define v3d-specific delays for NIR instructions 1474- broadcom/compiler: add a strategy to disable scheduling of general TMU reads 1475- broadcom/compiler: adjust register threshold for 2-thread compiles 1476- broadcom/compiler: add comment on why we don't use r5 with ldunifa 1477- broadcom/compiler: don't always assign r5 if available 1478- broadcom/compiler: remove unused functions 1479- v3dv: expose VK_EXT_image_drm_format_modifier 1480- v3dv: TFU destination must be UIF 1481- v3dv: return early on image to buffer blit copies if image is linear 1482- v3dv: don't expose image load/store features for linear images 1483- v3dv: fix temporary imports of semaphores and fences with multisync 1484- v3dv: don't signal semaphores/fences from a wait thread 1485- v3dv: fix semaphore wait from CPU job 1486- v3dv: lock around noop job submits 1487- v3dv: support importing external semaphores 1488- v3dv: implement VK_EXT_inline_uniform_block 1489- broadcom/compiler: allow ldunifa with indirect uniform loads 1490- v3dv: fix sampler array addressing in v3dv_descriptor_set_layout 1491- v3dv: drop unnecessary memset 1492- v3dv: add reference counting for descriptor set layouts 1493- broadcom/compiler: always enable per-quad on spill operations 1494- v3dv: fix limits for inline uniform blocks 1495- broadcom/compiler: prefer reconstruction over TMU spills when possible 1496- v3dv: fix bogus VkDrmFormatModifierProperties2EXT usage 1497 1498Ian Romanick (40): 1499 1500- gallivm/nir: Call nir_lower_bool_to_int32 after nir_opt_algebraic_late 1501- nir: Constify def parameter to nir_ssa_def_bits_used 1502- nir/search: Constify instr parameter to nir_search_expression::cond 1503- nir: All set-on-comparison opcodes can take all float types 1504- intel/fs: Don't optimize out 1.0*x and -1.0*x 1505- spriv: Produce correct result for GLSLstd450Step with NaN 1506- spirv: Produce correct result for GLSLstd450Modf with Inf 1507- spirv: Produce correct result for GLSLstd450Tanh with NaN 1508- nir: Properly handle various exceptional values in frexp 1509- nir: Produce correct results for atan with NaN 1510- glsl/lower_vector_derefs: Don't emit conditional assignments 1511- glsl: Use csel in do_vec_index_to_cond_assign 1512- glsl: Don't try to emit the "linear sequence" in lower_variable_index_to_cond_assign 1513- glsl/ir_builder: Eliminate unused conditional assignment builders 1514- glsl: Lower if to conditional select instead of conditional assignment 1515- glsl: Add ir_assignment constructor that takes just a write mask 1516- glsl: Remove the ability to read text IR with conditional assignments 1517- glsl: Eliminate unused conditional assignment constructor 1518- glsl: Don't clone assignment conditions 1519- glsl: Don't constant-fold the condition of an assignment 1520- glsl: Remove unused condition parameter from ir_assignment constructor 1521- glsl: Don't dead-built-in varying eliminate in the condition of an assignment 1522- glsl: Don't tree graft in the condition of an assignment 1523- glsl: Don't split arrays in the condition of an assignment 1524- glsl: Don't split structures in the condition of an assignment 1525- glsl: Don't lower vector indexing in the condition of an assignment 1526- glsl: Don't visit rvalues in the condition of an assignment 1527- glsl: Make ir_assignment::condition private 1528- glsl: Eliminate ir_assignment::condition 1529- Re-indentation after the previous commit 1530- nir: Add missing dependency on nir_opcodes.py 1531- i915g/ci: update piglit fails 1532- nir/algebraic: Optimize some cases of (sXX(a, b) != 0.0) 1533- i915g: Handle constants composed exclusively of 0 or ±1 specially 1534- i915g: Emit better code for SEQ(x, 0) and SNE(x, 0) 1535- iris/ci: Mark amd_performance_monitor tests as flakes. 1536- intel/fs: Force destination types on DP4A instructions 1537- nir: intel/compiler: Lower TXD on array surfaces on DG2+ 1538- intel/compiler: Fix sample_d messages on DG2 1539- intel/fs: Better handle constant sources of FS_OPCODE_PACK_HALF_2x16_SPLIT 1540 1541Icecream95 (31): 1542 1543- pan/mdg: Use util_logbase2 instead of C99 log2 1544- panfrost: Set PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION 1545- panfrost: Update point size limits to match hardware behaviour 1546- panfrost: Disable point size upper limit clamping 1547- pan/bi: Add interference between destinations 1548- pan/bi: Check dependencies of both destinations of instructions 1549- panfrost: Set dirty state in set_shader_buffers 1550- panfrost: Re-emit descriptors after resource shadowing 1551- pan/bi: Make disassembler build reproducibly 1552- pan/bi: Add documentation for bifrost_nir_lower_store_component 1553- panfrost: Improve comment for emit_fragment_job 1554- panfrost: Fix ubo_mask calculation 1555- pan/bi: Skip psuedo sources in ISA.xml 1556- pan/bi: Don't assign slots for the blend second source 1557- panfrost: Don't initialise the trampolines array 1558- panfrost: Optimise recalculation of max sampler view 1559- panfrost: Remove BO mapping from import 1560- util/hash_table: Remove Unicode byte order mark 1561- drm-shim: Add a function for mmap64 rather than using an alias 1562- drm-shim: Return fake render nodes in /dev/dri first 1563- drm-shim: Explicitly use off64_t for the offset to drm_shim_mmap 1564- drm-shim: Implement a shim function for close 1565- panfrost: Stop overallocating compressed textures 1566- panfrost: Fix pack_32_2x16 implementation 1567- pan/bi: Use texture index instead of sampler for message preloading 1568- nir/lower_tex: Copy more fields in lower_tex_to_txd and friends 1569- pan/mdg: Keep min_bound at 16 when alignment requires it 1570- pan/mdg: Use MAX2 to set min_alignment 1571- pan/mdg: Fix mask usage when filling before a spill 1572- pan/mdg: Return the instruction from mir_insert_instruction_*_scheduled 1573- pan/mdg: Fix multiple spilt writes in the same bundle 1574 1575Igor Torrente (13): 1576 1577- venus: Exposes VK_EXT_4444_formats extension 1578- venus: venus-protocol groundwork to VK_EXT_custom_border_color 1579- venus: add VK_EXT_custom_border_color extension 1580- venus: add macros to help with future extensions 1581- Venus: Add \`vn_physical_device_{features, properties}` for better organization 1582- Venus: Add VN_CMD_ENQUEUE macro with vkCmd* common code 1583- Venus: Adjust VN_CMD_ENQUEUE to set VN_COMMAND_BUFFER_STATE_INVALID 1584- Venus: add VN_CMD_ENQUEUE to vn_cmd_encode_memory_barriers 1585- venus: move vkGetCalibratedTimestamps to vn_protocol_driver_device.h 1586- venus: add VK_EXT_calibrated_timestamps extension 1587- venus: add VK_EXT_extended_dynamic_state2 extension 1588- venus: Update venus-protocol to add two new extensions 1589- venus: add VK_EXT_{conditional_rendering,index_type_uint8} extensions 1590 1591Ilia Mirkin (30): 1592 1593- glsl: simplify conditions for setting various allowed flags 1594- glsl: only validate xfb_buffer values when we have enhanced layouts 1595- st/mesa: only enable ARB_enhanced_layouts if there are xfb buffers 1596- rtasm: fix printf specifier for ptrdiff_t 1597- rtasm: add pcmpgtd operation 1598- translate: improve sse2 32-bit unsigned -> float conversion 1599- isaspec: fix gen_max to be 2^32-1 1600- isaspec: add gen-based leaf bitset separation 1601- freedreno/ir3: split up load/store/atomic by generation 1602- freedreno/a4xx: use correct macro for color 1603- freedreno/a4xx: make luminance formats renderable, add missing L8A8_SNORM 1604- freedreno/ir3: document GETINFO's x/y results 1605- nouveau: add dEQP/GLCTS run failure info for GF108/GT215 1606- mesa: enable GL_EXT_texture_sRGB_RG8 on desktop 1607- nvc0: disable EXT_texture_sRGB_RG8 1608- freedreno/ir3: remove bogus tg4 -> tex lowering pass 1609- nir: remove bogus logic to allow cube + offset to work 1610- freedreno/a4xx: move tex_type to header 1611- freedreno/a4xx: add swizzles to shader keys for tg4 workaround 1612- freedreno/a4xx: fix integer tg4 1613- freedreno/ir3: support a4xx in load/store buffer/image emission 1614- freedreno/ir3: support a4xx compute differences 1615- a4xx: add logic to emit image/ssbo state 1616- a4xx: add emission of compute state, and compute dispatch 1617- freedreno/a4xx: extend astc and tg4 workarounds to compute shaders 1618- freedreno/a4xx: improve condition for disabling early z 1619- freedreno/ir3: no need to count bits 16b at a time for a4xx 1620- freedreno/ir3: disable conversion folding on a4xx 1621- freedreno/a4xx: expose shaders and images, as well as ES 3.1 1622- freedreno: add a420 deqp-runner files 1623 1624Indrajit Das (1): 1625 1626- ac/gpu_info: disallow displayable DCC for Navi12 and Navi14 1627 1628Iván Briano (7): 1629 1630- anv: make the pointer valid before we assign stuff into it 1631- intel/compiler: remove what looks like a bad rebase 1632- anv: only advertise 64b atomic floats if 64b floats are supported 1633- intel/compiler: make CLUSTER_BROADCAST always deal with integers 1634- intel/fs: handle interpolation modes for at_sample and at_offset too 1635- vulkan/wsi/wayland: Fix double free on error condition 1636- anv: re-alloc push constants after secondary command buffers 1637 1638Jakob Bornecrantz (1): 1639 1640- vulkan-device-select: Don't leak xcb_query_extension_reply_t 1641 1642Faith Ekstrand (179): 1643 1644- vulkan,lavapipe: Simplify command recording code-gen 1645- zink: Re-interpret formats when using vkCmdClearColorImage() 1646- anv: Call vk_command_buffer_finish if create fails 1647- vulkan: Add a common vk_command_pool base struct 1648- anv: Use vk_command_pool 1649- radv: Use vk_command_pool 1650- panvk: Use vk_command_pool 1651- lavapipe: Use vk_command_pool 1652- v3dv: Use vk_command_pool 1653- turnip: Use vk_command_pool 1654- vulkan: Take a vk_command_pool in vk_command_buffer_init() 1655- vulkan/queue: Assert command buffers have the right queue family 1656- vulkan: Implement of a bunch of VkCommandPool functions 1657- anv: Don't use vk_alloc/free2 for command buffers 1658- anv: Drop anv_cmd_buffer::pool 1659- anv: Don't call DestroyCommandBuffers in AllocateCommandBuffers 1660- anv: Use the common vk_command_pool 1661- v3dv: Don't use vk_alloc/free2 for command buffers 1662- v3dv: Use the common command pool implementation 1663- vulkan: Rename vk_image_view::format to view_format 1664- vulkan: Add back vk_image_view::format 1665- anv: use vk_image_view::format for creating dynamic renderpasses 1666- anv: Don't assume depth/stencil attachments have depth 1667- panvk: Advertise VK_KHR_get_physical_device_properties2 1668- panvk: Advertise zero sparse format properties 1669- panvk: Non-destructively stub GetRenderAreaGranularity 1670- blorp: Add a binding_table_offset_to_pointer helper 1671- vulkan/cmd_queue: Re-flow MANUAL_COMMANDS 1672- vulkan/cmd_queue: Stop generating enqueue helpers for INTEL perf queries 1673- vulkan/cmd_queue: Generate enqueue entrypoints 1674- lavapipe: Reset the free_cmd_buffers list in TrimCommandPool 1675- vulkan,lavapipe: Move some enqueue helpers to common code 1676- vulkan/cmd_queue: Add a driver_free_cb hook 1677- vulkan/cmd_queue: Add a common vk_cmd_enqueue_CmdBindDescriptorSets 1678- lavapipe: Allocate descriptor set layouts with DEVICE scope 1679- lavapipe: Reference count pipeline layouts 1680- lavapipe: Use the common vk_enqueue_CmdBindDescriptorSets 1681- vulkan/cmd_queue: Properly support non-array pointer members 1682- lavapipe: Use the auto-generated vk_enqueue_BeginRendering 1683- anv: Allow MSAA resolve with different numbers of planes 1684- bifrost: Lower usub_borrow 1685- bifrost: Simplify derivatives a bit 1686- bifrost: Implement fine and coarse derivatives 1687- bifrost: Handle nir_op_frexp* and nir_op_ldexp 1688- bifrost: Constant fold after lower_explicit_io 1689- panvk: Stop advertising Vulkan 1.1 1690- panvk: Make panvk_image derive from vk_image 1691- panvk: Make panvk_image_view derive from vk_image_view 1692- panvk: Skip ZS setup if there is no depth/stencil attachment 1693- panvk: Rework texture, sampler, and image binding index calculation 1694- panvk: Use the correct integer border colors 1695- vulkan: Add a vk_shader_module_to_nir() helper 1696- anv: Use vk_shader_module_to_nir() 1697- panvk: Use vk_shader_module 1698- panvk: Use vk_shader_module_to_nir() 1699- panvk: Require 16B alignment for UBOs 1700- panvk: Fix SSBO buffer offsets 1701- panvk: Implement VK_EXT_vertex_attribute_divisor 1702- vulkan: Add a common vk_framebuffer struct 1703- anv: Convert to vk_framebuffer 1704- vulkan: Add a common vk_render_pass struct 1705- vulkan: Add a common CmdBegin/EndRederPass implementation 1706- vulkan: Add helpers for getting rendering info from a renderpass 1707- vulkan/render_pass: Provide self-dependeny information 1708- vulkan/render_pass: Support fragment shading rate 1709- vulkan/render_pass: Add an optimization for UNDEFINED+LOAD_OP_CLEAR 1710- vulkan/framebuffer: Add a flags field 1711- anv: Fix handling of null depth/stencil attachments with dynamic rendering 1712- anv: Better null surface state size for dynamic rendering 1713- anv/pass: Make unused color attachments VK_ATTACHMENT_UNUSED 1714- anv: Stop treating color input attachments specially 1715- anv/pipeline: Stop pretending we're the validator 1716- anv: Drop pipeline pass/subpass in favor of rendering_info 1717- anv: Convert to 100% dynamic rendering 1718- docs: Add the start of Vulkan runtime docs 1719- docs: Add high-level documentation for Vulkan render passes 1720- intel/guardband: