2018-11-01 18:56:54 +00:00
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
< html lang = "en" >
< head >
< meta http-equiv = "content-type" content = "text/html; charset=utf-8" >
< title > Mesa Release Notes< / title >
< link rel = "stylesheet" type = "text/css" href = "../mesa.css" >
< / head >
< body >
< div class = "header" >
2019-05-06 12:26:47 +01:00
The Mesa 3D Graphics Library
2018-11-01 18:56:54 +00:00
< / div >
< iframe src = "../contents.html" > < / iframe >
< div class = "content" >
< h1 > Mesa 19.0.0 Release Notes / TBD< / h1 >
< p >
Mesa 19.0.0 is a new development release. People who are concerned
with stability and reliability should stick with a previous release or
wait for Mesa 19.0.1.
< / p >
< p >
Mesa 19.0.0 implements the OpenGL 4.5 API, but the version reported by
glGetString(GL_VERSION) or glGetIntegerv(GL_MAJOR_VERSION) /
glGetIntegerv(GL_MINOR_VERSION) depends on the particular driver being used.
Some drivers don't support all the features required in OpenGL 4.5. OpenGL
4.5 is < strong > only< / strong > available if requested at context creation.
Compatibility contexts may report a lower version depending on each driver.
< / p >
< h2 > SHA256 checksums< / h2 >
< pre >
2019-03-13 19:09:08 +00:00
4c5b9c5227d37c1f6bdc786a6fa7ee7fbce40b2e8a87340c7d3234534ece3304 mesa-19.0.0.tar.gz
5a549dfb40ec31e5c36c47aadac04554cb2e2a8d144a046a378fc16da57e38f8 mesa-19.0.0.tar.xz
2018-11-01 18:56:54 +00:00
< / pre >
< h2 > New features< / h2 >
< ul >
2018-08-06 07:48:12 +01:00
< li > GL_AMD_texture_texture4 on all GL 4.0 drivers.< / li >
2018-11-02 15:48:26 +00:00
< li > GL_EXT_shader_implicit_conversions on all drivers (ES extension).< / li >
2018-08-06 05:29:00 +01:00
< li > GL_EXT_texture_compression_bptc on all GL 4.0 drivers (ES extension).< / li >
2018-08-06 03:50:54 +01:00
< li > GL_EXT_texture_compression_rgtc on all GL 3.0 drivers (ES extension).< / li >
2018-11-29 20:34:05 +00:00
< li > GL_EXT_render_snorm on gallium drivers (ES extension).< / li >
2018-11-28 01:42:44 +00:00
< li > GL_EXT_texture_view on drivers supporting texture views (ES extension).< / li >
2018-11-28 01:19:11 +00:00
< li > GL_OES_texture_view on drivers supporting texture views (ES extension).< / li >
2018-12-27 01:01:28 +00:00
< li > GL_NV_shader_atomic_float on nvc0 (Fermi/Kepler only).< / li >
2018-10-18 17:51:51 +01:00
< li > Shader-based software implementations of GL_ARB_gpu_shader_fp64, GL_ARB_gpu_shader_int64, GL_ARB_vertex_attrib_64bit, and GL_ARB_shader_ballot on i965.< / li >
2019-01-19 19:10:15 +00:00
< li > VK_ANDROID_external_memory_android_hardware_buffer on Intel< / li >
< li > Fixed and re-exposed VK_EXT_pci_bus_info on Intel and RADV< / li >
< li > VK_EXT_scalar_block_layout on Intel and RADV< / li >
< li > VK_KHR_depth_stencil_resolve on Intel< / li >
< li > VK_KHR_draw_indirect_count on Intel< / li >
< li > VK_EXT_conditional_rendering on Intel< / li >
< li > VK_EXT_memory_budget on RADV< / li >
2018-11-01 18:56:54 +00:00
< / ul >
< h2 > Bug fixes< / h2 >
< ul >
2019-03-13 17:40:19 +00:00
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=32211" > Bug 32211< / a > - [GLSL] lower_jumps with continue-statements in for-loops prevents loop unrolling< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=102349" > Bug 102349< / a > - nv4x crashing with plasmashell - gdb log included< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=102597" > Bug 102597< / a > - [Regression] mpv, high rendering times (two to three times higher)< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=104297" > Bug 104297< / a > - [i965] Downward causes GPU hangs and misrendering on Haswell< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=104602" > Bug 104602< / a > - [apitrace] Graphical artifacts in Civilization VI on RX Vega< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=105301" > Bug 105301< / a > - The big SKQP bug< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=106577" > Bug 106577< / a > - broken rendering with nine and nouveau (GM107)< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=106595" > Bug 106595< / a > - [RADV] Rendering distortions only when MSAA is enabled< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=107052" > Bug 107052< / a > - [Regression][bisected]. Crookz - The Big Heist Demo can't be launched despite the " true" flag in " drirc" < / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=107510" > Bug 107510< / a > - [GEN8+] up to 10% perf drop on several 3D benchmarks< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=107626" > Bug 107626< / a > - [SNB] The graphical corruption and GPU hang occur sometimes on the piglit test " arb_texture_multisample-large-float-texture" with parameter --fp16< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=107728" > Bug 107728< / a > - Wrong background in Sascha Willem's Multisampling Demo< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=107842" > Bug 107842< / a > - " invariant" qualifier on outputs of GLSL ES fragment shader causes compilation error.< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=107856" > Bug 107856< / a > - i965 incorrectly calculates the number of layers for texture views (assert)< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=108114" > Bug 108114< / a > - [vulkancts] new VK_KHR_16bit_storage tests fail.< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=108116" > Bug 108116< / a > - [vulkancts] stencil partial clear tests fail.< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=108245" > Bug 108245< / a > - RADV/Vega: Low mip levels of large BCn textures get corrupted by vkCmdCopyBufferToImage< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=108311" > Bug 108311< / a > - Query buffer object support is broken on r600.< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=108457" > Bug 108457< / a > - [OpenGL CTS] KHR-GL46.tessellation_shader.single.xfb_captures_data_from_correct_stage fails< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=108560" > Bug 108560< / a > - Mesa 32 is built without sse< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=108624" > Bug 108624< / a > - [regression][bisected] " nir: Copy propagation between blocks" regression< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=108630" > Bug 108630< / a > - [G965] piglit.spec.!opengl 1_2.tex3d-maxsize spins forever< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=108635" > Bug 108635< / a > - Mesa master commit 68dc591af16ebb36814e4c187e4998948103c99c causes XWayland to segfault< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=108636" > Bug 108636< / a > - test_optpass has use after free bug, failing with memory testing tools like address sanitizer< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=108713" > Bug 108713< / a > - Gallium: use after free with transform feedback< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=108734" > Bug 108734< / a > - Regression: [bisected] dEQP-GLES31.functional.tessellation.invariance.