1Mesa 19.1.0 Release Notes / June 11, 2019 2========================================= 3 4Mesa 19.1.0 is a new development release. People who are concerned with 5stability and reliability should stick with a previous release or wait 6for Mesa 19.1.1. 7 8Mesa 19.1.0 implements the OpenGL 4.5 API, but the version reported by 9glGetString(GL_VERSION) or glGetIntegerv(GL_MAJOR_VERSION) / 10glGetIntegerv(GL_MINOR_VERSION) depends on the particular driver being 11used. Some drivers don't support all the features required in OpenGL 124.5. OpenGL 4.5 is **only** available if requested at context creation. 13Compatibility contexts may report a lower version depending on each 14driver. 15 16SHA256 checksums 17---------------- 18 19:: 20 21 2a6c3af3a803389183168e449c536304cf03e0f82c4c9333077933543b9d02f3 mesa-19.1.0.tar.xz 22 23New features 24------------ 25 26- GL_ARB_parallel_shader_compile on all drivers. 27- GL_EXT_gpu_shader4 on all GL 3.1 drivers. 28- GL_EXT_shader_image_load_formatted on radeonsi. 29- GL_EXT_texture_buffer_object on all GL 3.1 drivers. 30- GL_EXT_texture_compression_s3tc_srgb on Gallium drivers and i965 (ES 31 extension). 32- GL_NV_compute_shader_derivatives on iris and i965. 33- GL_KHR_parallel_shader_compile on all drivers. 34- VK_EXT_buffer_device_address on Intel and RADV. 35- VK_EXT_depth_clip_enable on Intel and RADV. 36- VK_KHR_ycbcr_image_arrays on Intel. 37- VK_EXT_inline_uniform_block on Intel and RADV. 38- VK_EXT_external_memory_host on Intel. 39- VK_EXT_host_query_reset on Intel and RADV. 40- VK_KHR_surface_protected_capabilities on Intel and RADV. 41- VK_EXT_pipeline_creation_feedback on Intel and RADV. 42- VK_KHR_8bit_storage on RADV. 43- VK_AMD_gpu_shader_int16 on RADV. 44- VK_AMD_gpu_shader_half_float on RADV. 45- VK_NV_compute_shader_derivatives on Intel. 46- VK_KHR_shader_float16_int8 on Intel and RADV (RADV only supports 47 int8). 48- VK_KHR_shader_atomic_int64 on Intel. 49- VK_EXT_descriptor_indexing on Intel. 50- VK_KHR_shader_float16_int8 on Intel and RADV. 51- GL_INTEL_conservative_rasterization on iris. 52- VK_EXT_memory_budget on Intel. 53 54Bug fixes 55--------- 56 57- `Bug 81843 <https://bugs.freedesktop.org/show_bug.cgi?id=81843>`__ - 58 [SNB IVB HSW] ETC2 textures are not returned as compressed images 59- `Bug 99781 <https://bugs.freedesktop.org/show_bug.cgi?id=99781>`__ - 60 Some Unity games fail assertion on startup in 61 glXCreateContextAttribsARB 62- `Bug 100239 <https://bugs.freedesktop.org/show_bug.cgi?id=100239>`__ 63 - Incorrect rendering in CS:GO 64- `Bug 100316 <https://bugs.freedesktop.org/show_bug.cgi?id=100316>`__ 65 - Linking GLSL 1.30 shaders with invariant and deprecated variables 66 triggers an 'mismatching invariant qualifiers' error 67- `Bug 104272 <https://bugs.freedesktop.org/show_bug.cgi?id=104272>`__ 68 - [OpenGL CTS] [HSW] 69 KHR-GL46.direct_state_access.textures_compressed_subimage assert 70 fails 71- `Bug 104355 <https://bugs.freedesktop.org/show_bug.cgi?id=104355>`__ 72 - Ivy Bridge ignores component mappings in texture views 73- `Bug 104602 <https://bugs.freedesktop.org/show_bug.cgi?id=104602>`__ 74 - [apitrace] Graphical artifacts in Civilization VI on RX Vega 75- `Bug 107052 <https://bugs.freedesktop.org/show_bug.cgi?id=107052>`__ 76 - [Regression][bisected]. Crookz - The Big Heist Demo can't be 77 launched despite the "true" flag in "drirc" 78- `Bug 107505 <https://bugs.freedesktop.org/show_bug.cgi?id=107505>`__ 79 - [lars] 80 dEQP-GLES31.functional.geometry_shading.layered#render_with_default_layer_3d 81 failure 82- `Bug 107510 <https://bugs.freedesktop.org/show_bug.cgi?id=107510>`__ 83 - [GEN8+] up to 10% perf drop on several 3D benchmarks 84- `Bug 107563 <https://bugs.freedesktop.org/show_bug.cgi?id=107563>`__ 85 - [RADV] Broken rendering in Unity demos 86- `Bug 107987 <https://bugs.freedesktop.org/show_bug.cgi?id=107987>`__ 87 - [Debug mesa only]. Crash happens when calling drawArrays 88- `Bug 108250 <https://bugs.freedesktop.org/show_bug.cgi?id=108250>`__ 89 - [GLSL] layout-location-struct.shader_test fails to link 90- `Bug 108457 <https://bugs.freedesktop.org/show_bug.cgi?id=108457>`__ 91 - [OpenGL CTS] 92 KHR-GL46.tessellation_shader.single.xfb_captures_data_from_correct_stage 93 fails 94- `Bug 108540 <https://bugs.freedesktop.org/show_bug.cgi?id=108540>`__ 95 - vkAcquireNextImageKHR blocks when timeout=0 in Wayland 96- `Bug 108766 <https://bugs.freedesktop.org/show_bug.cgi?id=108766>`__ 97 - Mesa built with meson has RPATH entries 98- `Bug 108824 <https://bugs.freedesktop.org/show_bug.cgi?id=108824>`__ 99 - Invalid handling when GL buffer is bound on one context and 100 invalidated on another 101- `Bug 108841 <https://bugs.freedesktop.org/show_bug.cgi?id=108841>`__ 102 - [RADV] SPIRV's control flow attributes do not propagate to LLVM 103- `Bug 108879 <https://bugs.freedesktop.org/show_bug.cgi?id=108879>`__ 104 - [CIK] [regression] All opencl apps hangs indefinitely in 105 si_create_context 106- `Bug 108999 <https://bugs.freedesktop.org/show_bug.cgi?id=108999>`__ 107 - Calculating the scissors fields when the y is flipped (0 on top) 108 can generate negative numbers that will cause assertion failure later 109 on. 110- `Bug 109057 <https://bugs.freedesktop.org/show_bug.cgi?id=109057>`__ 111 - texelFetch from GL_TEXTURE_2D_MULTISAMPLE with integer format fails 112- `Bug 109107 <https://bugs.freedesktop.org/show_bug.cgi?id=109107>`__ 113 - gallium/st/va: change va max_profiles when using Radeon VCN 114 Hardware 115- `Bug 109216 <https://bugs.freedesktop.org/show_bug.cgi?id=109216>`__ 116 - 4-27% performance drop in Vulkan benchmarks 117- `Bug 109326 <https://bugs.freedesktop.org/show_bug.cgi?id=109326>`__ 118 - mesa: Meson configuration summary should be printed 119- `Bug 109328 <https://bugs.freedesktop.org/show_bug.cgi?id=109328>`__ 120 - [BSW BXT GLK] dEQP-VK.subgroups.arithmetic.subgroup regressions 121- `Bug 109391 <https://bugs.freedesktop.org/show_bug.cgi?id=109391>`__ 122 - LTO Build fails 123- `Bug 109401 <https://bugs.freedesktop.org/show_bug.cgi?id=109401>`__ 124 - [DXVK] Project Cars rendering problems 125- `Bug 109404 <https://bugs.freedesktop.org/show_bug.cgi?id=109404>`__ 126 - [ANV] The Witcher 3 shadows flickering 127- `Bug 109443 <https://bugs.freedesktop.org/show_bug.cgi?id=109443>`__ 128 - Build failure with MSVC when using Scons >= 3.0.2 129- `Bug 109451 <https://bugs.freedesktop.org/show_bug.cgi?id=109451>`__ 130 - [IVB,SNB] LINE_STRIPs following a TRIANGLE_FAN fail to use 131 primitive restart 132- `Bug 109543 <https://bugs.freedesktop.org/show_bug.cgi?id=109543>`__ 133 - After upgrade mesa to 19.0.0~rc1 all vulkan based application stop 134 working ["vulkan-cube" received SIGSEGV in 135 radv_pipeline_init_blend_state at 136 ../src/amd/vulkan/radv_pipeline.c:699] 137- `Bug 109561 <https://bugs.freedesktop.org/show_bug.cgi?id=109561>`__ 138 - [regression, bisected] code re-factor causing games to stutter or 139 lock-up system 140- `Bug 109573 <https://bugs.freedesktop.org/show_bug.cgi?id=109573>`__ 141 - dEQP-VK.spirv_assembly.instruction.graphics.module.same_module 142- `Bug 109575 <https://bugs.freedesktop.org/show_bug.cgi?id=109575>`__ 143 - Mesa-19.0.0-rc1 : Computer Crashes trying to run anything Vulkan 144- `Bug 109581 <https://bugs.freedesktop.org/show_bug.cgi?id=109581>`__ 145 - [BISECTED] Nothing is Rendered on Sascha Willem's "subpasses" demo 146- `Bug 109594 <https://bugs.freedesktop.org/show_bug.cgi?id=109594>`__ 147 - totem assert failure: totem: src/intel/genxml/gen9_pack.h:72: 148 \__gen_uint: La declaración \`v <= max' no se cumple. 149- `Bug 109597 <https://bugs.freedesktop.org/show_bug.cgi?id=109597>`__ 150 - wreckfest issues with transparent objects & skybox 151- `Bug 109601 <https://bugs.freedesktop.org/show_bug.cgi?id=109601>`__ 152 - [Regression] RuneLite GPU rendering broken on 18.3.x 153- `Bug 109603 <https://bugs.freedesktop.org/show_bug.cgi?id=109603>`__ 154 - nir_instr_as_deref: Assertion \`parent && parent->type == 155 nir_instr_type_deref' failed. 156- `Bug 109645 <https://bugs.freedesktop.org/show_bug.cgi?id=109645>`__ 157 - build error on arm64: tegra_screen.c:33: 158 /usr/include/xf86drm.h:41:10: fatal error: drm.h: No such file or 159 directory 160- `Bug 109646 <https://bugs.freedesktop.org/show_bug.cgi?id=109646>`__ 161 - New video compositor compute shader render glitches mpv 162- `Bug 109647 <https://bugs.freedesktop.org/show_bug.cgi?id=109647>`__ 163 - /usr/include/xf86drm.h:40:10: fatal error: drm.h: No such file or 164 directory 165- `Bug 109648 <https://bugs.freedesktop.org/show_bug.cgi?id=109648>`__ 166 - AMD Raven hang during va-api decoding 167- `Bug 109659 <https://bugs.freedesktop.org/show_bug.cgi?id=109659>`__ 168 - Missing OpenGL symbols in OSMesa Gallium when building with meson 169- `Bug 109698 <https://bugs.freedesktop.org/show_bug.cgi?id=109698>`__ 170 - dri.pc contents invalid when built with meson 171- `Bug 109717 <https://bugs.freedesktop.org/show_bug.cgi?id=109717>`__ 172 - [regression] Cull distance tests asserting 173- `Bug 109735 <https://bugs.freedesktop.org/show_bug.cgi?id=109735>`__ 174 - [Regression] broken font with mesa_vulkan_overlay 175- `Bug 109738 <https://bugs.freedesktop.org/show_bug.cgi?id=109738>`__ 176 - Child of Light shows only a black screen 177- `Bug 109739 <https://bugs.freedesktop.org/show_bug.cgi?id=109739>`__ 178 - Mesa build fails when vulkan-overlay-layer option is enabled 179- `Bug 109742 <https://bugs.freedesktop.org/show_bug.cgi?id=109742>`__ 180 - vdpau state tracker on nv92 started to hit assert after vl compute 181 work 182- `Bug 109743 <https://bugs.freedesktop.org/show_bug.cgi?id=109743>`__ 183 - Test fails: 184 piglit.spec.arb_sample_shading.arb_sample_shading-builtin-gl-sample-mask-mrt-alpha 185- `Bug 109747 <https://bugs.freedesktop.org/show_bug.cgi?id=109747>`__ 186 - Add framerate to vulkan-overlay-layer 187- `Bug 109759 <https://bugs.freedesktop.org/show_bug.cgi?id=109759>`__ 188 - [BISECTED][REGRESSION][IVB, HSW] Font rendering problem in OpenGL 189- `Bug 109788 <https://bugs.freedesktop.org/show_bug.cgi?id=109788>`__ 190 - vulkan-overlay-layer: Only installs 64bit version 191- `Bug 109810 <https://bugs.freedesktop.org/show_bug.cgi?id=109810>`__ 192 - nir_opt_copy_prop_vars.c:454: error: unknown field ‘ssa’ specified 193 in initializer 194- `Bug 109929 <https://bugs.freedesktop.org/show_bug.cgi?id=109929>`__ 195 - tgsi_to_nir.c:2111: undefined reference to 196 \`gl_nir_lower_samplers_as_deref' 197- `Bug 109944 <https://bugs.freedesktop.org/show_bug.cgi?id=109944>`__ 198 - [bisected] Android build test fails with: utils.c: error: use of 199 undeclared identifier 'PACKAGE_VERSION' 200- `Bug 109945 <https://bugs.freedesktop.org/show_bug.cgi?id=109945>`__ 201 - pan_assemble.c:51:46: error: passing argument 2 of ‘tgsi_to_nir’ 202 from incompatible pointer type [-Werror=incompatible-pointer-types] 203- `Bug 109980 <https://bugs.freedesktop.org/show_bug.cgi?id=109980>`__ 204 - [i915 CI][HSW] 205 spec@arb_fragment_shader_interlock@arb_fragment_shader_interlock-image-load-store 206 - fail 207- `Bug 109984 <https://bugs.freedesktop.org/show_bug.cgi?id=109984>`__ 208 - unhandled VkStructureType 209 VK_STRUCTURE_TYPE_RENDER_PASS_INPUT_ATTACHMENT_ASPECT_CREATE_INFO 210- `Bug 110134 <https://bugs.freedesktop.org/show_bug.cgi?id=110134>`__ 211 - SIGSEGV while playing large hevc video in mpv 212- `Bug 110143 <https://bugs.freedesktop.org/show_bug.cgi?id=110143>`__ 213 - Doom 3: BFG Edition - Steam and GOG.com - white flickering screen 214- `Bug 110201 <https://bugs.freedesktop.org/show_bug.cgi?id=110201>`__ 215 - [ivb] mesa 19.0.0 breaks rendering in kitty 216- `Bug 110211 <https://bugs.freedesktop.org/show_bug.cgi?id=110211>`__ 217 - If DESTDIR is set to an empty string, the dri drivers are not 218 installed 219- `Bug 110216 <https://bugs.freedesktop.org/show_bug.cgi?id=110216>`__ 220 - radv: Segfault when compiling compute shaders from Assassin's Creed 221 Odyssey (regression, bisected) 222- `Bug 110221 <https://bugs.freedesktop.org/show_bug.cgi?id=110221>`__ 223 - build error with meson 224- `Bug 110239 <https://bugs.freedesktop.org/show_bug.cgi?id=110239>`__ 225 - Mesa SIGABRT: src/intel/genxml/gen9_pack.h:72: \__gen_uint: 226 Assertion \`v <= max' failed 227- `Bug 110257 <https://bugs.freedesktop.org/show_bug.cgi?id=110257>`__ 228 - Major artifacts in mpeg2 vaapi hw decoding 229- `Bug 110259 <https://bugs.freedesktop.org/show_bug.cgi?id=110259>`__ 230 - radv: Sampling depth-stencil image in GENERAL layout returns 231 nothing but zero (regression, bisected) 232- `Bug 110291 <https://bugs.freedesktop.org/show_bug.cgi?id=110291>`__ 233 - Vega 64 GPU hang running Space Engineers 234- `Bug 110302 <https://bugs.freedesktop.org/show_bug.cgi?id=110302>`__ 235 - [bisected][regression] piglit egl-create-pbuffer-surface and 236 egl-gl-colorspace regressions 237- `Bug 110305 <https://bugs.freedesktop.org/show_bug.cgi?id=110305>`__ 238 - Iris driver fails ext_packed_depth_stencil-getteximage test 239- `Bug 110311 <https://bugs.freedesktop.org/show_bug.cgi?id=110311>`__ 240 - [IVB HSW SNB][regression][bisected] regressions on vec4 241 deqp/gl{es}cts tests 242- `Bug 110349 <https://bugs.freedesktop.org/show_bug.cgi?id=110349>`__ 243 - radv: Dragon Quest XI (DXVK) has a graphical glitch (regression, 244 bisected) 245- `Bug 110353 <https://bugs.freedesktop.org/show_bug.cgi?id=110353>`__ 246 - weird colors seen in valley 247- `Bug 110355 <https://bugs.freedesktop.org/show_bug.cgi?id=110355>`__ 248 - radeonsi: GTK elements become invisible in some applications (GIMP, 249 LibreOffice) 250- `Bug 110356 <https://bugs.freedesktop.org/show_bug.cgi?id=110356>`__ 251 - install_megadrivers.py creates new dangling symlink [bisected] 252- `Bug 110404 <https://bugs.freedesktop.org/show_bug.cgi?id=110404>`__ 253 - Iris fails piglit.spec.ext_transform_feedback.immediate-reuse test 254- `Bug 110422 <https://bugs.freedesktop.org/show_bug.cgi?id=110422>`__ 255 - AMD_DEBUG=forcedma will crash OpenGL aps with SIGFAULT on VegaM 256 8706G 257- `Bug 110441 <https://bugs.freedesktop.org/show_bug.cgi?id=110441>`__ 258 - [llvmpipe] complex-loop-analysis-bug regression 259- `Bug 110443 <https://bugs.freedesktop.org/show_bug.cgi?id=110443>`__ 260 - vaapi/vpp: wrong output for non 64-bytes align width (ex: 1200) 261- `Bug 110454 <https://bugs.freedesktop.org/show_bug.cgi?id=110454>`__ 262 - [llvmpipe] piglit arb_color_buffer_float-render GL_RGBA8_SNORM 263 failure with llvm-9 264- `Bug 110462 <https://bugs.freedesktop.org/show_bug.cgi?id=110462>`__ 265 - Epic Games Launcher renders nothing with "-opengl" option 266- `Bug 110474 <https://bugs.freedesktop.org/show_bug.cgi?id=110474>`__ 267 - [bisected][regression] vk cts fp16 arithmetic failures 268- `Bug 110497 <https://bugs.freedesktop.org/show_bug.cgi?id=110497>`__ 269 - [DXVK][Regression][Bisected][SKL] Project Cars 2 crashes with Bug 270 Splat when loading finishes 271- `Bug 110526 <https://bugs.freedesktop.org/show_bug.cgi?id=110526>`__ 272 - [CTS] dEQP-VK.ycbcr.{conversion,format}.\* fail 273- `Bug 110530 <https://bugs.freedesktop.org/show_bug.cgi?id=110530>`__ 274 - [CTS] dEQP-VK.ycbcr.format.g8_b8_r8_3plane_420\* reports VM faults 275 on Vega10 276- `Bug 110535 <https://bugs.freedesktop.org/show_bug.cgi?id=110535>`__ 277 - [bisected] [icl] GPU hangs on crucible 278 func.miptree.r8g8b8a8-unorm.aspect-color.view-2d.levels01.array01.extent-512x512.upload-copy-with-draw 279 tests 280- `Bug 110540 <https://bugs.freedesktop.org/show_bug.cgi?id=110540>`__ 281 - [AMD TAHITI XT] valve artifact broken 282- `Bug 110573 <https://bugs.freedesktop.org/show_bug.cgi?id=110573>`__ 283 - Mesa vulkan-radeon 19.0.3 system freeze and visual artifacts (RADV) 284- `Bug 110590 <https://bugs.freedesktop.org/show_bug.cgi?id=110590>`__ 285 - [Regression][Bisected] GTAⅣ under wine fails with GLXBadFBConfig 286- `Bug 110632 <https://bugs.freedesktop.org/show_bug.cgi?id=110632>`__ 287 - "glx: Fix synthetic error generation in \__glXSendError" broke wine 288 games on 32-bit 289- `Bug 110648 <https://bugs.freedesktop.org/show_bug.cgi?id=110648>`__ 290 - Dota2 will not open using vulkan since 19.0 series 291- `Bug 110655 <https://bugs.freedesktop.org/show_bug.cgi?id=110655>`__ 292 - VK_LAYER_MESA_OVERLAY_CONFIG=draw,fps renders sporadically 293- `Bug 110698 <https://bugs.freedesktop.org/show_bug.cgi?id=110698>`__ 294 - tu_device.c:900:4: error: initializer element is not constant 295- `Bug 110701 <https://bugs.freedesktop.org/show_bug.cgi?id=110701>`__ 296 - GPU faults in in Unigine Valley 1.0 297- `Bug 110721 <https://bugs.freedesktop.org/show_bug.cgi?id=110721>`__ 298 - graphics corruption on steam client with mesa 19.1.0 rc3 on polaris 299- `Bug 110761 <https://bugs.freedesktop.org/show_bug.cgi?id=110761>`__ 300 - Huge problems between Mesa and Electron engine apps 301- `Bug 110784 <https://bugs.freedesktop.org/show_bug.cgi?id=110784>`__ 302 - [regression][bisected] Reverting 'expose 0 shader binary formats 303 for compat profiles for Qt' causes get_program_binary failures on 304 Iris 305 306Changes 307------- 308 309Adam Jackson (1): 310 311- drisw: Try harder to probe whether MIT-SHM works 312 313Albert Pal (1): 314 315- Fix link release notes for 19.0.0. 316 317Alejandro Piñeiro (12): 318 319- blorp: introduce helper method blorp_nir_init_shader 320- nir, glsl: move pixel_center_integer/origin_upper_left to 321 shader_info.fs 322- nir/xfb: add component_offset at nir_xfb_info 323- nir_types: add glsl_varying_count helper 324- nir/xfb: adding varyings on nir_xfb_info and gather_info 325- nir/xfb: sort varyings too 326- nir_types: add glsl_type_is_struct helper 327- nir/xfb: handle arrays and AoA of basic types 328- nir/linker: use nir_gather_xfb_info 329- nir/linker: fix ARRAY_SIZE query with xfb varyings 330- nir/xfb: move varyings info out of nir_xfb_info 331- docs: document MESA_GLSL=errors keyword 332 333Alexander von Gluck IV (1): 334 335- haiku: Fix hgl dispatch build. Tested under meson/scons. 336 337Alexandros Frantzis (1): 338 339- virgl: Fake MSAA when max samples is 1 340 341Alok Hota (32): 342 343- swr/rast: update SWR rasterizer shader stats 344- gallium/swr: Param defaults for unhandled PIPE_CAPs 345- gallium/aux: add PIPE_CAP_MAX_VARYINGS to u_screen 346- swr/rast: Convert system memory pointers to gfxptr_t 347- swr/rast: Disable use of \__forceinline by default 348- swr/rast: Correctly align 64-byte spills/fills 349- swr/rast: Flip BitScanReverse index calculation 350- swr/rast: Move knob defaults to generated cpp file 351- swr/rast: FP consistency between POSH/RENDER pipes 352- swr/rast: Refactor scratch space variable names 353- swr/rast: convert DWORD->uint32_t, QWORD->uint64_t 354- swr/rast: simdlib cleanup, clipper stack space fixes 355- swr/rast: Add translation support to streamout 356- swr/rast: bypass size limit for non-sampled textures 357- swr/rast: Cleanup and generalize gen_archrast 358- swr/rast: Add initial SWTag proto definitions 359- swr/rast: Add string handling to AR event framework 360- swr/rast: Add general SWTag statistics 361- swr/rast: Fix autotools and scons codegen 362- swr/rast: Remove deprecated 4x2 backend code 363- swr/rast: AVX512 support compiled in by default 364- swr/rast: enforce use of tile offsets 365- swr/rast: add more llvm intrinsics 366- swr/rast: update guardband rects at draw setup 367- swr/rast: add SWR_STATIC_ASSERT() macro 368- swr/rast: add flat shading 369- swr/rast: add guards for cpuid on Linux 370- swr/rast: early exit on empty triangle mask 371- swr/rast: Cleanup and generalize gen_archrast 372- swr/rast: Add initial SWTag proto definitions 373- swr/rast: Add string handling to AR event framework 374- swr/rast: Add general SWTag statistics 375 376Alyssa Rosenzweig (192): 377 378- panfrost: Initial stub for Panfrost driver 379- panfrost: Implement Midgard shader toolchain 380- meson: Remove panfrost from default driver list 381- kmsro: Move DRM entrypoints to shared block 382- panfrost: Use u_pipe_screen_get_param_defaults 383- panfrost: Check in sources for command stream 384- panfrost: Include glue for out-of-tree legacy code 385- kmsro: Silence warning if missing 386- panfrost: Clean-up one-argument passing quirk 387- panfrost: Don't hardcode number of nir_ssa_defs 388- panfrost: Add kernel-agnostic resource management 389- panfrost: Remove if 0'd dead code 390- panfrost: Remove speculative if 0'd format bit code 391- panfrost: Elucidate texture op scheduling comment 392- panfrost: Specify supported draw modes per-context 393- panfrost: Fix build; depend on libdrm 394- panfrost: Backport driver to Mali T600/T700 395- panfrost: Identify MALI_OCCLUSION_PRECISE bit 396- panfrost: Implement PIPE_QUERY_OCCLUSION_COUNTER 397- panfrost: Don't align framebuffer dims 398- panfrost: Improve logging and patch memory leaks 399- panfrost: Fix various leaks unmapping resources 400- panfrost: Free imported BOs 401- panfrost: Swap order of tiled texture (de)alloc 402- panfrost: Cleanup mali_viewport (clipping) code 403- panfrost: Preserve w sign in perspective division 404- panfrost: Fix clipping region 405- panfrost: Stub out separate stencil functions 406- panfrost: Add pandecode (command stream debugger) 407- panfrost: Implement pantrace (command stream dump) 408- panfrost/midgard: Refactor tag lookahead code 409- panfrost/midgard: Fix nested/chained if-else 410- panfrost: Rectify doubleplusungood extended branch 411- panfrost/midgard: Emit extended branches 412- panfrost: Dynamically set discard branch targets 413- panfrost: Verify and print brx condition in disasm 414- panfrost: Use tiler fast path (performance boost) 415- panfrost/meson: Remove subdir for nondrm 416- panfrost/nondrm: Flag CPU-invisible regions 417- panfrost/nondrm: Make COHERENT_LOCAL explicit 418- panfrost/nondrm: Split out dump_counters 419- panfrost/midgard: Add fround(_even), ftrunc, ffma 420- panfrost: Decode render target swizzle/channels 421- panfrost: Add RGB565, RGB5A1 texture formats 422- panfrost: Identify 4-bit channel texture formats 423- panfrost: Expose perf counters in environment 424- panfrost/midgard: Allow flt to run on most units 425- panfrost: Import job data structures from v3d 426- panfrost: Decouple Gallium clear from FBD clear 427- panfrost: Cleanup cruft related to clears 428- panfrost/midgard: Don't force constant on VLUT 429- panfrost: Flush with offscreen rendering 430- panfrost/midgard: Promote smul to vmul 431- panfrost/midgard: Preview for data hazards 432- panfrost: List primitive restart enable bit 433- panfrost/drm: Cast pointer to u64 to fix warning 434- panfrost: Cleanup needless if in create_bo 435- panfrost: Combine has_afbc/tiled in layout enum 436- panfrost: Delay color buffer setup 437- panfrost: Determine framebuffer format bits late 438- panfrost: Allocate dedicated slab for linear BOs 439- panfrost: Support linear depth textures 440- panfrost: Document "depth-buffer writeback" bit 441- panfrost: Identify fragment_extra flags 442- util: Add a drm_find_modifier helper 443- v3d: Use shared drm_find_modifier util 444- vc4: Use shared drm_find_modifier util 445- freedreno: Use shared drm_find_modifier util 446- panfrost: Break out fragment to SFBD/MFBD files 447- panfrost: Remove staging SFBD for pan_context 448- panfrost: Remove staging MFBD 449- panfrost: Minor comment cleanup (version detection) 450- panfrost/mfbd: Implement linear depth buffers 451- panfrost/mfbd: Respect per-job depth write flag 452- panfrost: Comment spelling fix 453- panfrost: Allocate extra data for depth buffer 454- panfrost; Disable AFBC for depth buffers 455- panfrost: Compute viewport state on the fly 456- panfrost/midgard: Implement fpow 457- panfrost: Workaround buffer overrun with mip level 458- panfrost: Fix primconvert check 459- panfrost: Disable PIPE_CAP_TGSI_TEXCOORD 460- panfrost/decode: Respect primitive size pointers 461- panfrost: Replay more varying buffers 462- panfrost: Rewrite varying assembly 463- panfrost/midgard: Fix b2f32 swizzle for vectors 464- panfrost: Fix viewports 465- panfrost: Implement scissor test 466- panfrost/midgard: Add fcsel_i opcode 467- panfrost/midgard: Schedule ball/bany to vectors 468- panfrost/midgard: Add more ball/bany, iabs ops 469- panfrost/midgard: Map more bany/ball opcodes 470- panfrost/midgard: Lower bool_to_int32 471- panfrost/midgard: Lower f2b32 to fne 472- panfrost/midgard: Lower i2b32 473- panfrost/midgard: Implement b2i; improve b2f/f2b 474- panfrost/midgard: Lower source modifiers for ints 475- panfrost/midgard: Cleanup midgard_nir_algebraic.py 476- panfrost: Stub out ES3 caps/callbacks 477- panfrost/midgard: Add ult/ule ops 478- panfrost/midgard: Expand fge lowering to more types 479- panfrost/midgard: Handle i2b constant 480- panfrost/midgard: fpow is a two-part operation 481- panfrost: Preliminary work for mipmaps 482- panfrost: Fix vertex buffer corruption 483- panfrost/midgard: Disassemble \`cube\` texture op 484- panfrost/midgard: Add L/S op for writing cubemap coordinates 485- panfrost: Preliminary work for cubemaps 486- panfrost/decode: Decode all cubemap faces 487- panfrost: Include all cubemap faces in bitmap list 488- panfrost/midgard: Emit cubemap coordinates 489- panfrost: Implement command stream for linear cubemaps 490- panfrost: Extend tiling for cubemaps 491- panfrost: Implement missing texture formats 492- panfrost/decode: Print negative_start 493- panfrost: Clean index state between indexed draws 494- panfrost: Fix index calculation types and asserts 495- panfrost: Implement FIXED formats 496- panfrost: Remove support for legacy kernels 497- nir: Add "viewport vector" system values 498- panfrost: Implement system values 499- panfrost: Cleanup some indirection in pan_resource 500- panfrost: Respect box->width in tiled stores 501- panfrost: Size tiled temp buffers correctly 502- panfrost/decode: Add flags for tilebuffer readback 503- panfrost: Add tilebuffer load? branch 504- panfrost/midgard: Add umin/umax opcodes 505- panfrost/midgard: Add ilzcnt op 506- panfrost/midgard: Add ibitcount8 op 507- panfrost/midgard: Enable lower_find_lsb 508- panfrost: Remove "mali_unknown6" nonsense 509- panfrost/midgard: Drop dependence on mesa/st 510- panfrost: Cleanup indexed draw handling 511- nir: Add nir_lower_viewport_transform 512- panfrost/midgard: Use shared nir_lower_viewport_transform 513- panfrost: Track BO lifetime with jobs and reference counts 514- panfrost: Fixup vertex offsets to prevent shadow copy 515- panfrost/mdg: Use shared fsign lowering 516- panfrost/mdg/disasm: Print raw varying_parameters 517- panfrost/midgard: Pipe through varying arrays 518- panfrost/midgard: Implement indirect loads of varyings/UBOs 519- panfrost/midgard: Respect component of bcsel condition 520- panfrost/midgard: Remove useless MIR dump 521- panfrost: Respect backwards branches in RA 522- panfrost/midgard: Don't try to inline constants on branches 523- panfrost/midgard: imul can only run on \*mul 524- panfrost: Disable indirect outputs for now 525- panfrost: Use actual imov instruction 526- panfrost/midgard: Dead code eliminate MIR 527- panfrost/midgard: Track loop depth 528- panfrost/midgard: Fix off-by-one in successor analysis 529- panfrost/midgard: Remove unused mir_next_block 530- panfrost/midgard: Update integer op list 531- panfrost/midgard: Document sign-extension/zero-extension bits 532 (vector) 533- panfrost/midgard: Set integer mods 534- panfrost/midgard: Implement copy propagation 535- panfrost/midgard: Optimize MIR in progress loop 536- panfrost/midgard: Refactor opcode tables 537- panfrost/midgard: Add "op commutes?" property 538- panfrost/midgard: Remove assembler 539- panfrost/midgard: Reduce fmax(a, 0.0) to fmov.pos 540- panfrost/midgard: Extend copy propagation pass 541- panfrost/midgard: Optimize csel involving 0 542- panfrost/midgard: Copy prop for texture registers 543- panfrost/midgard: Identify inand 544- panfrost/midgard: Add new bitwise ops 545- Revert "panfrost/midgard: Extend copy propagation pass" 546- panfrost/midgard: Only copyprop without an outmod 547- panfrost/midgard: Fix regressions in -bjellyfish 548- panfrost/midgard: Fix tex propogation 549- panfrost/midgard: imov workaround 550- panfrost: Use fp32 (not fp16) varyings 551- panfrost/midgard: Safety check immediate precision degradations 552- panfrost: Workaround -bshadow regression 553- panfrost: Remove shader dump 554- panfrost/decode: Hit MRT blend shader enable bits 555- panfrost: Fix blend shader upload 556- panfrost/midgard: reg_mode_full -> reg_mode_32, etc 557- panfrost/midgard/disasm: Catch mask errors 558- panfrost/midgard/disasm: Extend print_reg to 8-bit 559- panfrost/midgard/disasm: Fill in .int mod 560- panfrost/midgard: Fix crash on unknown op 561- panfrost/midgard: Rename ilzcnt8 -> iclz 562- panfrost/midgard/disasm: Support 8-bit destination 563- panfrost/midgard/disasm: Print 8-bit sources 564- panfrost/midgard/disasm: Stub out 64-bit 565- panfrost/midgard/disasm: Handle dest_override generalized 566- panfrost: Support RGB565 FBOs 567- panfrost/midgard: Fix integer selection 568- panfrost/midgard: Fix RA when temp_count = 0 569- panfrost/midgard: Lower mixed csel (NIR) 570- panfrost/midgard: iabs cannot run on mul 571 572Alyssa Ross (1): 573 574- get_reviewer.pl: improve portability 575 576Amit Pundir (1): 577 578- mesa: android: freedreno: build libfreedreno_{drm,ir3} static libs 579 580Andre Heider (5): 581 582- iris: fix build with gallium nine 583- iris: improve PIPE_CAP_VIDEO_MEMORY bogus value 584- iris: add support for tgsi_to_nir 585- st/nine: enable csmt per default on iris 586- st/nine: skip position checks in SetCursorPosition() 587 588Andreas Baierl (2): 589 590- nir: add rcp(w) lowering for gl_FragCoord 591- lima/ppir: Add gl_FragCoord handling 592 593Andres Gomez (12): 594 595- mesa: INVALID_VALUE for wrong type or format in Clear*Buffer*Data 596- gitlab-ci: install distro's ninja 597- glsl: correctly validate component layout qualifier for dvec{3,4} 598- glsl/linker: always validate explicit location among inputs 599- glsl/linker: don't fail non static used inputs without matching 600 outputs 601- glsl/linker: simplify xfb_offset vs xfb_stride overflow check 602- Revert "glsl: relax input->output validation for SSO programs" 603- glsl/linker: location aliasing requires types to have the same width 604- docs: drop Andres Gomez from the release cycles 605- glsl/linker: always validate explicit locations for first and last 606 interfaces 607- docs/relnotes: add support for VK_KHR_shader_float16_int8 608- glsl/linker: check for xfb_offset aliasing 609 610Andrii Simiklit (5): 611 612- i965: consider a 'base level' when calculating width0, height0, 613 depth0 614- i965: re-emit index buffer state on a reset option change. 