Take min/max instead of total size 1721- anv: Calculate the real guardband based on render area 1722- anv: Move viewport/scissor emit to genX_cmd_buffer.c 1723- anv: Include scissors in viewport calculations 1724- panvk: Convert to the common sync/submit framework 1725- panvk: Re-arrange GetPhysicalDeviceFeatures2 1726- panvk: Add a 1.3 features struct 1727- panvk: Re-arrange GetPhysicalDeviceProperties2 1728- panvk: Move core properties into their respective core structs 1729- panvk: Implement VK_KHR_synchronization2 1730- panvk: Implement VK_KHR_copy_commands2 1731- panvk: Drop BindImage/BufferMemory 1732- panvk: Drop QueueBindSparse 1733- panvk: Only implement Get*MemoryRequirements2 1734- anv: Use layerCount for clears and transitions in BeginRendering 1735- vulkan: Add a 2 wrapper for vkGetPhysicalDeviceQueueFamilyProperties 1736- anv: Drop GetPhysicalDeviceQueueFamilyProperties 1737- radv: Drop GetPhysicalDeviceQueueFamilyProperties 1738- turnip: Drop tu_legacy.c 1739- v3dv: Drop GetPhysicalDeviceQueueFamilyProperties 1740- panvk: Drop GetPhysicalDeviceQueueFamilyProperties 1741- lavapipe: Use VK_OUTARRAY for GetPhysicalDeviceQueueFamilyProperties[2] 1742- lavapipe: Drop GetPhysicalDeviceQueueFamilyProperties 1743- vulkan/cmd_queue: Add a vk_cmd_queue_execute() helper 1744- vulkan/cmd_queue: Generate enqueue_if_not_primary entrypoints 1745- vulkan/cmd_queue: Auto-generate more vk_cmd_enqueue_unless_primary_Cmd* 1746- vulkan/runtime: Add emulated secondary command buffer support 1747- panvk: Hook up emulated secondary command buffers 1748- spirv: Properly mangle generic pointers 1749- nir/print: Add support for generic pointers 1750- intel/nir: Add optimizations to help OpenCL-style kernels 1751- intel/debug: Dump KERNEL source when INTEL_DEBUG=cs 1752- intel/compiler: Add code for compiling CL-style SPIR-V kernels 1753- intel/kernel: Implement some Intel built-in functions 1754- intel: Add a little OpenCL C compiler binary 1755- nir: Allow nir_var_mem_global variables 1756- spirv: Implement the function portion of the Linkage capability 1757- i915: Use the sin/cos lowering in nir_opt_algebraic.py 1758- vulkan: Add a vk_image_view_subresource_range helper 1759- lavapipe: Use vk_image_view 1760- lavapipe: Use vk_image_subresource_layer/level_count 1761- lavapipe: Delete render passes 1762- v3dv/queue: Rework multisync_free 1763- spirv/libclc: Add generic versions of arithmetic functions 1764- vulkan: Add more VU comments to justify framebuffer asserts 1765- vulkan/queue: Add a submit mode enum 1766- vulkan/queue: Rework submit thread enabling 1767- vulkan/queue: Rework vk_queue_submit() 1768- vulkan: Replace various uses of device->timeline_mode 1769- vulkan: Allow the driver to manually enable threaded submit 1770- intel/isl: Add a helper for swizzling color values 1771- anv: Disallow blending on swizzled formats 1772- anv: Generalize border color swizzles 1773- anv: Advertise two more formats 1774- util/timespec: Return overflow from timespec_add_[mn]sec() 1775- vulkan: Use timespec_add_nsec in vk_sync_timeline 1776- lavapipe: Switch to the common sync framework 1777- v3dv: Enable VK_EXT_debug_utils 1778- turnip: Enable VK_EXT_debug_utils 1779- lavapipe: Enable VK_EXT_debug_utils 1780- anv: Enable VK_EXT_debug_utils 1781- panvk: Enable VK_EXT_debug_report and VK_EXT_debug_utils 1782- lavapipe: Go back to manually signaling in lvp_AcquireNextImage2() 1783- iris: Account for BO offsets in iris_set_global_binding() 1784- iris: Handle range tracking for global bindings 1785- iris: Take offsets into account when mapping resources 1786- iris: Allow non-page-aligned userptr 1787- iris: Allow userptr on 1D and 2D images 1788- anv/formats: Relax usage checks if EXTENDED_USAGE_BIT is set 1789- vulkan: vk_object_base_init/finish have no unused parameters 1790- vulkan,docs: Document vk_object_base 1791- vulkan,docs: Document vk_instance 1792- vulkan,docs: Document vk_physical_device 1793- vulkan,docs: Document vk_device 1794- vulkan,docs: Add documentation for Vulkan dispatch 1795- nir: Dont set coord_components on txs 1796- nir/lower_int64: Fix [iu]mul_high handling 1797- intel/fs: Implement 16-bit [ui]mul_high 1798- intel/nir: Lower 8 and 16-bit bitwise unops 1799- iris: Don't leak scratch BOs 1800- vulkan/drm_syncobj: Implement WAIT_PENDING with a sync_file lookup 1801- v3dv: Stop directly setting vk_device::alloc 1802- v3dv: Put indirect compute CSD jobs in the job list 1803- v3dv: Don't use pthread functions on c11 mutexes 1804- v3dv: Destroy the device mutex on the teardown path 1805- v3dv: Switch to the common device lost tracking 1806- v3dv: Use util/os_time helpers 1807- v3dv: Add a condition variable for queries 1808- v3dv: Always wait on last_job_syncs if job->serialize 1809- v3dv: Switch to the common submit framework 1810- v3dv: Use the core version feature helpers 1811- v3dv: Use the core version property helpers 1812- v3dv: Add emulated timeline semaphore support 1813- iris: More gracefully fail in resource_from_user_memory 1814- nir/opcodes: fisfinite32 should return bool32 1815- util/set: Respect found in search_or_add_pre_hashed 1816- nir: Lower all bit sizes of usub_borrow 1817- vulkan: Set signals[i].stageMask = ALL_COMMANDS for QueueSubmit2 wrapping 1818- spirv: Handle Op*MulExtended for non-32-bit types 1819- vulkan/wsi: Set the right stage flags for semaphore waits 1820- nir/cf: Return a cursor from nir_cf_extract as well 1821- gallium/u_threaded_context: Use PIPE_MAX_SHADER_SAMPLER_VIEWS for sampler_buffers 1822- nir: Preserve metadata if remove_dead_derefs makes no progress 1823 1824Jesse Natalie (68): 1825 1826- microsoft/compiler: Lower mul_2x32_64 1827- microsoft/compiler: Only prep phis for the current function 1828- microsoft/compiler: Only treat tess level location as special if it's a patch constant 1829- microsoft/compiler: Add never-supported double ops to lower_doubles bitmask 1830- microsoft/compiler: It's possible to have doubles without int64 1831- microsoft/compiler: Fix splitdouble struct name 1832- microsoft/compiler: Fix make_double and split_double to respect swizzles 1833- microsoft/compiler: Fix dxil_nir_lower_double_math_instr pass for vectors 1834- microsoft/compiler: Set dx11_1_double_extensions flag for dfma/ddiv 1835- microsoft/compiler: Handle b2f64 1836- microsoft/compiler: Lower 64bit I/O to 32 and then run lower_pack 1837- microsoft/compiler: Handle I/O vars larger than a vec4 1838- microsoft/compiler: Process signatures before the shader code 1839- microsoft/compiler: Map I/O base locations to input IDs 1840- microsoft/compiler: Handle structs in I/O signatures 1841- d3d12: Update max input, output, and varying caps 1842- d3d12: Use a constant define for max anisotropy 1843- d3d12: Use the right constant for GS varying limits 1844- d3d12: Cache a modifyable copy of the nir options in d3d12_screen 1845- d3d12: Add a driver version to the screen to be used for workarounds 1846- d3d12: Get OPTIONS1 1847- d3d12: Add int64 support 1848- d3d12: Lower [de]construction of doubles via math ops into pack/unpack ops 1849- d3d12: Set lower full fp64 compiler options flag when needed 1850- d3d12: Handle structs in GS variants 1851- d3d12: Handle structs in TCS variants 1852- d3d12: ARB_gpu_shader_fp64 1853- docs: Update d3d12 features 1854- microsoft/compiler: Handle SV_ViewportArrayIndex 1855- microsoft/compiler: Set flag for VP/RT array index from VS/DS 1856- d3d12: Enable BPTC (BC6/BC7) 1857- d3d12: Bind 16 scissor rects when scissor disabled 1858- d3d12: Fix linkage for viewport index 1859- d3d12: Update depth invert to deal with multi-viewport 1860- d3d12: Fix location compares in MSAA disable 1861- d3d12: Update nir varying bitmasks when linking stages 1862- d3d12: Don't force a GS to be added for 'flat' sysvals 1863- d3d12: Don't add arrayed VS outputs when next stage uses per-vertex inputs 1864- d3d12: When adding new output varyings, write 0s 1865- d3d12: Allow RGB VS inputs without an alpha channel 1866- d3d12: GL4.2 1867- docs: Update d3d12 features 1868- tc: CPU storage needs to be freed with align_free 1869- d3d12: Fix set constant buffers 1870- d3d12: Fix range calculation for suballocated buffers in d3d12_bo_unmap 1871- d3d12: Always respect offsets when mapping a bo, not just when there's a range 1872- d3d12: Delete make_resource_writeable 1873- d3d12: Fix TBOs from suballocated buffers 1874- d3d12: Don't suballocate TBO buffers 1875- d3d12: Fix offset for buf/image copies with suballocated buffers 1876- d3d12: Actually suballocate and cache buffers 1877- d3d12: Add a buffer busy callback to the bufmgr 1878- d3d12: Use CPU storage in TC for buffers 1879- d3d12: Allow 8bit index buffer conversions by vbuf 1880- d3d12: Fix take_ownership semantic for constant buffers 1881- d3d12: When mapping a resource used in the current batch without blocking, at least flush 1882- d3d12: Forward wait condition from query -> result buffer 1883- d3d12: Move ID3D12Fence from context to screen 1884- d3d12: Add sampler's textures to batch bo tracking 1885- d3d12: Add residency info to d3d12_bo 1886- d3d12: Add a budget/usage callback to the screen 1887- d3d12: Implement residency management algorithm 1888- d3d12: Default newly-created resources to not-resident 1889- d3d12: Don't add a second dual-source output for Heaven 1890- driconf: Add Heaven entries for Windows .exe 1891- d3d12: Only force point sampling for emulated shadow samplers 1892- microsoft/compiler: Fill interpolation for sysval inputs to non-vertex shader 1893- microsoft/compiler: Unload DXIL validator library \*after* calling Release() 1894 1895Jiadong Zhu (1): 1896 1897- st/mesa: set GL_DYNAMIC_STORAGE_BIT for GL-VK interop buffers 1898 1899Jocelyn Falempe (2): 1900 1901- llvmpipe: fix color rendering on big endian. 1902- llvmpipe: remove unused array 1903 1904Jonathan Gray (10): 1905 1906- util: unbreak non-linux mips64 build 1907- util: fix u_print.cpp build on OpenBSD 1908- util: fix util_cpu_detect_once() build on OpenBSD 1909- radv: use MAJOR_IN_SYSMACROS for sysmacros.h include 1910- util/u_atomic: fix build on clang archs without 64-bit atomics 1911- util: fix build with clang 10 on mips64 1912- util: use correct type in sysctl argument 1913- util: fix msvc build 1914- intel/dev: sync ADL-S pci ids with linux 1915- intel/dev: add RPL-S pci ids from drm-intel-next 1916 1917Jordan Justen (13): 1918 1919- intel/fs: Assert that old pull-const code is not used if devinfo->has_lsc 1920- anv: Align GENERAL_STATE_POOL_MIN_ADDRESS to 2MiB 1921- anv: Align state pools to 2MiB on XeHP 1922- intel/dev: Add ATS-M pci-ids 1923- intel/dev: Add device info for RPL-P 1924- vulkan/wsi/x11: Use atomic_int with sent_image_count 1925- drm-uapi/i915_drm.h: Update from drm-next (2022-04-28) 1926- iris: Fix assertion meant to only target the clear-color stride 1927- intel/dev: Read hwconfig from i915 1928- intel_dev_info: Add --hwconfig command line parameter 1929- intel/dev: Add INTEL_PLATFORM_DG2_G12 1930- intel/dev: Add DG2 G12 PCI IDs 1931- intel/dev: Enable first set of DG2 PCI IDs 1932 1933Jose Maria Casanova Crespo (1): 1934 1935- v3dv/v3d: Fix copyright holder to Raspberry Pi Ltd 1936 1937José Expósito (2): 1938 1939- egl/wayland: fix crash in dri2_initialize_wayland_drm 1940- egl/wayland: fix crash in dri2_initialize_wayland_swrast 1941 1942Juan A. Suarez Romero (23): 1943 1944- v3dv/ci: Update failure list 1945- vc4/ci: update failing piglit tests 1946- vc4/ci: make piglit test mandatory 1947- vc4: remove redundant initialization 1948- v3d: enable texture filtering anisotropic 1949- v3d: rebind sampler view if resource changed the BO 1950- v3dv/ci: add test to skip list 1951- util/disk_cache: rename MESA_GLSL_CACHE envvar 1952- ci: use MESA_SHADER_CACHE envvar 1953- radv: change MESA_GLSL_CACHE envvar reference 1954- v3dv: change MESA_GLSL_CACHE envvar reference 1955- broadcom/ci: Update flake list 1956- v3d: add support for on-disk shader cache 1957- broadcom: add on-disk cache debug option 1958- broadcom: add line rasterization mode to packet definition 1959- v3dv: add subpixel precision definition 1960- v3dv: implement VK_EXT_line_rasterization 1961- v3d: fix some leaks in cache 1962- v3d: do not leak BO on query begin 1963- ci: enable v3dv arm64 jobs 1964- CODEOWNERS: add Broadcom maintainers 1965- ci: add Broadcom CI maintainer 1966- mesa: unref syncobj after wait_sync 1967 1968Kai Wasserbäch (2): 1969 1970- fix(FTBFS): clover: work around removal of PointerType::getElementType() 1971- fix(clover): FTBFS: Added missing include for ConstantInt for LLVM 15 1972 1973Karmjit Mahil (5): 1974 1975- pvr: Fix seg fault in vkAllocateDescriptorSets(). 1976- pvr: Add stricter type checking in pvr_csb_pack(). 1977- pvr: Add core count info and pvr_device_runtime_info. 1978- pvr: Add pvrsrvkm visibility test heap. 1979- pvr: Implement vkCreateQueryPool() and vkDestroyQueryPool(). 1980 1981Karol Herbst (1): 1982 1983- nir: fix nir_sweep for printf 1984 1985Kenneth Graunke (51): 1986 1987- iris: Make an iris_foreach_batch macro that skips unsupported batches 1988- anv: Increase maxUniformBufferRange to 2^30 when not using the sampler 1989- intel/genxml: Add missing MI_FLUSH_DW::Flush CCS field 1990- blorp: Assert that blorp_copy() on the blitter can handle it 1991- iris: Add support for flushing the blitter (hackily) 1992- ci/iris: Mark qbo tests as flakes 1993- iris: Allow IRIS_BATCH_BLITTER in iris_copy_region() 1994- iris: Use the hardware blitter for DRI PRIME blits 1995- iris: Disable PIPE_CAP_PREFER_BACK_BUFFER_REUSE 1996- anv: Increase maxBoundDescriptorSets to 32 1997- anv: Lower bufferImageGranularity to 1 from 64 1998- anv: Raise vertex input bindings and attributes limits slightly 1999- iris: Set MI_FLUSH_DW::PostSyncOperation correctly 2000- blorp: Add blorp_measure hooks to the blitter codepaths 2001- iris: Use more efficient binding table pointer formats on Icelake+. 