* start failing on r600< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=108805" > Bug 108805< / a > - i965 regressions from EXT_texture_sRGB_R8< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=108829" > Bug 108829< / a > - [meson] libglapi exports internal API< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=108868" > Bug 108868< / a > - [BYT IVB] Tesselation test regressions< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=108877" > Bug 108877< / a > - OpenGL CTS gl43 test cases were interrupted due to segment fault< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=108894" > Bug 108894< / a > - [anv] vkCmdCopyBuffer() and vkCmdCopyQueryPoolResults() write-after-write hazard< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=108909" > Bug 108909< / a > - Vkd3d test failure test_resolve_non_issued_query_data()< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=108910" > Bug 108910< / a > - Vkd3d test failure test_multisample_array_texture()< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=108911" > Bug 108911< / a > - Vkd3d test failure test_clear_render_target_view()< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=108914" > Bug 108914< / a > - blocky shadow artifacts in The Forest with DXVK, RADV_DEBUG=nohiz fixes this< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=108925" > Bug 108925< / a > - vkCmdCopyQueryPoolResults(VK_QUERY_RESULT_WAIT_BIT) for timestamps with large query count hangs< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=108936" > Bug 108936< / a > - [ILK,G45,G965] Regressions from texture-format enums rework< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=108943" > Bug 108943< / a > - Build fails on ppc64le with meson< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=108961" > Bug 108961< / a > - make check test_replace_src_bitsize failure< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=108974" > Bug 108974< / a > - make check DispatchSanity_test regression< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=108999" > Bug 108999< / a > - Calculating the scissors fields when the y is flipped (0 on top) can generate negative numbers that will cause assertion failure later on.< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109023" > Bug 109023< / a > - error: inlining failed in call to always_inline ‘ __m512 _mm512_and_ps(__m512, __m512)’ : target specific option mismatch< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109072" > Bug 109072< / a > - GPU hang in blender 2.80< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109075" > Bug 109075< / a > - radv: New D3D boolean optimizations cause GPU hang in Witcher 3< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109081" > Bug 109081< / a > - [bisected] [HSW] Regression in clipping.user_defined.clip_* vulkancts tests< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109086" > Bug 109086< / a > - Crash software mesa with gl_select render mode< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109107" > Bug 109107< / a > - gallium/st/va: change va max_profiles when using Radeon VCN Hardware< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109129" > Bug 109129< / a > - format_types.h:1220: undefined reference to `_mm256_cvtps_ph'< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109151" > Bug 109151< / a > - [KBL-G][vulkan] dEQP-VK.texture.explicit_lod.2d.sizes.31x55_nearest_linear_mipmap_nearest_repeat failed verification.< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109190" > Bug 109190< / a > - virgl: buffer flushing error with some dEQP tests [bisected]< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109202" > Bug 109202< / a > - nv50_ir.cpp:749:19: error: cannot use typeid with -fno-rtti< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109204" > Bug 109204< / a > - [regression, bisected] retroarch's crt-royale shader crash radv< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109229" > Bug 109229< / a > - glLinkProgram locks up for ~30 seconds< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109231" > Bug 109231< / a > - [nir] src/compiler/nir/nir_loop_analyze.c uninitialized variable< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109242" > Bug 109242< / a > - [RADV] The Witcher 3 system freeze< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109304" > Bug 109304< / a > - GfxBench AztecRuins Vulkan version Segfault< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109325" > Bug 109325< / a > - mesa: Need ability to retrieve command line of Meson configuration< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109328" > Bug 109328< / a > - [BSW BXT GLK] dEQP-VK.subgroups.arithmetic.subgroup regressions< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109353" > Bug 109353< / a > - [regression][bisected] " nir: Switch to using 1-bit Booleans for almost everything" regression with shared bools< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109401" > Bug 109401< / a > - [DXVK] Project Cars rendering problems< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109404" > Bug 109404< / a > - [ANV] The Witcher 3 shadows flickering< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109442" > Bug 109442< / a > - " make check" test anv_block_pool_no_free fails intermittently< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109443" > Bug 109443< / a > - Build failure with MSVC when using Scons > = 3.0.2< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109449" > Bug 109449< / a > - [snb] quakespasm triggers a segmentation fault.< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109451" > Bug 109451< / a > - [IVB,SNB] LINE_STRIPs following a TRIANGLE_FAN fail to use primitive restart< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109543" > Bug 109543< / a > - After upgrade mesa to 19.0.0~rc1 all vulkan based application stop working [" vulkan-cube" received SIGSEGV in radv_pipeline_init_blend_state at ../src/amd/vulkan/radv_pipeline.c:699]< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109561" > Bug 109561< / a > - [regression, bisected] code re-factor causing games to stutter or lock-up system< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109573" > Bug 109573< / a > - dEQP-VK.spirv_assembly.instruction.graphics.module.same_module< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109575" > Bug 109575< / a > - Mesa-19.0.0-rc1 : Computer Crashes trying to run anything Vulkan< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109581" > Bug 109581< / a > - [BISECTED] Nothing is Rendered on Sascha Willem's " subpasses" demo< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109594" > Bug 109594< / a > - totem assert failure: totem: src/intel/genxml/gen9_pack.h:72: __gen_uint: La declaración `v < = max' no se cumple.< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109597" > Bug 109597< / a > - wreckfest issues with transparent objects & skybox< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109601" > Bug 109601< / a > - [Regression] RuneLite GPU rendering broken on 18.3.x< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109603" > Bug 109603< / a > - nir_instr_as_deref: Assertion `parent & & parent-> type == nir_instr_type_deref' failed.< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109698" > Bug 109698< / a > - dri.pc contents invalid when built with meson< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109717" > Bug 109717< / a > - [regression] Cull distance tests asserting< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109735" > Bug 109735< / a > - [Regression] broken font with mesa_vulkan_overlay< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=109759" > Bug 109759< / a > - [BISECTED][REGRESSION][IVB, HSW] Font rendering problem in OpenGL< / li >
2018-11-01 18:56:54 +00:00
< / ul >
< h2 > Changes< / h2 >
2019-03-13 17:40:19 +00:00
< p > Adam Jackson (4):< / p >
< ul >
< li > glx: Demand success from CreateContext requests (v2)< / li >
< li > specs: Remove GLES profile interaction text from GLX_MESA_query_renderer< / li >
< li > specs: Remove GLX_RENDERER_ID_MESA from GLX_MESA_query_renderer< / li >
< li > specs: Bump GLX_MESA_query_renderer to version 9< / li >
< / ul >
< p > Aditya Swarup (1):< / p >
< ul >
< li > i965: Lift restriction in external textures for EGLImage support< / li >
< / ul >
< p > Alejandro Piñeiro (3):< / p >
< ul >
< li > nir: remove unused variable< / li >
< li > nir/xfb: don't assert when xfb_buffer/stride is present but not xfb_offset< / li >
< li > nir/xfb: distinguish array of structs vs array of blocks< / li >
< / ul >
< p > Alex Deucher (3):< / p >
< ul >
< li > pci_ids: add new vega10 pci ids< / li >
< li > pci_ids: add new vega20 pci id< / li >
< li > pci_ids: add new VegaM pci id< / li >
< / ul >
< p > Alex Smith (1):< / p >