615- util: clean the 24-bit unused field to avoid an issues 616- iris: make the TFB result visible to others 617- egl: return correct error code for a case req ver < 3 with 618 forward-compatible 619 620Antia Puentes (1): 621 622- nir/linker: Fix TRANSFORM_FEEDBACK_BUFFER_INDEX 623 624Anuj Phogat (7): 625 626- i965/icl: Add WA_2204188704 to disable pixel shader panic dispatch 627- anv/icl: Add WA_2204188704 to disable pixel shader panic dispatch 628- intel: Add Elkhart Lake device info 629- intel: Add Elkhart Lake PCI-IDs 630- iris/icl: Set Enabled Texel Offset Precision Fix bit 631- iris/icl: Add WA_2204188704 to disable pixel shader panic dispatch 632- intel: Add support for Comet Lake 633 634Axel Davy (49): 635 636- st/nine: Ignore window size if error 637- st/nine: Ignore multisample quality level if no ms 638- st/nine: Disable depth write when nothing gets updated 639- st/nine: Do not advertise support for D15S1 and D24X4S4 640- st/nine: Do not advertise CANMANAGERESOURCE 641- st/nine: Change a few advertised caps 642- Revert "d3dadapter9: Support software renderer on any DRI device" 643- st/nine: Fix D3DWindowBuffer_release for old wine nine support 644- st/nine: Use FLT_MAX/2 for RCP clamping 645- st/nine: Upload managed textures only at draw using them 646- st/nine: Upload managed buffers only at draw using them 647- st/nine: Fix buffer/texture unbinding in nine_state_clear 648- st/nine: Finish if nooverwrite after normal mapping 649- st/nine: Always return OK on SetSoftwareVertexProcessing 650- st/nine: Enable modifiers on ps 1.X texcoords 651- st/nine: Ignore nooverwrite for systemmem 652- st/nine: Fix SINCOS input 653- st/nine: Optimize surface upload with conversion 654- st/nine: Optimize volume upload with conversion 655- st/nine: rename \*_conversion to \*_internal 656- st/nine: Refactor surface GetSystemMemPointer 657- st/nine: Refactor volume GetSystemMemPointer 658- st/nine: Support internal compressed format for surfaces 659- st/nine: Support internal compressed format for volumes 660- st/nine: Add drirc option to use data_internal for dynamic textures 661- drirc: Add Gallium nine workaround for Rayman Legends 662- st/nine: Recompile optimized shaders based on b/i consts 663- st/nine: Control shader constant inlining with drirc 664- st/nine: Regroup param->rel tests 665- st/nine: Refactor param->rel 666- st/nine: Compact nine_ff_get_projected_key 667- st/nine: Compact pixel shader key 668- st/nine: use helper ureg_DECL_sampler everywhere 669- st/nine: Manually upload vs and ps constants 670- st/nine: Refactor shader constants ureg_src computation 671- st/nine: Make swvp_on imply IS_VS 672- st/nine: Refactor ct_ctor 673- st/nine: Track constant slots used 674- st/nine: Refactor counting of constants 675- st/nine: Prepare constant compaction in nine_shader 676- st/nine: Propagate const_range to context 677- st/nine: Cache constant buffer size 678- st/nine: Handle const_ranges in nine_state 679- st/nine: Enable computing const_ranges 680- st/nine: Use TGSI_SEMANTIC_GENERIC for fog 681- st/nine: Optimize a bit writeonly buffers 682- st/nine: Throttle rendering similarly for thread_submit 683- st/nine: Check discard_delayed_release is set before allocating more 684- d3dadapter9: Revert to old throttling limit value 685 686Bart Oldeman (1): 687 688- gallium-xlib: query MIT-SHM before using it. 689 690Bas Nieuwenhuizen (105): 691 692- radv: Only look at pImmutableSamples if the descriptor has a sampler. 693- amd/common: Add gep helper for pointer increment. 694- amd/common: Implement ptr->int casts in ac_to_integer. 695- radv: Fix the shader info pass for not having the variable. 696- amd/common: Use correct writemask for shared memory stores. 697- amd/common: Fix stores to derefs with unknown variable. 698- amd/common: Handle nir_deref_type_ptr_as_array for shared memory. 699- amd/common: handle nir_deref_cast for shared memory from integers. 700- amd/common: Do not use 32-bit loads for shared memory. 701- amd/common: Implement global memory accesses. 702- radv: Do not use the bo list for local buffers. 703- radv: Implement VK_EXT_buffer_device_address. 704- radv: Use correct num formats to detect whether we should be use 1.0 705 or 1. 706- radv: Sync ETC2 whitelisted devices. 707- radv: Clean up a bunch of compiler warnings. 708- radv: Handle clip+cull distances more generally as compact arrays. 709- radv: Implement VK_EXT_depth_clip_enable. 710- radv: Disable depth clamping even without 711 EXT_depth_range_unrestricted. 712- radv: Fix float16 interpolation set up. 713- radv: Allow interpolation on non-float types. 714- radv: Interpolate less aggressively. 715- turnip: Add driver skeleton (v2) 716- turnip: Fix up detection of device. 717- turnip: Gather some device info. 718- turnip: Remove abort. 719- turnip: Fix newly introduced warning. 720- turnip: Add buffer allocation & mapping support. 721- turnip: Report a memory type and heap. 722- turnip: Cargo cult the Intel heap size functionality. 723- turnip: Initialize memory type in requirements. 724- turnip: Disable more features. 725- turnip: Add 630 to the list. 726- turnip: Fix bo allocation after we stopped using libdrm_freedreno ... 727- turnip: Fix memory mapping. 728- turnip: Add image layout calculations. 729- turnip: Stop hardcoding the msm version check. 730- turnip: move tu_gem.c to tu_drm.c 731- turnip: Implement pipe-less param query. 732- turnip: Implement some format properties for RGBA8. 733- turnip: Remove some radv leftovers. 734- turnip: clean up TODO. 735- turnip: Implement some UUIDs. 736- turnip: Implement a slow bo list 737- turnip: Add a command stream. 738- turnip: Add msm queue support. 739- turnip: Make bo_list functions not static 740- turnip: Implement submission. 741- turnip: Fill command buffer 742- turnip: Shorten primary_cmd_stream name. 743- turnip: Add emit functions in a header. 744- turnip: Move stream functions to tu_cs.c 745- turnip: Add buffer memory binding. 746- turnip: Make tu6_emit_event_write shared. 747- turnip: Add tu6_rb_fmt_to_ifmt. 748- turnip: Implement buffer->buffer DMA copies. 749- turnip: Add image->buffer DMA copies. 750- turnip: Add buffer->image DMA copies. 751- turnip: Add todo for copies. 752- turnip: Fix GCC compiles. 753- turnip: Deconflict vk_format_table regeneration 754- gitlab-ci: Build turnip. 755- radeonsi: Remove implicit const cast. 756- radv: Allow fast clears with concurrent queue mask for some layouts. 757- vulkan/util: Handle enums that are in platform-specific headers. 758- vulkan: Update the XML and headers to 1.1.104 759- radv: Implement VK_EXT_host_query_reset. 760- radv: Use correct image view comparison for fast clears. 761- radv: Implement VK_EXT_pipeline_creation_feedback. 762- ac/nir: Return frag_coord as integer. 763- nir: Add access qualifiers on load_ubo intrinsic. 764- radv: Add non-uniform indexing lowering. 765- radv: Add bolist RADV_PERFTEST flag. 766- ac: Move has_local_buffers disable to radeonsi. 767- radv: Use local buffers for the global bo list. 768- radv: Support VK_EXT_inline_uniform_block. 769- radv: Add support for driconf. 770- vulkan/wsi: Add X11 adaptive sync support based on dri options. 771- radv: Add adaptive_sync driconfig option and enable it by default. 772- radv: Add logic for subsampled format descriptions. 773- radv: Add logic for multisample format descriptions. 774- radv: Add multiple planes to images. 775- radv: Add single plane image views & meta operations. 776- radv: Support different source & dest aspects for planar images in 777 blit2d. 778- radv: Add ycbcr conversion structs. 779- radv: Add support for image views with multiple planes. 780- radv: Allow mixed src/dst aspects in copies. 781- ac/nir: Add support for planes. 782- radv: Add ycbcr samplers in descriptor set layouts. 783- radv: Update descriptor sets for multiple planes. 784- radv: Add ycbcr lowering pass. 785- radv: Run the new ycbcr lowering pass. 786- radv: Add hashing for the ycbcr samplers. 787- radv: Add ycbcr format features. 788- radv: Add ycbcr subsampled & multiplane formats to csv. 789- radv: Enable YCBCR conversion feature. 790- radv: Expose VK_EXT_ycbcr_image_arrays. 791- radv: Expose Vulkan 1.1 for Android. 792- radv: Fix hang width YCBCR array textures. 793- radv: Set is_array in lowered ycbcr tex instructions. 794- radv: Restrict YUVY formats to 1 layer. 795- radv: Disable subsampled formats. 796- radv: Implement cosited_even sampling. 797- radv: Do not use extra descriptor space for the 3rd plane. 798- nir: Actually propagate progress in nir_opt_move_load_ubo. 799- radv: Prevent out of bound shift on 32-bit builds. 800 801Benjamin Gordon (1): 802 803- configure.ac/meson.build: Add options for library suffixes 804 805Benjamin Tissoires (1): 806 807- CI: use wayland ci-templates repo to create the base image 808 809Boyan Ding (3): 810 811- gk110/ir: Add rcp f64 implementation 812- gk110/ir: Add rsq f64 implementation 813- gk110/ir: Use the new rcp/rsq in library 814 815Boyuan Zhang (1): 816 817- st/va: reverse qt matrix back to its original order 818 819Brian Paul (51): 820 821- st/mesa: whitespace/formatting fixes in st_cb_texture.c 822- svga: assorted whitespace and formatting fixes 823- svga: fix dma.pending > 0 test 824- mesa: fix display list corner case assertion 825- st/mesa: whitespace fixes in st_sampler_view.c 826- st/mesa: line wrapping, whitespace fixes in st_cb_texture.c 827- st/mesa: whitespace fixes in st_texture.h 828- svga: init fill variable to avoid compiler warning 829- svga: silence array out of bounds warning 830- st/wgl: init a variable to silence MinGW warning 831- gallium/util: whitespace cleanups in u_bitmask.[ch] 832- gallium/util: add some const qualifiers in u_bitmask.c 833- pipebuffer: use new pb_usage_flags enum type 834- pipebuffer: whitespace fixes in pb_buffer.h 835- winsys/svga: use new pb_usage_flags enum type 836- st/mesa: move, clean-up shader variant key decls/inits 837- st/mesa: whitespace, formatting fixes in st_cb_flush.c 838- svga: refactor draw_vgpu10() function 839- svga: remove SVGA_RELOC_READ flag in SVGA3D_BindGBSurface() 840- pipebuffer: s/PB_ALL_USAGE_FLAGS/PB_USAGE_ALL/ 841- st/mesa: init hash keys with memset(), not designated initializers 842- intel/decoders: silence uninitialized variable warnings in 843 gen_print_batch() 844- intel/compiler: silence unitialized variable warning in 845 opt_vector_float() 846- st/mesa: move utility functions, macros into new st_util.h file 847- st/mesa: move around some code in st_context.c 848- st/mesa: add/improve sampler view comments 849- st/mesa: rename st_texture_release_sampler_view() 850- st/mesa: minor refactoring of texture/sampler delete code 851- docs: try to improve the Meson documentation (v2) 852- drisw: fix incomplete type compilation failure 853- gallium/winsys/kms: fix incomplete type compilation failure 854- nir: silence a couple new compiler warnings 855- docs: separate information for compiler selection and compiler 856 options 857- docs: link to the meson_options.txt file gitlab.freedesktop.org 858- st/mesa: implement "zombie" sampler views (v2) 859- st/mesa: implement "zombie" shaders list 860- st/mesa: stop using pipe_sampler_view_release() 861- svga: stop using pipe_sampler_view_release() 862- llvmpipe: stop using pipe_sampler_view_release() 863- swr: remove call to pipe_sampler_view_release() 864- i915g: remove calls to pipe_sampler_view_release() 865- gallium/util: remove pipe_sampler_view_release() 866- nir: fix a few signed/unsigned comparison warnings 867- st/mesa: fix texture deletion context mix-up issues (v2) 868- nir: use {0} initializer instead of {} to fix MSVC build 869- util: no-op \__builtin_types_compatible_p() for non-GCC compilers 870- docs: s/Aptril/April/ 871- llvmpipe: init some vars to NULL to silence MinGW compiler warnings 872- glsl: work around MinGW 7.x compiler bug 873- svga: add SVGA_NO_LOGGING env var (v2) 874- glsl: fix typo in #warning message 875 876Caio Marcelo de Oliveira Filho (61): 877 878- nir: keep the phi order when splitting blocks 879- i965: skip bit6 swizzle detection in Gen8+ 880- anv: skip bit6 swizzle detection in Gen8+ 881- isl: assert that Gen8+ don't have bit6_swizzling 882- intel/compiler: use 0 as sampler in emit_mcs_fetch 883- nir: fix example in opt_peel_loop_initial_if description 884- iris: Fix uses of gl_TessLevel\* 885- iris: Add support for TCS passthrough 886- iris: always include an extra constbuf0 if using UBOs 887- nir/copy_prop_vars: don't get confused by array_deref of vectors 888- nir/copy_prop_vars: add debug helpers 889- nir/copy_prop_vars: keep track of components in copy_entry 890- nir/copy_prop_vars: change test helper to get intrinsics 891- nir: nir_build_deref_follower accept array derefs of vectors 892- nir/copy_prop_vars: add tests for load/store elements of vectors 893- nir: fix MSVC build 894- st/nir: count num_uniforms for FS bultin shader 895- nir/copy_prop_vars: rename/refactor store_to_entry helper 896- nir/copy_prop_vars: use NIR_MAX_VEC_COMPONENTS 897- nir/copy_prop_vars: handle load/store of vector elements 898- nir/copy_prop_vars: add tests for indirect array deref 899- nir/copy_prop_vars: prefer using entries from equal derefs 900- nir/copy_prop_vars: handle indirect vector elements 901- anv: Implement VK_EXT_external_memory_host 902- nir: Add a pass to combine store_derefs to same vector 903- intel/nir: Combine store_derefs after vectorizing IO 904- intel/nir: Combine store_derefs to improve code from SPIR-V 905- nir: Handle array-deref-of-vector case in loop analysis 906- spirv: Add an execution environment to the options 907- intel/compiler: handle GLSL_TYPE_INTERFACE as GLSL_TYPE_STRUCT 908- spirv: Use interface type for block and buffer block 909- iris: Clean up compiler warnings about unused 910- nir: Take if_uses into account when repairing SSA 911- mesa: Extension boilerplate for NV_compute_shader_derivatives 912- glsl: Remove redundant conditions when asserting in_qualifier 913- glsl: Enable derivative builtins for NV_compute_shader_derivatives 914- glsl: Enable texture builtins for NV_compute_shader_derivatives 915- glsl: Parse and propagate derivative_group to shader_info 916- nir/algebraic: Lower CS derivatives to zero when no group defined 917- nir: Don't set LOD=0 for compute shader that has derivative group 918- intel/fs: Use TEX_LOGICAL whenever implicit lod is supported 919- intel/fs: Add support for CS to group invocations in quads 920- intel/fs: Don't loop when lowering CS intrinsics 921- intel/fs: Use NIR_PASS_V when lowering CS intrinsics 922- i965: Advertise NV_compute_shader_derivatives 923- gallium: Add PIPE_CAP_COMPUTE_SHADER_DERIVATIVES 924- iris: Enable NV_compute_shader_derivatives 925- spirv: Add support for DerivativeGroup capabilities 926- anv: Implement VK_NV_compute_shader_derivatives 927- docs: Add NV_compute_shader_derivatives to 19.1.0 relnotes 928- spirv: Add more to_string helpers 929- spirv: Tell which opcode or value is unhandled when failing 930- spirv: Rename vtn_decoration literals to operands 931- spirv: Handle SpvOpDecorateId 932- nir: Add option to lower tex to txl when shader don't support 933 implicit LOD 934- intel/fs: Don't handle texop_tex for shaders without implicit LOD 935- spirv: Properly handle SpvOpAtomicCompareExchangeWeak 936- intel/fs: Assert when brw_fs_nir sees a nir_deref_instr 937- anv: Fix limits when VK_EXT_descriptor_indexing is used 938- nir: Fix nir_opt_idiv_const when negatives are involved 939- nir: Fix clone of nir_variable state slots 940 941Carlos Garnacho (1): 942 943- wayland/egl: Ensure EGL surface is resized on DRI update_buffers() 944 945Chad Versace (17): 946 947- turnip: Drop Makefile.am and Android.mk 948- turnip: Fix indentation in function signatures 949- turnip: Fix result of vkEnumerate*LayerProperties 950- turnip: Fix result of vkEnumerate*ExtensionProperties 951- turnip: Use vk_outarray in all relevant public functions 952- turnip: Fix a real -Wmaybe-uninitialized 953- turnip: Fix indentation 954- turnip: Require DRM device version >= 1.3 955- turnip: Add TODO for Android logging 956- turnip: Use vk_errorf() for initialization error messages 957- turnip: Replace fd_bo with tu_bo 958- turnip: Add TODO file 959- turnip: Fix 'unused' warnings 960- turnip: Don't return from tu_stub funcs 961- turnip: Annotate vkGetImageSubresourceLayout with tu_stub 962- turnip: Fix error behavior for 963 VkPhysicalDeviceExternalImageFormatInfo 964- turnip: Use Vulkan 1.1 names instead of KHR 965 966Charmaine Lee (5): 967 968- svga: add svga shader type in the shader variant 969- svga: move host logging to winsys 970- st/mesa: purge framebuffers with current context after unbinding 971 winsys buffers 972- mesa: unreference current winsys buffers when unbinding winsys 973 buffers 974- svga: Remove unnecessary check for the pre flush bit for setting 975 vertex buffers 976 977Chenglei Ren (1): 978 979- anv/android: fix missing dependencies issue during parallel build 980 981Chia-I Wu (78): 982 983- egl: fix KHR_partial_update without EXT_buffer_age 984- turnip: add .clang-format 985- turnip: use msm_drm.h from inc_freedreno 986- turnip: remove unnecessary libfreedreno_drm dep 987- turnip: add wrappers around DRM_MSM_GET_PARAM 988- turnip: add wrappers around DRM_MSM_SUBMITQUEUE\_\* 989- turnip: constify tu_device in tu_gem\_\* 990- turnip: preliminary support for tu_QueueWaitIdle 991- turnip: run sed and clang-format on tu_cs 992- turnip: document tu_cs 993- turnip: add tu_cs_add_bo 994- turnip: minor cleanup to tu_cs_end 995- turnip: update cs->start in tu_cs_end 996- turnip: inline tu_cs_check_space 997- turnip: add more tu_cs helpers 998- turnip: build drm_msm_gem_submit_bo array directly 999- turnip: add tu_bo_list_merge 1000- turnip: add cmdbuf->bo_list to bo_list in queue submit 1001- turnip: preliminary support for tu_BindImageMemory2 1002- turnip: preliminary support for tu_image_view_init 1003- turnip: preliminary support for tu_CmdBeginRenderPass 1004- turnip: add tu_cs_reserve_space(_assert) 1005- turnip: emit HW init in tu_BeginCommandBuffer 1006- turnip: preliminary support for tu_GetRenderAreaGranularity 1007- turnip: add tu_tiling_config 1008- turnip: add internal helpers for tu_cs 1009- turnip: add tu_cs_{reserve,add}_entry 1010- turnip: specify initial size in tu_cs_init 1011- turnip: never fail tu_cs_begin/tu_cs_end 1012- turnip: add tu_cs_sanity_check 1013- turnip: provide both emit_ib and emit_call 1014- turnip: add tu_cs_mode 1015- turnip: add TU_CS_MODE_SUB_STREAM 1016- turnip: preliminary support for loadOp and storeOp 1017- turnip: add a more complete format table 1018- turnip: add functions to import/export prime fd 1019- turnip: advertise VK_KHR_external_memory_capabilities 1020- turnip: advertise VK_KHR_external_memory 1021- turnip: add support for VK_KHR_external_memory_{fd,dma_buf} 1022- turnip: fix VkClearValue packing 1023- turnip: preliminary support for fences 1024- turnip: respect color attachment formats 1025- turnip: mark IBs for dumping 1026- turnip: use 32-bit offset in tu_cs_entry 1027- turnip: more/better asserts for tu_cs 1028- turnip: add tu_cs_discard_entries 1029- turnip: tu_cs_emit_array 1030- turnip: fix tu_cs sub-streams 1031- turnip: simplify tu_cs sub-streams usage 1032- turnip: create a less dummy pipeline 1033- turnip: parse VkPipelineDynamicStateCreateInfo 1034- turnip: parse VkPipelineInputAssemblyStateCreateInfo 1035- turnip: parse VkPipelineViewportStateCreateInfo 1036- turnip: parse VkPipelineRasterizationStateCreateInfo 1037- turnip: parse VkPipelineDepthStencilStateCreateInfo 1038- turnip: parse VkPipeline{Multisample,ColorBlend}StateCreateInfo 1039- turnip: preliminary support for shader modules 1040- turnip: compile VkPipelineShaderStageCreateInfo 1041- turnip: parse VkPipelineShaderStageCreateInfo 1042- turnip: parse VkPipelineVertexInputStateCreateInfo 1043- turnip: add draw_cs to tu_cmd_buffer 1044- turnip: preliminary support for draw state binding 1045- turnip: preliminary support for tu_CmdDraw 1046- turnip: guard -Dvulkan-driver=freedreno 1047- turnip: preliminary support for tu_GetImageSubresourceLayout 1048- turnip: preliminary support for Wayland WSI 1049- vulkan/wsi: move modifier array into wsi_wl_swapchain 1050- vulkan/wsi: create wl_drm wrapper as needed 1051- vulkan/wsi: refactor drm_handle_format 1052- vulkan/wsi: add wsi_wl_display_drm 1053- vulkan/wsi: add wsi_wl_display_dmabuf 1054- vulkan/wsi: make wl_drm optional 1055- virgl: handle fence_server_sync in winsys 1056- virgl: hide fence internals from the driver 1057- virgl: introduce virgl_drm_fence 1058- virgl: fix fence fd version check 1059- virgl: clear vertex_array_dirty 1060- virgl: skip empty cmdbufs 1061 1062Chris Forbes (3): 1063 1064- glsl: add scaffolding for EXT_gpu_shader4 1065- glsl: enable noperspective|flat|centroid for EXT_gpu_shader4 1066- glsl: enable types for EXT_gpu_shader4 1067 1068Chris Wilson (19): 1069 1070- i965: Assert the execobject handles match for this device 1071- iris: fix import from dri2/3 1072- iris: IndexFormat = size/2 1073- iris: Set resource modifier on handle 1074- iris: Wrap userptr for creating bo 1075- iris: AMD_pinned_memory 1076- iris: Record reusability of bo on construction 1077- iris: fix memzone_for_address since multibinder changes 1078- iris: Tidy exporting the flink handle 1079- iris: Fix assigning the output handle for exporting for KMS 1080- iris: Merge two walks of the exec_bos list 1081- iris: Tag each submitted batch with a syncobj 1082- iris: Add fence support using drm_syncobj 1083- iris: Wire up EGL_IMG_context_priority 1084- iris: Use PIPE_BUFFER_STAGING for the query objects 1085- iris: Use coherent allocation for PIPE_RESOURCE_STAGING 1086- iris: Use streaming loads to read from tiled surfaces 1087- iris: Push heavy memchecker code to DEBUG 1088- iris: Adapt to variable ppGTT size 1089 1090Christian Gmeiner (12): 1091 1092- etnaviv: rs: mark used src resource as read from 1093- etnaviv: blt: mark used src resource as read from 1094- etnaviv: implement ETC2 block patching for HALTI0 1095- etnaviv: keep track of mapped bo address 1096- etnaviv: hook-up etc2 patching 1097- etnaviv: enable ETC2 texture compression support for HALTI0 GPUs 1098- etnaviv: fix resource usage tracking across different pipe_context's 1099- etnaviv: fix compile warnings 1100- st/dri: allow direct UYVY import 1101- etnaviv: shrink struct etna_3d_state 1102- nir: add lower_ftrunc 1103- etnaviv: use the correct uniform dirty bits 1104 1105Chuck Atkins (1): 1106 1107- meson: Fix missing glproto dependency for gallium-glx 1108 1109Connor Abbott (6): 1110 1111- nir/serialize: Prevent writing uninitialized state_slot data 1112- nir: Add a stripping pass for improved cacheability 1113- radeonsi/nir: Use nir stripping pass 1114- nir/search: Add automaton-based pre-searching 1115- nir/search: Add debugging code to dump the pattern matched 1116- nir/algebraic: Don't emit empty initializers for MSVC 1117 1118Daniel Schürmann (2): 1119 1120- nir: Define shifts according to SM5 specification. 1121- nir: Use SM5 properties to optimize shift(a\@32, iand(31, b)) 1122 1123Daniel Stone (2): 1124 1125- panfrost: Properly align stride 1126- vulkan/wsi/wayland: Respect non-blocking AcquireNextImage 1127 1128Danylo Piliaiev (13): 1129 1130- anv: Handle VK_ATTACHMENT_UNUSED in colorAttachment 1131- radv: Handle VK_ATTACHMENT_UNUSED in CmdClearAttachment 1132- anv: Fix VK_EXT_transform_feedback working with varyings packed in 1133 PSIZ 1134- anv: Fix destroying descriptor sets when pool gets reset 1135- anv: Treat zero size XFB buffer as disabled 1136- glsl: Cross validate variable's invariance by explicit invariance 1137 only 1138- i965,iris,anv: Make alpha to coverage work with sample mask 1139- intel/fs: Make alpha test work with MRT and sample mask 1140- st/mesa: Fix GL_MAP_COLOR with glDrawPixels GL_COLOR_INDEX 1141- iris: Fix assert when using vertex attrib without buffer binding 1142- intel/compiler: Do not reswizzle dst if instruction writes to flag 1143 register 1144- drirc: Add workaround for Epic Games Launcher 1145- anv: Do not emulate texture swizzle for INPUT_ATTACHMENT, 1146 STORAGE_IMAGE 1147 1148Dave Airlie (63): 1149 1150- virgl: enable elapsed time queries 1151- virgl: ARB_query_buffer_object support 1152- docs: update qbo support for virgl 1153- glsl: glsl to nir fix uninit class member. 1154- radv/llvm: initialise passes member. 1155- radv: remove alloc parameter from pipeline init 1156- iris: fix some hangs around null framebuffers 1157- iris: fix crash in sparse vertex array 1158- iris: add initial transform feedback overflow query paths (V3) 1159- iris: fix cube texture view 1160- iris: execute compute related query on compute batch. 