2002- iris: Rename surface_base_address to binder_address in a few places 2003- intel: Limit Wa_1607854226 to Gfx12.0 only 2004- intel: Use 3DSTATE_BINDING_TABLE_POOL_ALLOC exclusively on Gfx11+ 2005- iris: Restore flagging of dirty bindings in binder_realloc 2006- iris: Fix MOCS for copy regions 2007- nir: Print divergence status of SSA values if analysis was ever run. 2008- intel: Add INTEL_DEBUG=noccs alias for INTEL_DEBUG=norbc 2009- nir: Teach nir_divergence_analysis about Intel-specific intrinsics 2010- intel/compiler: Implement nir_intrinsic_last_invocation 2011- intel/compiler: Set divergence analysis options 2012- intel/compiler: Convert to LCSSA and use divergence analysis. 2013- intel/compiler: Use nir_opt_uniform_atomics() 2014- intel/decoder: Fix decoder handling of binding table pool alloc on XeHP 2015- iris: Properly tell the decoder about inherited binder addresses 2016- intel/compiler: Call inst->resize_sources before setting the sources 2017- anv: Stop updating STATE_BASE_ADDRESS on XeHP 2018- intel/genxml: Add SAMPLER_STATE::Allow Low Quality LOD Calculation field 2019- intel/genxml: Add new "Low Quality Filter" field on Gfx12+. 2020- intel/genxml: Delete SAMPLER_MODE register definition on Gfx12+ 2021- intel/genxml: Add SAMPLER_MODE bits for enabling Small PL on Icelake 2022- st/mesa: Make transcode_astc also check for non-SRGB format support 2023- st/mesa: Transcode ASTC to BC7 (BPTC) where possible 2024- iris: Use IRIS_DOMAIN_DEPTH_WRITE for read only depth/stencil. 2025- iris: Split out an IRIS_DOMAIN_SAMPLER_READ domain from OTHER_READ 2026- iris: Fix UBO cache tracking for the !indirect_ubos_use_sampler case 2027- iris: Add an iris_is_domain_l3_coherent helper. 2028- iris: Add a separate PIPE_CONTROL_L3_READ_ONLY_CACHE_INVALIDATE bit 2029- iris: Extend the cache tracker to handle L3 flushes and invalidates 2030- iris: Add pre-draw flushing for stream output targets 2031- iris: Use cache-tracker for draw count flushing 2032- iris: Emit flushes for push constant source buffers 2033- iris: Demote DC flush to HDC flush in cache tracker 2034- iris: Add VF_CACHE_INVALIDATE to IRIS_DOMAIN_OTHER_WRITE flush bits 2035- anv: Fix INTEL_DEBUG=bat on XeHP 2036- iris: Add FLUSH_HDC to PIPE_CONTROL_CACHE_FLUSH_BITS 2037- Revert "st/mesa: Transcode ASTC to BC7 (BPTC) where possible" 2038 2039Konstantin Seurer (16): 2040 2041- radv: Move common code to seperate file 2042- nir,spirv: Preserve ray_query_value 2043- radv: Lower ray queries 2044- radv: Enable KHR_ray_query 2045- radv: Use common GetPhysicalDeviceMemoryProperties 2046- venus: Use trivial common entrypoints 2047- radv: Fully implement ray primitive culling 2048- radv: Advertise ray primitive culling 2049- radv: Replace magic constants with enum values 2050- radv: Remove radv_util.c 2051- radv: Refactor radv_tex_aniso_filter 2052- radv: Refactor ray tracing support checks 2053- radv: Fix barriers with cp dma 2054- radv: Enable rt primitive culling for spirv2nir 2055- radv: Fix lowering ignore_ray_intersection 2056- ac/nir: Do not set cursor in try_extract_additions 2057 2058Krunal Patel (1): 2059 2060- frontend/va: Create decoder once the max_references is updated 2061 2062Lionel Landwerlin (104): 2063 2064- anv: flush utrace before at device destroy 2065- anv: add dynamic rendering traces 2066- intel/ds: fix compilation with perfetto 2067- nir: change intel dss_id intrinsic to topology_id 2068- intel/fs: rework dss_id opcode into generic opcode 2069- intel/fs: add support for Eu/Thread/Lane id 2070- intel/nir/rt: add a new number of SIMD lanes per DSS helper 2071- intel/fs: name sources for A64 opcodes 2072- intel/fs: add support for ACCESS_ENABLE_HELPER 2073- intel/fs: don't set allow_sample_mask for CS intrinsics 2074- intel/fs: load more fields from BVH instance leafs 2075- intel/compiler: tracker number of ray queries in prog_data 2076- intel/fs: limit FS dispatch to SIMD16 when using ray queries 2077- intel/nir: fix shader call lowering 2078- intel/nir: use a single intel intrinsic to deal with ray traversal 2079- nir/builder: add nir_ior_imm() helper 2080- intel: add a new intrinsic to get the shader stage from bindless shaders 2081- intel/nir/rt: load bvh_level value off mem_hit structure 2082- intel/fs: make trivial shader complete tracing operations with missing shaders 2083- intel/nir: document committed argument 2084- intel/nir/rt: make RT manipulation helpers helper invocations ready 2085- intel/nir/rt: add more helpers for ray queries 2086- nir/lower_shader_calls: consider relocated constants as rematerializable 2087- intel/nir: document RT builder 2088- intel/fs: lower ray query intrinsics 2089- anv: enable ray queries 2090- anv: fix conditional render for vkCmdDrawIndirectByteCountEXT 2091- anv: update limit for maxVertexInputBindingStride 2092- intel/dev: provide some default values for no_hw 2093- nir: fix lower_memcpy 2094- anv: add a custom AcquireNextImage2KHR func 2095- anv/genxml/intel/fs: fix binding shader record entry 2096- anv: fix fast clear type value with external images 2097- intel/fs: fix total_scratch computation 2098- anv: fix acceleration structure descriptor copies 2099- anv: don't lazy allocate surface states in descriptor sets 2100- anv: rename host only descriptor internal flag 2101- anv: make a couple of descriptor function private 2102- anv: silence compiler warning 2103- anv: silence compiler warnings 2104- anv: fix variable shadowing 2105- anv: zero-out anv_batch_bo 2106- anv: emit timestamp & availability using the same part of CS 2107- anv: flush tile cache with query copy command 2108- intel/nir/fs: replace COMPUTE || KERNEL by gl_shader_stage_is_compute() 2109- intel/fs: handle inline data on OpenCL style kernels 2110- meson: try to find clang-cpp before going through each module 2111- intel/clc: allow multiple CL files to be compiled together 2112- intel/clc: allow producing SPIRV files 2113- intel/clc: specify supported extensions 2114- intel/clc: disable tool prior to Gfx12.5 platforms 2115- ci: add clang/spirv-tools/llvm-spirv packages to fedora container 2116- ci: enable llvm on debian-release build 2117- ci: enable intel-clc on some platforms 2118- vulkan: move EXT_tooling_info implementation to runtime 2119- anv: implement EXT_tooling_info 2120- intel/dev: default to B stepping on DG2 for offline compiler 2121- intel/kernel: enable groups caps 2122- intel/kernel: enable linkage cap 2123- intel/clc: add option to printout kernel prog_data 2124- intel/clc: compile fix 2125- intel/clc: fixup shared memory offsets 2126- intel/clc: deal with SPIRV-Tools linker new behavior 2127- iris: don't synchronize BO for batch decoding 2128- anv: allow baking of 3DSTATE_DEPTH_BOUNDS in pipeline batch 2129- anv: fix dynamic state emission 2130- anv: fix VK_DYNAMIC_STATE_COLOR_WRITE_ENABLE_EXT state 2131- anv: fix color write enable interaction with color mask 2132- anv: use local dynamic pointer more 2133- anv: fix dynamic sample locations on Gen7/7.5 2134- anv: don't store sample location sample count 2135- intel/clc: fix missing pointer write 2136- anv: disable preemption on 3DPRIMITIVE on gfx12 2137- anv: remove unused 3DSTATE_DEPTH_BOUNDS fields 2138- blorp: disable depth bounds 2139- intel/nir: don't report progress on rayqueries if no queries 2140- nir/lower_shader_calls: name resume shaders 2141- intel/fs: fix metadata preserve on trace_ray intrinsic 2142- intel/fs: add a note on possible optimization of root node address 2143- anv: stop using old entrypoint/struct/enum names for 1.3 2144- intel: fix URB programming for GT1s 2145- anv: allow getting the address of the beginning of the batch 2146- nir/divergence: handle load_global_block_intel 2147- intel: fixup number of threads per EU on XeHP 2148- anv: fix acceleration structure descriptor template writes 2149- anv: skip acceleration structure in binding table emission 2150- anv: fix clflush usage on utrace copy batch 2151- iris: use new kernel uAPI to compute video memory 2152- anv: fix invalid utrace memcpy l3 config on gfx < 11 2153- anv: reset all dynamic state after secondary execution 2154- anv: add missing logic op set in pipeline dyn state 2155- Revert "anv: fix dynamic state emission" 2156- anv: rework sample location 2157- anv: never emit 3DSTATE_CPS in the pipeline batch 2158- anv: don't emit 3DSTATE_STREAMOUT in pipeline batch 2159- anv: don't emit 3DSTATE_WM in pipeline batch 2160- anv: don't emit 3DSTATE_BLEND_STATE_POINTERS in pipeline batch 2161- anv: don't emit 3DSTATE_DEPTH_BOUNDS in pipeline batch 2162- anv: don't emit 3DSTATE_VF_TOPOLOGY in pipeline batch 2163- anv: remove static_state_mask 2164- nir/cf: return cursor after insertion of cf_list 2165- nir/lower_shader_calls: don't use nop instructions as cursors 2166- nir/lower_shader_calls: don't insert code after break/continue 2167- nir/lower_shader_calls: put inserted instructions into a dummy block 2168 2169Louis-Francis Ratté-Boulianne (5): 2170 2171- vulkan/runtime: Add a level field to vk_command_buffer 2172- vulkan/cmd_queue: Add an initializer for the vk_cmd_queue object 2173- vulkan/runtime: Add a vk_cmd_queue object to vk_command_buffer 2174- dzn: Add a debug flag to enable D3D12 debug layer 2175- dzn: Add CI target for vulkan driver 2176 2177M Henning (2): 2178 2179- nouveau: Handle unaligned tlsBase during spills 2180- nouveau: Fix out-of-bounds access in AlgebraicOpt 2181 2182Marcin Ślusarz (20): 2183 2184- intel/compiler: fix array & struct IO lowering in mesh shaders 2185- anv: don't set color state when input state was requested 2186- intel/compiler: remove redundant code from fs_visitor::run_* 2187- anv: cleanup begin_subpass & end_subpass 2188- intel/compiler: ignore per-primitive attrs when calculating flat input mask 2189- intel/compiler: optimize flat inputs mask calculation 2190- intel/compiler: shift mesh urb read/write window when offset is too large 2191- intel/compiler: inject MUE initialization 2192- intel/compiler: handle ViewportIndex, PrimitiveID and Layer in MUE setup 2193- intel/compiler: mark some variables as per-primitive in FS if they come from MS 2194- anv: set number of viewports in clip state (mesh) 2195- anv: include Primitive Header in mesh shader per-primitive output 2196- anv: fix push constant lowering for task/mesh 2197- anv: initialize 3DMESH_1D.ExtendedParameter0 when ExtendedParameter0Present 2198- nir: remove gl_PrimitiveID output from MS when it's not used in FS 2199- anv: invalidate all metadata in anv_nir_lower_ubo_loads 2200- intel/compiler: invalidate all metadata in brw_nir_lower_intersection_shader 2201- intel/compiler: invalidate metadata in brw_nir_initialize_mue 2202- anv: update task/mesh distribution with the recommended values 2203- anv: disable streamout before emitting mesh shading state 2204 2205Marek Olšák (61): 2206 2207- gallium: add PIPE_RESOURCE_FLAG_UNMAPPABLE for shared unmappable buffers 2208- ci: bump piglit version 2209- gallivm: fix build with LLVM 15 2210- winsys/radeon: fix a hang due to introducing spi_cu_en 2211- amd: remove the _UMD suffix from register definitions 2212- amd: add a workaround for an SQ perf counter bug 2213- ac: add a gfx9 workaround for high priority compute 2214- ac/gpu_info: print units for some radeon_info fields 2215- ac/surface: don't set the display flag for 1D textures 2216- ac/llvm: remove unused function dpp_row_sl 2217- winsys/amdgpu: fix a warning of defining radeon_screen_create_t twice 2218- radeonsi: add assertions to check if buffer_map/texture_map calls are valid 2219- radeonsi: always set FLUSH_ON_BINNING_TRANSITION 2220- radeonsi: program SQ_THREAD_TRACE_CTRL.AUTO_FLUSH_MODE on gfx10.3 2221- radeonsi: change ACCUM_ISOLINE to 12 2222- radeonsi: document an unexpected behavior of PS_DONE 2223- radeonsi: reduce the max TBO/SSBO binding size to 512 MB to help 32-bit builds 2224- radeonsi: more fixes for si_buffer_from_winsys_buffer for GL-VK interop 2225- radeonsi: don't map buffers that VK made unmappable 2226- radeonsi: replace SI_RESOURCE_FLAG_UNMAPPABLE with PIPE_RESOURCE_FLAG_UNMAPPABLE 2227- radeonsi: remove bit gaps in SI_RESOURCE_FLAG_* 2228- radeonsi: increase the tesselation factor ring size 2229- radeonsi: fix the unaligned clear_buffer fallback with TC 2230- ac,radeonsi: rework and optimize how TMPRING_SIZE is set 2231- radeonsi: apply the LLVM discard bug workaround to LLVM 13 only 2232- ac/surface/tests: fix missing NUM_PKRS extraction in test_modifier 2233- ac/surface: add more elements to meta equations because HTILE can use them 2234- radeonsi: use SET_SH_REG_INDEX with index=3 for registers containing CU_EN 2235- ac/llvm: replace structured by vindex != NULL in ac_build_tbuffer_store 2236- ac/llvm: replace structured by vindex != NULL in ac_build_buffer_store_common 2237- radeonsi: move Arcturus code outside the gfx9 branch 2238- ac: update shadowed registers 2239- ac: set correct cache size per TCC for Yellow Carp 2240- amd: add support for gfx1036 and gfx1037 chips 