< ul >
< li > radv: Flush before vkCmdWriteTimestamp() if needed< / li >
< / ul >
< p > Alexander von Gluck IV (1):< / p >
< ul >
< li > egl/haiku: Fix reference to disp vs dpy< / li >
< / ul >
< p > Alok Hota (8):< / p >
< ul >
< li > swr/rast: Use gfxptr_t value in JitGatherVertices< / li >
< li > swr/rast: Add annotator to interleave isa text< / li >
< li > swr/rast: partial support for Tiled Resources< / li >
< li > swr/rast: Unaligned and translations in gathers< / li >
< li > swr/rast: Scope MEM_CLIENT enum for mem usages< / li >
< li > swr/rast: New execution engine per JIT< / li >
< li > swr/rast: Store cached files in multiple subdirs< / li >
< li > swr/rast: bypass size limit for non-sampled textures< / li >
< / ul >
< p > Alyssa Rosenzweig (1):< / p >
< ul >
< li > util: Fix warning in u_cpu_detect on non-x86< / li >
< / ul >
< p > Andre Heider (4):< / p >
< ul >
< li > st/nine: fix stack corruption due to ABI mismatch< / li >
< li > st/nine: plug thread related leaks< / li >
< li > st/nine: clean up thead shutdown sequence a bit< / li >
< li > d3dadapter9: use snprintf(..., "%s", ...) instead of strncpy< / li >
< / ul >
< p > Andres Gomez (8):< / p >
< ul >
< li > glsl/linker: complete documentation for assign_attribute_or_color_locations< / li >
< li > docs: update 18.3 and add 19.x cycles for the release calendar< / li >
< li > glsl: correct typo in GLSL compilation error message< / li >
< li > editorconfig: Add max_line_length property< / li >
< li > glsl/linker: specify proper direction in location aliasing error< / li >
< li > docs: complete the calendar and release schedule documentation< / li >
< li > bin/get-pick-list.sh: fix the oneline printing< / li >
< li > bin/get-pick-list.sh: fix redirection in sh< / li >
< / ul >
< p > Andrii Simiklit (9):< / p >
< ul >
< li > intel/tools: avoid 'unused variable' warnings< / li >
< li > compiler: avoid 'unused variable' warnings< / li >
< li > i965: avoid 'unused variable' warnings< / li >
< li > i965/batch: avoid reverting batch buffer if saved state is an empty< / li >
< li > intel/tools: make sure the binary file is properly read< / li >
< li > anv/pipeline: remove unnecessary null-pointer check< / li >
< li > intel/batch-decoder: fix vertex buffer size calculation for gen< 8< / li >
< li > intel/batch-decoder: fix a vb end address calculation< / li >
< li > i965: re-emit index buffer state on a reset option change.< / li >
< / ul >
< p > Anuj Phogat (7):< / p >
< ul >
< li > i965/icl: Set Error Detection Behavior Control Bit in L3CNTLREG< / li >
< li > anv/icl: Set Error Detection Behavior Control Bit in L3CNTLREG< / li >
< li > anv/icl: Disable prefetching of sampler state entries< / li >
< li > i965/icl: Fix L3 configurations< / li >
< li > i965/icl: Set use full ways in L3CNTLREG< / li >
< li > intel/icl: Set way_size_per_bank to 4< / li >
< li > anv/icl: Set use full ways in L3CNTLREG< / li >
< / ul >
< p > Axel Davy (12):< / p >
< ul >
< li > st/nine: Allow 'triple buffering' with thread_submit< / li >
< li > st/nine: Remove thread_submit warning< / li >
< li > st/nine: Use helper to release swapchain buffers later< / li >
< li > st/nine: Switch to presentation buffer if resize is detected< / li >
< li > st/nine: Fix volumetexture dtor on ctor failure< / li >
< li > st/nine: Bind src not dst in nine_context_box_upload< / li >
< li > st/nine: Add src reference to nine_context_range_upload< / li >
< li > st/nine: Increase the limit of cached ff shaders< / li >
< li > st/nine: Immediately upload user provided textures< / li >
< li > st/nine: Enable debug info if NDEBUG is not set< / li >
< li > st/nine: Ignore window size if error< / li >
< li > st/nine: Ignore multisample quality level if no ms< / li >
< / ul >
< p > Bart Oldeman (1):< / p >
< ul >
< li > gallium-xlib: query MIT-SHM before using it.< / li >
< / ul >
< p > Bas Nieuwenhuizen (41):< / p >
< ul >
< li > radv: Use structured intrinsics instead of indexing workaround for GFX9.< / li >
< li > vulkan: Allow storage images in the WSI.< / li >
< li > radv: Fix opaque metadata descriptor last layer.< / li >
< li > radv: Clamp gfx9 image view extents to the allocated image extents.< / li >
< li > radv: Align large buffers to the fragment size.< / li >
< li > radv/android: Mark android WSI image as shareable.< / li >
< li > radv/android: Use buffer metadata to determine scanout compat.< / li >
< li > radv: Check for shareable images in central place.< / li >
< li > radv: Remove redundant format check.< / li >
< li > radv: Fix multiview depth clears< / li >
< li > radv: Work around non-renderable 128bpp compressed 3d textures on GFX9.< / li >
< li > radv: Fix wrongly positioned paren.< / li >
< li > radv: Do a cache flush if needed before reading predicates.< / li >
< li > radv: Implement buffer stores with less than 4 components.< / li >
< li > anv/android: Do not reject storage images.< / li >
< li > radv: Remove device path.< / li >
< li > radv: Remove unused variable.< / li >
< li > amd/common: Add some parentheses to silence warning.< / li >
< li > radv: Fix rasterization precision bits.< / li >
< li > spirv: Fix matrix parameters in function calls.< / li >
< li > freedreno: Move register constant files to src/freedreno.< / li >
< li > radv: Only use 32 KiB per threadgroup on Stoney.< / li >
< li > radv: Set partial_vs_wave for pipelines with just GS, not tess.< / li >
< li > nir: Account for atomics in copy propagation.< / li >
< li > radv: Remove unused variable.< / li >
< li > radv/winsys: Set winsys bo priority on creation.< / li >
< li > radv/winsys: Add priority handling during submit.< / li >
< li > radv: Enable VK_EXT_memory_priority.< / li >
< li > radv: Fix the shader info pass for not having the variable.< / li >
< li > amd/common: Fix stores to derefs with unknown variable.< / li >
< li > amd/common: Add gep helper for pointer increment.< / li >
< li > amd/common: Handle nir_deref_type_ptr_as_array for shared memory.< / li >
< li > amd/common: handle nir_deref_cast for shared memory from integers.< / li >
< li > radv: Only look at pImmutableSamples if the descriptor has a sampler.< / li >
< li > amd/common: Use correct writemask for shared memory stores.< / li >
< li > radv: Sync ETC2 whitelisted devices.< / li >
< li > radv: Fix float16 interpolation set up.< / li >
< li > radv: Allow interpolation on non-float types.< / li >
< li > radv: Handle clip+cull distances more generally as compact arrays.< / li >
< li > radv: Fix rebase issue in 19.0 for float16 fix.< / li >
< li > radv: Interpolate less aggressively.< / li >
< / ul >
< p > Boyan Ding (3):< / p >
< ul >
< li > gk110/ir: Add rcp f64 implementation< / li >
< li > gk110/ir: Add rsq f64 implementation< / li >
< li > gk110/ir: Use the new rcp/rsq in library< / li >
< / ul >
< p > Brian Paul (3):< / p >
< ul >
< li > svga: add new gallium formats to the format conversion table< / li >
< li > mesa: fix display list corner case assertion< / li >
< li > svga: remove SVGA_RELOC_READ flag in SVGA3D_BindGBSurface()< / li >
< / ul >
< p > Bruce Cherniak (1):< / p >
< ul >
< li > gallium/swr: Fix multi-context sync fence deadlock.< / li >
< / ul >
< p > Caio Marcelo de Oliveira Filho (10):< / p >
< ul >
< li > nir: properly clear the entry sources in copy_prop_vars< / li >
< li > nir: properly find the entry to keep in copy_prop_vars< / li >
< li > nir: add a way to print the deref chain< / li >
< li > nir: remove dead code from copy_prop_vars< / li >
< li > nir: fix warning in nir_lower_io.c< / li >
< li > util: Helper to create sets and hashes with pointer keys< / li >
< li > src/compiler: use new hash table and set creation helpers< / li >
< li > src/intel: use new hash table and set creation helpers< / li >
< li > nir: check NIR_SKIP to skip passes by name< / li >
< li > gallium: Add PIPE_CAP_GLSL_TESS_LEVELS_AS_INPUTS< / li >
< / ul >
< p > Carlos Garnacho (1):< / p >