1161- iris: iris add load register reg32/64 1162- iris: add conditional render support 1163- iris: fix gpu calcs for timestamp queries 1164- iris/WIP: add broadwell support 1165- iris: limit gen8 to 8 samples 1166- iris: setup gen8 caps 1167- iris: add fs invocations query workaround for broadwell 1168- iris: handle qbo fragment shader invocation workaround 1169- st/mesa: add support for lowering fp64/int64 for nir drivers 1170- softpipe: fix texture view crashes 1171- nir/spirv: don't use bare types, remove assert in split vars for 1172 testing 1173- nir/deref: remove casts of casts which are likely redundant (v3) 1174- softpipe: fix 32-bit bitfield extract 1175- softpipe: handle 32-bit bitfield inserts 1176- softpipe: remove shadow_ref assert. 1177- softpipe: fix integer texture swizzling for 1 vs 1.0f 1178- nir/split_vars: fixup some more explicit_stride related issues. 1179- draw: bail instead of assert on instance count (v2) 1180- draw/gs: fix point size outputs from geometry shader. 1181- draw/vs: partly fix basevertex/vertex id 1182- softpipe: fix clears to only clear specified color buffers. 1183- softpipe/draw: fix vertex id in soft paths. 1184- softpipe: add indirect store buffer/image unit 1185- nir/deref: fix struct wrapper casts. (v3) 1186- nir: use proper array sizing define for vectors 1187- intel/compiler: use defined size for vector components 1188- iris: avoid use after free in shader destruction 1189- ddebug: add compute functions to help hang detection 1190- draw: add stream member to stats callback 1191- tgsi: add support for geometry shader streams. 1192- softpipe: add support for indexed queries. 1193- draw: add support to tgsi paths for geometry streams. (v2) 1194- softpipe: add support for vertex streams (v2) 1195- virgl: add support for missing command buffer binding. 1196- virgl: add support for ARB_multi_draw_indirect 1197- virgl: add support for ARB_indirect_parameters 1198- draw: fix undefined shift of (1 << 31) 1199- swrast: fix undefined shift of 1 << 31 1200- llvmpipe: fix undefined shift 1 << 31. 1201- virgl/drm: cleanup buffer from handle creation (v2) 1202- virgl/drm: handle flink name better. 1203- virgl/drm: insert correct handles into the table. (v3) 1204- intel/compiler: fix uninit non-static variable. (v2) 1205- nir: fix bit_size in lower indirect derefs. 1206- r600: reset tex array override even when no view bound 1207- spirv: fix SpvOpBitSize return value. 1208- nir: fix lower vars to ssa for larger vector sizes. 1209- util/tests: add basic unit tests for bitset 1210- util/bitset: fix bitset range mask calculations. 1211- kmsro: add \_dri.so to two of the kmsro drivers. 1212- glsl: init packed in more constructors. 1213- Revert "mesa: unreference current winsys buffers when unbinding 1214 winsys buffers" 1215 1216David Riley (3): 1217 1218- virgl: Store mapped hw resource with transfer object. 1219- virgl: Allow transfer queue entries to be found and extended. 1220- virgl: Re-use and extend queue transfers for intersecting buffer 1221 subdatas. 1222 1223David Shao (1): 1224 1225- meson: ensure that xmlpool_options.h is generated for gallium targets 1226 that need it 1227 1228Deepak Rawat (2): 1229 1230- winsys/drm: Fix out of scope variable usage 1231- winsys/svga/drm: Fix 32-bit RPCI send message 1232 1233Dominik Drees (1): 1234 1235- Add no_aos_sampling GALLIVM_PERF option 1236 1237Drew Davenport (1): 1238 1239- util: Don't block SIGSYS for new threads 1240 1241Dylan Baker (40): 1242 1243- bump version for 19.0 branch 1244- docs: Add relnotes stub for 19.1 1245- gallium: wrap u_screen in extern "C" for c++ 1246- automake: Add --enable-autotools to distcheck flags 1247- android,autotools,i965: Fix location of float64_glsl.h 1248- meson: remove build_by_default : true 1249- meson: fix style in intel/tools 1250- meson: remove -std=c++11 from intel/tools 1251- get-pick-list: Add --pretty=medium to the arguments for Cc patches 1252- meson: Add dependency on genxml to anvil 1253- meson/iris: Use current coding style 1254- docs: Add release notes for 19.0.0 1255- docs: Add SHA256 sums for 19.0.0 1256- docs: update calendar, add news item, and link release notes for 1257 19.0.0 1258- bin/install_megadrivers.py: Correctly handle DESTDIR='' 1259- bin/install_megadrivers.py: Fix regression for set DESTDIR 1260- docs: Add release notes for 19.0.1 1261- docs: Add SHA256 sums for mesa 19.0.1 1262- docs: update calendar, add news item and link release notes for 1263 19.0.1 1264- meson: Error if LLVM doesn't have rtti when building clover 1265- meson: Error if LLVM is turned off but clover it turned on 1266- docs: Add release notes for 19.0.2 1267- docs: Add sha256 sums for 19.0.2 1268- docs: update calendar, and news item and link release notes for 1269 19.0.2 1270- Delete autotools 1271- docs: drop most autoconf references 1272- ci: Delete autotools build jobs 1273- docs: add relnotes for 19.0.3 1274- docs: Add SHA256 sums for mesa 19.0.3 1275- docs: update calendar, and news item and link release notes for 1276 19.0.3 1277- meson: always define libglapi 1278- glsl: fix general_ir_test with mingw 1279- meson: switch gles1 and gles2 to auto options 1280- meson: Make shader-cache a trillean instead of boolean 1281- meson: make nm binary optional 1282- util/tests: Use define instead of VLA 1283- glsl/tests: define ssize_t on windows 1284- tests/vma: fix build with MSVC 1285- meson: Don't build glsl cache_test when shader cache is disabled 1286- meson: Force the use of config-tool for llvm 1287 1288Eduardo Lima Mitev (5): 1289 1290- freedreno/a6xx: Silence compiler warnings 1291- nir: Add ir3-specific version of most SSBO intrinsics 1292- ir3/nir: Add a new pass 'ir3_nir_lower_io_offsets' 1293- ir3/compiler: Enable lower_io_offsets pass and handle new SSBO 1294 intrinsics 1295- ir3/lower_io_offsets: Try propagate SSBO's SHR into a previous shift 1296 instruction 1297 1298El Christianito (1): 1299 1300- drirc: add Budgie WM to adaptive-sync blacklist 1301 1302Eleni Maria Stea (6): 1303 1304- i965: Faking the ETC2 compression on Gen < 8 GPUs using two miptrees. 1305- i965: Fixed the CopyImageSubData for ETC2 on Gen < 8 1306- i965: Enabled the OES_copy_image extension on Gen 7 GPUs 1307- i965: Removed the field etc_format from the struct intel_mipmap_tree 1308- i965: fixed clamping in set_scissor_bits when the y is flipped 1309- radv: consider MESA_VK_VERSION_OVERRIDE when setting the api version 1310 1311Elie Tournier (3): 1312 1313- virgl: Add a caps to advertise GLES backend 1314- virgl: Set PIPE_CAP_DOUBLES when running on GLES This is a lie but no 1315 known app use fp64. 1316- virgl: Return an error if we use fp64 on top of GLES 1317 1318Emil Velikov (30): 1319 1320- vc4: Declare the last cpu pointer as being modified in NEON asm. 1321- docs: add release notes for 18.3.3 1322- docs: add sha256 checksums for 18.3.3 1323- docs: update calendar, add news item and link release notes for 1324 18.3.3 1325- anv: wire up the state_pool_padding test 1326- docs: add release notes for 18.3.4 1327- docs: add sha256 checksums for 18.3.4 1328- docs: update calendar, add news item and link release notes for 1329 18.3.4 1330- egl/dri: de-duplicate dri2_load_driver\* 1331- meson: egl: correctly manage loader/xmlconfig 1332- loader: use loader_open_device() to handle O_CLOEXEC 1333- egl/android: bump the number of drmDevices to 64 1334- docs: mention "Allow commits from members who can merge..." 1335- egl/sl: split out swrast probe into separate function 1336- egl/sl: use drmDevice API to enumerate available devices 1337- egl/sl: use kms_swrast with vgem instead of a random GPU 1338- docs: add release notes for 18.3.5 1339- docs: add sha256 checksums for 18.3.5 1340- docs: update calendar, add news item and link release notes for 1341 18.3.5 1342- docs: add release notes for 18.3.6 1343- docs: add sha256 checksums for 18.3.6 1344- docs: update calendar, add news item and link release notes for 1345 18.3.6 1346- turnip: drop dead close(master_fd) 1347- vulkan/wsi: check if the display_fd given is master 1348- vulkan/wsi: don't use DUMB_CLOSE for normal GEM handles 1349- llvmpipe: add lp_fence_timedwait() helper 1350- llvmpipe: correctly handle waiting in llvmpipe_fence_finish 1351- egl/dri: flesh out and use dri2_create_drawable() 1352- mapi: add static_date offset to MaxShaderCompilerThreadsKHR 1353- mapi: correctly handle the full offset table 1354 1355Emmanuel Gil Peyrot (1): 1356 1357- docs: make bugs.html easier to find 1358 1359Eric Anholt (121): 1360 1361- v3d: Always enable the NEON utile load/store code. 1362- v3d: Fix a release build set-but-unused compiler warning. 1363- mesa: Skip partial InvalidateFramebuffer of packed depth/stencil. 1364- v3d: Fix image_load_store clamping of signed integer stores. 1365- nir: Move V3D's "the shader was TGSI, ignore FS output types" flag to 1366 NIR. 1367- v3d: Fix precompile of FRAG_RESULT_DATA1 and higher outputs. 1368- v3d: Store the actual mask of color buffers present in the key. 1369- v3d: Fix dumping of shaders with alpha test. 1370- v3d: Fix pack/unpack of VFPACK operand unpacks. 1371- v3d: Fix input packing of .l for rounding/fdx/fdy. 1372- v3d: Fix copy-propagation of input unpacks. 1373- v3d: Whitespace consistency fix. 1374- nir: Move panfrost's isign lowering to nir_opt_algebraic. 1375- v3d: Use the NIR lowering for isign instead of rolling our own. 1376- intel: Use the NIR lowering for isign. 1377- freedreno: Use the NIR lowering for isign. 1378- v3d: Clear the GMP on initialization of the simulator. 1379- v3d: Sync indirect draws on the last rendering. 1380- v3d: Use the early_fragment_tests flag for the shader's disable-EZ 1381 field. 1382- v3d: Fix incorrect flagging of ldtmu as writing r4 on v3d 4.x. 1383- v3d: Drop a perf note about merging unpack_half_*, which has been 1384 implemented. 1385- v3d: Drop our hand-lowered nir_op_ffract. 1386- v3d: Add a helper function for getting a nop register. 1387- v3d: Refactor bcsel and if condition handling. 1388- v3d: Do bool-to-cond for discard_if as well. 1389- v3d: Kill off vir_PF(), which is hard to use right. 1390- v3d: Fix f2b32 behavior. 1391- v3d: Fix the check for "is the last thrsw inside control flow" 1392- v3d: Add a function to describe what the c->execute.file check means. 1393- v3d: Stop tracking num_inputs for VPM loads. 1394- v3d: Delay emitting ldvpm on V3D 4.x until it's actually used. 1395- v3d: Emit a simpler negate for the iabs implementation. 1396- v3d: Move i2b and f2b support into emit_comparison. 1397- kmsro: Add the rest of the current set of tinydrm drivers. 1398- nir: Just return when asked to rewrite uses of an SSA def to itself. 1399- v3d: Fix vir_is_raw_mov() for input unpacks. 1400- v3d: Dump the VIR after register spilling if we were forced to. 1401- v3d: Rematerialize MOVs of uniforms instead of spilling them. 1402- v3d: Fix build of NEON code with Mesa's cflags not targeting NEON. 1403- v3d: Restrict live intervals to the blocks reachable from any def. 1404- v3d: Stop treating exec masking specially. 1405- nir: Improve printing of load_input/store_output variable names. 1406- v3d: Translate f2i(fround_even) as FTOIN. 1407- v3d: Move the stores for fixed function VS output reads into NIR. 1408- v3d: Fix temporary leaks of temp_registers and when spilling. 1409- v3d: Do uniform rematerialization spilling before dropping 1410 threadcount 1411- v3d: Switch implicit uniforms over to being any qinst->uniform != ~0. 1412- v3d: Add support for vir-to-qpu of ldunif instructions to a temp. 1413- v3d: Drop the old class bits splitting up the accumulators. 1414- v3d: Add support for register-allocating a ldunif to a QFILE_TEMP. 1415- v3d: Use ldunif instructions for uniforms. 1416- v3d: Eliminate the TLB and TLBU files. 1417- v3d: Drop the V3D 3.x vpm read dead code elimination. 1418- v3d: Include a count of register pressure in the RA failure dumps. 1419- st/dri: Set the PIPE_BIND_SHARED flag on create_image_with_modifiers. 1420- util: Add a DAG datastructure. 1421- vc4: Switch over to using the DAG datastructure for QIR scheduling. 1422- v3d: Reuse list_for_each_entry_rev(). 1423- vc4: Reuse list_for_each_entry_rev(). 1424- v3d: Use the DAG datastructure for QPU instruction scheduling. 1425- vc4: Switch the post-RA scheduler over to the DAG datastructure. 1426- v3d: Disable PIPE_CAP_BLIT_BASED_TEXTURE_TRANSFER. 1427- v3d: Fix leak of the mem_ctx after the DAG refactor. 1428- v3d: Fix leak of the renderonly struct on screen destruction. 1429- mesa/st: Make sure that prog_to_nir NIR gets freed. 1430- mesa/st: Fix leaks of TGSI tokens in VP variants. 1431- v3d: Always lay out shared tiled buffers with UIF_TOP set. 1432- v3d: Allow the UIF modifier with renderonly. 1433- v3d: Expose the dma-buf modifiers query. 1434- v3d: Rename v3d_tmu_config_data to v3d_unit_data. 1435- v3d: Move constant offsets to UBO addresses into the main uniform 1436 stream. 1437- v3d: Upload all of UBO[0] if any indirect load occurs. 1438- v3d: Remove some dead members of struct v3d_compile. 1439- egl: Add a 565 pbuffer-only EGL config under X11. 1440- dri3: Return the current swap interval from glXGetSwapIntervalMESA(). 1441- v3d: Add support for handling OOM signals from the simulator. 1442- v3d: Bump the maximum texture size to 4k for V3D 4.x. 1443- v3d: Don't try to use the TFU blit path if a scissor is enabled. 1444- v3d: Add some more new packets for V3D 4.x. 1445- st: Lower uniforms in st in the !PIPE_CAP_PACKED_UNIFORMS case as 1446 well. 1447- vc4: Don't forget to set the range when scalarizing our uniforms. 1448- vc4: Split UBO0 and UBO1 address uniform handling. 1449- vc4: Upload CS/VS UBO uniforms together. 1450- v3d: Add an optimization pass for redundant flags updates. 1451- nir: Drop comments about the constant_index slots for load/stores. 1452- nir: Drop remaining references to const_index in favor of the call to 1453 use. 1454- nir: Add a comment about how intrinsic definitions work. 1455- v3d: Add and use a define for the number of channels in a QPU 1456 invocation. 1457- v3d: Drop a note for the future about PIPE_CAP_PACKED_UNIFORMS. 1458- v3d: Include the number of max temps used in the shader-db output. 1459- v3d: Replace the old shader-db env var output with the 1460 ARB_debug_output. 1461- v3d: Add Compute Shader compilation support. 1462- v3d: Add missing base offset to CS shared memory accesses. 1463- v3d: Add missing dumping for the spill offset/size uniforms. 1464- v3d: Detect the correct number of QPUs and use it to fix the spill 1465 size. 1466- v3d: Use the new lower_to_scratch implementation for indirects on 1467 temps. 1468- v3d: Only look up the 3rd texture gather offset for non-arrays. 1469- v3d: Always set up the qregs for CSD payload. 1470- v3d: Fix an invalid reuse of flags generation from before a thrsw. 1471- v3d: Fix atomic cmpxchg in shaders on hardware. 1472- nir: Fix deref offset calculation for structs. 1473- nir: Use the nir_builder \_imm helpers in setting up deref offsets. 1474- gallium: Remove the pool pipebuffer manager. 1475- gallium: Remove the ondemand pipebuffer manager. 1476- gallium: Remove the "alt" pipebuffer manager interface. 1477- gallium: Remove the malloc pipebuffer manager. 1478- st/mesa: Don't set atomic counter size != 0 if MAX_SHADER_BUFFERS == 1479 0. 1480- v3d: Disable SSBOs and atomic counters on vertex shaders. 1481- v3d: Fill in the ignored segment size fields to appease new 1482 simulator. 1483- v3d: Apply the GFXH-930 workaround to the case where the VS loads 1484 attrs. 1485- v3d: Assert that we do request the normal texturing return data. 1486- v3d: Use \_mesa_hash_table_remove_key() where appropriate. 1487- vc4: Use \_mesa_hash_table_remove_key() where appropriate. 1488- v3d: Add a note about i/o indirection for future performance work. 1489- v3d: Don't try to update the shadow texture for separate stencil. 1490- Revert "v3d: Disable PIPE_CAP_BLIT_BASED_TEXTURE_TRANSFER." 1491- v3d: Re-add support for memory_barrier_shared. 1492- v3d: Fix detection of the last ldtmu before a new TMU op. 1493- v3d: Fix detection of TMU write sequences in register spilling. 1494- kmsro: Add support for V3D. 1495- vc4: Fall back to renderonly if the vc4 driver doesn't have v3d. 1496 1497Eric Engestrom (142): 1498 1499- wsi/display: add comment 1500- egl: use coherent variable names 1501- gitlab-ci: add ubuntu container 1502- gitlab-ci: add a meson vulkan build 1503- gitlab-ci: add a make vulkan build 1504- gitlab-ci: add a scons no-llvm build 1505- gitlab-ci: add scons llvm 3.5 build 1506- gitlab-ci: add scons SWR build 1507- gitlab-ci: add meson loader/classic DRI build 1508- gitlab-ci: add meson gallium SWR build 1509- gitlab-ci: add meson gallium RadeonSI build 1510- gitlab-ci: add meson gallium "other drivers" build 1511- gitlab-ci: add meson gallium ST Clover (LLVM 5.0) build 1512- gitlab-ci: add meson gallium ST Clover (LLVM 6.0) build 1513- gitlab-ci: add meson gallium ST Clover (LLVM 7.0) build 1514- gitlab-ci: add meson gallium ST "Other" build 1515- gitlab-ci: add make loaders/classic DRI build 1516- gitlab-ci: add make Gallium Drivers SWR build 1517- gitlab-ci: add make Gallium Drivers RadeonSI build 1518- gitlab-ci: add make Gallium Drivers "Other" build 1519- gitlab-ci: add make Gallium ST Clover LLVM-3.9 build 1520- gitlab-ci: add make Gallium ST Clover LLVM-4.0 build 1521- gitlab-ci: add make Gallium ST Clover LLVM-5.0 build 1522- gitlab-ci: add make Gallium ST Clover LLVM-6.0 build 1523- gitlab-ci: add make Gallium ST Clover LLVM-7 build 1524- gitlab-ci: add make Gallium ST Other build 1525- travis: remove unused linux code path 1526- travis: remove unused scons code path 1527- gitlab-ci: add meson glvnd build 1528- xvmc: fix string comparison 1529- xvmc: fix string comparison 1530- meson: add script to print the options before configuring a builddir 1531- driconf: drop unused macro 1532- travis: fix osx make build 1533- gitlab-ci: workaround docker bug for users with uppercase characters 1534- wsi: query the ICD's max dimensions instead of hard-coding them 1535- gitlab-ci: limit ninja to 4 threads max 1536- drm-uapi/README: remove explicit list of driver names 1537- drm-uapi: use local files, not system libdrm 1538- gbm: drop duplicate #defines 1539- st/dri: drop duplicate #define 1540- etnaviv: drop duplicate #define 1541- anv/tests: compile to something sensible in release builds 1542- util/tests: compile to something sensible in release builds 1543- gitlab-ci: use ccache to speed up builds 1544- tegra/meson: add missing dep_libdrm 1545- tegra/autotools: add missing libdrm cflags 1546- gitlab-ci: limit the automatic CI to master and MRs 1547- gitlab-ci: automatically run the CI on pushes to \`ci/\*\` branches 1548- anv: sort extensions alphabetically 1549- anv: sort vendors extensions after KHR and EXT 1550- anv: make sure the extensions stay sorted 1551- anv: drop unused imports 1552- anv: use anv_shader_bin_write_to_blob()'s return value 1553- gitlab-ci: always run the containers build 1554- dri_interface: add missing #include 1555- driinfo: add DTD to allow the xml to be validated 1556- meson/swr: replace hard-coded path with current_build_dir() 1557- egl/android: replace magic 0=CbCr,1=CrCb with simple enum 1558- vulkan: use VkBase{In,Out}Structure instead of a custom struct 1559- driconf: add DTD to allow the drirc xml (00-mesa-defaults.conf) to be 1560 validated 1561- gitlab-ci: install xmllint to validate 00-mesa-defaults.conf 1562- anv: simplify chained comparison 1563- anv: drop unused parameter 1564- anv: remove spaces around kwargs assignment 1565- anv: fix typo 1566- Revert "swr/rast: Archrast codegen updates" 1567- meson: avoid going back up the tree with include_directories() 1568- anv: use the platform defines in vk.xml instead of hard-coding them 1569- radv: use the platform defines in vk.xml instead of hard-coding them 1570- util: #define PATH_MAX when undefined (eg. Hurd) 1571- vulkan: import missing file from Khronos 1572- egl: fix libdrm-less builds 1573- vulkan: import vk_layer.h from Khronos 1574- gitlab-ci: drop job prefixes 1575- meson: fix with_dri2 definition for GNU Hurd 1576- meson: remove unused include_directories(vulkan) 1577- vulkan/util: use the platform defines in vk.xml instead of 1578 hard-coding them 1579- vulkan/overlay: fix missing var rename in previous commit 1580- meson: don't build libGLES*.so with GLVND 1581- autotools: don't build libGLES*.so with GLVND 1582- travis: fix meson build by letting \`auto\` do its job 1583- travis: drop unused vars 1584- travis: clean up 1585- gitlab-ci: only build the default (=latest) and oldest llvm versions 1586- gitlab-ci: autotools needs to be told which llvm version to use 1587- r600: cast pointer to expected type 1588- build: make passing an incorrect pointer type a hard error 1589- gitlab-ci: fix llvm version (7 doesn't have a ".0") 1590- hgl/meson: drop unused include directory 1591- glx/meson: use full include path for dri_interface.h 1592- android: fix missing backspace for line continuation 1593- panfrost: fix tgsi_to_nir() call 1594- panfrost: move #include to fix compilation 1595- gitlab-ci: add panfrost to the gallium drivers build 1596- wsi: deduplicate get_current_time() functions between display and x11 1597- wsi/display: s/#if/#ifdef/ to fix -Wundef 1598- wsi/wayland: fix pointer casting warning on 32bit 1599- wsi/x11: use WSI_FROM_HANDLE() instead of pointer casts 1600- turnip: use the platform defines in vk.xml instead of hard-coding 1601 them 1602- travis: fix osx meson build 1603- nir: const \`nir_call_instr::callee\` 1604- gitlab-ci: add clang build 1605- gitlab-ci: drop most autotools builds 1606- util/disk_cache: close fd in the fallback path 1607- egl: hide entrypoints that shouldn't be exported when using glvnd 1608- meson: strip rpath from megadrivers 1609- gallium/hud: fix memory leaks 1610- gallium/hud: prevent buffer overflow 1611- gallium/hud: fix rounding error in nic bps computation 1612- simplify LLVM version string printing 1613- util/process: document memory leak 1614- vk/util: remove unneeded array index 1615- bin: drop unused import from install_megadrivers.py 1616- meson: remove meson-created megadrivers symlinks 1617- gitlab-ci: build gallium extra hud 1618- gitlab-ci: add lima to the build 1619- delete autotools .gitignore files 1620- delete autotools input files 1621- docs: remove unsupported GL function name mangling 1622- docs: drop autotools python information 1623- docs: replace autotools intructions with meson equivalent 1624- docs: use past tense when talking about autotools 1625- docs: haiku can be built using meson 1626- egl: fixup autotools-specific wording 1627- util: add os_read_file() helper 1628- anv: add support for VK_EXT_memory_budget 1629- radv: update to use the new features struct names 1630- turnip: update to use the new features struct names 1631- gitlab-ci: build vulkan drivers in clang build 1632- util: move #include out of #if linux 1633- wsi/wayland: document lack of vkAcquireNextImageKHR timeout support 1634- egl: hard-code destroy function instead of passing it around as a 1635 pointer 1636- gitlab-ci: add scons windows build using mingw 1637- gitlab-ci: merge several meson jobs 1638- gitlab-ci: meson-gallium-radeonsi was a subset of 1639 meson-gallium-clover-llvm 1640- gitlab-ci: simplify meson job names 1641- gitlab-ci: merge meson-glvnd into meson-swr 1642- travis: fix syntax, and drop unused stuff 1643- util/os_file: always use the 'grow' mechanism 1644- meson: expose glapi through osmesa 1645- util/os_file: actually return the error read() gave us 1646 1647Erico Nunes (5): 1648 1649- lima/ppir: support ppir_op_ceil 1650- nir/algebraic: add lowering for fsign 1651- lima: enable nir fsign lowering in ppir 1652- lima/gpir: add limit of max 512 instructions 1653- lima/ppir: support nir_op_ftrunc 1654 1655Erik Faye-Lund (79): 1656 1657- mesa: expose NV_conditional_render on GLES 1658- st/mesa: remove unused header-file 1659- swr/codegen: fix autotools build 1660- virgl: remove unused variables 1661- virgl: remove unused variable 1662- virgl: remove unused variable 1663- virgl: remove unused variable 1664- virgl: do not allow compressed formats for buffers 1665- virgl: stricter usage of compressed 3d textures 1666- virgl: also destroy all read-transfers 1667- virgl: use debug_printf instead of fprintf 1668- virgl: unsigned int -> unsigned 1669- virgl: only warn about unchecked flags 1670- virgl: do not warn about display-target binding 1671- virgl: use debug_printf instead of fprintf 1672- virgl: remove pointless transfer-counter 1673- virgl: tmp_resource -> templ 1674- virgl: track full virgl_resource instead of just virgl_hw_res 1675- virgl: simplify virgl_texture_transfer_unmap logic 1676- virgl: make unmap queuing a bit more straight-forward 1677- virgl: check for readback on correct resource 1678- virgl: wait for the right resource 1679- virgl: return error if allocating resolve_tmp fails 1680- virgl: rewrite core of virgl_texture_transfer_map 1681- virgl: use pipe_box for blit dst-rect 1682- virgl: support write-back with staged transfers 1683- virgl: make sure bind is set for non-buffers 1684- gallium/util: support translating between uint and sint formats 1685- virgl: get readback-formats from host 1686- virgl: only blit if resource is read 1687- virgl: do color-conversion during when mapping transfer 1688- virgl: document potentially failing blit 1689- mesa/st: remove impossible error-check 1690- gallium/u_vbuf: support NULL-resources 1691- i915: support NULL-resources 1692- nouveau: support NULL-resources 1693- swr: support NULL-resources 1694- mesa/st: accept NULL and empty buffer objects 1695- mesa/st: remove always-false state 1696- softpipe: setup pixel_offset for all primitive types 1697- docs: normaize css-indent style 1698- docs: remove non-existent css attribute 1699- docs: remove long commented out css 1700- docs: add missing semicolon 1701- docs: avoid repeating the font 1702- docs: avoid repeating the color 1703- docs: remove spurious newline 1704- docs: use multiple background-images for header 1705- docs: simplify css-centering 1706- docs: do not hard-code header-height 1707- docs: properly escape '>' 1708- docs: properly escape ampersand 1709- docs: remove stray paragraph-close 1710- docs: use h2 instead of b-tag for headings 1711- docs: use dl/dd instead of blockquote for freedesktop link 1712- docs: open list-item before closing it 1713- docs: close paragraphs before lists 1714- docs: close lists 1715- docs: remove stray paragraph-close 1716- docs: close paragraphs before preformatted text 1717- docs: start paragraph before closing it 1718- docs: drop paragraph around preformatted text 1719- docs: fix incorrectly closed paragraph 1720- docs: don't pointlessly close and re-start definition lists 1721- docs: remove stray list-start 1722- docs: fixup bad paragraphing 1723- docs: add missing lists 1724- docs: fix closing of paragraphs 1725- docs: fixup list-item tags 1726- docs: fix closing of list-items 1727- docs: replace empty list with a none-paragraph 1728- docs: turn faq-index into an ordered list 1729- docs: drop centered heading for faq 1730- docs: reorder heading and notice 1731- meson: lift driver-collection out into parent build-file 1732- meson: give dri- and gallium-drivers separate vars 1733- meson: add build-summary 1734- docs: fixup mistake in contents 1735- draw: flush when setting stream-out targets 1736 1737Ernestas Kulik (2): 1738 1739- vc4: Fix leak in HW queries error path 1740- v3d: Fix leak in resource setup error path 1741 1742Francisco Jerez (6): 1743 1744- intel/dump_gpu: Disambiguate between BOs from different GEM handle 1745 spaces. 