2241- amd: update addrlib 2242- radeonsi: fix broken VK-GL buffer interop 2243- nir: validate write_mask for all intrinsics that have it 2244- nir: fix nir_io_semantics::gs_streams in nir_lower_io_to_scalar 2245- nir: add transform feedback info into nir_intrinsic_store_output 2246- nir: add shader_info::xfb_strides 2247- nir: scalarize transform feedback info in nir_lower_io_to_scalar 2248- nir: add nir_io_semantics::no_varying, no_sysval_output, and helpers 2249- nir: add nir_lower_io_passes() with new transform feedback 2250- nir: add nir_gather_xfb_info_from_intrinsics for lowered IO 2251- nir: add nir_print_xfb_info 2252- nir: pass nir_shader into nir_recompute_io_bases instead of func_impl 2253- gallium/util: add util_dump_stream_output_info 2254- gallium/aux: add helper nir_gather_stream_output_info 2255- st/mesa: do nir_lower_io() for inputs & outputs with transform feedback info 2256- ci: add point coord failures to d3d12 2257- ac/llvm: update LLVM processor names for gfx10.3 2258- ac: parse SET_SH_REG_INDEX packet 2259- ac: define PKT3_ATOMIC_MEM 2260- ac: add an environment variable that parses IBs in files 2261- radeonsi: fix register shadowing after the pm4 state size was decreased 2262- radeonsi: fix an assertion failure with register shadowing 2263- nir: fix an uninitialized variable valgrind warning in nir_group_loads 2264- ac/surface: fix an addrlib race condition on gfx9 2265- winsys/amdgpu: fix a mutex deadlock when we fail to create pipe_screen 2266- ac/llvm: set the correct cache policy for sparse buffer loads 2267- radeonsi: fix a crash when failing to create a context 2268 2269Mark Janes (1): 2270 2271- Revert "intel/fs: handle interpolation modes for at_sample and at_offset too" 2272 2273Martin Roukala (né Peres) (1): 2274 2275- ci/b2c: fix the generation of the IMAGE_UNDER_TEST variable 2276 2277Matt Coster (2): 2278 2279- pvr: Gate offline compiler build behind -Dtools=imagination 2280- pvr: ci: Initial freedesktop CI integration 2281 2282Matt Turner (8): 2283 2284- intel/perf: Don't print leading space from desc_units() 2285- intel/perf: Deduplicate perf counters 2286- intel/perf: Use a function to initialize perf counters 2287- intel/perf: Use slimmer intel_perf_query_counter_data struct 2288- intel/perf: Store indices to strings rather than pointers 2289- intel/perf: Mark intel_perf_counter_* enums as PACKED 2290- intel/perf: Fix mistake in description string 2291- intel/perf: Destination array calculation into function 2292 2293Matti Hamalainen (1): 2294 2295- aux/trace: fix dumping of pipe_texture_target 2296 2297Melissa Wen (1): 2298 2299- broadcom/simulator: enable multisync in the simulator 2300 2301Michael Olbrich (1): 2302 2303- crocus: export GEM handle with RDWR access rights 2304 2305Michel Dänzer (3): 2306 2307- ci: Use $CI_COMMIT_BRANCH 2308- ci: Use $CI_PIPELINE_SOURCE 2309- ci: Remove unused is-for-marge YAML anchor 2310 2311Michel Zou (4): 2312 2313- vulkan/wsi: drop unused wsi_create_win32_image 2314- lavapipe: set non-zero device/driver uuid 2315- lavapipe: fix i686 mingw build 2316- gallium: fix unused symbols warnings 2317 2318Mihai Preda (12): 2319 2320- radeonsi/tests: print PCI-id of GPU device under test 2321- radeonsi/tests: update piglit baseline on vega20 2322- radeonsi/tests: update glcts baseline on vega20 2323- radeonsi/tests: update piglit baseline on vega20 2324- amd/ac_gpu_info: fix warning on fread unused result 2325- radeonsi/tests: fix file left open in radeonsi-run-tests.py 2326- radeonsi/tests: add flakes option to radeonsi-run-tests.py 2327- radeonsi/tests: update baseline and flakes on vega20 2328- radeonsi: convert copy_image shader to NIR 2329- radeonsi: convert copy_image_1d_array shader to NIR 2330- radeonsi: merge the copy_image shader generators 2331- amd/llvm: Transition to LLVM "opaque pointers" 2332 2333Mike Blumenkrantz (538): 2334 2335- zink: disable PIPE_SHADER_CAP_FP16_CONST_BUFFERS 2336- llvmpipe: disable PIPE_SHADER_CAP_FP16_CONST_BUFFERS 2337- llvmpipe: ci updates 2338- zink: add VK_BUFFER_USAGE_CONDITIONAL_RENDERING_BIT_EXT for query binds 2339- zink: add synchronization for conditional render buffer 2340- zink: use scanout obj when returning resource param info 2341- zink: export PIPE_SHADER_CAP_TGSI_CONT_SUPPORTED 2342- zink: fix PIPE_CAP_TGSI_BALLOT export conditional 2343- zink: reject invalid draws 2344- zink: min/max blit region in coverage functions 2345- vk/sync: add asserts for timeline semaphore count matching 2346- gallium: add PIPE_CAP_CULL_DISTANCE_NOCOMBINE 2347- zink: export PIPE_CAP_CULL_DISTANCE_NOCOMBINE 2348- lavapipe: use util_pack_color_union() for generating clear colors 2349- zink: ci updates 2350- lavapipe: ci updates 2351- zink: ci updates 2352- zink: add Sample decorations to fragment shader inputs with sample shading 2353- zink: ci updates 2354- aux/draw: fix llvm tcs lane vec generation 2355- zink: anv (icl) ci updates 2356- zink: always set VkPipelineMultisampleStateCreateInfo::pSampleMask 2357- zink: ci updates 2358- zink: break out spirv shader dumping into separate function 2359- zink: make spirv_buffer_emit_word() return the word that was written 2360- zink: make spirv_builder_emit_exec_mode_literal() return the word for the param 2361- zink: store the tcs_vertices_out spirv word 2362- zink: store the tcs_vertices_out spirv word to the spirv_shader struct 2363- zink: split off CreateShaderModule into util function 2364- zink: store the spirv_shader to the zink_shader struct for generated tcs 2365- zink: add a tcs shader key 2366- zink: move pipeline tcs patch_vertices value to tcs shader key 2367- zink: implement generated tcs variants using spirv shortcut 2368- zink: ci updates 2369- zink: track internal conditional render state 2370- zink: always terminate conditional render when flushing a batch 2371- zink: restart conditional render when crossing batch boundary 2372- zink: activate conditional render for compute dispatch when necessary 2373- zink: ci updates 2374- zink: map R8G8B8X8_SRGB -> R8G8B8A8_SRGB 2375- zink: export PIPE_SHADER_CAP_INDIRECT_TEMP_ADDR 2376- zink: always invalidate streamout counter buffer if not resuming 2377- zink: ci updates 2378- zink: radv ci updates 2379- zink: radv ci updates 2380- zink: prune ci lists 2381- zink: allow null descriptor set layouts 2382- zink: remove zink_descriptor_util_init_null_set() 2383- zink: ci updates 2384- zink: add a flake channel 2385- lavapipe: EXT_depth_clip_control 2386- features: VK_EXT_depth_clip_control for lavapipe 2387- zink: update descriptor refs after starting renderpass 2388- zink: fix typo for image descriptor rebinds 2389- zink: add layout to sampler descriptor hash 2390- zink: skip extra descriptor lookups for images during barrier updates 2391- zink: set shader key size to 0 for non-generated tcs 2392- zink: fix cached descriptor set invalidation for array bindings 2393- zink: free push descriptor pools on deinit 2394- zink: don't free non-fbfetch dsl structs when switching to fbfetch 2395- nir/lower_psiz: create the store instruction more accurately 2396- nir/lower_psiz_mov: stop clobbering existing exports 2397- zink: update psiz handling to fix xfb output 2398- glx/drisw: handle GL_RESET_NOTIFICATION_STRATEGY 2399- zink: use VK_EXT_depth_clip_control when available 2400- zink: restrict clear flushing on sampler/image bind to compute binds 2401- zink: directly create surfaces for shader images 2402- zink: clamp 3d/array shader images to lower dimensionality using layer counts 2403- zink: more accurately clamp 3d fb surfaces to corresponding 2d target 2404- zink: add layer asserts for 3d imageview creation 2405- zink: remove a bunch of flakes 2406- zink: emit Aliased decorations for any image that isn't explicitly marked restrict 2407- zink: switch to u_foreach_bit for ntv image access decorations 2408- zink: remove loop from generated tcs 2409- zink: update query states before starting renderpass during draw 2410- zink: split out query suspending into util function 2411- zink: split primitives generated queries if xfb/gs states change 2412- zink: ci updates 2413- gallivm: fix debug prints for halfs 2414- gallivm: avoid division by zero when computing cube face 2415- lavapipe: heap-allocate rendering_state struct 2416- lavapipe: scan shaders for image/ssbo access and generate per-stage masks 2417- lavapipe: accurately set image/ssbo access based on shader usage 2418- zink: always update shader variants when rebinding a gfx program 2419- zink: add function for refcounting zink_program structs 2420- zink: use a fence for pipeline cache update jobs 2421- zink: wait on program cache fences before destroying programs 2422- zink: fix descriptor cache pointer array allocation 2423- zink: mark fbfetch push sets as non-cached 2424- zink: stop leaking descriptor sets 2425- zink: invalidate non-punted recycled descriptor sets that are not valid 2426- zink: hide descriptor debug behind #ifdef 2427- Revert "lavapipe: accurately set image/ssbo access based on shader usage" 2428- zink: run nir_lower_phis_to_scalar in optimization loop 2429- zink: fix 64bit float shader ops 2430- zink: ci updates 2431- llvmpipe: fix debug print iterating in set_framebuffer_state 2432- lavapipe: clamp clear attachments rects 2433- llvmpipe: clamp surface clear geometry 2434- lavapipe: update multisample state after blend state 2435- lavapipe: fix pipeline creation for blend and zs states 2436- aux/trace: more screen methods 2437- aux/trace: rzalloc the context struct 2438- aux/trace: dump clear colors as uints 2439- aux/trace: dump clear_texture colors 2440- aux/trace: dump more rasterizer state members 2441- aux/cso: stop tracing during cso_unbind() 2442- ci: add another stoney flake 2443- ci: more stoney flakes 2444- zink: use 64bit mask for xfb analysis 2445- zink: store the correct number of components for xfb packing outputs 2446- zink: correctly set xfb packed output offsets 2447- zink: fix xfb analysis variable finding for arrays 2448- zink: handle remaining xfb corner cases during analysis 2449- zink: store shader to ntv_context 2450- zink: handle spirv xfb insanity 2451- glsl: store OES/EXT point_size extension enablement to shader struct 2452- mesa/st: add a gl_program struct flag to skip psiz exports for xfb 2453- mesa/st: conditionally add pointsize outputs to ES tess/geom shaders 2454- nir/gather_info: check copy_deref instrs for writing outputs 2455- mesa/st: only add pointsize output if it doesn't exceed max component limit 2456- mesa: always set PointSizeEnabled for API_OPENGLES2 2457- mesa/st: simplify pointsize shader update conditional 2458- mesa/st: simplify pointsize precompile conditional 2459- mesa/st: precompile with API pointsize only if the shader doesn't have pointsize 2460- mesa/st: count FF shaders as needing psiz export for precompile 2461- mesa/st: check max output components for adding pointsize during precompile 2462- mesa/st: make export_point_size shader key clobber existing psiz 2463- zink: update radv fails 2464- zink: add another radv fail 2465- zink: lower dmod on AMD hardware 2466- lavapipe: add the full list of cts fails 2467- lavapipe: use the correct value for dynamic render resolve attachment indexing 2468- lavapipe: skip format checks for EXTENDED_USAGE 2469- lavapipe: ci updates 2470- lavapipe: run nir_opt_copy_prop_vars during optimization loop 2471- lavapipe: ci updates 2472- llvmpipe: fix occlusion queries with early depth test 2473- anv: fix xfb usage with rasterizer discard 2474- anv: fix CmdSetColorWriteEnableEXT for maximum rts 2475- anv: fix some dynamic rasterization discard cases in pipeline construction 2476- lavapipe: make device limits a physical device struct 2477- lavapipe: validate per-stage descriptor limits when creating pipeline layouts 2478- lavapipe: more descriptor validation 2479- zink: remove flake 2480- zink: anv ci updates 2481- zink: assert that the dynamic state array size is big enough 2482- zink: move dynamic state2 pipeline state to substruct in pipeline state 2483- zink: use dynamic rasterizer_discard state when possible 2484- lavapipe: always clone shader nir for shader states 2485- lavapipe: ref/unref pipeline layouts for pipeline creation 2486- lavapipe: implement vkGetDevice*MemoryRequirements 2487- lavapipe: set maxBufferSize for maintenance4 2488- lavapipe: maintenance4 2489- gallivm: implement nir_op_pack_32_4x8_split 2490- lavapipe: KHR_shader_integer_dot_product 2491- zink: add nir_var_function_temp support to ntv 2492- zink: add DOUBLE glsl type for streamout export 2493- zink: fix unreachable() location in ntv streamout info 2494- zink: export indirect io pipe caps 2495- lavapipe: strip unneeded scoped barriers 2496- llvmpipe: set nir_shader_compiler_options::use_scoped_barrier 2497- lavapipe: enable KHR_memory_model support 2498- gallium: add flag to draw info to indicate converted draws 2499- lavapipe: use VkFormatFeatureFlags2 in format detection 2500- lavapipe: KHR_format_feature_flags2 2501- llvmpipe: add handling for zeroing cs shared memory 2502- lavapipe: KHR_zero_initialize_workgroup_memory 2503- lavapipe: EXT_pipeline_creation_feedback 2504- lavapipe: EXT_pipeline_creation_cache_control 2505- gallivm: fix oob image detection for cube/1dArray/2dArray/cubeArray 2506- lavapipe: EXT_image_robustness 2507- zink: flag sample locations for re-set on batch flush 2508- zink: force-add usage when adding last-ref tracking 2509- zink: only update usage on buffer