< ul >
< li > wayland/egl: Ensure EGL surface is resized on DRI update_buffers()< / li >
< / ul >
< p > Carsten Haitzler (Rasterman) (2):< / p >
< ul >
< li > vc4: Use named parameters for the NEON inline asm.< / li >
< li > vc4: Declare the cpu pointers as being modified in NEON asm.< / li >
< / ul >
< p > Chad Versace (1):< / p >
< ul >
< li > i965: Fix -Wswitch on INTEL_COPY_STREAMING_LOAD< / li >
< / ul >
< p > Chia-I Wu (2):< / p >
< ul >
< li > meson: fix EGL/X11 build without GLX< / li >
< li > freedreno/drm: sync uapi again< / li >
< / ul >
< p > Christian Gmeiner (6):< / p >
< ul >
< li > nir: add lowering for ffloor< / li >
< li > etnaviv: drop redundant ctx function parameter< / li >
< li > meson: add etnaviv to the tools option< / li >
< li > etnaviv: extend etna_resource with an addressing mode< / li >
< li > etnaviv: update headers from rnndb< / li >
< li > etnaviv: add linear sampling support< / li >
< / ul >
< p > Connor Abbott (4):< / p >
< ul >
< li > Revert "radv: disable VK_SUBGROUP_FEATURE_VOTE_BIT"< / li >
< li > nir/algebraic: Rewrite bit-size inference< / li >
< li > nir/algebraic: Add unit tests for bitsize validation< / li >
< li > nir: Fixup algebraic test for variable-sized conversions< / li >
< / ul >
< p > Daniel Stone (1):< / p >
< ul >
< li > gbm: Clarify acceptable formats for gbm_bo< / li >
< / ul >
< p > Danylo Piliaiev (9):< / p >
< ul >
< li > i965: Fix calculation of layers array length for isl_view< / li >
< li > nir: add if opt opt_if_loop_last_continue()< / li >
< li > glsl/linker: Fix unmatched TCS outputs being reduced to local variable< / li >
< li > glsl: Make invariant outputs in ES fragment shader not to cause error< / li >
< li > glsl: Fix copying function's out to temp if dereferenced by array< / li >
< li > anv: Implement VK_KHR_draw_indirect_count for gen 7+< / li >
< li > anv: Implement VK_EXT_conditional_rendering for gen 7.5+< / li >
< li > anv: Fix VK_EXT_transform_feedback working with varyings packed in PSIZ< / li >
< li > anv: Fix destroying descriptor sets when pool gets reset< / li >
< / ul >
< p > Dave Airlie (19):< / p >
< ul >
< li > radv: apply xfb buffer offset at buffer binding time not later. (v2)< / li >
< li > radv: fix begin/end transform feedback with 0 counter buffers.< / li >
< li > virgl: fix vtest regression since fencing changes.< / li >
< li > spirv/vtn: handle variable pointers without offset lowering< / li >
< li > nir: move getting deref from var after we check deref type.< / li >
< li > nir: handle shared pointers in lowering indirect derefs.< / li >
< li > ac: avoid casting pointers on bcsel and stores< / li >
< li > radv: handle loading from shared pointers< / li >
< li > ac: handle cast derefs< / li >
< li > r600: make suballocator 256-bytes align< / li >
< li > virgl: fix undefined shift to use unsigned.< / li >
< li > virgl: fix const warning on debug flags.< / li >
< li > radv: use 3d shader for gfx9 copies if dst is 3d< / li >
< li > radv/xfb: fix counter buffer bounds checks.< / li >
< li > virgl/vtest: fix front buffer flush with protocol version 0.< / li >
< li > virgl: use primconvert provoking vertex properly< / li >
< li > dri_interface: add put shm image2 (v2)< / li >
< li > glx: add support for putimageshm2 path (v2)< / li >
< li > gallium: use put image shm2 path (v2)< / li >
< / ul >
< p > David Shao (1):< / p >
< ul >
< li > meson: ensure that xmlpool_options.h is generated for gallium targets that need it< / li >
< / ul >
< p > Dieter Nützel (1):< / p >
< ul >
< li > docs/features: Delete double nv50 entry and wrong enumeration< / li >
< / ul >
< p > Dylan Baker (48):< / p >
< ul >
< li > meson: link gallium nine with pthreads< / li >
< li > meson: Don't set -Wall< / li >
< li > meson: fix libatomic tests< / li >
< li > meson: Add tests to suites< / li >
< li > util: promote u_memory to src/util< / li >
< li > meson: Add nir_algebraic_parser_test to suites< / li >
< li > meson: Fix ppc64 little endian detection< / li >
< li > meson: remove duplicate definition< / li >
< li > meson: Add support for gnu hurd< / li >
< li > meson: Add toggle for glx-direct< / li >
< li > docs/meson: Recommend not using CFLAGS and friends< / li >
< li > travis: meson: use native files to override llvm-config< / li >
< li > travis: Don't try to read libdrm out of configure.ac< / li >
< li > travis: meson: enable unit tests< / li >
< li > docs: add note about using backticks for rbs in gitlab< / li >
< li > docs/install: Add meson to the main install page< / li >
< li > docs/meson: Update LLVM section with information about native files< / li >
< li > docs/install: Update python dependency section< / li >
< li > docs/autoconf: Mark autoconf as being replaced< / li >
< li > meson: Override C++ standard to gnu++11 when building with altivec on ppc64< / li >
< li > meson: Error out if building nouveau and using LLVM without rtti< / li >
< li > autotools: Remove tegra vdpau driver< / li >
< li > meson: Add a script to extract the cmd line used for meson< / li >
< li > meson: allow building dri driver without window system if osmesa is classic< / li >
< li > bin/meson-cmd-extract: Also handle cross and native files< / li >
< li > meson: fix swr KNL build< / li >
< li > meson: Fix compiler checks for SWR with ICC< / li >
< li > meson: Add warnings and errors when using ICC< / li >
< li > automake: Fix path to generated source< / li >
< li > automake: Add float64.glsl to dist tarball< / li >
< li > automake: Add include dir for nir src directory< / li >
< li > configure: Bump SWR LLVM requirement to 7< / li >
< li > automake: Add --enable-autotools to distcheck flags< / li >
< li > android,autotools,i965: Fix location of float64_glsl.h< / li >
< li > VERSION: bump to 19.0.0-rc1< / li >
< li > Version: Bump for rc2< / li >
< li > cherry-ignore: Add some patches< / li >
< li > Revert "intel/compiler: More peephole_select for pre-Gen6"< / li >
< li > Revert "nir/opt_peephole_select: Don't peephole_select expensive math instructions"< / li >
< li > Revert "intel/compiler: More peephole select"< / li >
< li > Bump version for 19.0-rc3< / li >
< li > version: bump for 19.0-rc4< / li >
< li > get-pick-list: Add --pretty=medium to the arguments for Cc patches< / li >
< li > meson: Add dependency on genxml to anvil< / li >
< li > Version: update to 19.0-rc5< / li >
< li > Bump version for rc6< / li >
< li > VERSION: bump version for rc7< / li >
< li > cherry-ignore: Update the cherry-ignore file< / li >
< / ul >
< p > Eduardo Lima Mitev (2):< / p >
< ul >
< li > freedreno/ir3: Make imageStore use num components from image format< / li >
< li > freedreno/ir3: Handle GL_NONE in get_num_components_for_glformat()< / li >
< / ul >
< p > Eleni Maria Stea (1):< / p >
< ul >
< li > i965: fixed clamping in set_scissor_bits when the y is flipped< / li >
< / ul >
< p > Elie Tournier (17):< / p >
< ul >
< li > glsl: Add "built-in" function to do abs(fp64)< / li >
< li > glsl: Add "built-in" functions to do neg(fp64)< / li >
< li > glsl: Add "built-in" function to do sign(fp64)< / li >
< li > glsl: Add "built-in" functions to do eq/ne(fp64, fp64)< / li >
< li > glsl: Add utility function to extract 64-bit sign< / li >
< li > glsl: Add "built-in" functions to do lt(fp64, fp64)< / li >
< li > glsl: Add "built-in" functions to do add(fp64, fp64)< / li >
< li > glsl: Add "built-in" functions to do mul(fp64, fp64)< / li >
< li > glsl: Add "built-in" functions to do fp64_to_uint(fp64)< / li >
< li > glsl: Add "built-in" functions to do uint_to_fp64(uint)< / li >
< li > glsl: Add "built-in" functions to do fp64_to_int(fp64)< / li >
< li > glsl: Add "built-in" functions to do int_to_fp64(int)< / li >
< li > glsl: Add "built-in" functions to do fp64_to_fp32(fp64)< / li >
< li > glsl: Add "built-in" functions to do fp32_to_fp64(fp32)< / li >
< li > glsl: Add "built-in" functions to do sqrt(fp64)< / li >