1746- intel/fs: Exclude control sources from execution type and region 1747 alignment calculations. 1748- intel/fs: Lower integer multiply correctly when destination stride 1749 equals 4. 1750- intel/fs: Cap dst-aligned region stride to maximum representable 1751 hstride value. 1752- intel/fs: Implement extended strides greater than 4 for IR source 1753 regions. 1754- intel/fs: Rely on undocumented unrestricted regioning for 32x16-bit 1755 integer multiply. 1756 1757Fritz Koenig (4): 1758 1759- freedreno: pass count to query_dmabuf_modifiers 1760- freedreno/a6xx: UBWC support 1761- freedreno: UBWC allocator 1762- freedreno/a6xx: Enable UBWC modifier 1763 1764Gert Wollny (35): 1765 1766- mesa/core: Enable EXT_texture_sRGB_R8 also for desktop GL 1767- radeonsi: release tokens after creating the shader program 1768- mesa: release references to image textures when a context is 1769 destroyed 1770- virgl: Enable mixed color FBO attachemnets only when the host 1771 supports it 1772- mesa/core: Enable EXT_depth_clamp for GLES >= 2.0 1773- nir: Add posibility to not lower to source mod 'abs' for ops with 1774 three sources 1775- mesa: Expose EXT_texture_query_lod and add support for its use 1776 shaders 1777- softpipe: Enable PIPE_CAP_MIXED_COLORBUFFER_FORMATS It seems 1778 softpipe actually supports this. This change enables the following 1779 piglits as passing without regressions in the gpu test set: 1780- virgl: Add a caps feature check version 1781- softpipe: Implement ATOMFADD and enable cap TGSI_ATOMFADD 1782- virgl: define MAX_VERTEX_STREAMS based on availability of TF3 1783- softpipe: Use mag texture filter also for clamped lod == 0 1784- softpipe: Don't use mag filter for gather op 1785- softpipe: raise number of bits used for X coordinate texture lookup 1786- softpipe: Add an extra code path for the buffer texel lookup 1787- softpipe: Enable PIPE_CAP_TEXTURE_BUFFER_OFFSET_ALIGNMENT 1788- Gallium: Add new CAP that indicated whether IO array definitions can 1789 be shriked 1790- virgl: Enable passing arrays as input to fragment shaders 1791- doc/features: Add a few extensions to the feature matrix 1792- softpipe: Factor gradient evaluation out of the lambda evaluation 1793- softpipe: Prepare handling explicit gradients 1794- softpipe: Pipe gather_comp through from st_tgsi_get_samples 1795- softpipe: Move selection of shadow values up and clean parameter list 1796- softpipe: tie in new code path for lod evaluation 1797- softpipe: keep input lod for explicite derivatives 1798- softpipe: evaluate cube the faces on a per sample bases 1799- softpipe: Factor out evaluation of the source indices 1800- softpipe: Add an per-input array for interpolator correctors to 1801 machine 1802- softpipe: Add (fake) support for TGSI_OPCODE_INTERP_SAMPLE 1803- softpipe: Add support for TGSI_OPCODE_INTERP_OFFSET 1804- softpipe: Add support for TGSI_OPCODE_INTERP_CENTROID 1805- softpipe: Increase the GLSL feature level 1806- doc: Update feature matrix 1807- softpipe/buffer: load only as many components as the the buffer 1808 resource type provides 1809- Revert "softpipe/buffer: load only as many components as the the 1810 buffer resource type provides" 1811 1812Greg V (3): 1813 1814- util: emulate futex on FreeBSD using umtx 1815- gallium/hud: add CPU usage support for FreeBSD 1816- gallium: enable dmabuf on BSD as well 1817 1818Grigori Goronzy (1): 1819 1820- glx: add support for GLX_ARB_create_context_no_error (v3) 1821 1822Guido Günther (4): 1823 1824- docs: Fix 19.0.x version numbers 1825- gallium: ddebug: Add missing fence related wrappers 1826- gallium/u_dump: util_dump_sampler_view: Dump u.tex.first_level 1827- gallium: trace: Add missing fence related wrappers 1828 1829Gurchetan Singh (44): 1830 1831- mesa/main: Expose EXT_texture_compression_s3tc_srgb 1832- i965: Set flag for EXT_texture_compression_s3tc_srgb 1833- st/mesa: expose EXT_texture_compression_s3tc_srgb 1834- docs: add GL_EXT_texture_compression_s3tc_srgb to release notes 1835- virgl: add ability to do finer grain dirty tracking 1836- virgl: use virgl_resource_dirty helper 1837- virgl: don't mark unclean after a flush 1838- virgl: track level cleanliness rather than resource cleanliness 1839- virgl: make alignment smaller when uploading index user buffers 1840- virgl: unmap uploader at flush time 1841- virgl: when creating / freeing transfers, pass slab pool directly 1842- virgl: add protocol for resource transfers 1843- virgl: use virgl_transfer in inline write 1844- virgl: limit command length to 16 bits 1845- virgl: keep track of number of computations 1846- virgl: pass virgl transfer to virgl_res_needs_flush_wait 1847- virgl: add extra checks in virgl_res_needs_flush_wait 1848- virgl: make winsys modifications for encoded transfers 1849- virgl: add encoder functions for new protocol 1850- virgl: introduce transfer queue 1851- virgl: use transfer queue 1852- virgl: use virgl_transfer_inline_write even less 1853- virgl/vtest: deprecate protocol version 1 1854- egl/sl: also allow virtgpu to fallback to kms_swrast 1855- virgl: use uint16_t mask instead of separate booleans 1856- configure.ac / meson: depend on libnativewindow when appropriate 1857- anv: move anv_GetMemoryAndroidHardwareBufferANDROID up a bit 1858- anv: fix build on Nougat 1859- egl/android: move droid_image_loader_extension down a bit 1860- egl/android: move droid_open_device_drm_gralloc down a bit 1861- egl/android: droid_open_device_drm_gralloc --> droid_open_device 1862- egl/android: refactor droid_load_driver a bit 1863- egl/android: plumb swrast option 1864- egl/android: use swrast option in droid_load_driver 1865- egl/android: use software rendering when appropriate 1866- egl/android: chose node type based on swrast and preprocessor flags 1867- virgl: wait after a flush 1868- virgl/vtest: execute a transfer_get when flushing the front buffer 1869- virgl/vtest: add utilities for receiving fds 1870- virgl/vtest: plumb support for shared memory 1871- virgl/vtest: receive and handle shared memory fd 1872- virgl/vtest: modify sending and receiving data for shared memory 1873- virgl/vtest: wait after issuing a transfer get 1874- virgl/vtest: bump up protocol version + support encoded transfers 1875 1876Guttula, Suresh (1): 1877 1878- st/va:Add support for indirect manner by returning 1879 VA_STATUS_ERROR_OPERATION_FAILED 1880 1881Hal Gentz (1): 1882 1883- glx: Fix synthetic error generation in \__glXSendError 1884 1885Heinrich (1): 1886 1887- gbm: Improve documentation of BO import 1888 1889Iago Toral Quiroga (39): 1890 1891- compiler/nir: add an is_conversion field to nir_op_info 1892- compiler/nir: add lowering option for 16-bit fmod 1893- compiler/nir: add lowering for 16-bit flrp 1894- compiler/nir: add lowering for 16-bit ldexp 1895- intel/compiler: add a NIR pass to lower conversions 1896- intel/compiler: split float to 64-bit opcodes from int to 64-bit 1897- intel/compiler: handle b2i/b2f with other integer conversion opcodes 1898- intel/compiler: assert restrictions on conversions to half-float 1899- intel/compiler: lower some 16-bit float operations to 32-bit 1900- intel/compiler: handle extended math restrictions for half-float 1901- intel/compiler: implement 16-bit fsign 1902- intel/compiler: drop unnecessary temporary from 32-bit fsign 1903 implementation 1904- intel/compiler: add instruction setters for Src1Type and Src2Type. 1905- intel/compiler: add new half-float register type for 3-src 1906 instructions 1907- intel/compiler: don't compact 3-src instructions with Src1Type or 1908 Src2Type bits 1909- intel/compiler: allow half-float on 3-source instructions since gen8 1910- intel/compiler: set correct precision fields for 3-source float 1911 instructions 1912- intel/compiler: fix ddx and ddy for 16-bit float 1913- intel/compiler: fix ddy for half-float in Broadwell 1914- intel/compiler: workaround for SIMD8 half-float MAD in gen8 1915- intel/compiler: split is_partial_write() into two variants 1916- intel/compiler: activate 16-bit bit-size lowerings also for 8-bit 1917- intel/compiler: rework conversion opcodes 1918- intel/compiler: ask for an integer type if requesting an 8-bit type 1919- intel/eu: force stride of 2 on NULL register for Byte instructions 1920- intel/compiler: generalize the combine constants pass 1921- intel/compiler: implement is_zero, is_one, is_negative_one for 1922 8-bit/16-bit 1923- intel/compiler: add a brw_reg_type_is_integer helper 1924- intel/compiler: fix cmod propagation for non 32-bit types 1925- intel/compiler: remove inexact algebraic optimizations from the 1926 backend 1927- intel/compiler: skip MAD algebraic optimization for half-float or 1928 mixed mode 1929- intel/compiler: implement SIMD16 restrictions for mixed-float 1930 instructions 1931- intel/compiler: also set F execution type for mixed float mode in BDW 1932- intel/compiler: validate region restrictions for half-float 1933 conversions 1934- intel/compiler: validate conversions between 64-bit and 8-bit types 1935- intel/compiler: validate region restrictions for mixed float mode 1936- compiler/spirv: move the check for Int8 capability 1937- anv/pipeline: support Float16 and Int8 SPIR-V capabilities in gen8+ 1938- anv/device: expose VK_KHR_shader_float16_int8 in gen8+ 1939 1940Ian Romanick (55): 1941 1942- nir: Silence zillions of unused parameter warnings in release builds 1943- intel/compiler: Silence warning about value that may be used 1944 uninitialized 1945- nir: Document some fields of nir_loop_terminator 1946- nir: Refactor code that checks phi nodes in opt_peel_loop_initial_if 1947- nir: Select phi nodes using prev_block instead of continue_block 1948- nir: Split ALU instructions in loops that read phis 1949- nir: Convert a bcsel with only phi node sources to a phi node 1950- spirv: Add missing break 1951- nir/algebraic: Convert some f2u to f2i 1952- nir/algebraic: Simplify comparison with sequential integers starting 1953 with 0 1954- intel/vec4: Emit constants for some ALU sources as immediate values 1955- nir/algebraic: Replace i2b used by bcsel or if-statement with 1956 comparison 1957- intel/fs: Relax type matching rules in cmod propagation from MOV 1958 instructions 1959- intel/fs: Handle OR source modifiers in algebraic optimization 1960- intel/fs: Refactor ALU source and destination handling to a separate 1961 function 1962- intel/fs: Emit logical-not of operands on Gen8+ 1963- intel/fs: Use De Morgan's laws to avoid logical-not of a logic result 1964 on Gen8+ 1965- intel/fs: Emit better code for b2f(inot(a)) and b2i(inot(a)) 1966- nir/algebraic: Replace a bcsel of a b2f sources with a b2f(!(a \|\| 1967 b)) 1968- intel/fs: Generate if instructions with inverted conditions 1969- nir/algebraic: Replace a-fract(a) with floor(a) 1970- intel/fs: Don't assert on b2f with a saturate modifier 1971- nir/algebraic: Optimize away an fsat of a b2f 1972- intel/compiler: Silence many unused parameter warnings in brw_eu.h 1973- intel/compiler: Silence unused parameter warning in 1974 brw_interpolation_map.c 1975- intel/fs: nir_op_extract_i8 extracts a byte, not a word 1976- intel/fs: Fix extract_u8 of an odd byte from a 64-bit integer 1977- nir/algebraic: Fix up extract_[iu]8 after loop unrolling 1978- nir/algebraic: Remove redundant extract_[iu]8 patterns 1979- nir/algebraic: Add missing 64-bit extract_[iu]8 patterns 1980- nir/algebraic: Add missing 16-bit extract_[iu]8 patterns 1981- nir/algebraic: Fix up extract_[iu]8 after loop unrolling 1982- nir/algebraic: Remove redundant extract_[iu]8 patterns 1983- nir/algebraic: Add missing 64-bit extract_[iu]8 patterns 1984- nir/algebraic: Add missing 16-bit extract_[iu]8 patterns 1985- nir: Add nir_const_value_negative_equal 1986- nir: Add nir_alu_srcs_negative_equal 1987- nir: Add partial redundancy elimination for compares 1988- intel/compiler: Use partial redundancy elimination for compares 1989- intel/fs: Eliminate dead code first 1990- intel/fs: Refactor code generation for nir_op_fsign to its own 1991 function 1992- intel/fs: Add a scale factor to emit_fsign 1993- intel/fs: Generate better code for fsign multiplied by a value 1994- nir/algebraic: Recognize open-coded copysign(1.0, a) 1995- nir/algebraic: Replace a pattern where iand with a Boolean is used as 1996 a bcsel 1997- nir/algebraic: Fix some 1-bit Boolean weirdness 1998- nir/algebraic: Strength reduce some compares of x and -x 1999- intel/fs: Add support for float16 to the fsign optimizations 2000- glsl: Silence may unused parameter warnings in glsl/ir.h 2001- intel/compiler: Don't have sepearate, per-Gen nir_options 2002- intel/compiler: Lower ffma on Gen4 and Gen5 2003- intel/fs: Fix D to W conversion in opt_combine_constants 2004- mesa: Add missing display list support for GL_FOG_COORDINATE_SOURCE 2005- nir: Saturating integer arithmetic is not associative 2006- Revert "nir: add late opt to turn inot/b2f combos back to bcsel" 2007 2008Icenowy Zheng (5): 2009 2010- lima: add dummy set_sample_mask function 2011- lima: make lima_context_framebuffer subtype of pipe_framebuffer_state 2012- lima: implement blit with util_blitter 2013- lima: lower bool to float when building shaders 2014- lima: add Android build 2015 2016Ilia Mirkin (14): 2017 2018- nv50,nvc0: add explicit settings for recent caps 2019- nvc0: add support for handling indirect draws with attrib conversion 2020- nvc0/ir: always use CG mode for loads from atomic-only buffers 2021- nvc0/ir: fix second tex argument after levelZero optimization 2022- nvc0: fix 3d images on kepler 2023- nv50,nvc0: use condition for occlusion queries when already complete 2024- nvc0: stick zero values for the compute invocation counts 2025- nvc0: we have 16k-sized framebuffers, fix default scissors 2026- swr: set PIPE_CAP_MAX_VARYINGS correctly 2027- mesa: add explicit enable for EXT_float_blend, and error condition 2028- st/mesa: enable GL_EXT_float_blend when possible 2029- i965: always enable EXT_float_blend 2030- nv50: disable compute 2031- glsl: fix recording of variables for XFB in TCS shaders 2032 2033Illia Iorin (1): 2034 2035- mesa/main: Fix multisample texture initialize 2036 2037James Zhu (12): 2038 2039- gallium/auxiliary/vl: Move dirty define to header file 2040- gallium/auxiliary/vl: Split vl_compositor graphic shaders from 2041 vl_compositor API 2042- gallium/auxiliary/vl: Rename csc_matrix and increase its size. 2043- gallium/auxiliary/vl: Add compute shader to support video compositor 2044 render 2045- gallium/auxiliary/vl: Add video compositor compute shader render 2046- gallium/auxiliary/vl: Fix transparent issue on compute shader with 2047 rgba 2048- gallium/auxiliary/vl: Increase shader_params size 2049- gallium/auxiliary/vl: Change grid setting 2050- gallium/auxiliary/vl: Change weave compute shader implementation 2051- gallium/auxiliary/vl: Fixed blur issue with weave compute shader 2052- gallium/auxiliary/vl: Fixed blank issue with compute shader 2053- gallium/auxiliary/vl: Add barrier/unbind after compute shader launch. 2054 2055Jan Vesely (2): 2056 2057- Partially revert "gallium: fix autotools build of pipe_msm.la" 2058- gallium/aux: Report error if loading of a pipe driver fails. 2059 2060Jan Zielinski (1): 2061 2062- swr/rast: fix 32-bit compilation on Linux 2063 2064Faith Ekstrand (212): 2065 2066- spirv: Replace vtn_constant_value with vtn_constant_uint 2067- spirv: Rework handling of spec constant workgroup size built-ins 2068- spirv: Handle constants and types before execution modes 2069- spirv: Handle OpExecutionModeId 2070- spirv: Support LocalSizeId and LocalSizeHintId execution modes 2071- intel/nir: Add global support to lower_mem_access_bit_sizes 2072- intel/fs/cse: Split create_copy_instr into three cases 2073- intel/fs: Properly handle 64-bit types in LOAD_PAYLOAD 2074- intel/fs: Do the grf127 hack on SIMD8 instructions in SIMD16 mode 2075- intel/fs: Implement load/store_global with A64 untyped messages 2076- intel/fs: Use SENDS for A64 writes on gen9+ 2077- intel/fs: Implement nir_intrinsic_global_atomic\_\* 2078- anv: Implement VK_EXT_buffer_device_address 2079- relnotes: Add VK_EXT_buffer_device_address 2080- nir/deref: Drop zero ptr_as_array derefs 2081- README: Drop the badges from the readme 2082- intel/fs: Use enumerated array assignments in fb read TXF setup 2083- nir/deref: Rematerialize parents in 2084 rematerialize_derefs_in_use_blocks 2085- nir: Silence a couple of warnings in release builds 2086- anv/blorp: Delete a pointless assert 2087- anv: Silence some compiler warnings in release builds 2088- intel/fs: Silence a compiler warning 2089- intel/fs: Bail in optimize_extract_to_float if we have modifiers 2090- nir/dead_cf: Inline cf_node_has_side_effects 2091- nir/dead_cf: Stop relying on liveness analysis 2092- compiler/types: Add a contains_64bit helper 2093- nir/xfb: Properly align 64-bit values 2094- nir: Rewrite lower_clip_cull_distance_arrays to do a lot less 2095 lowering 2096- nir/xfb: Work in terms of components rather than slots 2097- nir/xfb: Handle compact arrays in gather_xfb_info 2098- nir: Fix a compile warning 2099- nir/lower_clip_cull: Fix an incorrect assert 2100- iris: Don't lower image formats for write-only images 2101- iris/compute: Don't increment the grid size offset 2102- iris/compute: Zero out the last grid size on indirect dispatches 2103- iris: Configure the L3$ on the compute context 2104- iris: Don't set constant read lengths at upload time 2105- iris: Allocate buffer resources separately 2106- iris: Copy anv's MI_MATH helpers for multiplication and division 2107- nir/split_vars: Don't compact vectors unnecessarily 2108- nir/builder: Don't emit no-op swizzles 2109- intel/eu: Add an EOT parameter to send_indirect_[split]_message 2110- intel/fs: Add an enum type for logical sampler inst sources 2111- intel/fs: Re-order logical surface arguments 2112- intel/fs: Drop the fs_surface_builder 2113- intel/vec4: Drop dead code for handling typed surface messages 2114- intel/fs: Get rid of the IMAGE_SIZE opcode 2115- intel/compiler: Drop unused surface opcodes 2116- intel/schedule_instructions: Move some comments 2117- intel/compiler: Re-prefix non-logical surface opcodes with VEC4 2118- anv: Count surfaces for non-YCbCr images in 2119 GetDescriptorSetLayoutSupport 2120- spirv: OpImageQueryLod requires a sampler 2121- intel,nir: Lower TXD with min_lod when the sampler index is not < 16 2122- anv: Use an actual binding for gl_NumWorkgroups 2123- anv/pipeline: Drop anv_fill_binding_table 2124- anv/descriptor_set: Refactor alloc/free of descriptor sets 2125- anv: Rework arguments to anv_descriptor_set_write\_\* 2126- anv: Stop allocating buffer views for dynamic buffers 2127- anv: Count image param entries rather than images 2128- anv: Clean up descriptor set layouts 2129- anv: drop add_var_binding from anv_nir_apply_pipeline_layout.c 2130- anv: Refactor descriptor pushing a bit 2131- anv: Take references to push descriptor set layouts 2132- anv: Add a concept of a descriptor buffer 2133- spirv: Pull offset/stride from the pointer for OpArrayLength 2134- spirv: Use the generic dereference function for OpArrayLength 2135- spirv: Use the same types for resource indices as pointers 2136- anv: Implement VK_EXT_inline_uniform_block 2137- nir: Expose double and int64 op_to_options_mask helpers 2138- nir: Teach loop unrolling about 64-bit instruction lowering 2139- i965: Compile the fp64 program based on nir options 2140- intel/debug: Add a debug flag to force software fp64 2141- intel/nir: Drop an unneeded lower_constant_initializers call 2142- glsl/nir: Add a shared helper for building float64 shaders 2143- glsl/nir: Inline functions in float64_funcs_to_nir 2144- nir/inline_functions: Break inlining into a builder helper 2145- nir/deref: Expose nir_opt_deref_impl 2146- nir/lower_doubles: Inline functions directly in lower_doubles 2147- intel/nir: Move 64-bit lowering later 2148- st/nir: Move 64-bit lowering later 2149- nir/builder: Emit better code for iadd/imul_imm 2150- nir/builder: Cast array indices in build_deref_follower 2151- nir/builder: Add a build_deref_array_imm helper 2152- intel/nir: Move lower_mem_access_bit_sizes to postprocess_nir 2153- anv/pipeline: Move lower_explicit_io much later 2154- nir: Add a pass for lowering IO back to vector when possible 2155- intel/nir: Vectorize all IO 2156- anv: Ignore VkRenderPassInputAttachementAspectCreateInfo 2157- nir/loop_unroll: Fix out-of-bounds access handling 2158- glsl/list: Add a list variant of insert_after 2159- glsl/lower_vector_derefs: Don't use a temporary for TCS outputs 2160- anv: Stop using VK_TRUE/FALSE 2161- anv/pass: Flag the need for a RT flush for resolve attachments 2162- anv: Only set 3DSTATE_PS::VectorMaskEnable on gen8+ 2163- nir/algebraic: Add a couple optimizations for iabs and ishr 2164- nir/validate: Only require bare types to match for copy_deref 2165- nir/validate: Allow 32-bit boolean load/store intrinsics 2166- compiler/types: Add a new is_interface C wrapper 2167- compiler/types: Add a C wrapper to get full struct field data 2168- compiler/types: Add helpers to get explicit types for standard 2169 layouts 2170- nir/deref: Consider COHERENT decorated var derefs as aliasing 2171- nir: Rename nir_address_format_vk_index_offset to not be vk 2172- nir/lower_io: Add a new buffer_array_length intrinsic and lowering 2173- glsl: Don't lower vector derefs for SSBOs, UBOs, and shared 2174- glsl/nir: Set explicit types on UBO/SSBO variables 2175- glsl/nir: Handle unlowered SSBO atomic and array_length intrinsics 2176- glsl/nir: Add a pass to lower UBO and SSBO access 2177- i965: Stop setting LowerBuferInterfaceBlocks 2178- st/mesa: Let NIR lower UBO and SSBO access when we have it 2179- nir/builder: Add a vector extract helper 2180- nir: Add a new pass to lower array dereferences on vectors 2181- intel/nir: Lower array-deref-of-vector UBO and SSBO loads 2182- anv: Implement VK_EXT_host_query_reset 2183- anv,radv: Implement VK_KHR_surface_capability_protected 2184- Revert "nir: const \`nir_call_instr::callee`" 2185- anv: Bump maxComputeWorkgroupInvocations 2186- nir: Constant values are per-column not per-component 2187- anv,radv,turnip: Lower TG4 offsets with nir_lower_tex 2188- spirv: Drop inline tg4 lowering 2189- nir/lower_io: Add a bounds-checked 64-bit global address format 2190- nir: Add a lowering pass for non-uniform resource access 2191- nir: Add texture sources and intrinsics for bindless 2192- nir: Add access flags to deref and SSBO atomics 2193- spirv: Handle the NonUniformEXT decoration 2194- Revert "anv/radv: release memory allocated by glsl types during 2195 spirv_to_nir" 2196- nir: Lock around validation fail shader dumping 2197- nir/algebraic: Drop some @bool specifiers 2198- nir/algebraic: Add some logical OR and AND patterns 2199- vc4: Prefer nir_src_comp_as_uint over nir_src_as_const_value 2200- nir/search: Search for all combinations of commutative ops 2201- nir: Get rid of nir_register::is_packed 2202- nir: Get rid of global registers 2203- intel/common: Add a MI command builder 2204- intel/common: Add unit tests for gen_mi_builder 2205- anv: Use gen_mi_builder for CmdDrawIndirectByteCount 2206- anv: Use gen_mi_builder for computing resolve predicates 2207- anv: Use gen_mi_builder for indirect draw parameters 2208- anv: Use gen_mi_builder for indirect dispatch 2209- anv: Use gen_mi_builder for conditional rendering 2210- anv: Use gen_mi_builder for queries 2211- anv: Move mi_memcpy and mi_memset to gen_mi_builder 2212- anv/cmd_buffer: Use gen_mi_sub instead of gen_mi_add with a negative 2213- intel/common: Support bigger right-shifts with mi_builder 2214- anv/pipeline: Fix MEDIA_VFE_STATE::PerThreadScratchSpace on gen7 2215- nir: Add a pass for selectively lowering variables to scratch space 2216- intel/nir: Take a nir_tex_instr and src index in brw_texture_offset 2217- nir/builder: Add a nir_imm_zero helper 2218- nir/print: Use nir_src_as_int for array indices 2219- nir/constant_folding: Get rid of a bit size switch statement 2220- spirv: Drop some unneeded bit size switch statements 2221- nir/load_const_to_scalar: Get rid of a bit size switch statement 2222- nir/validate: Require unused bits of nir_const_value to be zero 2223- vulkan: Update the XML and headers to 1.1.106 2224- anv: Update to use the new features struct names 2225- nir/algebraic: Move the template closer to the render function 2226- nir/algebraic: Use a cache to avoid re-emitting structs 2227- intel/mi_builder: Re-order an initializer 2228- intel/mi_builder: Disable mem_mem tests on IVB 2229- nir: Drop "struct" from some nir\_\* declarations 2230- nir: Rework nir_src_as_alu_instr to not take a pointer 2231- nir: Add a nir_src_as_intrinsic() helper 2232- anv: Re-sort the GetPhysicalDeviceFeatures2 switch statement 2233- anv: Drop some unneeded ANV_FROM_HANDLE for physical devices 2234- intel/fs: Account for live range lengths in spill costs 2235- anv: Make all VkDeviceMemory BOs resident permanently 2236- anv: Put image params in the descriptor set buffer on gen8 and 2237 earlier 2238- anv: Add a #define for the max binding table size 2239- anv/pipeline: Sort bindings by most used first 2240- anv/pipeline: Add skeleton support for spilling to bindless 2241- nir/lower_io: Expose some explicit I/O lowering helpers 2242- intel/nir: Re-run int64 lowering in postprocess_nir 2243- anv: Add a has_a64_buffer_access to anv_physical_device 2244- anv: Lower some SSBO operations in apply_pipeline_layout 2245- anv: Implement SSBOs bindings with GPU addresses in the descriptor BO 2246- anv: Implement VK_KHR_shader_atomic_int64 2247- intel,nir: Lower TXD with a bindless sampler 2248- intel/fs: Add support for bindless texture ops 2249- anv: Count the number of planes in each descriptor binding 2250- anv: Use write_image_view to initialize immutable samplers 2251- anv: Pass the plane into lower_tex_deref 2252- anv: Use bindless textures and samplers 2253- intel/fs: Add support for bindless image load/store/atomic 2254- anv: Use bindless handles for images 2255- anv: Put binding flags in descriptor set layouts 2256- anv: Implement VK_EXT_descriptor_indexing 2257- nir: Add helpers for getting the type of an address format 2258- anv/nir: Add a central helper for figuring out SSBO address formats 2259- anv: Ignore descriptor binding flags if bindingCount == 0 2260- anv: Rework the descriptor set layout create loop 2261- anv,radv: Update release notes for newly implemented extensiosn 2262- nir: Use the NIR_SRC_AS\_ macro to define nir_src_as_deref 2263- anv/descriptor_set: Unlink sets from the pool in set_destroy 2264- anv/descriptor_set: Destroy sets before pool finalization 2265- anv/descriptor_set: Only vma_heap_finish if we have a descriptor 2266 buffer 2267- anv/descriptor_set: Properly align descriptor buffer to a page 2268- anv: Better handle 32-byte alignment of descriptor set buffers 2269- anv/descriptor_set: Don't fully destroy sets in pool destroy/reset 2270- nir/algebraic: Optimize integer cast-of-cast 2271- util/bitset: Return an actual bool from test macros 2272- anv: Stop including POS in FS input limits 2273- anv,i965: Stop warning about incomplete gen11 support 2274- nir: Add a SSA type gathering pass 2275- intel/fs/ra: Only add dest interference to sources that exist 2276- intel/fs/ra: Stop adding RA interference to too many SENDS nodes 2277- anv: Emulate texture swizzle in the shader when needed 2278- anv: Stop forcing bindless for images 2279- anv: Only consider minSampleShading when sampleShadingEnable is set 2280- iris: Don't assume UBO indices are constant 2281- intel/fs,vec4: Use g0 as the header for MFENCE 2282- intel/fs: Do a stalling MFENCE in endInvocationInterlock() 2283- nir/dead_cf: Call instructions aren't dead 2284- nir/propagate_invariant: Don't add NULL vars to the hash table 2285 2286Jian-Hong Pan (1): 2287 2288- intel: Fix the description of Coffeelake pci-id 0x3E98 2289 2290Jiang, Sonny (1): 2291 2292- va: use a compute shader for the blit 2293 2294John Stultz (3): 2295 2296- mesa: android: freedreno: Fix build failure due to path change 2297- mesa: Makefile.sources: Add 2298 ir3_nir_lower_load_barycentric_at_sample/offset to Makefile.sources 2299- mesa: Makefile.sources: Add nir_lower_fb_read.c to Makefile.sources 2300 list 2301 2302Jon Turney (1): 2303 2304- meson: Force '.so' extension for DRI drivers 2305 2306Jonathan Marek (22): 2307 2308- nir: add missing vec opcodes in lower_bool_to_float 2309- freedreno: a2xx: fix fast clear 2310- freedreno: a2xx: don't write 4th vertex in mem2gmem 2311- freedreno: a2xx: add use_hw_binning function 2312- freedreno: a2xx: fix fast clear for some gmem configurations 2313- freedreno: a2xx: fix mipmapping for NPOT textures 2314- freedreno: use renderonly path for buffers allocated with modifiers 2315- freedreno: catch failing fd_blit and fallback to software blit 2316- mesa: add GL_AMD_compressed_ATC_texture support 2317- gallium: add ATC format support 2318- llvmpipe, softpipe: no support for ATC textures 2319- st/mesa: add ATC support 2320- freedreno: a3xx: add GL_AMD_compressed_ATC_texture support 2321- freedreno: a2xx: add GL_AMD_compressed_ATC_texture support 2322- svga: add new ATC formats to the format conversion table 2323- freedreno: a2xx: fix builtin blit program compilation 2324- freedreno: a2xx: disable PIPE_CAP_PACKED_UNIFORMS 2325- freedreno: a2xx: use nir_lower_io for TGSI shaders 2326- freedreno: a2xx: enable batch reordering 2327- freedreno: a2xx: same gmem2mem sequence for all tiles 2328- nir: improve convert_yuv_to_rgb 2329- freedreno/ir3: fix input ncomp for vertex shaders 2330 2331Jordan Justen (22): 2332 2333- iris: Set num_uniforms in bytes 2334- iris/compute: Set mask bits on PIPELINE_SELECT 2335- iris: Add IRIS_DIRTY_CONSTANTS_CS 2336- iris: Add iris_restore_compute_saved_bos 2337- iris/compute: Add MEDIA_STATE_FLUSH following WALKER 2338- iris/compute: Flush compute batches 2339- iris/compute: Get group counts from grid->grid 2340- iris/program: Don't try to push ubo ranges for compute 2341- iris/compute: Wait on compute batch when mapping 2342- iris/compute: Provide binding table entry for gl_NumWorkGroups 2343- iris/compute: Flush compute batch on memory-barriers 2344- iris/compute: Push subgroup-id 2345- iris/compute: Support indirect compute dispatch 2346- iris: Emit default L3 config for the render pipeline 2347- genxml/gen_bits_header.py: Use regex to strip no alphanum chars 2348- genxml: Remove extra space in gen4/45/5 field name 2349- iris: Add gitlab-ci build testing 2350- iris: Always use in-tree i915_drm.h 2351- nir: Add int64/doubles options into nir_shader_compiler_options 2352- intel/compiler: Move int64/doubles lowering options 2353- scons: Generate float64_glsl.h for glsl_to_nir fp64 lowering 2354- intel/genxml: Support base-16 in value & start fields in 2355 gen_sort_tags.py 2356 2357Jose Maria Casanova Crespo (4): 2358 2359- iris: Enable ARB_shader_draw_parameters support 2360- glsl: fix typos in comments "transfor" -> "transform" 2361- glsl: TCS outputs can not be transform feedback candidates on GLES 2362- iris: setup EdgeFlag Vertex Element when needed. 