rebind if rebinds occurred 2510- zink: set vbo resource usage on bind 2511- zink: add some nice docs for batch usage and tracking 2512- features: fix some vk extension listings 2513- Revert "features: fix some vk extension listings" 2514- lavapipe: store number of immutable samplers to pipeline layout 2515- lavapipe: break out resolves into separate functions 2516- lavapipe: zalloc lvp_image_view structs 2517- zink: store vertices statistics query to context 2518- zink: handle conversion for vertices statistics query with LINE_LOOP draws 2519- nir/lower_is_helper_invocation: create load_helper_invocation instr with bitsize=1 2520- lavapipe: run some shader passes for demote handling 2521- lavapipe: EXT_shader_demote_to_helper_invocation 2522- lavapipe: extend demote->discard pass to handle terminate 2523- lavapipe: KHR_shader_terminate_invocation 2524- lavapipe: don't emit compute states during draw 2525- lavapipe: zalloc pipeline layout structs 2526- lavapipe: add a stream uploader to rendering_state and queue objects 2527- lavapipe: save pipeline stages that push constants are active on 2528- lavapipe: use stream uploader for push constant upload 2529- lavapipe: remove unused struct member 2530- lavapipe: implement EXT_inline_uniform_block 2531- lavapipe: EXT_inline_uniform_block 2532- llvmpipe: fix gl_NumSubgroups 2533- llvmpipe: fix subgroup id construction 2534- llvmpipe: fix variable naming insanity in cs generator 2535- lavapipe: EXT_subgroup_size_control 2536- lavapipe: add a GetPhysicalDeviceToolPropertiesEXT stub 2537- lavapipe 1.3 2538- docs: update lavapipe features and relnotes 2539- lavapipe: fix typo in set_event execution 2540- lavapipe: add sync2 cmdbuf method implementations 2541- lavapipe: add QueueSubmit2 implementation 2542- lavapipe: KHR_synchronization2 2543- zink: hook up EXT_color_write_enable 2544- zink: force disable rasterization discard if primgen query is active 2545- zink: start a unified driver workarounds struct 2546- zink: disable color_write_enable on ANV 2547- zink: use EXT_color_write_enable when possible 2548- zink: add an alternate path for EXT_color_write_enable usage 2549- zink: use EXT_color_write_enable to mask out primgen+rasterizer_discard output 2550- zink: use the right query type for primitives generated 2551- zink: ci updates 2552- zink: ci updates 2553- zink: create compute pipeline after updating shader variants 2554- zink: use the current compute shader, not the base one 2555- zink: add RADV to list of broken drivers for EXT_color_write_enable 2556- draw: fix gs vertex stream counting 2557- draw: fix nonzero stream primitives generated queries 2558- nir/lower_tex: add txp lowering option for arrays 2559- zink: lower txp for cube and ms textures 2560- zink: lower txp for array textures 2561- zink: update radv ci 2562- zink: fix up color_write_enable workaround 2563- zink: flush clears before toggling color write 2564- zink: update anv icl ci list 2565- mesa/st: add special casing for pointsize constant updating during validate 2566- zink: update radv fails 2567- zink: add anv cts skips from waiver 2568- zink: add another radv flake 2569- zink: use z24_in_z32f support and radv ci updates 2570- radv: fix CmdSetColorWriteEnableEXT(attachmentCount==MAX_RTS) 2571- zink: add a couple flakes 2572- lavapipe: run optimize loop before krangling pipeline layout 2573- zink: run piglit's gpu profile 2574- vulkan: update spec to 1.3.210 2575- lavapipe: EXT_primitives_generated_query 2576- lavapipe: support KHR_pipeline_library 2577- lavapipe: implement EXT_graphics_pipeline_library 2578- lavapipe: display EXT_graphics_pipeline_library 2579- doc: update extensions for lavapipe 2580- zink: fix error logging for 2d z/s checking 2581- zink: break out CmdSetColorWriteEnableEXT to util function 2582- zink: remove anv workaround for broken color writes 2583- zink: update radv baseline 2584- zink: more radv fails 2585- zink: add in radv passes to baseline 2586- intel/isl: fix 2d view of 3d textures 2587- crocus: assert that 3d samplerview base_array_layer is zero 2588- iris: assert that samplerview base_array_layer is zero for hw < skl 2589- vulkan: update more headers to 1.3.210 2590- lavapipe: fix shader indexing of sampler arrays with const array index 2591- lavapipe: fix xfb availability query copying 2592- lavapipe: allow timeline progress in GetSemaphoreCounterValue 2593- lavapipe: enforce monotonic timeline incrementing 2594- zink: force push descriptors cache update if hashing detects changes 2595- llvmpipe: handle sampling from 2d views of 3d images 2596- zink: add driver workaround for broken EXT_depth_clip_control 2597- mesa/st: rework atom flagging when pointsize changes 2598- mesa/st: always flag last vertex stage constants for upload on pointsize change 2599- mesa/st: rework pointsize constant uploads 2600- mesa/st: don't add pointsize to ES programs if it already exists 2601- zink: update radv ci baseline 2602- zink: remove radv cwrite driver workaround 2603- lavapipe: add an env var to enable poisoning memory allocations 2604- zink: set LVP_POISON_MEMORY for ci 2605- lavapipe: set LVP_POISON_MEMORY for ci 2606- aux/trace: dump format in set_shader_images 2607- vulkan: spec update to 1.3.211 2608- vulkan: check 3D image type for VK_IMAGE_CREATE_2D_VIEW_COMPATIBLE_BIT_EXT 2609- lavapipe: expose VK_EXT_image_2d_view_of_3d 2610- anv: expose VK_EXT_image_2d_view_of_3d 2611- docs: update features for VK_EXT_image_2d_view_of_3d 2612- zink: fix barrier generation for ssbo descriptors 2613- zink: only uncommit sparse pages that have been committed 2614- zink: merge stencil test case for draw-time dynamic state 2615- zink: always set stencil dynamic states before draw 2616- zink: set Geometry capability for fs if geometry inputs are read 2617- zink: handle 1bit xor as OpLogicalNotEqual 2618- zink: clamp min viewport width to 1 2619- zink: apply fb attachment layout to dummy attachments 2620- zink: convert all 64bit vertex attribs to 32bit 2621- zink: fix max geometry input component advertising 2622- zink: prune shader i/o 2623- zink: unset resource layout+access when doing storage setup 2624- zink: use EXT_image_2d_view_of_3d 2625- zink: refactor zink_descriptor_util_image_layout_eval 2626- zink: remove commented code 2627- zink: further simplify zs case for zink_descriptor_util_image_layout_eval 2628- zink: add a renderpass flag for mixed zs layout 2629- zink: add a ctx param to zink_descriptor_util_image_layout_eval 2630- zink: delete some code in get_layout_for_binding() 2631- zink: use store op NONE when necessary for depth usage 2632- zink: update samplerview layouts for zs attachments during renderpass prep 2633- zink: use GENERAL layout for mixed zs fb attachments 2634- zink: switch warn_missing_feature to mesa_logw 2635- zink: add a param to warn_missing_feature() macro 2636- zink: fix warning text in missing feature macro 2637- zink: rework missing feature warnings 2638- zink: run shader optimize loop during initial create 2639- zink: prune shader i/o more aggressively 2640- zink: use local variable more consistently in producer shader i/o assign 2641- zink: use local variable in consumer shader i/o assign to match producer usage 2642- zink: simplify shader i/o assignment 2643- st/manager: update framebuffer size if texture has been resized 2644- zink: move flush queue init down a little further 2645- zink: put screen param into flush queue global data 2646- zink: move update_framebuffer_state() higher up in file 2647- zink: move blit src/dst decls up in function 2648- zink: change early returns in zink_blit to gotos 2649- zink: use two submits for every queue submit 2650- zink: add VK_KHR_swapchain_mutable_format 2651- zink: add fail logging for drmPrimeFDToHandle 2652- zink: split surface creation more to allow disabling caching 2653- zink: move drirc handling up 2654- zink: pass index to unbind_fb_surface 2655- zink: move variable decl up in unbind_fb_surface 2656- zink: check whether clear is enabled before applying in unbind 2657- zink: add kopper api 2658- zink: it's kopperin' time 2659- zink: export PIPE_CAP_DEVICE_RESET_STATUS_QUERY 2660- driconf: add override for Xwayland 2661- zink: handle zombie swapchains 2662- zink: ci fixup 2663- zink: update ci list 2664- radv: improve failure logging for amdgpu on init 2665- zink: rename a variable 2666- zink: add a gently mangled version of the d3d12 cubemap -> array compiler pass 2667- zink: support nir_op_imod 2668- zink: specify struct member name when copying inline uniforms for gfx variants 2669- zink: handle shader key variants that have nonseamless cubemaps 2670- zink: set nonseamless hint for sampler states 2671- zink: create an array view for all cube samplerviews 2672- zink: handle nonseamless cube sampler binding 2673- zink: run the cubemap -> array compiler pass if the shader key is set 2674- zink: export PIPE_CAP_SEAMLESS_CUBE_MAP_PER_TEXTURE 2675- zink: ci updates 2676- egl: implement more hooks for swrast 2677- zink: handle deferred swapchain resource flushing 2678- zink: only apply swapchain behavior in flush_resource for swapchain images 2679- zink: handle swapchain readbacks when a present is pending 2680- kopper: add a dmabuf-free image interface for use with sw drivers 2681- zink: set nir_shader_compiler_options::has_txs 2682- nir/lower_tex: avoid adding invalid LOD to RECT textures 2683- zink: allow lod for RECT sampler types 2684- zink: adds refs to user index buffers when tc is not active 2685- lavapipe: KHR_swapchain_mutable_format 2686- zink: add error logging for SRGB framebuffer without KHR_swapchain_mutable_format 2687- zink: ci updates 2688- docs: update features for lavapipe 2689- mesa: add a bool indicating when pointsize == 1.0 2690- nir/lower_point_size_mov: fix check for overwriting existing pointsize 2691- mesa/st: declare added pointsize var as hidden 2692- mesa/st: always inject a 1.0 pointsize for vertex stages 2693- mesa/st: also add pointsize to fixedfunction vertex shaders as needed 2694- mesa/st: only use constant upload pointsize variants if pointsize != 1.0 2695- mesa/st: only flag pointsize constant uploads if they're needed 2696- mesa/st: don't precompile the pointsize upload variant anymore 2697- mesa/st: simplify st_can_add_pointsize_to_program iterator 2698- zink: handle SUBPASS_MS in ntv 2699- zink: handle multisampled fbfetch 2700- zink: don't emit SpvCapabilityStorageImageMultisample for fbfetch 2701- zink: only trigger deferred present barrier if swapchain has acquired 2702- zink: unset deferred present barrier on flush 2703- zink: only get swapchain present semaphore on batch flush after acquire 2704- zink: only get swapchain present semaphore on batch flush if not presented 2705- zink: clamp out partial texels when creating bufferviews 2706- zink: bitcast InterpolateAtOffset offset to fvec 2707- zink: hook up sync2 extension 2708- zink: add handling for !sync2 in renderpass dependencies 2709- zink: don't rely on implicit access for generated barriers 2710- zink: don't generate VK_ACCESS_SHADER_READ_BIT barrier for vertex inputs 2711- zink/kopper: don't use generated include in kopper interface 2712- ci: disable v3dv arm64 jobs 2713- mesa: set PointSizeIsOne on context creation 2714- mesa/st: fix pointsize adding check 2715- mesa/st: handle copy_deref cases for adding pointsize 2716- nir/lower_point_size_mov: handle case where gl_Position isn't written 2717- mesa/st: handle adding pointsize when gl_Position is never written 2718- zink: ci updates 2719- zink: clamp cube size queries to 2 return components 2720- zink only use zs-specific layout for zs attachments 2721- zink: remove compiled conditional for lavapipe usage 2722- zink: add a self-dependency for fbfetch renderpasses 2723- zink: rework texture_barrier hook 2724- zink: force texture barriers when performing in-renderpass clears 2725- zink: only do swapchain update during fb setup if swapchain is active 2726- zink: update radv ci baseline 2727- zink: reorganize radv ci baseline 2728- zink: update radv piglit baseline 2729- zink: add issue notes for remaining radv fails 2730- zink: refactor copy_region path in zink_blit to util function 2731- zink: try copy region first for non-resolve blits 2732- zink: prune unused st-injected pointsize exports 2733- zink: handle 0 ubos and 0 ssbos in pipeline layout 2734- zink: create pipeline layout if only bindless descriptor set is used 2735- radv: set read/write without format flags for supported texel buffers 2736- zink: finish up radv piglit baseline updates 2737- features: mark off ARB_seamless_cubemap_per_texture for zink 2738- kopper: print better error message if loader not detected 2739- egl: don't make LIBGL_ALWAYS_SOFTWARE and MESA_LOADER_DRIVER_OVERRIDE=zink exclusive 2740- zink: ZINK_USE_LAVAPIPE -> LIBGL_ALWAYS_SOFTWARE 2741- aux/trace: clean up some zink+lavapipe tracing awfulness 2742- zink: rework choose_pdev to (finally) be competent 2743- zink: use the calculated last struct member idx for ssbo size in ntv 2744- zink: avoid creating ssbo variable types with multiple runtime arrays 2745- zink: set optimal tiling on swapchain images 2746- nir/lower_tex: fix rect queries with lower_rect set 2747- mesa/st: set normalized coords for RECT samplers if rects are unsupported 2748- zink: remove tcs patch slot map 2749- zink: reject resource creation if format features don't match attachment 2750- zink: use mixed zs renderpass for depth read/write 