< li > glsl: Add "built-in" functions to do trunc(fp64)< / li >
< li > glsl: Add "built-in" functions to do round(fp64)< / li >
< / ul >
< p > Emil Velikov (81):< / p >
< ul >
< li > mesa: bump version to 19.1.0-devel< / li >
< li > docs: add 19.0.0-devel release notes template< / li >
< li > docs: mention EXT_shader_implicit_conversions< / li >
< li > egl: add EGL_EXT_device_base entrypoints< / li >
< li > egl/glvnd: correctly report errors when vendor cannot be found< / li >
< li > docs/releasing.html: polish cherry-picking/testing text< / li >
< li > docs/submittingpatches.html: correctly handle the < p> tag< / li >
< li > docs: document the staging branch and add reference to it< / li >
< li > bin/get-pick-list.sh: simplify git oneline printing< / li >
< li > bin/get-pick-list.sh: prefix output with "[stable] "< / li >
< li > bin/get-pick-list.sh: handle "typod" usecase.< / li >
< li > bin/get-pick-list.sh: handle the fixes tag< / li >
< li > bin/get-pick-list.sh: tweak the commit sha matching pattern< / li >
< li > bin/get-pick-list.sh: flesh out is_sha_nomination< / li >
< li > bin/get-pick-list.sh: handle fixes tag with missing colon< / li >
< li > bin/get-pick-list.sh: handle unofficial "broken by" tag< / li >
< li > bin/get-pick-list.sh: use test instead of [ ]< / li >
< li > bin/get-pick-list.sh: handle reverts prior to the branchpoint< / li >
< li > travis: drop unneeded x11proto-xf86vidmode-dev< / li >
< li > glx: make xf86vidmode mandatory for direct rendering< / li >
< li > travis: adding missing x11-xcb for meson+vulkan< / li >
< li > egl/wayland: bail out when drmGetMagic fails< / li >
< li > egl/wayland: plug memory leak in drm_handle_device()< / li >
< li > docs: update 18.3.0 release notes< / li >
< li > docs: add sha256 checksums for 18.3.0< / li >
< li > docs: update calendar, add news item and link release notes for 18.3.0< / li >
< li > freedreno: drop duplicate MKDIR_GEN declaration< / li >
< li > freedreno: add the missing _la in libfreedreno_ir3_la< / li >
< li > amd/addrlib: drop si_ci_vi_merged_enum.h from the list< / li >
< li > docs: add release notes for 18.3.1< / li >
< li > docs: add sha256 checksums for 18.3.1< / li >
< li > docs: update calendar, add news item and link release notes for 18.3.1< / li >
< li > glx: mandate xf86vidmode only for "drm" dri platforms< / li >
< li > bin/get-pick-list.sh: rework handing of sha nominations< / li >
< li > bin/get-pick-list.sh: warn when commit lists invalid sha< / li >
< li > meson: don't require glx/egl/gbm with gallium drivers< / li >
< li > pipe-loader: meson: reference correct library< / li >
< li > TODO: glx: meson: build dri based glx tests, only with -Dglx=dri< / li >
< li > glx: meson: drop includes from a link-only library< / li >
< li > glx: meson: wire up the dispatch-index-check test< / li >
< li > glx/test: meson: assorted include fixes< / li >
< li > configure: add CXX11_CXXFLAGS to LLVM_CXXFLAGS< / li >
< li > travis: flip to distro xenial, drop sudo false< / li >
< li > travis: meson: print the configured state< / li >
< li > travis: printout llvm-config --version< / li >
< li > travis: meson: use FOO_DRIVERS directly< / li >
< li > travis: meson: add unwind handling< / li >
< li > travis: meson: explicitly control the DRI loaders< / li >
< li > travis: meson: add explicit handling to gallium ST< / li >
< li > travis: meson: port gallium build combinations over< / li >
< li > docs: add release notes for 18.3.2< / li >
< li > docs: add sha256 checksums for 18.3.2< / li >
< li > docs: update calendar, add news item and link release notes for 18.3.2< / li >
< li > freedreno: automake: ship ir3_nir_trig.py in the tarball< / li >
< li > mesa: correctly use os.path.join in our python scripts< / li >
< li > Revert "mesa/main: remove ARB suffix from glGetnTexImage"< / li >
< li > mapi: sort static entrypoints numerically< / li >
< li > mapi: add all _glapi_table entrypoints to static_data.py< / li >
< li > genCommon.py: Fix typo in _LIBRARY_FEATURE_NAMES.< / li >
< li > mapi: move genCommon.py to src/mapi/new< / li >
< li > mapi/new: import mapi scripts from glvnd< / li >
< li > mapi/new: sort by slot number< / li >
< li > mapi/new: use the static_data offsets in the new generator< / li >
< li > mapi/new: reinstate _NO_HIDDEN suffixes in the new generator< / li >
< li > mapi/new: split out public_entries handling< / li >
< li > mapi/new: don't print info we don't need for ES1/ES2< / li >
< li > mapi/new: fixup the GLDEBUGPROCKHR typedef to the non KHR one< / li >
< li > mapi/new: remove duplicate GLvoid/void substitution< / li >
< li > autotools: wire the new generator for es1 and es2< / li >
< li > meson: wire the new generator for es1 and es2< / li >
< li > scons: wire the new generator for es1 and es2< / li >
< li > Revert "mapi/new: sort by slot number"< / li >
< li > mapi/es*api: remove GL_OES_EGL_image entrypoints< / li >
< li > mapi/es*api: remove GL_EXT_multi_draw_arrays entrypoints< / li >
< li > mapi/es2api: remove no longer present entrypoints< / li >
< li > mapi: remove old, unused ES* generator code< / li >
< li > mapi: remove machinery handling CSV files< / li >
< li > mapi: print function declarations for shared glapi< / li >
< li > vc4: Declare the last cpu pointer as being modified in NEON asm.< / li >
< li > anv: wire up the state_pool_padding test< / li >
< li > meson: egl: correctly manage loader/xmlconfig< / li >
< / ul >
< p > Eric Anholt (171):< / p >
< ul >
< li > v3d: Fix a copy-and-paste comment in the simulator code.< / li >
< li > v3d: Fix a typo in a comment in job handling.< / li >
< li > v3d: Drop #if 0-ed out v3d_dump_to_file().< / li >
< li > v3d: Respect user-passed strides for BO imports.< / li >
< li > v3d: Take advantage of _mesa_hash_table_remove_key() in the simulator.< / li >
< li > v3d: Use the TLB R/B swapping instead of recompiles when available.< / li >
< li > v3d: Update the TLB config for depth writes on V3D 4.2.< / li >
< li > vc4: Drop the winsys_stride relayout in the simluator< / li >
< li > v3d: Maintain a mapping of the GEM buffer in the simulator.< / li >
< li > v3d: Remove the special path for simulaton of the submit ioctl.< / li >
< li > vc4: Take advantage of _mesa_hash_table_remove_key() in the simulator.< / li >
< li > vc4: Maintain a separate GEM mapping of BOs in the simulator.< / li >
< li > vc4: Use the normal simulator ioctl path for CL submit as well.< / li >
< li > gbm: Move gbm_format_canonicalize() to the core.< / li >
< li > gbm: Introduce a helper function for printing GBM format names.< / li >
< li > egl: Improve the debugging of gbm format matching in DRI configs.< / li >
< li > v3d: Fix double-swapping of R/B on V3D 4.1< / li >
< li > v3d: Don't try to set PF flags on a LDTMU operation< / li >
< li > vc4: Make sure we make ro scanout resources for create_with_modifiers.< / li >
< li > vc4: Don't return a vc4 BO handle on a renderonly screen.< / li >
< li > glx: Remove an old DEFAULT_DRIVER_DIR default.< / li >
< li > glx: Move DRI extensions pointer loading to driOpenDriver().< / li >
< li > egl: Move loader_set_logger() up to egl_dri2.c.< / li >
< li > loader: Stop using a local definition for an in-tree header< / li >
< li > loader: Factor out the common driver opening logic from each loader.< / li >
< li > egl: Print the actual message to the console from _eglError().< / li >
< li > gallium: Fix uninitialized variable warning in compute test.< / li >
< li > gallium: Remove unused variable in u_tests.< / li >
< li > v3d: Add renderonly support.< / li >
< li > v3d: Add support for RGBA_SRGB along with BGRA_SRGB.< / li >
< li > v3d: Add missing OES_half_float_linear support.< / li >
< li > v3d: Use combined input/output segments.< / li >
< li > v3d: Add the V3D TFU submit interface to the simulator.< / li >
< li > v3d: Use the TFU to do generatemipmap.< / li >