2363 2364José Fonseca (1): 2365 2366- scons: Workaround failures with MSVC when using SCons 3.0.[2-4]. 2367 2368Juan A. Suarez Romero (22): 2369 2370- anv/cmd_buffer: check for NULL framebuffer 2371- nir: move ALU instruction before the jump instruction 2372- nir: remove jump from two merging jump-ending blocks 2373- genxml: add missing field values for 3DSTATE_SF 2374- anv: advertise 8 subpixel precision bits 2375- nir/spirv: return after emitting a branch in block 2376- anv: destroy descriptor sets when pool gets reset 2377- nir: deref only for OpTypePointer 2378- anv: advertise 8 subtexel/mipmap precision bits 2379- nir/xfb: do not use bare interface type 2380- meson: Add dependency on genxml to anvil genfiles 2381- Revert "intel/compiler: split is_partial_write() into two variants" 2382- spirv: add missing SPV_EXT_descriptor_indexing capabilities 2383- radv: enable descriptor indexing capabilities 2384- anv: enable descriptor indexing capabilities 2385- Update version to 19.1.0-rc1 2386- Update version to 19.1.0-rc2 2387- cherry-ignore: radeonsi: update buffer descriptors in all contexts 2388 after buffer invalidation 2389- Update version to 19.1.0-rc3 2390- Update version to 19.1.0-rc4 2391- Update version to 19.1.0-rc5 2392- Update version to 19.1.0 2393 2394Julien Isorce (5): 2395 2396- gallium: add resource_get_info to pipe_screen 2397- radeonsi: implement resource_get_info 2398- st/va: properly set stride and offset in vlVaDeriveImage 2399- r600: implement resource_get_info 2400- st/va: check resource_get_info nullity in vlVaDeriveImage 2401 2402Józef Kucia (3): 2403 2404- mesa: Fix GL_NUM_DEVICE_UUIDS_EXT 2405- radv: Fix driverUUID 2406- radv: clear vertex bindings while resetting command buffer 2407 2408Karol Herbst (82): 2409 2410- nvc0/ir: replace cvt instructions with add to improve shader 2411 performance 2412- gk104/ir: Use the new rcp/rsq in library 2413- gm107/ir: add fp64 rcp 2414- gm107/ir: add fp64 rsq 2415- gallium: add PIPE_CAP_MAX_VARYINGS 2416- st/mesa: require RGBA2, RGB4, and RGBA4 to be renderable 2417- glsl_type: initialize offset and location to -1 for glsl_struct_field 2418- nir/opt_if: don't mark progress if nothing changes 2419- clover: update ICD table to support everything up to 2.2 2420- nir: replace magic numbers with M_PI 2421- nir/spirv: improve parsing of the memory model 2422- nir: add support for address bit sized system values 2423- nir/vtn: add support for SpvBuiltInGlobalLinearId 2424- nir/spirv: initial handling of OpenCL.std extension opcodes 2425- prog_to_nir: fix write from vps to FOG 2426- nvc0: print the shader type when dumping headers 2427- nv50/ir: move common converter code in base class 2428- nv50/ir: add lowering helper 2429- nouveau: add support for nir 2430- nouveau: fix nir and TGSI shader cache collision 2431- nv50/ir/nir: run some passes to make the conversion easier 2432- nv50/ir/nir: track defs and provide easy access functions 2433- nv50/ir/nir: add nir type helper functions 2434- nv50/ir/nir: run assignSlots 2435- nv50/ir/nir: add loadFrom and storeTo helpler 2436- nv50/ir/nir: parse NIR shader info 2437- nv50/ir/nir: implement nir_load_const_instr 2438- nv50/ir/nir: add skeleton for nir_intrinsic_instr 2439- nv50/ir/nir: implement nir_alu_instr handling 2440- nv50/ir/nir: implement nir_intrinsic_load_uniform 2441- nv50/ir/nir: implement nir_intrinsic_store_(per_vertex\_)output 2442- nv50/ir/nir: implement load_(interpolated\_)input/output 2443- nv50/ir/nir: implement intrinsic_discard(_if) 2444- nv50/ir/nir: implement loading system values 2445- nv50/ir/nir: implement nir_ssa_undef_instr 2446- nv50/ir/nir: implement nir_instr_type_tex 2447- nv50/ir/nir: add skeleton getOperation for intrinsics 2448- nv50/ir/nir: implement vote and ballot 2449- nv50/ir/nir: implement variable indexing 2450- nv50/ir/nir: implement geometry shader nir_intrinsics 2451- nv50/ir/nir: implement nir_intrinsic_load_ubo 2452- nv50/ir/nir: implement ssbo intrinsics 2453- nv50/ir/nir: implement images 2454- nv50/ir/nir: add memory barriers 2455- nv50/ir/nir: implement load_per_vertex_output 2456- nv50/ir/nir: implement intrinsic shader_clock 2457- nv50/ir/nir: handle user clip planes for each emitted vertex 2458- nv50ir/nir: move immediates before use 2459- glsl: add packed for struct types 2460- glsl: add cl_size and cl_alignment 2461- nir/lower_locals_to_regs: cast array index to 32 bit 2462- nir/spirv: handle kernel function parameters 2463- nir/spirv: support physical pointers 2464- nir: add support for gather offsets 2465- nv50/ir/nir: support gather offsets 2466- nir/lower_tex: Add support for tg4 offsets lowering 2467- nir/print: fix printing the image_array intrinsic index 2468- nir/validate: validate that tex deref sources are actually derefs 2469- v3d: prefer using nir_src_comp_as_int over nir_src_as_const_value 2470- panfrost/midgard: use nir_src_is_const and nir_src_as_uint 2471- glsl/standalone: add GLES3.1 and GLES3.2 compatibility 2472- nir: move brw_nir_rewrite_image_intrinsic into common code 2473- glsl_to_nir: handle bindless textures 2474- glsl/nir: fetch the type for images from the deref instruction 2475- glsl/nir: add support for lowering bindless images_derefs 2476- nv50/ir/nir: handle bindless texture 2477- nv50/ir/nir: add support for bindless images 2478- nvc0/nir: enable bindless texture 2479- lima: add bool parameter to type_size function 2480- amd/nir: some cleanups 2481- radv: use nir constant helpers 2482- intel/nir: use nir_src_is_const and nir_src_as_uint 2483- freedreno/ir3: use nir_src_as_uint in a few places 2484- lima: use nir_src_as_float 2485- nir/builder: Move nir_imm_vec2 from blorp into the builder 2486- nir/loop_analyze: use nir_const_value.b for boolean results, not u32 2487- spirv: reduce array size in vtn_handle_constant 2488- nir: make nir_const_value scalar 2489- vtn: handle bitcast with pointer src/dest 2490- nir: Add a nir_builder_alu variant which takes an array of components 2491- nir: Add nir_op_vec helper 2492- spirv/cl: support vload/vstore 2493 2494Kasireddy, Vivek (3): 2495 2496- nir/lower_tex: Add support for XYUV lowering 2497- dri: Add XYUV8888 format 2498- i965: Add support for sampling from XYUV images 2499 2500Kenneth Graunke (872): 2501 2502- st/mesa: Set pipe_image_view::shader_access in PBO readpixels. 2503- st/nir: Move varying setup code to a helper function. 2504- st/nir: Make new helpers for constructing built-in NIR shaders. 2505- st/mesa: Add a NIR version of the drawpixels/bitmap VS copy shader. 2506- st/mesa: Add NIR versions of the drawpixels Z/stencil fragment 2507 shaders. 2508- st/mesa: Add NIR versions of the clear shaders. 2509- st/mesa: Add a NIR version of the OES_draw_texture built-in shaders. 2510- st/mesa: Add NIR versions of the PBO upload/download shaders. 2511- program: Use u_bit_scan64 in prog_to_nir. 2512- program: Extend prog_to_nir handle system values. 2513- nir: Record info->fs.pixel_center_integer in lower_system_values 2514- compiler: Mark clip/cull distance arrays as compact before lowering. 2515- nir: Bail on clip/cull distance lowering if GLSL IR already did it. 2516- nir: Avoid clip/cull distance lowering multiple times. 2517- nir: Avoid splitting compact arrays into per-element variables. 2518- st/nir: Call nir_lower_clip_cull_distance_arrays(). 2519- gallium: Add a PIPE_CAP_NIR_COMPACT_ARRAYS capability bit. 2520- nouveau: Silence unhandled cap warnings 2521- st/mesa: Limit GL_MAX_[NATIVE\_]PROGRAM_PARAMETERS_ARB to 2048 2522- glsl: Allow gl_nir_lower_samplers*() without a gl_shader_program 2523- glsl: Don't look at sampler uniform storage for internal vars 2524- i965: Call nir_lower_samplers for ARB programs. 2525- st/nir: Pull sampler lowering into a helper function. 2526- st/nir: Lower sampler derefs for builtin shaders. 2527- st/nir: Use sampler derefs in built-in shaders. 2528- program: Make prog_to_nir create texture/sampler derefs. 2529- nir: Use sampler derefs in drawpixels and bitmap lowering. 2530- nir: Gather texture bitmasks in gl_nir_lower_samplers_as_deref. 2531- i965: Drop unnecessary 'and' with prog->SamplerUnits 2532- i965: Use info->textures_used instead of prog->SamplersUsed. 2533- mesa: Advertise EXT_float_blend in ES 3.0+ contexts. 2534- anv: Put MOCS in the correct location 2535- spirv: Eliminate dead input/output variables after translation. 2536- nir: Don't reassociate add/mul chains containing only constants 2537- compiler: Make is_64bit(GL_*) helper more broadly available 2538- mesa: Align doubles to a 64-bit starting boundary, even if packing. 2539- radeonsi: Go back to using llvm.pow intrinsic for nir_op_fpow 2540- st/mesa: Copy VP TGSI tokens if they exist, even for NIR shaders. 2541- nir: Don't forget if-uses in new nir_opt_dead_cf liveness check 2542- iris: Initial commit of a new 'iris' driver for Intel Gen8+ GPUs. 2543- iris: viewport state, sort of 2544- iris: port over batchbuffer updates 2545- iris: initial render state upload 2546- iris: packing with valgrind. 2547- iris: merge pack 2548- iris: initial gpu state, merges 2549- iris: RASTER + SF + some CLIP, fix DIRTY vs. NEW 2550- iris: scissors 2551- iris: SF_CLIP_VIEWPORT 2552- iris: Surfaces! 2553- iris: sampler views 2554- iris: stipples and vertex elements 2555- iris: framebuffers 2556- iris: don't segfault on !old_cso 2557- iris: fix SF_CL length 2558- iris: a bit of depth 2559- iris: some draw info, vbs, sample mask 2560- iris: fix crash - CSO binding can be NULL (when destroying context) 2561- iris: COLOR_CALC_STATE 2562- iris: sampler states 2563- iris: emit 3DSTATE_SAMPLER_STATE_POINTERS 2564- iris: basic push constant alloc 2565- iris: some program code 2566- iris: linear resources 2567- iris: maps 2568- iris: shader debug log 2569- iris: drop unused field 2570- iris: make an ice->render_batch field 2571- iris: disable execbuf for now 2572- iris: delete iris_pipe.c, shuffle code around 2573- iris: init the batch! 2574- iris: fix/rework line stipple 2575- iris: actually save VBs 2576- iris: msaa sample count packing problems 2577- iris: fix prim type 2578- iris: fix bogus index buffer reference 2579- iris: draw->restart_index is uninitialized if PR is not enabled 2580- iris: parse INTEL_DEBUG 2581- iris: reworks, FS compile pieces 2582- iris: import program cache code 2583- iris: do the FS...asserts because we don't lower uniforms yet 2584- iris: lower io 2585- iris: make iris_batch target a particular ring 2586- iris: kill iris_new_batch 2587- iris: move MAX defines to iris_batch.h 2588- iris: bit of SBA code 2589- iris: flag SBA updates when instruction BO changes 2590- iris: try and have an iris address 2591- iris: so, sba then. 2592- iris: reference VB BOs 2593- iris: VB addresses 2594- iris: DEBUG=bat 2595- iris: VB fixes 2596- iris: actually APPEND commands, not stomp over the top and never incr 2597- iris: actually flush the commands 2598- iris: actually advance forward when emitting commands 2599- iris: initialize dirty bits to ~0ull 2600- iris: hack to stop crashing on samplers for now 2601- iris: fix indentation 2602- iris: fix assert 2603- iris: fix VBs 2604- iris: vertex packet fixes 2605- iris: fix VF instancing length so we don't get garbage in batch 2606- iris: 3DPRIMITIVE fields 2607- iris: bind_state -> compute state 2608- iris: scissor slots 2609- iris: some shader bits 2610- iris: promote iris_program_cache_item to iris_compiled_shader 2611- iris: actually save derived state 2612- iris: emit shader packets 2613- iris: convert IRIS_DIRTY\_\* to #defines 2614- iris: don't forget about TE 2615- iris: reorganize commands to match brw 2616- iris: initial gpu state 2617- iris: WM. 2618- iris: index buffer BO 2619- iris: more comes from bits filled in 2620- iris: drop const from prog data parameters 2621- iris: softpin some things 2622- iris: use vtbl to avoid multiple symbols, fix state base address 2623- iris: fix SBA 2624- iris: move key pop to state module 2625- iris: bits of WM key 2626- iris: shuffle comments 2627- iris: no NEW_SBA 2628- iris: rewrite program cache to use u_upload_mgr 2629- iris: actually destroy the cache 2630- iris: actually softpin at an address 2631- iris: actually set KSP offsets 2632- iris: URB configs. 2633- iris: dummy constants 2634- iris: blend state 2635- iris: alpha testing in PSB 2636- iris: basic SBE code 2637- iris: warning fixes 2638- iris: fix silly unused batch with addr macro 2639- iris: render targets! 2640- iris: don't do samplers for disabled stages 2641- iris: smaller blend state 2642- iris: actually pin the instruction cache buffers 2643- iris: compctrl 2644- iris: more sketchy SBE 2645- iris: fix dmabuf retval comparisons 2646- iris: more SF CL VPs 2647- iris: catastrophic state pointer mistake 2648- iris: fix extents 2649- iris: write DISABLES are not write ENABLES...whoops 2650- iris: sample mask...not 0. 2651- iris: uniform bits...badly 2652- iris: warn if execbuf fails 2653- iris: NOOP pad batches correctly 2654- iris: decode batches if they fail to submit 2655- iris: enable a few more formats 2656- iris: set strides on transfers 2657- iris: stop adding 9 to our varyings 2658- iris: bufmgr updates. 2659- iris: some thinking about binding tables 2660- iris: Soft-pin the universe 2661- iris: fix icache memzone 2662- iris: dump gtt offset in dump_validation_list 2663- iris: Also set SUPPORTS_48B? Not sure if necessary. 2664- iris: more uploaders 2665- iris: rewrite to use memzones and not relocs 2666- iris: set EXEC_OBJECT_WRITE 2667- iris: include p_defines.h in iris_bufmgr.h 2668- iris: binders 2669- iris: hook up batch decoder 2670- iris: binder fixes 2671- iris: decoder fixes 2672- iris: update vb BO handling now that we have softpin 2673- iris: validation dumping improvements 2674- iris: canonicalize addresses. 2675- iris: delete more trash 2676- iris: allocate SURFACE_STATEs up front and stop streaming them 2677- iris: same treatment for sampler views 2678- iris: assemble SAMPLER_STATE table at bind time 2679- iris: fix a scissor bug 2680- iris: SBA once at context creation, not per batch 2681- iris: TES stash 2682- iris: isv freeing fixes 2683- iris: set sampler views 2684- iris: decoder fixes 2685- iris: better BT asserts 2686- iris: increase allocator alignment 2687- iris: fix index 2688- iris: port bug fix from i965 2689- iris: fixes from i965 2690- iris: fixes 2691- iris: crazy pipe control code 2692- iris: bo reuse 2693- iris: vma fixes - don't free binder address 2694- iris: vma - fix assert 2695- iris: better SBE 2696- iris: fix texturing! 2697- iris: Move get_command_space to iris_batch.c 2698- iris: Defines for base addresses rather than numbers everywhere 2699- iris: pull in newer comments 2700- iris: copy over i965's cache tracking 2701- iris: move bo_offset_from_sba 2702- iris: bits of blorp code 2703- iris: more blitting code to make readpixels work 2704- iris: drop bogus binder free 2705- iris: fix sampler view crashes 2706- iris: more blorp 2707- iris: fix blorp prog data crashes 2708- iris: add INTEL_DEBUG=reemit 2709- iris: drop the 48b printout, we never use anything else 2710- iris: hacky flushing for now 2711- iris: linear staging buffers - fast CPU access... 2712- iris: make blorp pin the binder 2713- iris: blorp URB 2714- iris: no more drawing rectangle in blorp 2715- iris: assert surf init 2716- iris: some depth stuff :( 2717- iris: bump GL version to 4.2 2718- iris: uniforms for VS 2719- iris: proper length for VE packet? 2720- iris: proper # of uniforms 2721- iris: properly reject formats, fixes RGB32 rendering with texture 2722 float 2723- iris: blorp bug fixes 2724- iris: delete growing code and just die for now 2725- iris: just turn batch reset_and_clear_caches into reset 2726- iris: chaining not growing 2727- iris: caps 2728- iris: fix batch chaining... 2729- iris: fix decoding and undo testing code 2730- iris: Lower the max number of decoded VBO lines 2731- iris: fix whitespace 2732- iris: fix 3DSTATE_VERTEX_ELEMENTS length 2733- iris: more depth stuffs... 2734- iris: fix VF INSTANCING length 2735- iris: util_copy_framebuffer_state (ported from Rob's v3d patches) 2736- iris: transfers 2737- iris: flush always 2738- iris: maybe slightly less boats uniforms 2739- iris: fix constant packet length to match i965 2740- iris: better ubo handling 2741- iris: completely rewrite binder 2742- iris: have more than one const_offset 2743- iris: make surface states for cbufs 2744- iris: fill out pull constant buffers 2745- iris: fix pull bufs that aren't the first user upload 2746- iris: use u_transfer helpers for now 2747- iris: better VFI 2748- iris: fix release builds 2749- iris: drop assert for now 2750- iris: disable \__gen_validate_value in release mode 2751- iris: allow mapped buffers during execution (faster) 2752- iris: comment about reemitting and flushing 2753- iris: state cleaning 2754- iris: untested index buffer upload 2755- iris: delete some pointless STATIC_ASSERTS 2756- iris: untested SAMPLER_STATE pin BO fix 2757- iris: put back the always flush - fixes some things :( 2758- iris: save pointers to streamed state resources 2759- iris: fix the validation list on new batches 2760- iris: flag DIRTY_WM properly 2761- iris: bindings dirty tracking 2762- iris: some dirty fixes 2763- iris: clear dirty 2764- iris: plug leaks 2765- iris: more leak fixes 2766- iris: pc fixes 2767- iris: remove 4 bytes of padding in iris_compiled_shader 2768- iris: rzalloc iris_compiled_shader so memcmp works even if padding 2769 creeps in 2770- iris: don't leak sampler state table resources 2771- iris: don't leak keyboxes when searching for an existing program 2772- iris: indentation 2773- iris: use pipe resources not direct BOs 2774- iris: clean up some warnings so I can see through the noise 2775- iris: print binder utilization in INTEL_DEBUG=submit 2776- iris: redo VB CSO a bit 2777- iris: print refcounts in INTEL_DEBUG=submit 2778- iris: support signed vertex buffer offsets 2779- iris: fix major refcounting bug with resources 2780- iris: fix caps so tests run again 2781- iris: avoid crashing on unbound constant resources 2782- iris: emit 3DSTATE_SBE_SWIZ 2783- iris: max VP index 2784- iris: fix viewport counts and settings 2785- iris: fix num viewports to be based on programs 2786- iris: fix VP iteration 2787- iris: scissor count fixes 2788- iris: actually init num_viewports 2789- iris: print second batch size separately 2790- iris: don't always flush 2791- iris: Handle batch submission failure "better" 2792- iris: bad inherited comments 2793- iris: colorize batchbuffer failures to make them stand out 2794- iris: iris - fix QWord aligned endings after batch chaining rework 2795- iris: tidy comments about mirroring modes 2796- iris: Disable unsupported mirror clamp modes 2797- iris: fix fragcoord ytransform 2798- iris: better boxing on maps 2799- iris: clears 2800- iris: rework DEBUG_REEMIT 2801- iris: shader dirty bits 2802- iris: clear fix 2803- iris: fall back to u_generate_mipmap 2804- iris: implement copy image 2805- iris: lightmodel flat 2806- iris: maybe-flush before blorp operations 2807- iris: fix provoking vertex ordering 2808- iris: larger polygon offset 2809- iris: TES uniform fixes 2810- iris: geometry shader support 2811- iris: don't emit garbage 3DSTATE_VERTEX_BUFFERS when there aren't any 2812- iris: fix 3DSTATE_VERTEX_ELEMENTS / VF_INSTANCING for 0 elements 2813- iris: fix GS dispatch mode 2814- iris: depth clears 2815- iris: null surface for unbound textures 2816- iris: state ref tuple 2817- iris: don't include binder in surface VMA range 2818- iris: border color memory zone :( 2819- iris: implement border color, fix other sampler nonsense 2820- iris: dead pointer 2821- iris: just malloc one iris_genx_state instead of a bunch of oddball 2822 pieces 2823- iris: SBE change stash 2824- iris: fix zoffset asserts with 2DArray/Cube 2825- iris: rename map->stride 2826- iris: actually set cube bit properly 2827- iris: keep DISCARD_RANGE 2828- iris: actually handle array layers in blits 2829- iris: comment out l/a/i/la 2830- iris: fix clip flagging on fb changes 2831- iris: fix depth bounds clamp enables 2832- iris: don't crash on shader perf logs 2833- iris: slab allocate transfers 2834- iris: rearrange iris_resource.h 2835- iris: Implement 3DSTATE_SO_DECL_LIST 2836- iris: SO buffers 2837- iris: streamout 2838- iris: set even if no outputs 2839- iris: bother setting program_string_id... 2840- iris: fix SO_DECL_LIST 2841- iris: actually pin the buffers 2842- iris: fix sample mask for MSAA-off 2843- iris: disable 6x MSAA support 2844- iris: multislice transfer maps 2845- iris: fix CC_VIEWPORT 2846- iris: draw indirect support? 2847- iris: save query type 2848- iris: bits of multisample program key 2849- iris: s/hwcso/state/g 2850- iris: bind state helper function 2851- iris: NOS mechanics 2852- iris: record FS NOS 2853- iris: fix crash 2854- iris: fix sampler views of TBOs 2855- iris: fix texture buffer stride 2856- iris: TES program key inputs 2857- iris: compile a TCS...don't bother with passthrough yet 2858- iris: don't emit SO_BUFFERS and SO_DECL_LIST unless streamout is 2859 enabled 2860- iris: vertex ID, instance ID 2861- iris: fix SGVS when there are no valid vertex elements 2862- iris: fill out MAX_PATCH_VERTICES 2863- iris: assert about passthrough shaders to make this easier to detect 2864- iris: fix EmitNoIndirect 2865- iris: fix Z24 2866- iris: reemit blend state for alpha test function changes 2867- iris: point sprite enables 2868- iris: hack around samples confusion 2869- iris: fix blorp filters 2870- iris: expose more things that we already support 2871- iris: fix msaa flipping filters 2872- iris: export get_shader_info 2873- iris: implement set_shader_buffers 2874- iris: emit binding table for atomic counters and SSBOs 2875- iris: shorten loop 2876- iris: unbind compiled shaders if none are present 2877- iris: fix TBO alignment to match 965 2878- iris: enable SSBOs 2879- iris: fix SSBO indexing 2880- iris: fix for disabling ssbos 2881- iris: update bindings when changing programs 2882- iris: drop unused bo parameter 2883- iris: implement texture/memory barriers 2884- iris: Don't reserve new binding table section unless things are dirty 2885- iris: update a todo comment 2886- iris: BIG OL' HACK for UBO updates 2887- iris: enable texture gather 2888- iris: Avoid croaking when trying to create FBO surfaces with bad 2889 formats 2890- iris: fix GS output component limit 2891- iris: drop pipe_shader_state 2892- iris: fix sample mask 2893- iris: cube arrays are cubes too 2894- iris: we don't support textureGatherOffsets, need it lowered 2895- iris: add minor comments 2896- iris: comment everything 2897- iris: sync bugfixes from brw_bufmgr 2898- iris: remember to set bo->userptr 2899- iris: rename ring to engine 2900- iris: simplify batch len qword alignment 2901- iris: get angry about execbuf failures 2902- iris: fill out more caps 2903- iris: depth or stencil fixes 2904- iris: clear stencil 2905- iris: actually emit stencil packets 2906- iris: allow S8 as a stencil format 2907- iris: WTF transfers 2908- iris: use u_transfer_helper for depth stencil packing/unpacking 2909- iris: drop stencil handling now that u_transfer_helper does it 2910- iris: refcounting, who needs it? 2911- iris: actually do stencil blits 2912- iris: say no to more formats 2913- iris: deal with Marek's new MSAA caps 2914- iris: we can do multisample Z resolves 2915- iris: Convert RGBX to RGBA for rendering. 