2751- zink: unconditionally set line width on rasterizer state change 2752- zink: support restart with PIPE_PRIM_LINES_ADJACENCY if ext is available 2753- zink: fix extended restart prim types without dynamic state2 2754- zink: make a kopper debug print into an error 2755- zink: rework zink_kopper_update() assert 2756- zink: set VK_QUERY_RESULT_WAIT_BIT when copying to qbo 2757- zink: fix synchronization when drawing from streamout 2758- zink: fix xfb counter buffer barriers 2759- zink: remove xfb_barrier flag 2760- zink: handle device-local unsynchronized maps 2761- util/draw: fix map size of indirect buffer in util_draw_indirect_read 2762- util/draw: handle draw_count=0 when reading indirect parameters 2763- util/draw: fix indirect draw count readback 2764- zink: move the kopper present fence to the displaytarget object 2765- wgl: support GL 4.6 2766- zink: fix tcs control barriers for use without vk memory model 2767- zink: fix semantics mask for compute control barriers 2768- zink: add synchronization for buffer clears 2769- mesa/st: clamp GL_RENDERBUFFER to GL_TEXTURE_2D for sparse queries 2770- glsl/nir: set new_style_shadow for sparse tex ops as necessary 2771- zink: fix group memory barrier emission 2772- vulkan: bump layer api versions to current vk header version 2773- kopper: always fetch and store drawable info 2774- kopper: move drawable geometry updating up in function 2775- kopper: store whether screen has dmabuf support 2776- kopper: copy a bunch of code for texture_from_pixmap 2777- kopper: add DISPLAY_TARGET bind for depth buffer 2778- zink: fix/improve swapchain surface info updating 2779- zink: fix up swapchain depth buffer geometry during fb update 2780- zink: ci update 2781- drisw: remove dead code 2782- zink: add a util function for finding swapchain resource 2783- zink: flag swapchains when updating fails 2784- zink: handle dead swapchains in acquire 2785- zink: unset image layout when killing swapchain 2786- zink: rework swapchain object check for acquires 2787- zink: fix return for zink_kopper_acquire 2788- zink: handle swapchain acquire failures more directly 2789- zink: flag swapchain resources as swapchains 2790- zink/kopper: add a mechanism for checking swapchain status 2791- kopper: add some error logging for pixmap->texture failure 2792- kopper: check for modifiers to use modifier functionality 2793- zink/kopper: add a function for determining if running on software 2794- kopper: check whether zink is using sw 2795- kopper: rename a confusing variable 2796- kopper: implement texture-from-pixmap for software (non-dmabuf) 2797- zink: fix dmabuf plane returns 2798- zink: export fd info for all 2d images 2799- zink: ignore KMS handle types 2800- zink: remove drm_fd 2801- llvmpipe: disable statistic queries during draw if queries are disabled 2802- llvmpipe: disable compute statistics queries if queries are disabled 2803- wgl: always set alpha on kopper windows 2804- wgl: don't auto-load zink before software drivers 2805- zink: add supported present modes to kopper displaytarget 2806- lavapipe: lower quad_broadcast intrinsics 2807- kopper: invalidate drawables when resizing textures in place 2808- kopper: pass the current context to dri_flush 2809- nir/gather_info: flag fbfetch on subpass image loads 2810- llvmpipe: never infer early zs tests when fbfetch is active 2811- lavapipe: enqueue pipeline destruction 2812- zink: fix null buffer/surface formats 2813- zink: add more image usage for null surfaces 2814- zink: use descriptor surfaces for notemplates ref updating 2815- kopper: remove unused zink include 2816- zink: remove unused headers/struct members/linkage 2817- zink: outdent some code 2818- zink: add a flag to indicate whether a resource is exportable 2819- zink: break out resource bind-adding into separate function for reuse 2820- zink: introduce copy contexts 2821- zink: create a copy context for the screen on init 2822- Revert "zink: export fd info for all 2d images" 2823- zink: always check for fb rebinds when starting renderpass 2824- zink: unset pipe_resource::next pointers when creating resource copies 2825- zink: add a bind flag to indicate a resource is exported as a dmabuf 2826- zink: fix conditional for modifier usage 2827- zink: force memory exports if dmabuf bind is specified 2828- zink: add a LINEAR drm modifier if rebinding to add dmabuf export 2829- zink: rebind resources for export as needed 2830- zink: create images with modifiers any time there is an import handle 2831- zink: pass sparse backing page offset to binding function 2832- zink: pass sparse bind bo offset through for texture binds 2833- zink: set all usage flags when querying sparse features 2834- zink: fix multisample conditional in sparse image query 2835- zink: bump number of image binds that can be batched to 50 2836- zink: semi-handle 1D sparse texture rewrites for drivers that don't support them 2837- zink: ci regressions 2838- util/blitter: fix sampler restore with 0 saved samplers 2839- gallivm: fix ssbo oob reads 2840- lavapipe: fix CmdPushDescriptorSetWithTemplateKHR with refcounting 2841- st/draw_feedback: set constant buffer stride 2842- gallivm/draw: fix oob ubo reads 2843- llvmpipe: always set ssbo data pointers for draw 2844- gallivm: fix oob txf swizzling 2845- zink: fix up sparse texture sampling for shadow samplers 2846- zink: clamp out min_lod operands for explicit lod ops 2847- zink: fix sparse binding for arrayed textures 2848- zink: set sparse flag in cubemap lowering 2849- zink: fix sparse texture depth calcs for arrayed textures 2850- zink: remove implicit query resets 2851- zink: remove refs from shader images 2852- zink: reuse local res pointer in set_shader_images 2853- zink: simplify dumb update flagging in set_shader_images 2854- zink: read shader image r/w usage from incoming data struct 2855- zink: copy incoming shader image struct after doing updates 2856- zink: stop leaking shader image surfaces 2857- zink: fix 3d shader image miplevel calc for depth 2858- zink: directly reuse surface ivci when rebinding 2859- lavapipe: ignore depth/stencil states if attachment is not present 2860- lavapipe: fix pipeline handling for dynamic render zs state with pipeline library 2861- spirv: fix barrier scope assert 2862- zink: never create a sw context unless sw is requested 2863- zink: only reject non-blittable resource creation if it has no other binds 2864- zink: add separate error message for push descriptor set alloc fail 2865- zink: add extra validation for resource creation 2866- zink: remove input attachment usage when pruning surface usage 2867- zink: add extended usage and retry if resource creation fails 2868- zink: fix surface/bufferview cache comparisons 2869- zink: force render target usage for sampler-only resources 2870- zink: clamp 1D_ARRAY sparse textures to 2D as needed 2871- zink: handle PIPE_BUFFER sparse texture queries 2872- zink: fix non-dynamic vertex stride update flagging 2873 2874Mykhailo Skorokhodov (1): 2875 2876- intel/fs: Enable b2f(inot(a)) and b2i(inot(a)) optimization for Gfx12+ 2877 2878Nanley Chery (24): 2879 2880- iris: Don't fast clear with the view format 2881- iris: Drop format param from fast_clear_color 2882- iris: Drop res param from surf_state_offset_for_aux 2883- iris: Add and use iris_surface_state::aux_usages 2884- iris: Add and use use_surface_state 2885- iris: Add and use fill_surface_states 2886- iris: Inline some surface_state.cpu references 2887- iris: Avoid making some invalid CCS surface states 2888- intel/isl: Add format assertions for surfaces using CCS 2889- iris: Use iris_sample_with_depth_aux more often 2890- iris: Compute aux.possible_usages from aux.usage 2891- iris: Drop the iris_resource aux usage bit fields 2892- anv: Don't disable HiZ/MCS in anv_BindImageMemory2 2893- anv: Fall back to HiZ when disabling CCS on HiZ+CCS 2894- anv: Add a perf warning in anv_BindImageMemory2 2895- Revert "anv: Require the local heap for CCS on XeHP" 2896- anv: Change a parameter of the implicit layout fn 2897- anv: Refactor anv_image_init_from_create_info 2898- anv: Disable aux if the explicit modifier lacks it 2899- Revert "anv: Disable CCS_E for some 8/16bpp copies on TGL+" 2900- ci/anv: Changes from enabling 8/16bpp CCS more 2901- iris: Return a 64B stride for clear color plane 2902- intel/isl: Add a score for I915_FORMAT_MOD_4_TILED 2903- isl,iris: Add DG2 CCS modifier support for XeHP 2904 2905Neha Bhende (1): 2906 2907- svga: store shared_mem_size in svga_compute_shader instead of svga_context 2908 2909Oleksandr Gabrylchuk (1): 2910 2911- venus: Implement guest vram blob type. 2912 2913Omar Akkila (7): 2914 2915- venus: Implement VK_EXT_extended_dynamic_state commands 2916- venus: Advertise VK_EXT_extended_dynamic_state support 2917- lavapipe: Lift fence check into dedicated function 2918- ci: uprev Fossilize 2919- ci: uprev vkd3d-proton to v2.6 2920- ci: cherry-pick deqp fix for zlib dependency 2921- ci: bump VK-GL-CTS to 1.3.1.1 2922 2923Otavio Pontes (1): 2924 2925- nir: Use proper macro to set bits of variable correctly 2926 2927Paulo Zanoni (5): 2928 2929- iris: handle IRIS_MEMZONE_BINDER with a real vma_heap like the others 2930- iris: have a single border color pool per bufmgr 2931- iris: use the same VM for every context 2932- iris: fix register spilling on compute shaders on XeHP 2933- iris: fix race condition during busy tracking 2934 2935Pavel Ondračka (23): 2936 2937- r300: fix deadcode elimination in loops with breaks 2938- r300: fix transformation of abs modifiers with negate 2939- r300: use nir lowering for sin and cos on R300 and R400 2940- r300: transform vs sin and cos input to [-PI,PI] range in NIR 2941- r300: transform fs sin and cos input to [0,1) range in NIR 2942- r300: only print shader statistics when compilation succeeds 2943- r300: use %u specifiers when printing unsigned stats values 2944- r300: report number of loops in shader statistics 2945- r300: add predicate instructions to statistics of vertex shaders 2946- r300: remove some dead logic in tex pair scheduling 2947- r300: schedule TEX instructions before OUT instructions 2948- r300: set PVS_LAST_VTX_SRC_INST properly to last input read 2949- r300: respect output_semantic_index when writing colors 2950- r300: don't move position output to the end when duplicating it for WPOS 2951- r300: optimize single write scenarios in rc_copy_output 2952- r300: restructure r300_vertex_shader 2953- r300: move r300_init_vs_outputs to r300_translate_vertex_shader 2954- r300: only output wpos in vertex shaders when needed 2955- r300: set PVS_XYZW_VALID_INST properly to last position write 2956- r300: don't assume position is always OUT[0] in rc_copy_output 2957- r300: move pointer dereference after a NULL check 2958- nine: check hardware support before using vertex texture 2959- r300: set PIPE_BIND_CONSTANT_BUFFER for const_uploader 2960 2961Pierre-Eric Pelloux-Prayer (39): 2962 2963- radeonsi: limit loop unrolling for LLVM < 13 2964- gallium/dri: add missing PIPE_BIND_DRI_PRIME handling 2965- gallium: rename PIPE_BIND_DRI_PRIME 2966- radeonsi: create prime buffers as uncached 2967- radeonsi/blit: relax conditions to use sdma copy for prime buffers 2968- radeonsi: add SI_PROFILE_CLAMP_DIV_BY_ZERO 2969- radeonsi: use SI_PROFILE_CLAMP_DIV_BY_ZERO for viewperf 2970- radeonsi/drirc: use force_gl_vendor for Maya 2971- radeonsi: change rounding mode to round to even 2972- radeonsi/tests: always add the --gpu argument 2973- radeonsi: use 1 shader compilation thread if NIR_PRINT is used 2974- gallium/u_threaded: late alloc cpu_storage 2975- radeonsi: enable tc cpu_storage by default 2976- gallium/tc: warn if an app is incompatible with cpu_storage 2977- util/slab: add slab_zalloc 2978- gallium/tc: zero alloc transfers 2979- radeonsi: don't clear framebuffer.state before dcc decomp 2980- radeonsi: replace opencoded slab_zalloc 2981- r600: replace opencoded slab_zalloc 2982- zink: replace opencoded slab_zalloc 2983- virgl: replace opencoded slab_zalloc 2984- vc4: replace opencoded slab_zalloc 2985- v3d: replace opencoded slab_zalloc 2986- iris: replace opencoded slab_zalloc 2987- lima: replace opencoded slab_zalloc 2988- freedreno: replace opencoded slab_zalloc 2989- etnaviv: replace opencoded slab_zalloc 2990- d3d12: replace opencoded slab_zalloc 2991- crocus: replace opencoded slab_zalloc 2992- glsl/nir/linker: update shader_storage_blocks_write_access for SPIR-V 2993- glsl/nir/linker: fix shader_storage_blocks_write_access 2994- drirc: enable radeonsi_zerovram for Black Geyser 2995- docs: document useful radeonsi env variables 2996- radeonsi: drop LLVM global instruction selector 2997- ac: remove LLVM 4.0 workaround 2998- ac/surface: adjust gfx9.pitch[*] based on surf->blk_w 2999- radeonsi: don't use wave32 for GE on gfx10 if culling is used 3000- amd: fix ac_build_mbcnt_add in wave32 mode 3001- glx: set max values for pbuffer width / height 3002 3003Qiang Yu (10): 3004 3005- radeonsi: workaround Specviewperf13 Catia hang on GFX9 3006- radeonsi: fix depth stencil multi sample texture blit 3007- glx: fix pbuffer refcount init 3008- glx: merge drawable release to the same function 3009- glx: keep native window glx drawable by driconf option 3010- driconf: add Abaqus configs 3011- util/util_vertex_state_cache: remove error check when deinit 3012- nir/linker: disable varying from uniform lowering by default 3013- lima: enable nir lower_varying_from_uniform 3014- nir/linker: set varying from uniform as flat 3015 3016Rajnesh Kanwal (6): 3017 3018- pvr: Implement vkCreateSampler and vkDestroySampler APIs. 3019- pvr: Use vk_common_GetDeviceQueue API. 3020- pvr: Check if the buffer/image was bound before unbinding. 