< li > v3d: Update simulator cache flushing code to match the kernel better.< / li >
< li > v3d: Create a state uploader for packing our shaders together.< / li >
< li > v3d: Put default vertex attribute values into the state uploader as well.< / li >
< li > v3d: Re-use the wrap mode uniform on V3D 3.3.< / li >
< li > v3d: Make an array for frag/vert texture state in the context.< / li >
< li > v3d: Don't forget to flush writes to UBOs.< / li >
< li > v3d: Convert to using nir_src_as_uint() from const_value derefs.< / li >
< li > v3d: Fix a comment typo< / li >
< li > v3d: Return the right gl_SampleMaskIn[] value.< / li >
< li > v3d: Fix handling of texture first_layer offsets for 3D textures.< / li >
< li > v3d: Avoid confusing auto-indenting in TEXTURE_SHADER_STATE packing< / li >
< li > v3d: Split most of TEXTURE_SHADER_STATE setup out of sampler views.< / li >
< li > v3d: Garbage collect unused uniforms code.< / li >
< li > v3d: Simplify VIR uniform dumping using a temporary.< / li >
< li > v3d: Add VIR dumping of TMU config p0/p1.< / li >
< li > v3d: Fix a leak of the transfer helper on screen destroy.< / li >
< li > vc4: Fix a leak of the transfer helper on screen destroy.< / li >
< li > v3d: Fix a leak of the disassembled instruction string during debug dumps.< / li >
< li > tfu< / li >
< li > shader-packing< / li >
< li > nir: Add some more consts to the nir_format_convert.h helpers.< / li >
< li > nir: Pull some of intel's image load/store format conversion to nir_format.h< / li >
< li > intel: Simplify the half-float packing in image load/store lowering.< / li >
< li > mesa/st: Expose compute shaders when NIR support is advertised.< / li >
< li > nir: Print the format of image variables.< / li >
< li > Revert "intel: Simplify the half-float packing in image load/store lowering."< / li >
< li > nir: Move intel's half-float image store lowering to to nir_format.h.< / li >
< li > v3d: Don't forget to wait for our TFU job before rendering from it.< / li >
< li > v3d: Set up the right stride for raster TFU.< / li >
< li > v3d: Don't forget to bump the number of writes when doing TFU ops.< / li >
< li > v3d: Add support for using the TFU to do some blits.< / li >
< li > v3d: Add support for texturing from linear.< / li >
< li > v3d: Add safety checks for resource_create().< / li >
< li > v3d: Make sure that a thrsw doesn't split a multop from its umul24.< / li >
< li > v3d: Add missing flagging of SYNCB as a TSY op.< / li >
< li > v3d: Add support for draw indirect for GLES3.1.< / li >
< li > v3d: Avoid assertion failures when removing end-of-shader instructions.< / li >
< li > v3d: Move uinfo-> data[] dereference to the top of v3d_write_uniforms().< / li >
< li > v3d: Move uniform pretty-printing to its own helper function.< / li >
< li > v3d: Use the uniform pretty-printer in v3d_write_uniforms()'s debug code.< / li >
< li > v3d: Do uniform pretty-printing in the QPU dump.< / li >
< li > v3d: Drop in a bunch of notes about performance improvement opportunities.< / li >
< li > vc4: Use the original bit size when scalarizing uniform loads.< / li >
< li > v3d: Use the original bit size when scalarizing uniform loads.< / li >
< li > vc4: Reuse nir_format_convert.h in our blend lowering.< / li >
< li > v3d: Fix the argument type for vir_BRANCH().< / li >
< li > nir: Fix clamping of uints for image store lowering.< / li >
< li > v3d: Put the dst bo first in the list of BOs for TFU calls.< / li >
< li > v3d: Fix check for TFU job completion in the simulator.< / li >
< li > v3d: Don't try to create shadow tiled temporaries for 1D textures.< / li >
< li > v3d: Remove dead prototypes for load/store utile functions.< / li >
< li > v3d: Implement texture_subdata to reduce teximage upload copies.< / li >
< li > vc4: Move the utile load/store functions to a header for reuse by v3d.< / li >
< li > v3d: Add a fallthrough path for utile load/store of 32 byte lines.< / li >
< li > v3d: Load and store aligned utiles all at once.< / li >
< li > docs: Add a note that MRs should still include any r-b or a-b tags.< / li >
< li > docs: Add an encouraging note about providing reviews and acks.< / li >
< li > v3d: Fix simulator mode on i915 render nodes.< / li >
< li > v3d: Drop shadow comparison state from shader variant key.< / li >
< li > v3d: Hook up perf_debug() output to GL_ARB_debug output as well.< / li >
< li > vc4: Hook up perf_debug() output to GL_ARB_debug_output as well.< / li >
< li > gallium/ttn: Fix setup of outputs_written.< / li >
< li > v3d: Fix uniform pretty printing assertion failure with branches.< / li >
< li > v3d: Add a "precompile" debug flag for shader-db.< / li >
< li > v3d: Hook up some shader-db output to GL_ARB_debug_output.< / li >
< li > v3d: Drop unused count_nir_instrs() helper.< / li >
< li > v3d: Drop incorrect dependency for flpop.< / li >
< li > v3d: Move "does this instruction have flags" from sched to generic helpers.< / li >
< li > v3d: Don't generate temps for comparisons.< / li >
< li > v3d: Dead-code eliminate unused flags updates.< / li >
< li > v3d: Add a note for a potential performance win on multop/umul24.< / li >
< li > v3d: Force sampling from base level for tg4.< / li >
< li > v3d: Add support for non-constant texture offsets.< / li >
< li > v3d: Add support for requesting the sample offsets.< / li >
< li > v3d: Add support for textureSize() on MSAA textures.< / li >
< li > v3d: Add support for gl_HelperInvocation.< / li >
< li > v3d: Fix segfault when failing to compile a program.< / li >
< li > v3d: Don't forget to include RT writes in precompiles.< / li >
< li > v3d: Simplify the emission of comparisons for the bcsel optimization.< / li >
< li > v3d: Move the "Find the ALU instruction generating our bool" out of bcsel.< / li >
< li > v3d: Don't try to fold non-SSA-src comparisons into bcsels.< / li >
< li > v3d: Fold comparisons for IF conditions into the flags for the IF.< / li >
< li > v3d: Handle dynamically uniform IF statements with uniform control flow.< / li >
< li > v3d: Refactor compiler entrypoints.< / li >
< li > v3d: Reinstate the new shader-db output after v3d_compile() refactor.< / li >
< li > v3d: Fix up VS output setup during precompiles.< / li >
< li > v3d: Remove dead switch cases and comments from v3d_nir_lower_io.< / li >
< li > v3d: Do UBO loads a vector at a time.< / li >
< li > v3d: Stop scalarizing our uniform loads.< / li >
< li > nir: Allow nir_format_unpack_int/sint to unpack larger values.< / li >
< li > nir: Add nir_lower_tex options to lower sampler return formats.< / li >
< li > v3d: Use the core tex lowering.< / li >
< li > nir: Add nir_lower_tex support for Broadcom's swizzled TG4 results.< / li >
< li > v3d: Enable GL_ARB_texture_gather on V3D 4.x.< / li >
< li > nir: Make nir_deref_instr_build/get_const_offset actually use size_align.< / li >
< li > glsl: Fix buffer overflow with an atomic buffer binding out of range.< / li >
< li > v3d: Add support for flushing dirty TMU data at job end.< / li >
< li > v3d: Add support for the early_fragment_tests flag.< / li >
< li > v3d: Add support for GL_ARB_framebuffer_no_attachments.< / li >
< li > v3d: Fix txf_ms 2D_ARRAY array index.< / li >
< li > v3d: Add an isr to the simulator to catch GMP violations.< / li >
< li > v3d: Add support for matrix inputs to the FS.< / li >
< li > v3d: Drop the GLSL version level.< / li >
< li > v3d: Add SSBO/atomic counters support.< / li >
< li > v3d: Add support for shader_image_load_store.< / li >
< li > v3d: Add support for CS workgroup/invocation id intrinsics.< / li >
< li > v3d: Add support for CS shared variable load/store/atomics.< / li >
< li > v3d: Add support for CS barrier() intrinsics.< / li >
< li > v3d: SHARED but not necessarily SCANOUT buffers on RO must be linear.< / li >
< li > v3d: If the modifier is not known on BO import, default to linear for RO.< / li >
< li > v3d: Restructure RO allocations using resource_from_handle.< / li >