2916- iris: disallow RGB32 formats too 2917- iris: Fix tiled memcpy for cubes...and for array slices 2918- iris: blorp blit multiple slices 2919- iris: assert depth is 1 in resource_copy_region 2920- iris: call maybe_flush for each blorp operation 2921- iris: implement ARB_clear_texture 2922- iris: last VUE map NOS, handle > 16 FS inputs 2923- iris: drop dead assignments 2924- iris: drop pwrite 2925- iris: port non-bucket alignment bugfix 2926- iris: don't emit SBE all the time 2927- iris: rename pipe to base 2928- iris: Drop bogus sampler state saving 2929- iris: move iris_shader_state from ice->shaders.state to 2930 ice->state.shaders 2931- iris: Move things to iris_shader_state 2932- iris: Move iris_sampler_view declaration to iris_resource.h 2933- iris: track depth/stencil writes enabled 2934- iris: use consistent copyright formatting 2935- iris: Move cache tracking to iris_resolve.c 2936- iris: proper cache tracking 2937- iris: precompute hashes for cache tracking 2938- iris: Reduce binder alignment from 64 to 32 2939- iris: reenable R32G32B32 texture buffers 2940- iris: z_res -> s_res 2941- iris: implement get_sample_position 2942- iris: fix line-aa-width 2943- iris: try to hack around binder issue 2944- iris: fix sampler state setting 2945- iris: big old hack for tex-miplevel-selection 2946- iris: use linear for 1D textures 2947- iris: handle level/layer in direct maps 2948- iris: fix crash when binding optional shader for the first time 2949- iris: Skip primitive ID overrides if the shader wrote a custom value 2950- iris: fix blend state memcpy 2951- iris: new caps 2952- iris: use Eric's new caps helper 2953- iris: Allow inlining of require/get_command_space 2954- iris: skip over whole function if dirty == 0 2955- iris: don't unconditionally emit 3DSTATE_VF / 3DSTATE_VF_TOPOLOGY 2956- iris: fix constant buffer 0 to be absolute 2957- iris: set EXEC_OBJECT_CAPTURE on all driver internal buffers 2958- iris: fix null FB and unbound tex surface state addresses 2959- iris: Support multiple binder BOs, update Surface State Base Address 2960- iris: fix SO offset writes for multiple streams 2961- iris: update comments for multibinder 2962- iris: move binder pinning outside the dirty == 0 check 2963- iris: re-pin binding table contents if we didn't re-emit them 2964- iris: enable ARB_enhanced_layouts 2965- iris: refactor LRIs in context setup 2966- iris: initialize "don't suck" bits, as Ben likes to call them 2967- iris: totally untested icelake support 2968- iris: refactor program CSO stuff 2969- iris: silence const warning 2970- iris: fix context restore of 3DSTATE_CONSTANT ranges 2971- iris: properly re-pin stencil buffers 2972- iris: delete bogus comment 2973- iris: inherit the index buffer properly 2974- iris: use 0 for TCS passthrough program string ID 2975- iris: rw_bo for pipe controls 2976- iris: LRM/SRM/SDI hooks 2977- iris: initial query code 2978- iris: gen10+ workarounds and break fix 2979- iris: results write 2980- iris: flush batch when asking for result via QBO 2981- iris: fix random failures via CS stall...but why? 2982- iris: gpr0 to bool 2983- iris: play chicken with timer queries for now 2984- iris: pipeline stats 2985- iris: primitives generated query support 2986- iris: drop explicit pinning 2987- iris: timestamps 2988- iris: ...and SO prims emitted queries 2989- iris: glGet timestamps, more correct timestamps 2990- iris: Need to \| 1 when asking for timestamps 2991- iris: 36-bit overflow fixes 2992- iris: early return properly 2993- iris: better query file comment 2994- iris: magic number 36 -> #define 2995- iris: Enable ARB_shader_vote 2996- iris: just mark snapshots_landed from the CPU 2997- iris: drop a bunch of pipe_sampler_state stuff we don't need 2998- iris: vma_free bo->size, not bo_size 2999- iris: don't mark contains_draw = false when chaining batches 3000- iris: fix Z32_S8 depth sampling 3001- iris: stencil texturing 3002- iris: force persample interp cap 3003- iris: pipe to scs -> iris_pipe.h 3004- iris: inline stage_from_pipe to avoid unused warnings 3005- iris: add gen11 to genX_call 3006- iris: Allow PIPE_CONTROL with Stall at Scoreboard and RT flush 3007- iris: rework format translation apis 3008- iris: Use R/RG instead of I/L/A when sampling 3009- iris: enable I/L formats 3010- iris: X32_S8X24 :/ 3011- iris: set the binding table size 3012- iris: lower storage image derefs 3013- iris: implement set_shader_images hook 3014- iris: bother with BTIs 3015- iris: set image access correctly 3016- iris: actually set image access 3017- iris: null for non-existent cbufs 3018- iris: move images next to textures in binding table 3019- iris: advertise GL_ARB_shader_texture_image_samples 3020- iris: Enable fb fetch 3021- iris: initial compute caps 3022- iris: yes 3023- iris: drop dead format //'s 3024- iris: drop XXX's about swizzling 3025- iris: little bits of compute basics 3026- iris: drop XXX that Jordan handled 3027- iris: drop unnecessary #ifdefs 3028- iris: leave XXX about unnecessary binding table uploads 3029- iris: bail if SLM is needed 3030- iris: fix whitespace 3031- iris: XXX for compute state tracking :/ 3032- iris: rewrite grid surface handling 3033- iris: better dirty checking 3034- iris: don't let render/compute contexts stomp each other's dirty bits 3035- iris: hack to avoid memorybarriers out the wazoo 3036- iris: do PIPELINE_SELECT for render engine, add flushes, GLK hacks 3037- iris: fix SBA flushing by refactoring code 3038- iris: try and avoid pointless compute submissions 3039- iris: fix UBOs with bindings that have an offset 3040- iris: flag CC_VIEWPORT when changing num viewports 3041- iris: fix SF_CLIP_VIEWPORT array indexing with multiple VPs 3042- iris: Fix texture buffer / image buffer sizes. 3043- iris: Clamp UBO and SSBO access to the actual BO size, for safety 3044- iris: Move snapshots_landed to the front. 3045- iris: Fix off by one in scissoring, empty scissors, default scissors 3046- iris: Fall back to 1x1x1 null surface if no framebuffer supplied 3047- iris: SO_DECL_LIST fix 3048- iris: Fix refcounting of grid surface 3049- iris: delete dead code 3050- iris: fix overhead regression from "don't stomp each other's dirty 3051 bits" 3052- iris: allow binding a null vertex buffer 3053- iris: Flag constants dirty on program changes 3054- iris: Disable a PIPE_CONTROL workaround on Icelake 3055- iris: Enable ARB_shader_stencil_export 3056- iris: Enable A8/A16_UNORM in an inefficient manner 3057- iris: Drop B5G5R5X1 support 3058- iris: Use at least 1x1 size for null FB surface state. 3059- iris: Cross-link iris_batches so they can potentially flush each 3060 other 3061- iris: cross batch flushing 3062- iris: Don't leak the compute batch 3063- iris: Actually create/destroy HW contexts 3064- iris: Enable msaa_map transfer helpers 3065- iris: tidy more warnings 3066- iris: implement scratch space! 3067- iris: Fix MSAA smooth points 3068- iris: Fix TextureBarrier 3069- iris: Fix multiple RTs with non-independent blending 3070- iris: partial set_query_active_state 3071- iris: Print the batch name when decoding 3072- iris: Clone the NIR 3073- iris: Defer cbuf0 upload to draw time 3074- iris: drop unnecessary param[] setup from iris_setup_uniforms 3075- iris: add param domain defines 3076- iris: fill out params array with built-ins, like clip planes 3077- iris: only bother with params if there are any... 3078- iris: lower user clip planes 3079- iris: hook up key stuff for clip plane lowering 3080- iris: fix system value remapping 3081- iris: dodge backend UCP lowering 3082- iris: bypass params and do it ourselves 3083- iris: actually upload clip planes. 3084- iris: fix num clip plane consts 3085- iris: fix more uniform setup 3086- iris: drop iris_setup_push_uniform_range 3087- iris: enable push constants if we have sysvals but no uniforms 3088- iris: regather info so we get CLIP_DIST slots, not CLIP_VERTEX 3089- iris: don't support pull constants. 3090- iris: don't trip on param asserts 3091- iris: drop param stuffs 3092- iris: don't forget to upload CS consts 3093- iris: fix sysval only binding tables 3094- iris: only clip lower if there's something to clip against 3095- iris: leave another TODO 3096- iris: Fix SourceAlphaBlendFactor 3097- iris: "Fix" transfer maps of buffers 3098- iris: Fix independent alpha blending. 3099- iris: more TODO 3100- iris: scissored and mirrored blits 3101- iris: more todo notes 3102- iris: Fix TCS/TES slot unification 3103- iris: properly pin stencil buffers 3104- iris: Fix SLM 3105- iris: Use iris_use_pinned_bo rather than add_exec_bo directly 3106- iris: Combine iris_use_pinned_bo and add_exec_bo 3107- iris: Avoid cross-batch synchronization on read/reads 3108- iris: Avoid synchronizing due to the workaround BO 3109- iris: replace vestiges of fence fds with newer exec_fence API 3110- iris: Drop vestiges of throttling code 3111- iris: Hang on to the last batch's sync-point, so we can wait on it 3112- iris: Add wait fences to properly sync between render/compute 3113- iris: leave a TODO 3114- iris: flush the compute batch too if border pool is redone 3115- iris: put render batch first in fence code 3116- iris: Put batches in an array 3117- iris: PIPE_CONTROL workarounds for GPGPU mode 3118- iris: RT flush for memorybarrier with texture bit 3119- iris: update comment 3120- iris: Enable ctx->Const.UseSTD430AsDefaultPacking 3121- iris: Lie about indirects 3122- iris: Fix buffer -> buffer copy_region 3123- iris: Fix VIEWPORT/LAYER in stream output info 3124- iris: Do the 48-bit vertex buffer address invalidation workaround 3125- iris: drop long dead XXX comment 3126- iris: Track a binding history for buffer resources 3127- iris: add iris_flush_and_dirty_for_history 3128- iris: Flush for history at various moments 3129- iris: Re-pin even if nothing is dirty 3130- iris: fix prototype warning 3131- iris: export iris_upload_shader 3132- iris: fix comment location 3133- iris: Use wrappers for create_xs_state rather than a switch statement 3134- iris: rework program cache interface 3135- iris: Enable precompiles 3136- iris: Use program's num textures not the state tracker's bound 3137- iris: drop pull constant binding table entry 3138- iris: add assertions about binding table starts 3139- iris: add an extra BT assert from Chris Wilson 3140- iris: actually flush for storage images 3141- iris: fix some SO overflow query bugs and tidy the code a bit 3142- iris: drop key_size_for_cache 3143- iris: for BLORP, only use the predicate enable bit when USE_BIT 3144- iris: check query first 3145- iris: fix conditional compute, don't stomp predicate for pipelined 3146 queries 3147- iris: Rework tiling/modifiers handling 3148- iris: Fix failed to compile TCS message 3149- iris: Destroy transfer helper on screen teardown 3150- iris: Destroy the border color pool 3151- iris: Unref unbound_tex resource 3152- iris: Fix IRIS_MEMZONE_COUNT to exclude the border color pool 3153- iris: Destroy the bufmgr 3154- iris: Stop leaking iris_uncompiled_shaders like mad 3155- iris: move some non-buffer case code in a bit 3156- iris: Don't bother considering if the underlying surface is a cube 3157- iris: fix alpha channel for RGB BC1 formats 3158- iris: fix dma buf import strides 3159- iris: CS stall for stream out -> VB 3160- iris: make clipper statistics dynamic 3161- iris: reject all clipping when we can't use streamout render disabled 3162- iris: omask can kill 3163- iris: reemit SBE when sprite coord origin changes 3164- iris: re-pin inherited streamout buffers 3165- iris: Fix NOS mechanism 3166- iris: fix overhead regression from flushing for storage images 3167- iris: fix set_sampler_views to not unbind, be better about bounds 3168- iris: Fix set_sampler_views with start > 0 3169- iris: Replace num_textures etc with a bitmask we can scan 3170- iris: Drop continues in resolve 3171- iris: Fix clear dimensions 3172- iris: Clamp viewport extents to the framebuffer dimensions 3173- iris: Enable guardband clipping 3174- iris: Fix primitive generated query active flag 3175- iris: Always do rasterizer discard in clipper 3176- iris: override alpha to one src1 blend factors 3177- iris: handle PatchVerticesIn as a system value. 3178- iris: rewrite set_vertex_buffer and VB handling 3179- iris: Reorder LRR parameters to have dst first. 3180- iris: Add \_MI_ALU helpers that don't paste 3181- iris: Don't bother packing 3DSTATE_SO_BUFFER at create time 3182- iris: Move iris_stream_output_target def to iris_context.h 3183- iris: only get space for one offset in stream output targets 3184- iris: Implement DrawTransformFeedback() 3185- iris: drop unnecessary genx->streamout field 3186- iris: Fix for PIPE_CAP_SIGNED_VERTEX_BUFFER_OFFSET 3187- iris: Fix the prototype for iris_bo_alloc_tiled 3188- iris: don't print the pointer in INTEL_DEBUG=submit 3189- iris: Use a surface state fill helper 3190- iris: Make a alloc_surface_state helper 3191- iris: whitespace fixes 3192- iris: Track blend enables, save outbound for resolve code 3193- iris: always pin the binder...in the compute context, too. 3194- iris: delete finished comments 3195- iris: pin and re-pin the scratch BO 3196- iris: more dead comments 3197- iris: only mark depth/stencil as writable if writes are actually 3198 enabled 3199- iris: better MOCS 3200- iris: Fix scratch space allocation on Icelake. 3201- iris: Only resolve inputs for actual shader stages 3202- iris: Add a more long term TODO about timebase scaling 3203- iris: Fix compute scratch pinning 3204- iris: Delete bogus comment about cube array counting. 3205- iris: Fix framebuffer layer count 3206- iris: Don't enable push constants just because there are system 3207 values 3208- iris: Don't make duplicate system values 3209- iris: Fill out brw_image_params for storage images on Broadwell 3210- iris: Fix surface states for Gen8 lowered-to-untype images 3211- iris: Leave a comment about why Broadwell images are broken 3212- iris: Implement multi-slice copy_region 3213- iris: Flush the render cache in flush_and_dirty_for_history 3214- iris: Handle PIPE_TRANSFER_DISCARD_WHOLE_RESOURCE somewhat 3215- iris: Don't check other batches for our batch BO 3216- iris: Drop a dead comment 3217- iris: Delete genx->bound_vertex_buffers 3218- iris: Fix Broadwell WaDividePSInvocationCountBy4 3219- iris: Use new PIPE_STAT_QUERY enums rather than hardcoded numbers. 3220- iris: Switch to the new PIPELINE_STATISTICS_QUERY_SINGLE capability 3221- iris: fail to create screen for older unsupported HW 3222- iris: Allow sample mask of 0 3223- iris: Don't enable smooth points when point sprites are enabled 3224- iris: Assert about blits with color masking 3225- iris: Pay attention to blit masks 3226- iris: CS stall on VF cache invalidate workarounds 3227- iris: Fix SO issue with INTEL_DEBUG=reemit, set fewer bits 3228- iris: Don't whack SO dirty bits when finishing a BLORP op 3229- iris: Fix memzone_for_address for the surface and binder zones 3230- iris: Do binder address allocations per-context, not globally. 3231- iris: Zero the compute predicate when changing the render condition 3232- iris: Remap stream output indexes back to VARYING_SLOT_*. 3233- iris: Enable PIPE_CAP_COMPACT_ARRAYS 3234- iris: Drop comment about ISP_DIS 3235- iris: Drop dead state_size hash table 3236- iris: Unreference some more things on state module teardown 3237- iris: minor tidying 3238- iris: Fix bug in bound vertex buffer tracking 3239- iris: Implement ALT mode for ARB_{vertex,fragment}_shader 3240- iris: Add a timeout_nsec parameter, rename check_syncpt to 3241 wait_syncpt 3242- iris: Fix accidental busy-looping in query waits 3243- iris: Use READ_ONCE and WRITE_ONCE for snapshots_landed 3244- iris: Make a iris_batch_reference_signal_syncpt helper function. 3245- iris: Add PIPE_CAP_MAX_VARYINGS 3246- iris: rework num textures to util_lastbit 3247- iris: Stop chopping off the first nine characters of the renderer 3248 string 3249- iris: Drop XXX about alpha testing 3250- iris: Set 3DSTATE_WM::ForceThreadDispatchEnable 3251- iris: Set HasWriteableRT correctly 3252- iris: Drop XXX about checking for swizzling 3253- iris: Move create and bind driver hooks to the end of iris_program.c 3254- iris: Make an IRIS_MAX_MIPLEVELS define 3255- iris: Simplify iris_get_depth_stencil_resources 3256- iris: Add missing depth cache flushes 3257- iris: Always emit at least one BLEND_STATE 3258- iris: Add iris_resource fields for aux surfaces 3259- iris: Fill out res->aux.possible_usages 3260- iris: Fill out SURFACE_STATE entries for each possible aux usage 3261- iris: create aux surface if needed 3262- iris: Initial import of resolve code 3263- iris: blorp using resolve hooks 3264- iris: add some draw resolve hooks 3265- iris: actually use the multiple surf states for aux modes 3266- iris: try to fix copyimage vs copybuffers 3267- iris: be sure to skip buffers in resolve code 3268- iris: resolve before transfer maps 3269- iris: pin the buffers 3270- iris: store modifier info in res 3271- iris: Make blit code use actual aux usages 3272- iris: consider framebuffer parameter for aux usages 3273- iris: Resolves for compute 3274- iris: disable aux for external things 3275- iris: some initial HiZ bits 3276- iris: don't use hiz for MSAA buffers 3277- iris: Set program key fields for MCS 3278- iris: make surface states for CCS_D too 3279- iris: do flush for buffers still 3280- iris: Allow disabling aux via INTEL_DEBUG options 3281- iris: Fix aux usage in render resolve code 3282- iris: Only resolve compute resources for compute shaders 3283- iris: Enable auxiliary buffer support 3284- iris: Enable -msse2 and -mstackrealign 3285- Revert "iris: Enable auxiliary buffer support" 3286- vulkan: Fix 32-bit build for the new overlay layer 3287- mesa: Fix RGBBuffers for renderbuffers with sized internal formats 3288- iris: Drop RGBX -> RGBA for storage image usages 3289- iris: Properly allow rendering to RGBX formats. 3290- i965: Implement threaded GL support. 3291- tgsi_to_nir: use sampler variables and derefs 3292- iris: Fix MOCS for blits and clears 3293- isl: Add a swizzle parameter to isl_buffer_fill_state() 3294- iris: Plumb through ISL_SWIZZLE_IDENTITY in buffer surface emitters 3295- iris: Defer uploading sampler state tables until draw time 3296- iris: Properly support alpha and luminance-alpha formats 3297- iris: Drop PIPE_CAP_BUFFER_SAMPLER_VIEW_RGBA_ONLY 3298- iris: Spruce up "are we using this engine?" checks for flushing 3299- iris: Export a copy_region helper that doesn't flush 3300- iris: Use copy_region and staging resources to avoid transfer stalls 3301- Revert MR 369 (Fix extract_i8 and extract_u8 for 64-bit integers) 3302- iris: Fix backface stencil write condition 3303- iris: Rework default tessellation level uploads 3304- iris: Fix TES gl_PatchVerticesIn handling. 3305- iris: Move depth/stencil flushes so they actually do something 3306- iris: Refactor depth/stencil buffer pinning into a helper. 3307- iris: Fix write enable in pinning of depth/stencil resources 3308- i965: Move some genX infrastructure to genX_boilerplate.h. 3309- i965: Rename ISP_DIS to INDIRECT_STATE_POINTERS_DISABLE. 3310- i965: Use genxml for emitting PIPE_CONTROL. 3311- i965: Reimplement all the PIPE_CONTROL rules. 3312- intel/fs: Fix opt_peephole_csel to not throw away saturates. 3313- iris: Don't mutate box in transfer map code 3314- iris: Don't flush the batch for unsynchronized mappings 3315- iris: Slightly better bounds on buffer sizes 3316- gallium: Add PIPE_BARRIER_UPDATE_BUFFER and UPDATE_TEXTURE bits. 3317- nvc0: Skip new update barrier bits 3318- nir: Record non-vector/scalar varyings as unmovable when compacting 3319- iris: Fix util_vma_heap_init size for IRIS_MEMZONE_SHADER 3320- iris: Skip input resolve handling if bindings haven't changed 3321- iris: Skip framebuffer resolve tracking if framebuffer isn't dirty 3322- iris: Skip resolves and flushes altogether if unnecessary 3323- iris: Fix batch chaining map_next increment. 3324- iris: Actually advertise some modifiers 3325- st/nir: Free the GLSL IR after linking. 3326- st/mesa: Fix blitting from GL_DEPTH_STENCIL to GL_STENCIL_INDEX 3327- iris: Fix blits with S8_UINT destination 3328- iris: Print the memzone name when allocating BOs with INTEL_DEBUG=buf 3329- iris: Save/restore MI_PREDICATE_RESULT, not MI_PREDICATE_DATA. 3330- iris: Silence unused variable warnings in release mode 3331- gallium/util: Add const to u_range_intersect 3332- iris: Actually pin the scratch BO. 3333- glsl: Set location on structure-split sampler uniform variables 3334- intel: Emit 3DSTATE_VF_STATISTICS dynamically 3335- iris: Actually mark blorp_copy_buffer destinations as written. 3336- iris: Preserve all PIPE_TRANSFER flags in xfer->usage 3337- iris: Fix FLUSH_EXPLICIT handling with staging buffers. 3338- iris: Make shader_perf_log print to stderr if INTEL_DEBUG=perf is set 3339- i965: Move program key debugging to the compiler. 3340- iris: Print the reason for shader recompiles. 3341- iris: Move iris_debug_recompile calls before uploading. 3342- iris: Change vendor and renderer strings 3343- iris: Add texture cache flushing hacks for blit and 3344 resource_copy_region 3345- iris: Be less aggressive at postdraw work skipping 3346- iris: Add mechanism for iris-specific driconf options 3347- iris: Enable the dual_color_blend_by_location driconf option. 3348- iris: Track bound and writable SSBOs 3349- Revert "glsl: Set location on structure-split sampler uniform 3350 variables" 3351- i965: Ignore uniform storage for samplers or images, use binding info 3352- i965: Tidy bogus indentation left by previous commit 3353- iris: Mark constants dirty on transfer unmap even if no flushes occur 3354- iris: Track bound constant buffers 3355- iris: Rework UBOs and SSBOs to use pipe_shader_buffer 3356- iris: Rework image views to store pipe_image_view. 3357- iris: Make a gl_shader_stage -> pipe_shader_stage helper function 3358- iris: Make memzone_for_address non-static 3359- iris: Replace buffer backing storage and rebind to update addresses. 3360- iris: Make a resource_is_busy() helper 3361- iris: Track valid data range and infer unsynchronized mappings. 3362- iris: Make some offset math helpers take a const isl_surf pointer 3363- iris: Fix DrawTransformFeedback math when there's a buffer offset 3364- iris: Prefer staging blits when destination supports CCS_E. 3365- iris: Actually put Mesa in GL_RENDERER string 3366- iris: Split iris_flush_and_dirty_for_history into two helpers. 3367- iris: Enable GL_AMD_depth_clamp_separate 3368- iris: Advertise EXT_texture_sRGB_R8 support 3369- iris: Some tidying for preemption support 3370- iris: Silence unused function warning 3371- iris: Fix zeroing of transform feedback offsets in strange cases. 3372- glsl/list: Add an exec_list_is_singular() helper. 3373- nir: Add a new nir_cf_list_is_empty_block() helper. 3374- intel/fs: Don't emit empty ELSE blocks. 3375- iris: Set XY Clipping correctly. 3376- iris: Only enable GL_AMD_depth_clamp_separate on Gen9+ 3377- iris: Fix imageBuffer and PBO download. 3378- iris: Disable dual source blending when shader doesn't handle it 3379- iris: Resolve textures used by the program, not merely bound textures 3380- iris: Fix 4GB memory zone heap sizes. 3381- iris: leave the top 4Gb of the high heap VMA unused 3382- iris: Force VMA alignment to be a multiple of the page size. 3383- iris: Delete bucketing allocators 3384- i965: Fix BRW_MEMZONE_LOW_4G heap size. 3385- i965: Force VMA alignment to be a multiple of the page size. 3386- i965: leave the top 4Gb of the high heap VMA unused 3387- i965: Fix memory leaks in brw_upload_cs_work_groups_surface(). 3388- iris: Use full ways for L3 cache setup on Icelake. 3389- egl/x11: calloc dri2_surf so it's properly zeroed 3390 3391Kevin Strasser (1): 3392 3393- egl/dri: Avoid out of bounds array access 3394 3395Khaled Emara (1): 3396 3397- freedreno: PIPE_CAP_SHADER_BUFFER_OFFSET_ALIGNMENT unreachable 3398 statement 3399 3400Khem Raj (1): 3401 3402- winsys/svga/drm: Include sys/types.h 3403 3404Kishore Kadiyala (1): 3405 3406- android: static link with libexpat with Android O+ 3407 3408Konstantin Kharlamov (1): 3409 3410- mapi: work around GCC LTO dropping assembly-defined functions 3411 3412Kristian Høgsberg (49): 3413 3414- st/nir: Use src/ relative include path for autotools 3415- freedreno/a6xx: Emit blitter dst with OUT_RELOCW 3416- freedreno/a6xx: Use tiling for all resources 3417- freedreno/a6xx: regen headers 3418- freedreno/a6xx: Drop render condition check in blitter 3419- freedreno: Log number of draw for sysmem passes 3420- freedreno/a6xx: Use the right resource for separate stencil stride 3421- freedreno/a6xx: Combine emit_blit and fd6_blit 3422- freedreno: Consolidate u_blitter functions in freedreno_blitter.c 3423- freedreno: Don't tell the blitter what it can't do 3424- freedreno/a6xx: Move blit check so as to restore comment 3425- freedreno/a6xx: Support some depth/stencil blits on blitter 3426- freedreno/a6xx: Support y-inverted blits 3427- freedreno/a6xx: Add format argument to fd6_tex_swiz() 3428- freedreno/a6xx: Fall back to masked RGBA blits for depth/stencil 3429- freedreno/a6xx: Clean up mixed use of swap and swizzle for texture 3430 state 3431- freedreno/a6xx: Update headers 3432- freedreno/a6xx: Front facing needs UNK3 bit 3433- freedreno/a6xx: Fix point coord 3434- .mailmap: Add a few more alises for myself 3435- freedreno: Update headers 3436- freedreno/a6xx: Copy stencil as R8_UINT 3437- freedreno/a6xx: Support MSAA resolve blits on blitter 3438- freedreno/a6xx: Only output MRT control for used framebuffers 3439- freedreno/a6xx: Don't zero SO buffer addresses 3440- freedreno: Fix a couple of warnings 3441- turnip: Only get bo offset when we need to mmap 3442- freedreno: Use c_vis_args and no_override_init_args 3443- freedreno/a6xx: Remove extra parens 3444- freedreno/ir3: Track whether shader needs derivatives 3445- freedreno/ir3: Fix operand order for DSX/DSY 3446- st/glsl_to_nir: Calculate num_uniforms from NumParameterValues 3447- freedreno/ir3: Enable PIPE_CAP_PACKED_UNIFORMS 3448- freedreno/ir3: Push UBOs to constant file 3449- freedreno/ir3: Don't access beyond available regs 3450- freedreno/ir3: Add workaround for VS samgq 3451- freedreno/ir3: Mark ir3_context_error() as NORETURN 3452- freedreno/a2xx: Fix redundant if statement 3453- freedreno: Use enum values from matching enum 3454- freedreno/a6xx: Add helper for incrementing regid 3455- freedreno: Fix format string warning 3456- .gitignore: Remove autotool artifacts 3457- tgsi: Mark tgsi_strings_check() unused 3458- glsl_to_nir: Initialize debug variable 3459- nir_opcodes.py: Saturate to expression that doesn't overflow 3460- ralloc: Fully qualify non-virtual destructor call 3461- egl/dri2: Mark potentially unused 'display' variable with 3462 MAYBE_UNUSED 3463- gallium/auxiliary/vl: Fix a couple of warnings 3464- freedreno/drm: Quiet pointer to u64 conversion warning 3465 3466Leo Liu (6): 3467 3468- st/va: fix the incorrect max profiles report 3469- st/va/vp9: set max reference as default of VP9 reference number 3470- vl/dri3: remove the wait before getting back buffer 3471- radeon/vcn: add H.264 constrained baseline support 3472- radeon/vcn/vp9: search the render target from the whole list 3473- winsys/amdgpu: add VCN JPEG to no user fence group 3474 3475Lepton Wu (2): 3476 3477- virgl: close drm fd when destroying virgl screen. 3478- virgl: Set bind when creating temp resource. 