3021- amd: Use common u_format.h implementation for vk_format_get_component_bits. 3022- vulkan: Move common format function to vulkan/util/vk_format.h 3023- pvr: Remove logic to set vk_device::alloc. 3024 3025Renato Pereyra (2): 3026 3027- venus: Increase the base sleep of vn_relax 3028- Revert "venus: Increase the base sleep of vn_relax" 3029 3030Rhys Perry (74): 3031 3032- aco/tests: add a bunch more building helpers 3033- aco/tests: implement sub-dword program inputs 3034- aco: don't combine fneg/fabs of different bit-size 3035- aco: don't apply omod/clamp of different bit-size 3036- aco: don't combine add/mul of different bit-size 3037- aco: fix neg(mul)/abs(mul) optimization with different bit-size 3038- aco: add test for optimizations with casts 3039- aco: don't encode src2 for v_writelane_b32_e64 3040- radv: fix R_02881C_PA_CL_VS_OUT_CNTL with mixed cull/clip distances 3041- nir/validate: don't validate the GC list by default 3042- nir/validate: don't add instrs not present in shader to shader_gc_list 3043- anv: Enable nir_opt_access 3044- ac/nir: remove TCS nir_var_shader_out memory barrier 3045- radv: include disable_aniso_single_level and adjust_frag_coord_z in key 3046- aco: remove vcc hint from branch definitions 3047- aco/ra: add get_reg_phi() helper 3048- aco/ra: fix register allocation of branch definitions 3049- aco: add validate_instr_defs() 3050- aco: fix branch definition validation 3051- aco/tests: add test for branch definition RA 3052- aco: rework removal of jumps over branches 3053- aco/insert_exec_mask: fix top-level to-exact with non-global exact mask 3054- aco/insert_exec_mask: use get_exec_op 3055- aco/insert_exec_mask: optimize top-level transition to exact before demote 3056- aco: split and recombine unaligned sgpr inputs 3057- radv,aco,ac/llvm: fix indirect dispatches on the compute queue on GFX7-10 3058- aco: fix fp16 opcode definitions 3059- aco: improve support for v_fma_mix 3060- aco: refactor selection of mad/fma 3061- aco: use v_fma_mix to combine mul/add/fma input conversions 3062- aco: combine add/mul as v_fma_mix into fma 3063- aco: apply clamp to v_fma_mix 3064- aco: use v_fma_mix to combine mul/add/fma output conversions 3065- aco/tests: add v_fma_mix tests 3066- nir: add load_{scalar,vector}_arg_amd and load_smem_amd intrinsics 3067- aco: implement load_{scalar,vector}_arg_amd and load_smem_amd 3068- ac/llvm: implement implement load_{scalar,vector}_arg_amd and load_smem_amd 3069- radv: move radv_declare_shader_args() out of shader_variant_compile() 3070- nir: allow bindless image/texture/sampler handles to be vectors 3071- ac/llvm: remove deref requirement for image fmask loads 3072- ac/llvm: implement nir_intrinsic_bindless_image_sparse_load 3073- ac/llvm: remove deref chasing for tg4 integer workaround 3074- ac/llvm: implement nir_tex_src_{texture,sampler}_handle 3075- radv,aco: lower vulkan_resource_index in NIR 3076- radv,aco: lower buffer descriptor loads in NIR 3077- radv,aco: lower texture descriptor loads in NIR 3078- radv,aco: lower image descriptor loads in NIR 3079- aco: fix RA validation of 16-bit fma_mix operands 3080- aco: don't use v_mad_mix on GFX9 if 16-bit denormals must be preserved 3081- radv: allow inline push constants in more situations 3082- radv: allow holes in inline push constants 3083- radv,aco: implement 64-bit inline push constants 3084- radv: increase inline push constant limit if we can inline all constants 3085- radv: don't require robust vectorization for nir_var_mem_global 3086- aco: use vcc for 64-bit vgpr addition 3087- aco: use saddr for global access with sgpr address 3088- aco: don't expand smem/mubuf global loads 3089- nir: add _amd global access intrinsics 3090- ac/llvm: implement _amd global access intrinsics 3091- aco: implement _amd global access intrinsics 3092- aco: increase global_load_params.max_const_offset_plus_one 3093- radv,ac/nir: lower global access to _amd global access intrinsics 3094- aco: remove old global access intrinsics 3095- nir: add load_shared2_amd and store_shared2_amd 3096- nir/opt_load_store_vectorize: fix broken indentation 3097- nir/opt_load_store_vectorize: create load_shared2_amd/store_shared2_amd 3098- aco: fix signedness of DS_instruction::offset0/1 3099- aco: handle read2st64/write2st64 in optimizer 3100- aco: implement load_shared2_amd/store_shared2_amd 3101- ac/llvm: implement load_shared2_amd/store_shared2_amd 3102- radv: use load_shared2_amd/store_shared2_amd 3103- ac/nir: properly handle large global access constant offsets 3104- radv: fix clearing of TRUNC_COORD with tg4 and immutable samplers 3105- aco/ra: fix vgpr_limit 3106 3107Rob Clark (28): 3108 3109- freedreno/registers: Add a couple regs we need for kernel 3110- gallium/dri: Extend image extension to support in-fence 3111- android: Push in-fence-fd down to driver 3112- egl+libsync: Add helper to complain about invalid fence fd's 3113- freedreno: Update uapi header 3114- freedreno/drm: Add SYSPROF param 3115- freedreno/perfetto+fdperf: Set SYSPROF param 3116- gallium/dri: Add missing in_fence_fd initialization 3117- turnip: Don't call getenv() directly 3118- mesa: Fix discard_framebuffer for fbo vs winsys 3119- mesa: Easier shader capture for android 3120- freedreno/ir3: Remove unused define 3121- freedreno/ir3: Don't try re-swapping cat3 srcs 3122- freedreno/drm: Split msm backend into subdir 3123- freedreno/drm: Add fd_device_open() helper 3124- freedreno/drm: Rework bo creation path 3125- freedreno/drm: Add FD_BO_NOMAP hint 3126- freedreno/drm: Add FD_BO_SHARED hint 3127- freedreno/drm: Add fd_bo_upload() 3128- freedreno/drm: Avoid CPU_PREP ioctl if bo is idle 3129- freedreno/drm: Move submit_queue to base 3130- freedreno/drm: Move bo idx to base 3131- freedreno/drm: Move ring_pool slab parent to base 3132- freedreno/drm: Extract out "softpin" submit/ringbuffer base class 3133- freedreno/drm: Reorder device destroy 3134- freedreno/drm: Add virtio backend 3135- pipe-loader: Try loading freedreno for virtgpu device 3136- freedreno/ci: Update a306 expectations 3137 3138Rohan Garg (4): 3139 3140- anv: Refactor descriptor copy 3141- anv: Handle VK_DESCRIPTOR_POOL_CREATE_HOST_ONLY_BIT_VALVE for descriptor sets 3142- anv: Drop dead code in anv_UpdateDescriptorSets 3143- iris: set a default EDSC flag 3144 3145Roman Stratiienko (2): 3146 3147- panfrost: Don't crash on panfrost_bo_create() with size==0 invocation 3148- android: Set max platform-sdk-version to 10000 3149 3150Sagar Ghuge (2): 3151 3152- intel/fs: Add Wa_22013689345 3153- intel/fs: Add Wa_14014435656 3154 3155Samuel Pitoiset (141): 3156 3157- radv: stop setting streamout state when a new pipeline is bound 3158- radv: add reference counting for descriptor set layouts 3159- Revert "radv: re-apply "Do not access set layout during vkCmdBindDescriptorSets."" 3160- radv: allow RADV_FORCE_VRS with pipeline VRS declared as dynamic 3161- radv: rewrite RADV_FORCE_VRS directly in NIR 3162- radv: do not force per-vertex VRS if there is no pixel shader 3163- radv: only emit the per-vertex VRS state if the pipeline forced it 3164- radv/ci: update CI lists for CTS 1.3.1.0 3165- radv/winsys: fix initializing debug/perftest options if multiple instances 3166- radv: fix allocating/uploading the trap handler shader 3167- radv: make the trap handler shader BO resident 3168- radv: fix finding shaders by PC 3169- radv: remove useless NULL checks in vkBind{Buffer,Image}Memory2() 3170- radv: remove set but unused radv_buffer::shareable 3171- nir: add nir_intrinsic_load_vrs_rates_amd 3172- radv: declare a new shader argument for loading the VRS rates 3173- ac/llvm: implement nir_intrinsic_load_vrs_rates_amd 3174- aco: implement nir_intrinsic_load_vrs_rates_amd 3175- radv: rework RADV_FORCE_VRS to make it more dynamic 3176- radv: only re-emit the per-vertex VRS rates if necessary 3177- radv: rename RADV_FORCE_VRS_NONE to RADV_FORCE_VRS_1x1 and accept 1x1 3178- radv: add RADV_FORCE_VRS_CONFIG_FILE to configure per-vertex VRS 3179- radv: allow applications to dynamically change RADV_FORCE_VRS 3180- radv: allow to force per-vertex VRS if the config file is present 3181- radv: enable radv_disable_aniso_single_level for The Evil Within 1&2 3182- radv: allow to force per-vertex VRS in the tessellation stage 3183- radv: do not enable per-vertex VRS if the FS uses gl_FragCoord 3184- ci: upgrade to libdrm 2.4.110 3185- meson: bump libdrm_amdgpu version to 2.4.110 3186- include/drm-uapi: update amdgpu_drm.h for new CTX OP to set/get stable pstates 3187- ac: add ac_gpu_info::has_stable_pstate 3188- radv/winsys: add support for new CTX OP to set/get stable pstates 3189- radv: set profile_peak when capturing with SQTT 3190- radv,aco,llvm: lower adjusting vertex alpha in NIR 3191- radv: fix build on BSD 3192- radv: initialize more depth/stencil states earlier 3193- radv: initialize VGT_GS_OUT_PRIM_TYPE earlier 3194- radv: remove useless radv_blend_state::single_cb_enable field 3195- radv: initialize extra state for internal pipelines at one place 3196- radv,drirc: move RADV workarounds to 00-radv-defaults.conf 3197- radv: remove color exports in presence of holes 3198- radv: disable DCC for Fable Anniversary, Dragons Dogma, GTA IV and more 3199- radv,aco: do not lower nir_op_pack_{unorm,snorm}_2x16 3200- nir: introduce nir_pack_{sint,uint}_2x16 instructions 3201- aco: implement nir_op_pack_{uint,sint}_2x16 3202- ac/nir: implement nir_op_pack_{uint,sint}_2x16 3203- radv/ci: skip dEQP-VK.renderpass2.depth_stencil_resolve.*_samplemask 3204- radv/ci: update list of expected failures 3205- radv/ci: remove unused files 3206- radv,aco,llvm: lower post shuffle vertex in NIR 3207- aco: always emit vk_cvt_pkrtz_f16_f32 for nir_op_pack_half_2x16_split 3208- radv: optimize the number of loaded components for VS inputs in NIR 3209- radv: fix the CS regalloc hang workaround on GFX6 and few GFX7 chips 3210- radv: rework the CS regalloc hang workaround 3211- amd: add PKT3_LOAD_SH_REG_INDEX 3212- radv: fix indirect dispatches on the compute queue on GFX10.3+ 3213- radv: stop waiting for DMA to be idle for all transfer operations 3214- radv: update inputs_read when lowering the view index 3215- radv: fix compatibility with VK_IMAGE_CREATE_EXTENDED_USAGE_BIT 3216- radv: remove useless check in radv_cmd_buffer_upload_data() 3217- radv: remove unnecessary NULL check in TrimCommandPool() 3218- radv: remove unnecessary check in FreeCommandBuffers() 3219- radv: move waiting for events to CmdWaitEvents2KHR() 3220- radv: stop zeroing radv_sample_locations_state in barriers 3221- radv: export the pipeline hash via VK_KHR_pipeline_executable_properties 3222- radv: remove VK_AMD_shader_info support 3223- radv: fix returning empty drmFormatModifierTilingFeatures 3224- radv/winsys: remove old comment about zerovram 3225- radv: only clear VRAM for app and descriptor BOs when set via drirc 3226- radv: fix missing destruction of the inotify thread 3227- radv: do not compute the cache UUID for LLVM if it's not used 3228- radv: enable radv_disable_aniso_single_level for DXVK/vkd3d 3229- radv,aco: lower color exports in NIR 3230- radv: lower load_sample_mask_in in NIR 3231- radv: lower adjusting gl_FragCoord.z for VRS in NIR 3232- radv: implement VK_EXT_depth_clip_control 3233- radv: advertise VK_EXT_depth_clip_control 3234- radv: remove now unused radv_nir_compiler_options::layout 3235- radv: fix mismatch between radv_GetPhysicalDeviceMemoryProperties*() 3236- radv: drop EXT or KHR suffixes for stuff promoted in Vulkan 1.3 3237- radv: lower has_multiview_view_index in NIR 3238- radv: save/restore the stencil write mask during internal driver operations 3239- radv: suspend/resume queries during internal driver operations 3240- radv: rework the workaround that disables DCC for incompatible copies 3241- radv: convert the meta depth decompression path to dynamic rendering 3242- radv: convert the meta fast clear flush path to dynamic rendering 3243- radv: convert the meta blit 2d path to dynamic rendering 3244- radv: convert the meta resolve color FS path to dynamic rendering 3245- radv: convert the meta resolve depth/stencil FS path to dynamic rendering 3246- radv: convert the meta resolve HW path to dynamic rendering 3247- radv: convert the meta blit path to dynamic rendering 3248- radv: convert the meta clear path to dynamic rendering 3249- radv: remove now unused radv_cmd_buffer_{begin,end}_render_pass() 3250- radv: use the common vk_framebuffer 3251- radv: enable VK_EXT_separate_stencil_usage 3252- radv: fix cleaning the image view for CmdCopyImageToBuffer() 3253- radv: save/restore the stencil reference during internal driver operations 3254- radv/ci: update CI lists against CTS 1.3.1.1 3255- radv: save/restore more dynamic states during internal driver operations 3256- radv: use radv_dynamic_state for saving/restoring meta operations 3257- radv: only declare dynamic states that are used by internal operations 3258- radv: enable VK_KHR_pipeline_library 3259- radv: fix dynamic raster discard with VK_EXT_depth_clip_control 3260- radv: add a new helper to determine if rasterization is enabled 3261- radv: mark all states declared dynamic at pipeline creation 3262- radv: do not check if VkPipelineRenderingCreateInfo is NULL 3263- radv: use radv_pipeline_has_ds_attachments() more 3264- radv: remove unused parameters in radv_get_{wave,ballot_bit}_size() 3265- radv: add radv_generate_pipeline_key() for common graphics/compute keys 3266- radv: add a new helper to initialize various type of pipelines 3267- radv: add radv_is_vrs_enabled() 3268- radv: assert that the arg is declared when used in get_scalar_arg() 3269- radv: lower ycbcr textures just before applying the pipeline layout 3270- radv: remove more references to the pipeline layout during compilation 3271- radv: copy the spirv module for debugging after compilation 3272- radv: drop the module reference in radv_can_dump_shader() 3273- radv: drop the module reference in radv_can_dump_shader_stats() 3274- radv: drop the module reference for enable_mrt_output_nan_fixup 3275- radv: stop passing the module to the compiler debug callback 3276- radv: stop relying on shader modules after SPIRV->NIR 3277- radv: re-order shader stages directly in radv_create_shaders() 3278- radv: rework pipeline and shaders creation feedback 3279- radv: add missing multi inclusion define to radv_shader_args.h 3280- radv: introduce new radv_pipeline_stage structure 3281- radv: replace convert_rt_stage() by vk_to_mesa_shader_stage() 3282- radv: store the shader sha1 to radv_pipeline_stage 3283- radv: drop the remaining uses of shader modules 3284- radv: remove unused radv_pipeline_layout::size field 3285- radv: add few helpers to deal with pipeline layout 3286- Revert "radv: Disable NGG for GS with suboptimal output vertex count." 