< li > v3d: Don't leak the GPU fd for renderonly usage.< / li >
< li > vc4: Don't leak the GPU fd for renderonly usage.< / li >
< li > gallium: Enable unit tests as actual meson unit tests.< / li >
< li > gallium: Fix comment about possible colorspaces.< / li >
< li > gallium: Make sure we return is_unorm/is_snorm for compressed formats.< / li >
< li > v3d: Rename gallium-local limits defines from VC5 to V3D.< / li >
< li > v3d: Fix overly-large vattr_sizes structs.< / li >
< li > v3d: Avoid duplicating limits defines between gallium and v3d core.< / li >
< li > v3d: Drop maximum number of texture units down to 16.< / li >
< li > v3d: Fix BO stats accounting for imported buffers.< / li >
< li > v3d: Flush blit jobs immediately after generating them.< / li >
< li > v3d: Fix release-build warning about utile_h.< / li >
< li > v3d: Fix stencil sampling from packed depth/stencil.< / li >
< li > v3d: Fix stencil sampling from a separate-stencil buffer.< / li >
< li > v3d: Use the symbolic names for wrap modes from the XML.< / li >
< li > v3d: Move the sampler state to the long-lived state uploader.< / li >
< li > v3d: Create separate sampler states for the various blend formats.< / li >
< li > pl111: Rename the pl111 driver to "kmsro".< / li >
< li > kmsro: Extend to include hx8357d.< / li >
< li > vc4: Enable NEON asm on meson cross-builds.< / li >
< li > v3d: Fix the autotools build.< / li >
< li > mesa: Skip partial InvalidateFramebuffer of packed depth/stencil.< / li >
< li > v3d: Fix image_load_store clamping of signed integer stores.< / li >
< li > v3d: Use the early_fragment_tests flag for the shader's disable-EZ field.< / li >
< li > v3d: Fix the check for "is the last thrsw inside control flow"< / li >
< li > st/dri: Set the PIPE_BIND_SHARED flag on create_image_with_modifiers.< / li >
< / ul >
< p > Eric Engestrom (47):< / p >
< ul >
< li > wsi/wayland: use proper VkResult type< / li >
< li > wsi/wayland: only finish() a successfully init()ed display< / li >
< li > REVIEWERS: add include path for EGL< / li >
< li > REVIEWERS: add Emil as EGL reviewer< / li >
< li > REVIEWERS: add Vulkan reviewer group< / li >
< li > xmlpool: update translation po files< / li >
< li > meson: only run vulkan's meson.build when building vulkan< / li >
< li > gbm: remove unnecessary meson include< / li >
< li > meson: fix wayland-less builds< / li >
< li > gbm: add new entrypoint to symbols check< / li >
< li > egl: add missing glvnd entrypoint for EGL_ANDROID_blob_cache< / li >
< li > egl: fix bad rebase< / li >
< li > gbm: add missing comma between strings< / li >
< li > glapi: add missing visibility args< / li >
< li > anv: correctly use vulkan 1.0 by default< / li >
< li > vulkan/utils: s/VERSION/PACKAGE_VERSION/< / li >
< li > build: stop defining unused VERSION< / li >
< li > wsi/display: fix mem leak when freeing swapchains< / li >
< li > vulkan/wsi: fix s/,/;/ typo< / li >
< li > meson: skip asm check when asm is disabled< / li >
< li > anv: add unreachable() for VK_EXT_fragment_density_map< / li >
< li > mesa: drop unused & deprecated lib< / li >
< li > loader: deduplicate logger function declaration< / li >
< li > docs: add meson cross compilation instructions< / li >
< li > docs: format code blocks a bit nicely< / li >
< li > docs: fix the meson aarch64 cross-file< / li >
< li > docs: advertise distro-provided meson cross-files< / li >
< li > anv: drop unneeded KHR suffix< / li >
< li > wsi: drop unneeded KHR suffix< / li >
< li > radv: remove a few more unnecessary KHR suffixes< / li >
< li > egl: add missing includes< / li >
< li > egl: remove unused include< / li >
< li > travis: avoid using unset llvm-config< / li >
< li > egl: fix python lib deprecation warning< / li >
< li > docs: explain how to see what meson options exist< / li >
< li > travis: fix autotools build after --enable-autotools switch addition< / li >
< li > configure: EGL requirements only apply if EGL is built< / li >
< li > egl: finalize EGL_MESA_query_driver< / li >
< li > egl: update headers from Khronos< / li >
< li > egl: add glvnd entrypoints for EGL_MESA_query_driver< / li >
< li > travis: bump libdrm to 2.4.97< / li >
< li > egl/glvnd: sync egl.xml from Khronos< / li >
< li > anv: drop always-successful VkResult< / li >
< li > meson/vdpau: add missing soversion< / li >
< li > xvmc: fix string comparison< / li >
< li > xvmc: fix string comparison< / li >
< li > egl: fix libdrm-less builds< / li >
< / ul >
< p > Erik Faye-Lund (70):< / p >
< ul >
< li > glsl: add has_implicit_conversions()-helper< / li >
< li > glsl: add has_implicit_uint_to_int_conversion()-helper< / li >
< li > glsl: fall back to inexact function-match< / li >
< li > mesa/glsl: add support for EXT_shader_implicit_conversions< / li >
< li > glsl: do not allow implicit casts of unsized array initializers< / li >
< li > mesa: expose NV_conditional_render on GLES< / li >
< li > mesa/main: fixup make check after NV_conditional_render for gles< / li >
< li > Revert "mesa/main: fixup make check after NV_conditional_render for gles"< / li >
< li > Revert "mesa: expose NV_conditional_render on GLES"< / li >
< li > mesa/main: correct requirement for EXT_occlusion_query_boolean< / li >
< li > mesa/main: correct year for EXT_occlusion_query_boolean< / li >
< li > mesa/main: use non-prefixed enums for consistency< / li >
< li > mesa/main: simplify pipeline-statistics query validation< / li >
< li > mesa/main: fix validation of GL_SAMPLES_PASSED< / li >
< li > mesa/main: fix validation of GL_ANY_SAMPLES_PASSED< / li >
< li > mesa/main: fix validation of GL_ANY_SAMPLES_PASSED_CONSERVATIVE< / li >
< li > mesa/main: fix validation of GL_TIME_ELAPSED< / li >
< li > mesa/main: fix validation of transform-feedback queries< / li >
< li > mesa/main: fix validation of transform-feedback overflow queries< / li >
< li > mesa/main: fix validation of ARB_query_buffer_object< / li >
< li > mesa/main: fix validation of GL_TIMESTAMP< / li >
< li > mesa/main: remove overly strict query-validation< / li >
< li > mesa/main: remove ARB suffix from glGetnTexImage< / li >
< li > mesa/main: remove bogus error for zero-sized images< / li >
< li > mesa/main: factor out tex-image error-checking< / li >
< li > mesa/main: factor out common error-checking< / li >
< li > mesa/main: check cube-completeness in common code< / li >
< li > mesa/main: fix incorrect depth-error< / li >
< li > mesa/main: fixup requirements for GL_PRIMITIVES_GENERATED< / li >
< li > mesa/main: make _mesa_has_tessellation return bool< / li >
< li > mesa/main: rename format-check function< / li >
< li > mesa/main: clean up S3_s3tc check< / li >
< li > mesa/main: clean up OES_texture_float_linear check< / li >
< li > mesa/main: clean up ES2_compatibility check< / li >
< li > mesa/main: clean up integer texture check< / li >
< li > mesa/main: use _mesa_has_FOO_bar for compressed format checks< / li >
< li > mesa/main: do not allow s3tc enums on gles1< / li >
< li > mesa/main: do not allow etc2 enums on gles1< / li >
< li > mesa/main: do not allow astc enums on gles1< / li >
< li > mesa/main: do not allow depth-texture enums on gles1< / li >
< li > mesa/main: do not allow stencil-texture enums on gles1< / li >
< li > mesa/main: do not allow ARB_texture_rgb10_a2ui enums before gles3< / li >
< li > mesa/main: do not allow integer-texture enums before gles3< / li >
< li > mesa/main: do not allow ARB_depth_buffer_float enums before gles3< / li >
< li > mesa/main: do not allow EXT_packed_float enums before gles3< / li >
< li > mesa/main: do not allow rg-textures enums before gles3< / li >
< li > mesa/main: do not allow EXT_texture_shared_exponent enums before gles3< / li >
< li > mesa/main: do not allow MESA_ycbcr_texture enums on gles< / li >
< li > mesa/main: do not allow type_2_10_10_10_REV enums before gles3< / li >