3479 3480Lionel Landwerlin (127): 3481 3482- anv: assert that color attachment are valid 3483- radv: assert that colorAttachment is valid for CmdClearAttachment 3484- i965: scale factor changes should trigger recompile 3485- vulkan: Update the XML and headers to 1.1.101 3486- anv: implement VK_EXT_depth_clip_enable 3487- build: move imgui out of src/intel/tools to be reused 3488- imgui: bump copy 3489- imgui: make sure our copy of imgui doesn't clash with others in the 3490 same process 3491- vulkan: add an overlay layer 3492- intel: fix urb size for CFL GT1 3493- anv: add support for INTEL_DEBUG=bat 3494- Revert "anv: add support for INTEL_DEBUG=bat" 3495- intel/aub_viewer: printout 48bits addresses 3496- intel/aub_viewer: silence compiler warning 3497- intel/aub_viewer: silence more compiler warnings 3498- vulkan/overlay: fix missing installation of layer 3499- vulkan/overlay: fix includes 3500- imgui: update commit 3501- imgui: update memory editor 3502- vulkan/overlay: install layer binary in libdir 3503- intel/compiler: use correct swizzle for replacement 3504- vulkan/overlay: fix min/max computations 3505- vulkan/overlay: rework option parsing 3506- vulkan/overlay: add support for fps output in file 3507- anv: add support for INTEL_DEBUG=bat 3508- vulkan: update headers/registry to 1.1.102 3509- anv: update supported patch version 3510- radv: set num_components on vulkan_resource_index intrinsic 3511- vulkan/util: make header available from c++ 3512- vulkan/util: generate instance/device dispatch tables 3513- vulkan/overlay: drop dependency on validation layer headers 3514- intel/decoders: add address space indicator to get BOs 3515- intel/decoders: handle decoding MI_BBS from ring 3516- intel/decoders: limit number of decoded batchbuffers 3517- intel/aub_read: reuse defines from gen_context 3518- intel/aub_write: split comment section from HW setup 3519- intel/aub_write: write header in init 3520- intel/aub_write: break execlist write in 2 3521- intel/aub_write: switch to use i915_drm engine classes 3522- intel/aub_write: log mmio writes 3523- intel/aub_write: store the physical page allocator in struct 3524- intel/aub_write: turn context images arrays into functions 3525- intel/aub_write: factorize context image/pphwsp/ring creation 3526- iris: fix decoder call 3527- iris: fix decode_get_bo callback 3528- intel/error2aub: build a list of BOs before writing them 3529- intel/error2aub: identify buffers by engine 3530- intel/error2aub: strenghten batchbuffer identifier marker 3531- intel/error2aub: parse other buffer types 3532- intel/error2aub: annotate buffer with their address space 3533- intel/error2aub: store engine last ring buffer head/tail pointers 3534- intel/error2aub: write GGTT buffers into the aub file 3535- intel/error2aub: add a verbose option 3536- intel/error2aub: deal with GuC log buffer 3537- intel/error2aub: support older style engine names 3538- vulkan: factor out wsi dependencies 3539- anv: implement VK_EXT_pipeline_creation_feedback 3540- vulkan/overlay: properly register layer object with loader 3541- vulkan/overlay: silence validation layer warnings 3542- vulkan/overlay: check return value of swapchain get images 3543- vulkan/overlay: improve error reporting 3544- i965: perf: sklgt2: update a priority for register programming 3545- i965: perf: sklgt2: update compute metrics config 3546- i965: perf: sklgt2: update memory write config 3547- i965: perf: add PMA stall metrics 3548- i965: perf: chv: fixup counters names 3549- i965: perf: hsw: drop register programming not needed on HSW 3550- i965: perf: sklgt2: drop programming of an unused NOA register 3551- i965: perf: add Icelake metrics 3552- i965: perf: enable Icelake metrics 3553- i965: perf: add ring busyness metric for cfl gt2 3554- i965: perf: update render basic configs for big core gen9/gen10 3555- anv: implement VK_KHR_swapchain revision 70 3556- intel: add dependency on genxml generated files 3557- genxml: add a sorting script 3558- genxml: sort xml files using new script 3559- anv: don't use default pipeline cache for hits for 3560 VK_EXT_pipeline_creation_feedback 3561- anv: store heap address bounds when initializing physical device 3562- anv: leave the top 4Gb of the high heap VMA unused 3563- i965: store device revision in gen_device_info 3564- i965: extract performance query metrics 3565- i965: move mdapi data structure to intel/perf 3566- i965: move OA accumulation code to intel/perf 3567- i965: move brw_timebase_scale to device info 3568- i965: move mdapi result data format to intel/perf 3569- i965: move mdapi guid into intel/perf 3570- intel/perf: stub gen10/11 missing definitions 3571- i965: perf: add mdapi pipeline statistics queries on gen10/11 3572- intel/perf: drop counter size field 3573- intel/perf: constify accumlator parameter 3574- iris: implement WaEnableStateCacheRedirectToCS 3575- i965: implement WaEnableStateCacheRedirectToCS 3576- anv: implement WaEnableStateCacheRedirectToCS 3577- anv: fix uninitialized pthread cond clock domain 3578- intel/devinfo: fix missing num_thread_per_eu on ICL 3579- intel/devinfo: add basic sanity tests on device database 3580- anv: limit URB reconfigurations when using blorp 3581- intel: workaround VS fixed function issue on Gen9 GT1 parts 3582- anv: fix argument name for vkCmdEndQuery 3583- i965: fix icelake performance query enabling 3584- Revert "anv: limit URB reconfigurations when using blorp" 3585- vulkan/util: generate a helper function to return pNext struct sizes 3586- vulkan/overlay: update help printout 3587- vulkan/overlay: record stats in command buffers and accumulate on 3588 exec/submit 3589- vulkan/overlay: add pipeline statistic & timestamps support 3590- vulkan/overlay: add no display option 3591- vulkan/overlay: add a margin to the size of the window 3592- vulkan/overlay: record all select metrics into output file 3593- vulkan/overlay: add a frame counter option 3594- vulkan/overlay: make overlay size configurable 3595- vulkan/overlay: make overriden functions static 3596- vulkan/overlay: add TODO list 3597- anv: fix crash when application does not provide push constants 3598- anv: rework queries writes to ensure ordering memory writes 3599- anv: fix use after free 3600- anv: Use corresponding type from the vector allocation 3601- vulkan/overlay: keep allocating draw data until it can be reused 3602- nir: fix lower_non_uniform_access pass 3603- vulkan/overlay-layer: fix cast errors 3604- vulkan/overlay: fix truncating error on 32bit platforms 3605- nir: lower_non_uniform_access: iterate over instructions safely 3606- vulkan/overlay: fix timestamp query emission with no pipeline stats 3607- vulkan: fix build dependency issue with generated files 3608- anv: fix apply_pipeline_layout pass for arrays of YCbCr descriptors 3609- nir/lower_non_uniform: safely iterate over blocks 3610- intel/perf: fix EuThreadsCount value in performance equations 3611- intel/perf: improve dynamic loading config detection 3612 3613Lubomir Rintel (3): 3614 3615- kmsro: Extend to include armada-drm 3616- gallivm: guess CPU features also on ARM 3617- gallivm: disable NEON instructions if they are not supported 3618 3619Lucas Stach (3): 3620 3621- etnaviv: don't flush own context when updating resource use 3622- etnaviv: flush all pending contexts when accessing a resource with 3623 the CPU 3624- etnaviv: only try to construct scanout resource when on KMS winsys 3625 3626Marek Olšák (121): 3627 3628- radeonsi: enable dithered alpha-to-coverage for better quality 3629- radeonsi: merge & rename texture BO metadata functions 3630- radeonsi: unify error paths in si_texture_create_object 3631- winsys/amdgpu: remove amdgpu_drm.h definitions 3632- r600: add -Wstrict-overflow=0 to meson to silence the warning 3633- radeonsi: fix a comment typo in si_fine_fence_set 3634- gallium: allow more PIPE_RESOURCE\_ driver flags 3635- meson: drop the xcb-xrandr version requirement 3636- radeonsi: handle render_condition_enable in 3637 si_compute_clear_render_target 3638- radeonsi: fix crashing performance counters (division by zero) 3639- radeonsi: initialize textures using DCC to black when possible 3640- radeonsi: clear allocator_zeroed_memory with SDMA 3641- radeonsi: make allocator_zeroed_memory unmappable and use bigger 3642 buffers 3643- radeonsi: don't leak an index buffer if draw_vbo fails 3644- radeonsi: use local ws variable in si_need_dma_space 3645- gallium/u_threaded: fix EXPLICIT_FLUSH for flush offsets > 0 3646- radeonsi: fix EXPLICIT_FLUSH for flush offsets > 0 3647- winsys/amdgpu: don't drop manually added fence dependencies 3648- winsys/amdgpu: unify fence list code 3649- winsys/amdgpu: use a separate fence list for syncobjs 3650- winsys/amdgpu: remove occurence of INDIRECT_BUFFER_CONST 3651- winsys/amdgpu: clean up IB buffer size computation 3652- winsys/amdgpu: cs_check_space sets the minimum IB size for future IBs 3653- radeonsi: add AMD_DEBUG env var as an alternative to R600_DEBUG 3654- radeonsi: use MEM instead of MEM_GRBM in COPY_DATA.DST_SEL 3655- radeonsi: add driconf option radeonsi_enable_nir 3656- radeonsi: always enable NIR for Civilization 6 to fix corruption 3657- driconf: add Civ6Sub executable for Civilization 6 3658- st/mesa: always unmap the uploader in st_atom_array.c 3659- gallium/u_threaded: always unmap const_uploader 3660- gallium/u_upload_mgr: allow use of FLUSH_EXPLICIT with persistent 3661 mappings 3662- radeonsi: use SDMA for uploading data through const_uploader 3663- tgsi: don't set tgsi_info::uses_bindless_images for constbufs and hw 3664 atomics 3665- radeonsi: always use compute rings for clover on CI and newer (v2) 3666- gallium/u_tests: use a compute-only context to test GCN compute ring 3667- gallium: add pipe_grid_info::last_block 3668- omx: clean up enc_LoadImage_common 3669- omx: add a compute path in enc_LoadImage_common 3670- radeonsi: fix assertion failure by using the correct type 3671- mesa: implement ARB/KHR_parallel_shader_compile 3672- gallium: implement ARB/KHR_parallel_shader_compile 3673- util/queue: move thread creation into a separate function 3674- util/queue: add ability to kill a subset of threads 3675- util/queue: hold a lock when reading num_threads in util_queue_finish 3676- util/queue: add util_queue_adjust_num_threads 3677- radeonsi: implement ARB/KHR_parallel_shader_compile callbacks 3678- radeonsi: don't use PFP_SYNC_ME with compute-only contexts 3679- docs/relnotes: document parallel_shader_compile changes in 19.1.0, 3680 not 19.0.0 3681- amd/addrlib: fix uninitialized values for 3682 Addr2ComputeDccAddrFromCoord 3683- radeonsi/gfx9: add support for PIPE_ALIGNED=0 3684- radeonsi: add ability to bind images as image buffers 3685- radeonsi: add support for displayable DCC for 1 RB chips 3686- radeonsi: add support for displayable DCC for multi-RB chips 3687- radeonsi: enable displayable DCC on Ravens 3688- gallium: add writable_bitmask parameter into set_shader_buffers 3689- glsl: remember which SSBOs are not read-only and pass it to gallium 3690- radeonsi: set exact shader buffer read/write usage in CS 3691- tegra: fix the build after the set_shader_buffers change 3692- radeonsi: fix a crash when unbinding sampler states 3693- glsl: fix shader_storage_blocks_write_access for SSBO block arrays 3694- Revert "glsl: fix shader_storage_blocks_write_access for SSBO block 3695 arrays" 3696- glsl: allow the #extension directive within code blocks for the dri 3697 option 3698- mesa: don't overwrite existing shader files with 3699 MESA_SHADER_CAPTURE_PATH 3700- radeonsi: set AC_FUNC_ATTR_READNONE for image opcodes where it was 3701 missing 3702- ac: use the common helper ac_apply_fmask_to_sample 3703- ac: fix incorrect bindless atomic code in visit_image_atomic 3704- radeonsi: enable GL_EXT_shader_image_load_formatted 3705- nir: optimize gl_SampleMaskIn to gl_HelperInvocation for radeonsi 3706 when possible 3707- winsys/amdgpu: don't set GTT with GDS & OA placements on APUs 3708- radeonsi/gfx9: use the correct condition for the DPBB + QUANT_MODE 3709 workaround 3710- radeonsi: use CP DMA for the null const buffer clear on CIK 3711- tgsi/scan: add uses_drawid 3712- ac: add radeon_info::marketing_name, replacing the winsys callback 3713- ac: add radeon_info::is_pro_graphics 3714- ac: add ac_get_i1_sgpr_mask 3715- ac: add REWIND and GDS registers to register headers 3716- winsys/amdgpu: make IBs writable and expose their address 3717- winsys/amdgpu: reorder chunks, make BO_HANDLES first, IB and FENCE 3718 last 3719- winsys/amdgpu: enable chaining for compute IBs 3720- winsys/amdgpu: clean up and remove nonsensical assertion 3721- radeonsi: add si_cp_copy_data 3722- radeonsi: add helper si_get_minimum_num_gfx_cs_dwords 3723- radeonsi: delay adding BOs at the beginning of IBs until the first 3724 draw 3725- gallium: document conservative rasterization flags 3726- st/dri: simplify throttling code 3727- gallium: replace DRM_CONF_THROTTLE with PIPE_CAP_MAX_FRAMES_IN_FLIGHT 3728- gallium: replace DRM_CONF_SHARE_FD with PIPE_CAP_DMABUF 3729- gallium: replace drm_driver_descriptor::configuration with 3730 driconf_xml 3731- gallium: set PIPE_CAP_MAX_FRAMES_IN_FLIGHT to 2 for all drivers 3732- gallium: add PIPE_CAP_PREFER_COMPUTE_BLIT_FOR_MULTIMEDIA 3733- util: fix a compile failure in u_compute.c on windows 3734- mesa: enable glGet for EXT_gpu_shader4 3735- glsl: add \`unsigned int\` type for EXT_GPU_shader4 3736- glsl: apply some 1.30 and other rules to EXT_gpu_shader4 as well 3737- glsl: add builtin variables for EXT_gpu_shader4 3738- glsl: add arithmetic builtin functions for EXT_gpu_shader4 3739- glsl: add texture builtin functions for EXT_gpu_shader4 3740- glsl: allow "varying out" for fragment shader outputs with 3741 EXT_gpu_shader4 3742- mesa: expose EXT_texture_buffer_object 3743- mesa: only allow EXT_gpu_shader4 in the compatibility profile 3744- st/mesa: expose EXT_gpu_shader4 if GLSL 1.40 is supported 3745- glsl: handle interactions between EXT_gpu_shader4 and texture 3746 extensions 3747- radeonsi: add BOs after need_cs_space 3748- radeonsi/gfx9: set that window_rectangles always roll the context 3749- radeonsi/gfx9: rework the gfx9 scissor bug workaround (v2) 3750- radeonsi: remove dirty slot masks from scissor and viewport states 3751- glsl: fix shader_storage_blocks_write_access for SSBO block arrays 3752 (v2) 3753- radeonsi: don't ignore PIPE_FLUSH_ASYNC 3754- mesa: rework error handling in glDrawBuffers 3755- mesa: fix pbuffers because internally they are front buffers 3756- st/mesa: don't flush the front buffer if it's a pbuffer 3757- radeonsi: use new atomic LLVM helpers 3758- radeonsi: set sampler state and view functions for compute-only 3759 contexts 3760- st/dri: decrease input lag by syncing sooner in SwapBuffers 3761- glsl: fix and clean up NV_compute_shader_derivatives support 3762- st/mesa: fix 2 crashes in st_tgsi_lower_yuv 3763- radeonsi: remove old_va parameter from si_rebind_buffer by 3764 remembering offsets 3765- radeonsi: update buffer descriptors in all contexts after buffer 3766 invalidation 3767- radeonsi: fix a regression in si_rebind_buffer 3768- u_blitter: don't fail mipmap generation for depth formats containing 3769 stencil 3770- ac: fix a typo in ac_build_wg_scan_bottom 3771 3772Mario Kleiner (1): 3773 3774- drirc: Add sddm-greeter to adaptive_sync blacklist. 3775 3776Mark Janes (5): 3777 3778- mesa: properly report the length of truncated log messages 3779- mesa: rename logging functions to reflect that they format strings 3780- mesa: add logging function for formatted string 3781- intel/common: move gen_debug to intel/dev 3782- intel/tools: Remove redundant definitions of INTEL_DEBUG 3783 3784Mateusz Krzak (2): 3785 3786- panfrost: cast bo_handles pointer to uintptr_t first 3787- panfrost: use os_mmap and os_munmap 3788 3789Mathias Fröhlich (22): 3790 3791- st/mesa: Reduce array updates due to current changes. 3792- mesa: Track buffer object use also for VAO usage. 3793- st/mesa: Invalidate the gallium array atom only if needed. 3794- mesa: Implement helper functions to map and unmap a VAO. 3795- mesa: Factor out \_mesa_array_element. 3796- mesa: Use \_mesa_array_element in dlist save. 3797- mesa: Replace \_ae_{,un}map_vbos with \_mesa_vao_{,un}map_arrays 3798- mesa: Remove \_ae_{,un}map_vbos and dependencies. 3799- mesa: Use mapping tools in debug prints. 3800- vbo: Fix basevertex handling in display list compiles. 3801- vbo: Fix GL_PRIMITIVE_RESTART_FIXED_INDEX in display list compiles. 3802- mesa: Add assert to \_mesa_primitive_restart_index. 3803- mesa: Factor out index function that will have multiple use. 3804- mesa: Use glVertexAttrib*NV functions for fixed function attribs. 3805- mesa: Implement \_mesa_array_element by walking enabled arrays. 3806- mesa: Rip out now unused gl_context::aelt_context. 3807- mesa: Remove the now unused \_NEW_ARRAY state change flag. 3808- mesa: Constify static const array in api_arrayelt.c 3809- mesa: Remove the \_glapi_table argument from \_mesa_array_element. 3810- mesa: Set CurrentSavePrimitive in vbo_save_NotifyBegin. 3811- mesa: Correct the is_vertex_position decision for dlists. 3812- mesa: Leave aliasing of vertex and generic0 attribute to the dlist 3813 code. 3814 3815Matt Turner (7): 3816 3817- intel/compiler/test: Set devinfo->gen = 7 3818- intel/compiler: Avoid propagating inequality cmods if types are 3819 different 3820- intel/compiler/test: Add unit test for mismatched signedness 3821 comparison 3822- intel/compiler: Add commas on final values of compaction table arrays 3823- intel/compiler: Use SIMD16 instructions in fs saturate prop unit test 3824- intel/compiler: Add unit tests for sat prop for different exec sizes 3825- intel/compiler: Improve fix_3src_operand() 3826 3827Matthias Lorenz (1): 3828 3829- vulkan/overlay: Add fps counter 3830 3831Mauro Rossi (6): 3832 3833- android: intel/isl: remove redundant building rules 3834- android: anv: fix generated files depedencies (v2) 3835- android: anv: fix libexpat shared dependency 3836- android: nouveau: add support for nir 3837- android: fix LLVM version string related building errors 3838- draw: fix building error in draw_gs_init() 3839 3840Maya Rashish (1): 3841 3842- configure: fix test portability 3843 3844Michel Dänzer (19): 3845 3846- loader/dri3: Use strlen instead of sizeof for creating VRR property 3847 atom 3848- gitlab-ci: Re-use docker image from the main repo in forked repos 3849- gitlab-ci: List some longer-running jobs before others of the same 3850 stage 3851- gitlab-ci: Use 8 CPU cores in autotools job 3852- gitlab-ci: Make sure clang job actually uses ccache 3853- gitlab-ci: Only pull/push cache contents in build+test stage jobs 3854- gitlab-ci: Automatically retry jobs after runner system failure 3855- gitlab-ci: Run CI pipeline for all branches in the main repository 3856- gitlab-ci: Use Debian stretch instead of Ubuntu bionic 3857- gitlab-ci: Use HTTPS for APT repositories 3858- gitlab-ci: Use Debian packages instead of pip ones for meson and 3859 scons 3860- gitlab-ci: Install most packages from Debian buster 3861- gitlab-ci: Remove unneded (stuff from) APT command lines 3862- gitlab-ci: Remove unused Debian packages from Docker image 3863- gitlab-ci: Use clang 8 instead of 7 3864- gitlab-ci: Drop unused clang 5/6 packages 3865- gitlab-ci: Do not use subshells for compiling dependencies 3866- gitlab-ci: Use LLVM 3.4 from Debian jessie for scons-llvm job 3867- gitlab-ci: Use meson buildtype debug instead of default 3868 debugoptimized 3869 3870Mike Blumenkrantz (6): 3871 3872- iris: support INTEL_NO_HW environment variable 3873- gallium: add pipe cap for inner_coverage conservative raster mode 3874- st/mesa: indicate intel extension support for inner_coverage based on 3875 cap 3876- iris: add support for INTEL_conservative_rasterization 3877- iris: add preemption support on gen9 3878- iris: enable preemption support for gen10 3879 3880Nanley Chery (3): 3881 3882- i965: Rename intel_mipmap_tree::r8stencil\_\* -> ::shadow\_\* 3883- anv: Fix some depth buffer sampling cases on ICL+ 3884- anv/cmd_buffer: Initalize the clear color struct for CNL+ 3885 3886Nataraj Deshpande (1): 3887 3888- anv: Fix check for isl_fmt in assert 3889 3890Neha Bhende (2): 3891 3892- st/mesa: Fix topogun-1.06-orc-84k-resize.trace crash 3893- draw: fix memory leak introduced 7720ce32a 3894 3895Nicolai Hähnle (9): 3896 3897- amd/surface: provide firstMipIdInTail for metadata surface 3898 calculations 3899- radeonsi: add si_debug_options for convenient adding/removing of 3900 options 3901- util/u_log: flush auto loggers before starting a new page 3902- ddebug: set thread name 3903- ddebug: log calls to pipe->flush 3904- ddebug: dump driver state into a separate file 3905- ddebug: expose some helper functions as non-inline 3906- radeonsi: add radeonsi_aux_debug option for aux context debug dumps 3907- radeonsi: add radeonsi_sync_compile option 3908 3909Oscar Blumberg (3): 3910 3911- intel/fs: Fix memory corruption when compiling a CS 3912- radeonsi: Fix guardband computation for large render targets 3913- glsl: Fix function return typechecking 3914 3915Patrick Lerda (1): 3916 3917- lima/ppir: fix pointer referenced after a free 3918 3919Patrick Rudolph (1): 3920 3921- d3dadapter9: Support software renderer on any DRI device 3922 3923Philipp Zabel (1): 3924 3925- etnaviv: fill missing offset in etna_resource_get_handle 3926 3927Pierre Moreau (12): 3928 3929- include/CL: Update to the latest OpenCL 2.2 headers 3930- clover: Avoid warnings from new OpenCL headers 3931- clover: Remove the TGSI backend as unused 3932- clover: Add an helper for checking if an IR is supported 3933- clover/api: Rework the validation of devices for building 3934- clover/api: Fail if trying to build a non-executable binary 3935- clover: Disallow creating libraries from other libraries 3936- clover: Validate program and library linking options 3937- clover: Move device extensions definitions to core/device.cpp 3938- clover: Move platform extensions definitions to clover/platform.cpp 3939- clover: Only use devices supporting IR_NATIVE 3940- clover: Fix indentation issues 3941 3942Pierre-Eric Pelloux-Prayer (1): 3943 3944- radeonsi: init sctx->dma_copy before using it 3945 3946Plamena Manolova (3): 3947 3948- i965: Disable ARB_fragment_shader_interlock for platforms prior to 3949 GEN9 3950- isl: Set ClearColorConversionEnable. 3951- i965: Re-enable fast color clears for GEN11. 3952 3953Qiang Yu (9): 3954 3955- u_math: add ushort_to_float/float_to_ushort 3956- u_dynarray: add util_dynarray_grow_cap 3957- gallium/u_vbuf: export u_vbuf_get_minmax_index 3958- drm-uapi: add lima_drm.h 3959- gallium: add lima driver 3960- lima/gpir: fix compile fail when two slot node 3961- lima/gpir: fix alu check miss last store slot 3962- lima: fix lima_blit with non-zero level source resource 3963- lima: fix render to non-zero level texture 3964 3965Rafael Antognolli (45): 3966 3967- iris: Store internal_format when getting resource from handle. 3968- iris: Skip msaa16 on gen < 9. 3969- iris: Flush before hiz_exec. 3970- iris: Pin HiZ buffers when rendering. 3971- iris: Avoid leaking if we fail to allocate the aux buffer. 3972- iris/clear: Pass on render_condition_enabled. 3973- iris: Skip resolve if there's no context. 3974- iris: Flag ALL_DIRTY_BINDINGS on aux state change. 3975- iris: Add resolve on iris_flush_resource. 3976- iris: Convert RGBX to RGBA always. 3977- iris: Enable auxiliary buffer support again 3978- iris: Enable HiZ for multisampled depth surfaces. 3979- iris: Make intel_hiz_exec public. 3980- iris: Allocate buffer space for the fast clear color. 3981- iris: Use the clear depth when emitting 3DSTATE_CLEAR_PARAMS. 3982- iris: Fast clear depth buffers. 3983- iris: Add helper to convert fast clear color. 3984- iris: Add function to update clear color in surface state. 3985- iris: Bring back check for srgb and fast clear color. 3986- intel/isl: Add isl_format_has_color_component() function. 3987- intel/blorp: Make swizzle_color_value public. 3988- iris: Implement fast clear color. 3989- iris: Add iris_resolve_conditional_render(). 3990- iris: Stall on the CPU and resolve predication during fast clears. 3991- iris: Track fast clear color. 3992- iris: Let blorp update the clear color for us. 3993- i965/blorp: Remove unused parameter from blorp_surf_for_miptree. 3994- iris: Only update clear color for gens 8 and 9. 3995- iris/gen8: Re-emit the SURFACE_STATE if the clear color changed. 3996- iris: Manually apply fast clear color channel overrides. 3997- iris: Do not allocate clear_color_bo for gen8. 3998- iris: Add aux.sampler_usages. 3999- iris: Enable fast clears on gen8. 4000- intel/fs: Only propagate saturation if exec_size is the same. 4001- intel/fs: Move the scalar-region conversion to the generator. 4002- intel/fs: Add a lowering pass for linear interpolation. 4003- intel/fs: Remove fs_generator::generate_linterp from gen11+. 4004- intel/isl: Resize clear color buffer to full cacheline 4005- intel/genxml: Update MI_ATOMIC genxml definition. 4006- intel/blorp: Make blorp update the clear color in gen11. 4007- iris: Do not advertise multisampled image load/store. 4008- iris: Support sRGB fast clears even if the colorspaces differ. 4009- iris: Use the linear version of the surface format during fast 4010 clears. 4011- iris: Update the surface state clear color address when available. 4012- iris: Enable fast clear colors on gen11. 