3287- radv: fix initializing pipeline_key::topology for GFX9 and older 3288- radv: only apply enable_mrt_output_nan_fixup for 32-bit float MRTs 3289- aco: fix load_barycentric_at_{sample,offset} on GFX6-7 3290- nir: fix marking XFB varyings as always active IO 3291- nir: mark XFB varyings as unmoveable to prevent them to be remapped 3292- radv: fix handling divisor == 0 with dynamic vertex input state 3293- radv: allow to disable sinking of load inputs for FS via drirc 3294- radv: enable radv_disable_sinking_load_input_fs for Grid Autosport 3295- radv: re-emit dynamic line stipple state if the primitive topology changed 3296- radv: disable instance packing to fix pipeline query stats 3297- radv: disable DCC for Senra Kagura Shinovi Versus 3298 3299Shirish S (1): 3300 3301- radeonsi: allocate protected buffer only if required 3302 3303Shmerl (1): 3304 3305- docs/features: Mark VK_KHR_ray_query in progress 3306 3307Sidney Just (6): 3308 3309- wgl: add a flag to determine if running on zink 3310- wgl: add zink to the list of auto-loaded drivers 3311- zink: support VK_KHR_win32_surface 3312- kopper: add win32 loader interface 3313- zink: support win32 wsi 3314- wgl: support kopper 3315 3316Simon Ser (4): 3317 3318- vulkan/wsi/wayland: switch from alpha/opaque bools to bitfield 3319- vulkan/wsi/wayland: introduce wsi_wl_display_add_vk_format_modifier 3320- vulkan/wsi/wayland: de-duplicate wsi_wl_display_add_wl_shm_format 3321- vulkan/wsi/wayland: ensure added formats have flags 3322 3323Stefan Dirsch (1): 3324 3325- meson: restore private requires to libdrm in dri.pc file 3326 3327Sviatoslav Peleshko (2): 3328 3329- mesa: flush bitmap caches when changing scissors or window rects state 3330- anv: workaround apps that assume full subgroups without specifying it 3331 3332Tales Lelo da Aparecida (1): 3333 3334- zink: validate and log errors on vulkan calls 3335 3336Tapani Pälli (9): 3337 3338- intel/genxml: add PIPE_CONTROL field for L3 read only cache invalidation 3339- anv: invalidate L3 read only cache when VF cache is invalidated 3340- iris: invalidate L3 read only cache when VF cache is invalidated 3341- iris: fix a leak on surface states 3342- mesa/st: always use DXT5 when transcoding ASTC format 3343- iris: setup internal_format for memory object resources 3344- mesa: check for valid internalformat with glTex[Sub]Image 3345- ci: update various ci result files 3346- isl: disable mcs (and mcs+ccs) for color msaa on DG2 3347 3348Thierry Reding (2): 3349 3350- tegra: Use private reference count for sampler views 3351- tegra: Use private reference count for resources 3352 3353Thomas Debesse (1): 3354 3355- gallium/clover: pass -no-opaque-pointers to Clang 3356 3357Thomas H.P. Andersen (1): 3358 3359- pvr: fix overlapping comparison 3360 3361Thong Thai (7): 3362 3363- radeonsi: add check for graphics to si_try_normal_clear 3364- gallium: add parameters for encoder format conversion (EFC) support 3365- frontends/va: add encoder format conversion (EFC) support 3366- frontends/va: zero-copy efc 3367- radeon: add EFC support to only VCN2.0 devices 3368- radeonsi: add option to disable EFC 3369- frontends/va: fix decode issues introduced by efc change 3370 3371Timothy Arceri (5): 3372 3373- glsl/nir: free GLSL IR right after we convert to NIR 3374- glsl: fix needs_lowering() call in varying packing pass 3375- glsl/st: use nir pass to lower indirect rather than GLSL IR 3376- nir: fix setting varying from uniform as flat 3377- nir: fix sorting before assigning varying driver locations 3378 3379Timur Kristóf (38): 3380 3381- compiler: Extract num_mesh_vertices_per_primitive function. 3382- spirv: Create PRIMITIVE_INDICES for NV_mesh_shader on-demand. 3383- radv: Disable IB2 on compute queues. 3384- radv: Don't disturb dynamic primitive topology with mesh shading. 3385- ac/nir/ngg: Fix mixed up primitive ID after culling. 3386- radv: Better exclude special MS outputs from driver location assignment. 3387- aco: Allow 1-byte loads and stores with load/store_buffer_amd 3388- aco: Fix workgroup_id.y and .z for NV_mesh_shader. 3389- aco: Fix multiview view index for mesh shaders. 3390- ac/nir: Move LDS area for primitive count to the beginning. 3391- ac/nir: Properly handle when mesh API workgroup size is smaller than HW. 3392- ac/nir: Sanitize mesh shader primitive indices using umin. 3393- ac/nir: Make sure to exclude special outputs from arrayed output masks. 3394- ac/nir: Refactor mesh shader output code to smaller functions. 3395- radv: Refactor mesh shader draws and add num_workgroups. 3396- nir: Add new variable mode for task/mesh payload. 3397- spirv: Use task_payload mode for generic task outputs and mesh inputs. 3398- aco: Add storage class for Task Shader payload. 3399- aco: Support task_payload with barriers, refactor allowed storage class. 3400- aco: Support memory modes properly with load/store_buffer_amd. 3401- ac/nir: Use vertex count minus 1 to determine max index in mesh shaders. 3402- ac/nir: Reuse existing nir_builder for emit_ms_finale. 3403- ac/nir: Store mesh shader API and HW workgroup size in lowering state. 3404- ac/nir: Fix workgroup ID in mesh shader waves other than the first. 3405- ac/nir: Properly invalidate mesh shader metadata. 3406- aco: Remove superfluous code for mesh shader workgroup ID. 3407- ac/nir: Extract final mesh shader output counts to a separate function. 3408- nir: Fix handling of NV_mesh_shader PRIMITIVE_INDICES output. 3409- nir: Fix lowering terminology of compute system values: "from"->"to". 3410- nir: Extract lower_id_to_index into a separate function. 3411- nir: Introduce workgroup_index and ability to lower workgroup_id to it. 3412- ac: Query the amdgpu MEC firmware version. 3413- radv: Use correct buffer offset for conditional rendering. 3414- radv: Disable NGG for GS with suboptimal output vertex count. 3415- aco: Fix VOP2 instruction format in visit_tex. 3416- nir: Handle out of bounds access in nir_vectorize_tess_levels. 3417- radv: Fix gs_vgpr_comp_cnt for NGG VS without passthrough mode. 3418- radv: Only use TES vertex offset 2 for triangles and quads. 3419 3420Tomeu Vizoso (21): 3421 3422- iris/ci: Enable Whiskey Lake boards by default 3423- ci: Allow disabling the whole of the Collabora farm 3424- ci: Disable jobs to the Collabora lab 3425- Revert "ci: Disable jobs to the Collabora lab" 3426- ci/freedreno: Disable a618 jobs 3427- ci/freedreno: Update checksum for GolfWithYourFriends trace 3428- ci/iris: Increase console timeout for perf jobs 3429- ci/freedreno: Increase console timeout for perf jobs 3430- ci/panfrost: Disable some jobs due to a lab failure 3431- Revert "ci/panfrost: Disable some jobs due to a lab failure" 3432- ci/freedreno: Reduce concurrency when replaying traces on a630 3433- Revert "ci/freedreno: Disable a618 jobs" 3434- ci: Disable Link Power Management with RTL8153 3435- ci: Disable Google's lab 3436- ci: Use CI_PROJECT_NAME instead of hardcoding 'mesa' 3437- ci: Allow specifying a different kernel in LAVA jobs 3438- ci: Add env var to add packages to install in rootfs 3439- ci: Add env var to add packages to install in debian/arm_build image 3440- ci: Allow local installations to build additional stuff into the rootfs 3441- ci: Move most stuff out of root .gitlab-ci.yml 3442- ci: Disable jobs to the Collabora lab 3443 3444Vadym Shovkoplias (4): 3445 3446- iris: Do not apply SCANOUT allocation flags for SHARED-only requests 3447- anv: implement EXT_depth_clip_control 3448- anv: fix EXT_depth_clip_control 3449- anv: Fix geometry flickering issue when compute and 3D passes are combined 3450 3451Victor Hermann Chiletto (1): 3452 3453- radv: always check entry count in descriptor pool when allocating 3454 3455Vinod Koul (1): 3456 3457- freedreno/registers: update dsi registers to support dsc 3458 3459Vinson Lee (1): 3460 3461- pvr: Remove duplicate variable queue_create. 3462 3463Vitalii.Lomaka (1): 3464 3465- intel/batch-decoder: Fix uninitialized scalar variables 3466 3467Xaver Hugl (1): 3468 3469- gbm: improve documentation about the lifetime of resources 3470 3471Xiaohui Gu (1): 3472 3473- iris: Mark a dirty update when vs_needs_sgvs_element value changed 3474 3475Yevhenii Kolesnikov (1): 3476 3477- nir: Remove single-source phis before opt_if_loop_last_continue 3478 3479Yiwei Zhang (23): 3480 3481- venus: init renderer_info at renderer creation (part 1) 3482- venus: init renderer_info at renderer creation (part 2) 3483- venus: no roundtrip needed for shmem backed by BLOB_MEM_HOST3D 3484- turnip: advertise VK_EXT_queue_family_foreign 3485- venus: properly destroy deferred ahb image before real image creation 3486- venus: deep copy format list info for deferred image creation 3487- venus: clean up android wsi and ahb image builder 3488- venus: pass necessary format list at ahb image format query 3489- venus: add necessary format list for ahb image creation 3490- venus: clean up physical device features and properties 3491- venus: group extensions promoted to 1.3 3492- venus: update to latest venus protocol 3493- venus: add VK_EXT_line_rasterization support 3494- venus: add VK_EXT_provoking_vertex support 3495- venus: add VK_EXT_image_robustness support 3496- venus: update protocol to remove redundant decoders 3497- venus: let vn_android use vn_BindImageMemory2 and directly use reqs 3498- virgl: fake modifier plane count query support 3499- venus: workaround an ANGLE assumption on FORMAT_IMPLEMENTATION_DEFINED 3500- venus: store extension mask in renderer info 3501- venus: add cs helper stubs to be used by protocol 3502- venus: update protocol for mask helper and ignore renderer unknown pNext 3503- venus: prepare and feed renderer protocol info into cs 3504 3505Yogesh Mohan Marimuthu (4): 3506 3507- radeonsi: move clamp, alpha test from si_export_mrt_color() to new function 3508- radeonsi: prepare clamp, alpha test before mrtz prepare 3509- vulkan/device_select: for vulkan 1.0 use vid/did for boot_vga 3510- vulkan/device_select: add has_vulkan11 flag with has_pci_bus flag 3511 3512Yonggang Luo (20): 3513 3514- glx/egl: improve dri null screen related error messages. 3515- util: trim trailing space for files src/util/\**/* 3516- llvmpipe: Revise u_sse.h to remove unused _mm_shuffle_epi8 inline function 3517- gallium: Remove unused macro PIPE_ARCH_SSSE3 3518- util: Fixes unused parameter warnings 3519- c11: Fixes unused parameter warnings 3520- meson: Add predefined macro -D__MSVCRT_VERSION__=0x0700 only in mingw environment without _UCRT 3521- vtn: Fixes compiling error for mingw/ucrt by using setjmp/longjmp function instead compiler builtin 3522- nir: Move the define of snprintf to header nir.h 3523- util: Should not use ASSERTED in util_thread_get_time_nano 3524- vulkan/microsoft: Remove \`override_options: ['cpp_std=c++latest']` option for visual studio 3525- ci: Improve vs2019 mesa_build.ps1 for remove the need of cmd.exe 3526- util: Rename pipe_debug_message to util_debug_message 3527- util: Rename pipe_debug_callback to util_debug_callback 3528- pipe: place \`struct util_debug_callback` at the proper place in p_context.h 3529- util: Rename pipe_debug_type to util_debug_type 3530- util: Rename PIPE_DEBUG_TYPE to UTIL_DEBUG_TYPE 3531- util: Getting u_debug.h not depends on pipe/* 3532- util: Add tests for u_printf.h 3533- util: Convert util/u_printf.cpp to util/u_printf.c 3534 3535Zoltán Böszörményi (2): 3536 3537- crocus: enable GL46 tests for HSW in ci 3538- crocus: Enable compat profile the same way as core profile 3539 3540jiadozhu (1): 3541 3542- radeonsi: fix crash in flush_resource when used with buffers 3543 3544shansheng.wang (1): 3545 3546- frontends/va: fix coredump as creating surface with VAConfigAttrib 3547 3548wingdeans (1): 3549 3550- r600: Fix small leak in SfnLog 3551 3552xperia64 (1): 3553 3554- wgl: Add driver_zink as a dependency of the wgl frontend 3555