4013 4014Ray Zhang (1): 4015 4016- glx: fix shared memory leak in X11 4017 4018Rhys Kidd (1): 4019 4020- iris: Fix assertion in iris_resource_from_handle() tiling usage 4021 4022Rhys Perry (28): 4023 4024- nvc0: add compute invocation counter 4025- radv: bitcast 16-bit outputs to integers 4026- radv: ensure export arguments are always float 4027- ac/nir: implement 8-bit nir_load_const_instr 4028- ac/nir: fix 64-bit nir_op_f2f16_rtz 4029- ac/nir: make ac_build_clamp work on all bit sizes 4030- ac/nir: make ac_build_isign work on all bit sizes 4031- ac/nir: make ac_build_fdiv support 16-bit floats 4032- ac/nir: implement half-float nir_op_frcp 4033- ac/nir: implement half-float nir_op_frsq 4034- ac/nir: implement half-float nir_op_ldexp 4035- ac/nir: fix 16-bit ssbo stores 4036- ac/nir: implement 8-bit push constant, ssbo and ubo loads 4037- ac/nir: implement 8-bit ssbo stores 4038- ac/nir: add 8-bit types to glsl_base_to_llvm_type 4039- ac/nir: implement 8-bit conversions 4040- radv: enable VK_KHR_8bit_storage 4041- ac/nir: implement 16-bit pack/unpack opcodes 4042- radv: lower 16-bit flrp 4043- ac: add 16-bit support to ac_build_ddxy() 4044- nir,ac/nir: fix cube_face_coord 4045- gallium: add support for formatted image loads 4046- mesa, glsl: add support for EXT_shader_image_load_formatted 4047- st/mesa: add support for EXT_shader_image_load_formatted 4048- vc4: fix build 4049- ac,ac/nir: use a better sync scope for shared atomics 4050- radv: fix set_output_usage_mask() with composite and 64-bit types 4051- ac/nir: mark some texture intrinsics as convergent 4052 4053Rob Clark (135): 4054 4055- freedreno: fix release tarball 4056- freedreno: more fixing release tarball 4057- freedreno/a6xx: small compiler warning fix 4058- freedreno/ir3: fix varying packing vs. tex sharp edge 4059- freedreno/a6xx: move stream-out emit to helper 4060- freedreno/a6xx: clean up some open-coded bits 4061- freedreno/ir3: split out image helpers 4062- freedreno/ir3: split out a4xx+ instructions 4063- freedreno/ir3: fix ncomp for \_store_image() src 4064- freedreno/ir3: add image/ssbo <-> ibo/tex mapping 4065- freedreno/ir3: add a6xx instruction encoding 4066- freedreno/ir3: add a6xx+ SSBO/image support 4067- freedreno/ir3: HIGH reg w/a for a6xx 4068- freedreno/a6xx: border-color offset helper 4069- freedreno/a6xx: image/ssbo state emit 4070- freedreno/a6xx: compute support 4071- freedreno/a6xx: cache flush harder 4072- freedreno/a6xx: fix helper_invocation (sampler mask/id) 4073- freedreno/ir3: handle quirky atomic dst for a6xx 4074- freedreno/ir3: fix legalize for vecN inputs 4075- freedreno/ir3: fix crash in compile fail case 4076- freedreno/a6xx: 3d and cube image fixes 4077- freedreno: fix crash w/ masked non-SSA dst 4078- freedreno/ir3: rename put_dst() 4079- freedreno/ir3/a6xx: fix load_ssbo barrier type. 4080- freedreno/ir3: sync instr/disasm and add ldib encoding 4081- freedreno/ir3/a6xx: use ldib for ssbo reads 4082- freedreno/a6xx: samplerBuffer fixes 4083- freedreno/a6xx: enable tiled images 4084- freedreno: fix race condition 4085- freedreno/ir3: don't hardcode wrmask 4086- freedreno/a6xx: fix border-color offset 4087- freedreno/a6xx: cube image fix 4088- freedreno/a6xx: fix hangs with large shaders 4089- freedreno/ir3: use nopN encoding when possible 4090- freedreno/a6xx: fix ssbo alignment 4091- freedreno/ir3/a6xx: fix non-ssa atomic dst 4092- freedreno/a6xx: fix DRAW_IDX_INDIRECT max_indicies 4093- freedreno/a6xx: vertex_id is not \_zero_based 4094- freedreno/ir3/a6xx: fix atomic shader outputs 4095- freedreno/ir3: gsampler2DMSArray fixes 4096- freedreno/ir3: include nopN in expanded instruction count 4097- freedreno/ir3: add Sethi–Ullman numbering pass 4098- freedreno/ir3: track register pressure in sched 4099- freedreno: fix ir3_cmdline build 4100- freedreno/a6xx: remove astc_srgb workaround 4101- freedreno/a6xx: refactor fd6_tex_swiz() 4102- freedreno/a6xx: fix border-color swizzles 4103- freedreno/a6xx: perfcntrs 4104- freedreno/ir3: fix ir3_cmdline harder 4105- freedreno/ir3: turn on [iu]mul_high 4106- freedreno/a6xx: more bcolor fixes 4107- freedreno/ir3/cp: fix ldib bug 4108- freedreno/ir3/a6xx: fix ssbo comp_swap 4109- freedreno/ir3 better cat6 encoding detection 4110- freedreno/ir3/ra: fix half-class conflicts 4111- freedreno/ir3: fix sam.s2en decoding 4112- freedreno/ir3: fix sam.s2en encoding 4113- freedreno/ir3: fix regmask for merged regs 4114- nir: move gls_type_get_{sampler,image}_count() 4115- freedreno/ir3: find # of samplers from uniform vars 4116- freedreno/ir3: enable indirect tex/samp (sam.s2en) 4117- freedreno/ir3: optimize sam.s2en to sam 4118- freedreno/ir3: additional lowering 4119- freedreno/ir3: fix bit_count 4120- freedreno/ir3: dynamic UBO indexing vs 64b pointers 4121- freedreno/ir3: rename has_kill to no_earlyz 4122- freedreno/ir3: disable early-z for SSBO/image writes 4123- gallium: add PIPE_CAP_ESSL_FEATURE_LEVEL 4124- mesa/st: use ESSL cap top enable gpu_shader5 4125- freedreno: add ESSL cap 4126- docs: update freedreno status 4127- freedreno/a6xx: small cleanup 4128- freedreno/ir3: sched fix 4129- freedreno/ir3: reads/writes to unrelated arrays are not dependent 4130- freedreno/ir3: align const size to vec4 4131- nir: print var name for load_interpolated_input too 4132- nir: add lower_all_io_to_elements 4133- freedreno/ir3: re-indent comment 4134- freedreno/ir3: rework varying packing 4135- freedreno/ir3: add pass to move varying loads 4136- freedreno/ir3: convert to "new style" frag inputs 4137- gallium/docs: clarify set_sampler_views (v2) 4138- iris: fix set_sampler_view 4139- freedreno/ir3: fix const assert 4140- freedreno/drm: update for robustness 4141- freedreno: add robustness support 4142- compiler: rename SYSTEM_VALUE_VARYING_COORD 4143- freedreno/ir3: fix rgetpos decoding 4144- freedreno/ir3: more emit-cat5 fixes 4145- freedreno/ir3: cleanup instruction builder macros 4146- freedreno: update generated headers 4147- freedreno/ir3: lower load_barycentric_at_sample 4148- freedreno/ir3: lower load_barycentric_at_offset 4149- freedreno/ir3: remove bogus assert 4150- freedreno/ir3: rename frag_vcoord -> ij_pixel 4151- freedreno/a6xx: add VALIDREG/CONDREG helper macros 4152- freedreno/ir3: fix load_interpolated_input slot 4153- freedreno: wire up core sample-shading support 4154- freedreno/ir3: sample-shading support 4155- freedreno/a6xx: sample-shading support 4156- docs/features: update GL too 4157- freedreno/ir3: switch fragcoord to sysval 4158- freedreno/a6xx: small texture emit cleanup 4159- freedreno/a6xx: pre-bake UBWC flags in texture-view 4160- freedreno/ir3: fixes for half reg in/out 4161- freedreno/ir3: fix shader variants vs UBO analysis 4162- freedreno/ir3: fix lowered ubo region alignment 4163- freedreno/ir3: add IR3_SHADER_DEBUG flag to disable ubo lowering 4164- freedreno/ir3: add some ubo range related asserts 4165- nir: rework tex instruction printing 4166- nir: fix lower_wpos_ytransform in load_frag_coord case 4167- nir: add pass to lower fb reads 4168- freedreno/drm: expose GMEM_BASE address 4169- freedreno/ir3: fb read support 4170- freedreno/a6xx: KHR_blend_equation_advanced support 4171- freedreno/a6xx: smaller hammer for fb barrier 4172- docs: mark KHR_blend_equation_advanced done on a6xx 4173- nir: fix nir tex print harder 4174- freedreno/ir3: remove assert 4175- freedreno/a6xx: OUT_RELOC vs OUT_RELOCW fixes 4176- freedreno: update generated headers 4177- freedreno/a6xx: UBWC fixes 4178- freedreno/a6xx: UBWC support for images 4179- freedreno: mark imported resources as valid 4180- freedreno/a6xx: buffer resources cannot be compressed 4181- freedreno: move UBWC color offset to fd_resource_offset() 4182- freedreno: add ubwc_enabled helper 4183- freedreno/a6xx: deduplicate a few lines 4184- freedreno: remove unused forward struct declaration 4185- freedreno/ir3: fix rasterflat/glxgears 4186- freedreno/ir3: set more barrier bits 4187- freedreno/a6xx: fix GPU crash on small render targets 4188- freedreno/a6xx: fix issues with gallium HUD 4189- freedreno/a6xx: fix hangs with newer sqe fw 4190 4191Rob Herring (2): 4192 4193- kmsro: Add lima renderonly support 4194- kmsro: Add platform support for exynos and sun4i 4195 4196Rodrigo Vivi (1): 4197 4198- intel: Add more PCI Device IDs for Coffee Lake and Ice Lake. 4199 4200Roland Scheidegger (2): 4201 4202- gallivm: fix bogus assert in get_indirect_index 4203- gallivm: fix saturated signed add / sub with llvm 9 4204 4205Romain Failliot (1): 4206 4207- docs: changed "Done" to "DONE" in features.txt 4208 4209Ross Burton (1): 4210 4211- Revert "meson: drop GLESv1 .so version back to 1.0.0" 4212 4213Ryan Houdek (1): 4214 4215- panfrost: Adds Bifrost shader disassembler utility 4216 4217Sagar Ghuge (10): 4218 4219- iris: Don't allocate a BO per query object 4220- nir/glsl: Add another way of doing lower_imul64 for gen8+ 4221- glsl: [u/i]mulExtended optimization for GLSL 4222- nir/algebraic: Optimize low 32 bit extraction 4223- spirv: Allow [i/u]mulExtended to use new nir opcode 4224- iris: Refactor code to share 3DSTATE_URB\_\* packet 4225- iris: Track last VS URB entry size 4226- iris: Flag fewer dirty bits in BLORP 4227- intel/fs: Remove unused condition from opt_algebraic case 4228- intel/compiler: Fix assertions in brw_alu3 4229 4230Samuel Iglesias Gonsálvez (4): 4231 4232- isl: remove the cache line size alignment requirement 4233- isl: the display engine requires 64B alignment for linear surfaces 4234- radv: don't overwrite results in VkGetQueryPoolResults() when queries 4235 are not available 4236- radv: write availability status vkGetQueryPoolResults() when the data 4237 is not available 4238 4239Samuel Pitoiset (147): 4240 4241- radv/winsys: fix hash when adding internal buffers 4242- radv: fix build 4243- radv: bail out when no image transitions will be performed 4244- radv: remove unused radv_render_pass_attachment::view_mask 4245- radv: remove useless MAYBE_UNUSED in CmdBeginRenderPass() 4246- radv: add radv_cmd_buffer_begin_subpass() helper 4247- radv: move subpass image transitions to 4248 radv_cmd_buffer_begin_subpass() 4249- radv: store the list of attachments for every subpass 4250- radv: use the new attachments array when starting subpasses 4251- radv: determine the last subpass id for every attachments 4252- radv: handle final layouts at end of every subpass and render pass 4253- radv: move some render pass things to radv_render_pass_compile() 4254- radv: add radv_render_pass_add_subpass_dep() helper 4255- radv: track if subpasses have color attachments 4256- radv: handle subpass dependencies correctly 4257- radv: accumulate all ingoing external dependencies to the first 4258 subpass 4259- radv: execute external subpass barriers after ending subpasses 4260- radv: drop useless checks when resolving subpass color attachments 4261- radv: do not set preserveAttachments for internal render passes 4262- radv: don't flush src stages when dstStageMask == BOTTOM_OF_PIPE 4263- radv: fix compiler issues with GCC 9 4264- radv: gather more info about push constants 4265- radv: gather if shaders load dynamic offsets separately 4266- radv: keep track of the number of remaining user SGPRs 4267- radv: add support for push constants inlining when possible 4268- radv: fix using LOAD_CONTEXT_REG with old GFX ME firmwares on GFX8 4269- radv/winsys: fix BO list creation when RADV_DEBUG=allbos is set 4270- radv: always export gl_SampleMask when the fragment shader uses it 4271- ac: make use of ac_build_expand_to_vec4() in visit_image_store() 4272- radv: use MAX_{VBS,VERTEX_ATTRIBS} when defining max vertex input 4273 limits 4274- radv: store vertex attribute formats as pipeline keys 4275- radv: reduce the number of loaded channels for vertex input fetches 4276- radv: fix radv_fixup_vertex_input_fetches() 4277- radv: fix invalid element type when filling vertex input default 4278 values 4279- ac: add ac_build_llvm8_tbuffer_load() helper 4280- ac: use new LLVM 8 intrinsic when loading 16-bit values 4281- radv: write the alpha channel of MRT0 when alpha coverage is enabled 4282- radv: remove unused variable in gather_push_constant_info() 4283- radv: fix writing the alpha channel of MRT0 when alpha coverage is 4284 enabled 4285- radv: fix clearing attachments in secondary command buffers 4286- radv: fix out-of-bounds access when copying descriptors BO list 4287- radv: don't copy buffer descriptors list for samplers 4288- rav: use 32_AR instead of 32_ABGR when alpha coverage is required 4289- radv: allocate enough space in cmdbuf when starting a subpass 4290- radv: properly align the fence and EOP bug VA on GFX9 4291- radv: enable lower_mul_2x32_64 4292- Revert "radv: execute external subpass barriers after ending 4293 subpasses" 4294- radv: fix pointSizeRange limits 4295- radv: set the maximum number of IBs per submit to 192 4296- ac: rework typed buffers loads for LLVM 7 4297- radv: store more vertex attribute infos as pipeline keys 4298- radv: use typed buffer loads for vertex input fetches 4299- ac: add ac_build_{struct,raw}_tbuffer_load() helpers 4300- ac: use the raw tbuffer version for 16-bit SSBO loads 4301- radv: always initialize HTILE when the src layout is UNDEFINED 4302- radv: always load 3 channels for formats that need to be shuffled 4303- ac: use llvm.amdgcn.fract intrinsic for nir_op_ffract 4304- radv: fix binding transform feedback buffers 4305- ac: make use of ac_get_store_intr_attribs() where possible 4306- ac/nir: set attrib flags for SSBO and image store operations 4307- ac: add ac_build_buffer_store_format() helper 4308- ac/nir: remove one useless check in visit_store_ssbo() 4309- ac/nir: use new LLVM 8 intrinsics for SSBO atomic operations 4310- ac/nir: use ac_build_buffer_load() for SSBO load operations 4311- ac/nir: use ac_build_buffer_store_dword() for SSBO store operations 4312- ac: use new LLVM 8 intrinsics in ac_build_buffer_load() 4313- ac: add ac_build_{struct,raw}_tbuffer_store() helpers 4314- ac: use new LLVM 8 intrinsic when storing 16-bit values 4315- ac: use new LLVM 8 intrinsics in ac_build_buffer_store_dword() 4316- ac: add various int8 definitions 4317- ac: add ac_build_tbuffer_load_byte() helper 4318- ac: add ac_build_tbuffer_store_byte() helper 4319- radv: add missing initializations since 4320 VK_EXT_pipeline_creation_feedback 4321- ac: add f16_0 and f16_1 constants 4322- ac: add 16-bit support fo fsign 4323- ac: add 16-bit support to fract 4324- ac: fix 16-bit shifts 4325- ac: fix incorrect argument type for tbuffer.{load,store} with LLVM 7 4326- nir: use generic float types for frexp_exp and frexp_sig 4327- spirv,nir: lower frexp_exp/frexp_sig inside a new NIR pass 4328- nir: add nir_{load,store}_deref_with_access() helpers 4329- spirv: propagate the access flag for store and load derefs 4330- ac: use llvm.amdgcn.fmed3 intrinsic for nir_op_fmed3 4331- ac: add ac_build_frexp_mant() helper and 16-bit/32-bit support 4332- ac: add ac_build_frex_exp() helper ans 16-bit/32-bit support 4333- radv: do not lower frexp_exp and frexp_sig 4334- radv: enable VK_AMD_gpu_shader_int16 4335- radv: skip updating depth/color metadata for conditional rendering 4336- radv: do not always initialize HTILE in compressed state 4337- ac: fix return type for llvm.amdgcn.frexp.exp.i32.64 4338- ac/nir: fix nir_op_b2i16 4339- ac: fix ac_build_bit_count() for 16-bit integer type 4340- ac: fix ac_build_bitfield_reverse() for 16-bit integer type 4341- ac: fix ac_find_lsb() for 16-bit integer type 4342- ac: fix ac_build_umsb() for 16-bit integer type 4343- ac/nir: add support for nir_op_b2i8 4344- ac: add 8-bit support to ac_build_bit_count() 4345- ac: add 8-bit support to ac_find_lsb() 4346- ac: add 8-bit support to ac_build_umsb() 4347- ac: add 8-bit and 64-bit support to ac_build_bitfield_reverse() 4348- radv: partially enable VK_KHR_shader_float16_int8 4349- nir: do not pack varying with different types 4350- ac/nir: fix intrinsic names for atomic operations with LLVM 9+ 4351- radv: fix getting the vertex strides if the bindings aren't 4352 contiguous 4353- ac/nir: fix nir_op_b2f16 4354- radv: enable VK_AMD_gpu_shader_half_float 4355- wsi: allow to override the present mode with MESA_VK_WSI_PRESENT_MODE 4356- ac/nir: make use of ac_build_imax() where possible 4357- ac/nir: make use of ac_build_imin() where possible 4358- ac/nir: make use of ac_build_umin() where possible 4359- ac: add ac_build_umax() and use it where possible 4360- ac: add ac_build_ddxy_interp() helper 4361- ac: add ac_build_load_helper_invocation() helper 4362- ac/nir: remove useles LLVMGetUndef for nir_op_pack_64_2x32_split 4363- ac/nir: remove useless integer cast in 4364 adjust_sample_index_using_fmask() 4365- ac/nir: remove useless integer cast in visit_image_load() 4366- ac/nir: remove some useless integer casts for ALU operations 4367- spirv: add SpvCapabilityFloat16 support 4368- radv: enable VK_KHR_shader_float16_int8 4369- radv: set ACCESS_NON_READABLE on stores for copy/fill/clear meta 4370 shaders 4371- radv: enable shaderInt8 on SI and CIK 4372- radv: sort the shader capabilities alphabetically 4373- ac/nir: use new LLVM 8 intrinsics for SSBO atomics except cmpswap 4374- ac/nir: add 64-bit SSBO atomic operations support 4375- radv: add VK_KHR_shader_atomic_int64 but disable it for now 4376- ac: add support for more types with struct/raw LLVM intrinsics 4377- ac: use struct/raw load intrinsics for 8-bit/16-bit int with LLVM 9+ 4378- ac: use struct/raw store intrinsics for 8-bit/16-bit int with LLVM 9+ 4379- ac/nir: only use the new raw/struct image atomic intrinsics with LLVM 4380 9+ 4381- ac/nir: only use the new raw/struct SSBO atomic intrinsics with LLVM 4382 9+ 4383- ac/nir: use the new raw/struct SSBO atomic intrisics for comp_swap 4384- radv: add VK_NV_compute_shader_derivates support 4385- radv: add missing VEGA20 chip in radv_get_device_name() 4386- radv: do not need to force emit the TCS regs on Vega20 4387- radv: fix color conversions for normalized uint/sint formats 4388- radv: implement a workaround for VK_EXT_conditional_rendering 4389- ac: tidy up ac_build_llvm8_tbuffer_{load,store} 4390- radv: set WD_SWITCH_ON_EOP=1 when drawing primitives from a stream 4391 output buffer 4392- radv: only need to force emit the TCS regs on Vega10 and Raven1 4393- radv: fix radv_get_aspect_format() for D+S formats 4394- radv: apply the indexing workaround for atomic buffer operations on 4395 GFX9 4396- radv: fix setting the number of rectangles when it's dyanmic 4397- radv: add a workaround for Monster Hunter World and LLVM 7&8 4398- radv: allocate more space in the CS when emitting events 4399- radv: do not use gfx fast depth clears for layered depth/stencil 4400 images 4401- radv: fix alpha-to-coverage when there is unused color attachments 4402- radv: fix setting CB_SHADER_MASK for dual source blending 4403 4404Sergii Romantsov (4): 4405 4406- dri: meson: do not prefix user provided dri-drivers-path 4407- d3d: meson: do not prefix user provided d3d-drivers-path 4408- i965,iris/blorp: do not blit 0-sizes 4409- glsl: Fix input/output structure matching across shader stages 4410 4411Sonny Jiang (1): 4412 4413- radeonsi: use compute for clear_render_target when possible 4414 4415Tapani Pälli (42): 4416 4417- nir: add option to use scaling factor when sampling planes YUV 4418 lowering 4419- dri: add P010, P012, P016 for 10bit/12bit/16bit YUV420 formats 4420- intel/compiler: add scale_factors to sampler_prog_key_data 4421- i965: add P0x formats and propagate required scaling factors 4422- drirc/i965: add option to disable 565 configs and visuals 4423- mesa: return NULL if we exceed MaxColorAttachments in 4424 get_fb_attachment 4425- anv: anv: refactor error handling in anv_shader_bin_write_to_blob() 4426- iris: add Android build 4427- nir: initialize value in copy_prop_vars_block 4428- nir: use nir_variable_create instead of open-coding the logic 4429- android: add liblog to libmesa_intel_common build 4430- android: make libbacktrace optional on USE_LIBBACKTRACE 4431- iris: add libmesa_iris_gen8 library to the build 4432- util: fix a warning when building against clang7 headers 4433- anv: retain the is_array state in create_plane_tex_instr_implicit 4434- anv: toggle on support for VK_EXT_ycbcr_image_arrays 4435- anv: use anv_gem_munmap in block pool cleanup 4436- anv: call blob_finish when done with it 4437- nir: free dead_ctx in case of no progress 4438- anv: destroy descriptor sets when pool gets destroyed 4439- anv: release memory allocated by bo_heap when descriptor pool is 4440 destroyed 4441- anv: release memory allocated by glsl types during spirv_to_nir 4442- anv: revert "anv: release memory allocated by glsl types during 4443 spirv_to_nir" 4444- i965: remove scaling factors from P010, P012 4445- isl: fix automake build when sse41 is not supported 4446- android: Build fixes for OMR1 4447- iris: initialize num_cbufs 4448- iris: mark switch case fallthrough 4449- anv/radv: release memory allocated by glsl types during spirv_to_nir 4450- st/mesa: fix compilation warning on storage_flags_to_buffer_flags 4451- st/mesa: fix warnings about implicit conversion on enumeration type 4452- spirv: fix a compiler warning 4453- st/nir: run st_nir_opts after 64bit ops lowering 4454- iris: move variable to the scope where it is being used 4455- iris: move iris_flush_resource so we can call it from get_handle 4456- iris: handle aux properly in iris_resource_get_handle 4457- egl: setup fds array correctly when exporting dmabuf 4458- compiler/glsl: handle case where we have multiple users for types 4459- android/iris: fix driinfo header filename 4460- nir: use braces around subobject in initializer 4461- glsl: use empty brace initializer 4462- anv: expose VK_EXT_queue_family_foreign on Android 4463 4464Thomas Hellstrom (5): 4465 4466- winsys/svga: Add an environment variable to force host-backed 4467 operation 4468- winsys/svga: Enable the transfer_from_buffer GPU command for vgpu10 4469- svga: Avoid bouncing buffer data in malloced buffers 4470- winsys/svga: Update the drm interface file 4471- winsys/svga: Don't abort on EBUSY errors from execbuffer 4472 4473Timo Aaltonen (1): 4474 4475- util/os_misc: Add check for PIPE_OS_HURD 4476 4477Timothy Arceri (72): 4478 4479- st/glsl_to_nir: remove dead local variables 4480- ac/radv/radeonsi: add ac_get_num_physical_sgprs() helper 4481- radv: take LDS into account for compute shader occupancy stats 4482- util: move BITFIELD macros to util/macros.h 4483- st/glsl_to_nir: call nir_remove_dead_variables() after lowing local 4484 indirects 4485- nir: add support for marking used patches when packing varyings 4486- nir: add glsl_type_is_32bit() helper 4487- nir: add is_packing_supported_for_type() helper 4488- nir: rewrite varying component packing 4489- nir: prehash instruction in nir_instr_set_add_or_rewrite() 4490- nir: turn ssa check into an assert 4491- nir: turn an ssa check in nir_search into an assert 4492- nir: remove simple dead if detection from nir_opt_dead_cf() 4493- radeonsi/nir: set input_usage_mask properly 4494- radeonsi/nir: set colors_read properly 4495- radeonsi/nir: set shader_buffers_declared properly 4496- st/nir: use NIR for asm programs 4497- nir: remove non-ssa support from nir_copy_prop() 4498- nir: clone instruction set rather than removing individual entries 4499- nir: allow nir_lower_phis_to_scalar() on more src types 4500- radeonsi: fix query buffer allocation 4501- glsl: fix shader cache for packed param list 4502- radeonsi/nir: move si_lower_nir() call into compiler thread 4503- glsl: rename is_record() -> is_struct() 4504- glsl: rename get_record_instance() -> get_struct_instance() 4505- glsl: rename record_location_offset() -> struct_location_offset() 4506- glsl: rename record_types -> struct_types 4507- nir: rename glsl_type_is_struct() -> glsl_type_is_struct_or_ifc() 4508- glsl/freedreno/panfrost: pass gl_context to the standalone compiler 4509- glsl: use NIR function inlining for drivers that use glsl_to_nir() 4510- i965: stop calling nir_lower_returns() 4511- radeonsi/nir: stop calling nir_lower_returns() 4512- st/glsl: start spilling out common st glsl conversion code 4513- anv: add support for dumping shader info via VK_EXT_debug_report 4514- nir: add guess trip count support to loop analysis 4515- nir: add new partially_unrolled bool to nir_loop 4516- nir: add partial loop unrolling support 4517- nir: calculate trip count for more loops 4518- nir: unroll some loops with a variable limit 4519- nir: simplify the loop analysis trip count code a little 4520- nir: add helper to return inversion op of a comparison 4521- nir: add get_induction_and_limit_vars() helper to loop analysis 4522- nir: pass nir_op to calculate_iterations() 4523- nir: find induction/limit vars in iand instructions 4524- st/glsl_to_nir: fix incorrect arrary access 4525- radeonsi/nir: call some more var optimisation passes 4526- ac/nir_to_llvm: add assert to emit_bcsel() 4527- nir: only override previous alu during loop analysis if supported 4528- nir: fix opt_if_loop_last_continue() 4529- nir: add support for user defined loop control 4530- spirv: make use of the loop control support in nir 4531- nir: add support for user defined select control 4532- spirv: make use of the select control support in nir 4533- Revert "ac/nir: use new LLVM 8 intrinsics for SSBO atomic operations" 4534- nir: propagate known constant values into the if-then branch 4535- Revert "nir: propagate known constant values into the if-then branch" 4536- nir/radv: remove restrictions on opt_if_loop_last_continue() 4537- nir: initialise some variables in opt_if_loop_last_continue() 4538- nir/i965/freedreno/vc4: add a bindless bool to type size functions 4539- ac/nir_to_llvm: make get_sampler_desc() more generic and pass it the 4540 image intrinsic 4541- ac/nir_to_llvm: add image bindless support 4542- nir: fix packing components with arrays 4543- radeonsi/nir: fix scanning of bindless images 4544- st/mesa/radeonsi: fix race between destruction of types and shader 4545 compilation 4546- nir: fix nir_remove_unused_varyings() 4547- radeonsi/nir: create si_nir_opts() helper 4548- radeonsi/nir: call radeonsi nir opts before the scan pass 4549- util/drirc: add workarounds for bugs in Doom 3: BFG 4550- radeonsi: add config entry for Counter-Strike Global Offensive 4551- Revert "glx: Fix synthetic error generation in \__glXSendError" 4552- Revert "st/mesa: expose 0 shader binary formats for compat profiles 4553 for Qt" 4554- st/glsl: make sure to propagate initialisers to driver storage 4555 4556Timur Kristóf (19): 4557 4558- radeonsi/nir: Use uniform location when calculating const_file_max. 4559- iris: implement clearing render target and depth stencil 4560- nir: Add ability for shaders to use window space coordinates. 4561- tgsi_to_nir: Fix the TGSI ARR translation by converting the result to 4562 int. 4563- tgsi_to_nir: Fix TGSI LIT translation by using flt. 4564- tgsi_to_nir: Make the TGSI IF translation code more readable. 4565- tgsi_to_nir: Split to smaller functions. 4566- nir: Move nir_lower_uniforms_to_ubo to compiler/nir. 4567- nir: Add multiplier argument to nir_lower_uniforms_to_ubo. 4568- freedreno: Plumb pipe_screen through to irX_tgsi_to_nir. 4569- tgsi_to_nir: Produce optimized NIR for a given pipe_screen. 4570- tgsi_to_nir: Restructure system value loads. 4571- tgsi_to_nir: Extract ttn_emulate_tgsi_front_face into its own 4572 function. 4573- tgsi_to_nir: Support FACE and POSITION properly. 4574- tgsi_to_nir: Improve interpolation modes. 4575- tgsi_to_nir: Set correct location for uniforms. 4576- radeonsi/nir: Only set window_space_position for vertex shaders. 4577- iris: Face should be a system value. 4578- gallium: fix autotools build of pipe_msm.la 4579 4580Tobias Klausmann (1): 4581 4582- vulkan/util: meson build - add wayland client include 4583 4584Tomasz Figa (1): 4585 4586- llvmpipe: Always return some fence in flush (v2) 4587 4588Tomeu Vizoso (19): 4589 4590- panfrost: Add gem_handle to panfrost_memory and panfrost_bo 4591- panfrost: Add backend targeting the DRM driver 4592- panfrost/midgard: Add support for MIDGARD_MESA_DEBUG 4593- panfrost: Add support for PAN_MESA_DEBUG 4594- panfrost: Set bo->size[0] in the DRM backend 4595- panfrost: Set bo->gem_handle when creating a linear BO 4596- panfrost: Adapt to uapi changes 4597- panfrost: Fix sscanf format options 4598- panfrost: Set the GEM handle for AFBC buffers 4599- panfrost: Also tell the kernel about the checksum_slab 4600- panfrost: Pass the context BOs to the kernel so they aren't unmapped 4601 while in use 4602- panfrost: Wait for last job to finish in force_flush_fragment 4603- panfrost: split asserts in pandecode 4604- panfrost: Guard against reading past end of buffer 4605- panfrost/ci: Initial commit 4606- panfrost/midgard: Skip register allocation if there's no work to do 4607- panfrost/midgard: Skip liveness analysis for instructions without 4608 dest 4609- panfrost: Fix two uninitialized accesses in compiler 4610- panfrost: Only take the fast paths on buffers aligned to block size 4611 4612Toni Lönnberg (8): 4613 4614- intel/genxml: Only handle instructions meant for render engine when 4615 generating headers 4616- intel/genxml: Media instructions and structures for gen6 4617- intel/genxml: Media instructions and structures for gen7 4618- intel/genxml: Media instructions and structures for gen7.5 4619- intel/genxml: Media instructions and structures for gen8 4620- intel/genxml: Media instructions and structures for gen9 4621- intel/genxml: Media instructions and structures for gen10 4622- intel/genxml: Media instructions and structures for gen11 4623 4624Topi Pohjolainen (2): 4625 4626- intel/compiler/icl: Use tcs barrier id bits 24:30 instead of 24:27 4627- intel/compiler/fs/icl: Use dummy masked urb write for tess eval 4628 4629Vasily Khoruzhick (2): 4630 4631- lima: use individual tile heap for each GP job. 4632- lima: add support for depth/stencil fbo attachments and textures 4633 4634Vinson Lee (5): 4635 4636- gallium/auxiliary/vl: Fix duplicate symbol build errors. 4637- nir: Fix anonymous union initialization with older GCC. 4638- swr: Fix build with llvm-9.0. 4639- gallium: Fix autotools build with libxatracker.la. 4640- freedreno: Fix GCC build error. 4641 4642Vivek Kasireddy (1): 4643 4644- drm-uapi: Update headers from drm-next 4645 4646Xavier Bouchoux (1): 4647 4648- nir/spirv: Fix assert when unsampled OpTypeImage has unknown 'Depth' 4649 4650Yevhenii Kolesnikov (1): 4651 4652- i965: Fix allow_higher_compat_version workaround limited by OpenGL 4653 3.0 4654 4655coypu (1): 4656 4657- gbm: don't return void 4658 4659davidbepo (1): 4660 4661- drirc: add Waterfox to adaptive-sync blacklist 4662 4663grmat (1): 4664 4665- drirc: add Spectacle, Falkon to a-sync blacklist 4666 4667pal1000 (1): 4668 4669- scons: Compatibility with Scons development version string 4670 4671suresh guttula (3): 4672 4673- vl: Add cropping flags for H264 4674- radeon/vce:Add support for frame_cropping_flag of 4675 VAEncSequenceParameterBufferH264 4676- st/va/enc: Add support for frame_cropping_flag of 4677 VAEncSequenceParameterBufferH264 4678