< li > mesa/main: do not allow floating-point texture enums on gles1< / li >
< li > mesa/main: do not allow snorm-texture enums before gles3< / li >
< li > mesa/main: do not allow sRGB texture enums before gles3< / li >
< li > mesa/main: do not allow EXT_texture_sRGB_R8 enums before gles3< / li >
< li > mesa/main: split float-texture support checking in two< / li >
< li > mesa/main: require EXT_texture_type_2_10_10_10_REV for gles3< / li >
< li > mesa/main: require EXT_texture_sRGB for gles3< / li >
< li > mesa/st: do not probe for the same texture-formats twice< / li >
< li > mesa/main: do not require float-texture filtering for es3< / li >
< li > mesa/main: correct validation for GL_RGB565< / li >
< li > mesa/main: fix up _mesa_has_rg_textures for gles2< / li >
< li > virgl: force linear texturing support< / li >
< li > virgl: simplify virgl_hw_set_vertex_buffers< / li >
< li > virgl: simplify virgl_hw_set_index_buffer< / li >
< li > virgl: wrap vertex element state in a struct< / li >
< li > virgl: work around bad assumptions in virglrenderer< / li >
< li > anv/meson: make sure tests link with -msse2< / li >
< li > anv/autotools: make sure tests link with -msse2< / li >
< li > docs: add note about sending merge-requests from forks< / li >
< li > mapi: drop unneeded gl_dispatch_stub declarations< / li >
< li > virgl: remove unused variable< / li >
< / ul >
< p > Ernestas Kulik (2):< / p >
< ul >
< li > vc4: Fix leak in HW queries error path< / li >
< li > v3d: Fix leak in resource setup error path< / li >
< / ul >
< p > Francisco Jerez (14):< / p >
< ul >
< li > intel/fs: Prevent emission of IR instructions not aligned to their own execution size.< / li >
< li > intel/fs: Handle source modifiers in lower_integer_multiplication().< / li >
< li > intel/fs: Implement quad swizzles on ICL+.< / li >
< li > intel/fs: Fix bug in lower_simd_width while splitting an instruction which was already split.< / li >
< li > intel/eu/gen7: Fix brw_MOV() with DF destination and strided source.< / li >
< li > intel/fs: Respect CHV/BXT regioning restrictions in copy propagation pass.< / li >
< li > intel/fs: Constify fs_inst::can_do_source_mods().< / li >
< li > intel/fs: Introduce regioning lowering pass.< / li >
< li > intel/fs: Remove existing lower_conversions pass.< / li >
< li > intel/fs: Remove nasty open-coded CHV/BXT 64-bit workarounds.< / li >
< li > intel/fs: Remove FS_OPCODE_UNPACK_HALF_2x16_SPLIT opcodes.< / li >
< li > intel/fs: Promote execution type to 32-bit when any half-float conversion is needed.< / li >
< li > intel/fs: Exclude control sources from execution type and region alignment calculations.< / li >
< li > intel/fs: Implement extended strides greater than 4 for IR source regions.< / li >
< / ul >
< p > Fritz Koenig (2):< / p >
< ul >
< li > freedreno: drm_fourcc.h header include< / li >
< li > freedreno: add query for dmabuf modifiers< / li >
< / ul >
< p > Gert Wollny (30):< / p >
< ul >
< li > mesa/core: Add definitions and translations for EXT_texture_sRGB_R8< / li >
< li > Gallium: Add format PIPE_FORMAT_R8_SRGB< / li >
< li > mesa/st: Add support for EXT_texture_sRGB_R8< / li >
< li > virgl/vtest-winsys: Use virgl version of bind flags< / li >
< li > r600: Add support for EXT_texture_sRGB_R8< / li >
< li > mesa: Reference count shaders that are used by transform feedback objects< / li >
< li > virgl: Add command and flags to initiate debugging on the host (v2)< / li >
< li > nir: Allow to skip integer ops in nir_lower_to_source_mods< / li >
< li > i965: Correct L8_UNORM_SRGB table entry< / li >
< li > i965: be more specific about FBO completeness errors< / li >
< li > i965: Force zero swizzles for unused components in GL_RED and GL_RG< / li >
< li > i965: Add support for and expose EXT_texture_sRGB_R8< / li >
< li > virgl: Use file descriptor instead of un-allocated object< / li >
< li > i965:use FRAMEBUFFER_UNSUPPORTED instead of FRAMEBUFFER_INCOMPLETE_DIMENSIONS< / li >
< li > r600: Only set context streamout strides info from the shader that has outputs< / li >
< li > r600: clean up the GS ring buffers when the context is destroyed< / li >
< li > glsl: free or reuse memory allocated for TF varying< / li >
< li > virgl,vtest: Initialize return value< / li >
< li > virgl: Don't try handling server fences when they are not supported< / li >
< li > i965: Explicitely handle swizzles for MESA_FORMAT_R_SRGB8< / li >
< li > i965: Set the FBO error state INCOMPLETE_ATTACHMENT only for SRGB_R8< / li >
< li > autotools: Deprecate the use of autotools< / li >
< li > Gallium: Add new CAPS to indicate whether a driver can switch SRGB write< / li >
< li > virgl: Set sRGB write control CAP based on host capabilities< / li >
< li > mesa:main: Add flag for EXT_sRGB to gl_extensions< / li >
< li > i965: Set flag for EXT_sRGB< / li >
< li > mesa/st: rework support for sRGB framebuffer attachements< / li >
< li > mesa/main: Use flag for EXT_sRGB instead of EXT_framebuffer_sRGB where possible< / li >
< li > mesa/main/version: Lower the requirements for GLES 3.0< / li >
< li > mesa/main: Expose EXT_sRGB_write_control< / li >
< / ul >
< p > Guido Günther (2):< / p >
< ul >
< li > etnaviv: Make sure rs alignment checks match< / li >
< li > etnaviv: fix typo in cflush_all description< / li >
< / ul >
< p > Gurchetan Singh (18):< / p >
< ul >
< li > egl: add missing #include < stddef.h> in egldevice.h< / li >
< li > virgl: quadruple command buffer size< / li >
< li > virgl: avoid large inline transfers< / li >
< li > virgl: don't mark buffers as unclean after a write< / li >
< li > virgl: texture_transfer_pool --> transfer_pool< / li >
< li > virgl: remove unnessecary code< / li >
< li > virgl: move texture metadata to common code< / li >
< li > virgl: move virgl_resource_layout to common code< / li >
< li > virgl: move vrend_get_tex_image_offset to common code< / li >
< li > virgl: store layer_stride in metadata< / li >
< li > virgl: consolidate transfer code< / li >
< li > virgl: make transfer code with PIPE_BUFFER targets< / li >
< li > virgl: make virgl_buffers use resource helpers< / li >
< li > virgl: modify how we handle GL_MAP_FLUSH_EXPLICIT_BIT< / li >
< li > virgl: move resource metadata into base resource< / li >
< li > virgl: move resource creation / import / destruction to common code< / li >
< li > virgl: don't flush an empty range< / li >
< li > virgl: remove empty file< / li >
< / ul >
< p > Hanno Böck (1):< / p >
< ul >
< li > glsl/test: Fix use after free in test_optpass.< / li >
< / ul >
< p > Hyunjun Ko (1):< / p >
< ul >
< li > freedreno: implements get_sample_position< / li >
< / ul >
< p > Iago Toral Quiroga (22):< / p >
< ul >
< li > intel/compiler: fix node interference of simd16 instructions< / li >
< li > nir/constant_folding: fix incorrect bit-size check< / li >
< li > nir/from_ssa: fix bit-size of temporary register< / li >
< li > Revert "nir/builder: Assert that intN_t immediates fit"< / li >
< li > intel/compiler: fix indentation style in opt_algebraic()< / li >
< li > intel/compiler: fix register allocation in opt_peephole_sel< / li >
< li > intel/compiler: do not copy-propagate strided regions to ddx/ddy arguments< / li >
< li > intel/compiler: move nir_lower_bool_to_int32 before nir_lower_locals_to_regs< / li >
< li > compiler/nir: add a nir_b2f() helper< / li >
< li > compiler/nir: add nir_fadd_imm() and nir_fmul_imm() helpers< / li >
< li > compiler/spirv: handle 16-bit float in radians() and degrees()< / li >
< li > compiler/spirv: implement 16-bit asin< / li >
< li > compiler/spirv: implement 16-bit acos< / li >
< li > compiler/spirv: implement 16-bit atan< / li >
< li > compiler/spirv: implement 16-bit atan2< / li >
< li > compiler/spirv: implement 16-bit exp and log< / li >
< li > compiler/spirv: implement 16-bit hyperbolic trigonometric functions< / li >
< li > compiler/spirv: implement 16-bit frexp< / li >
< li > compiler/spirv: use 32-bit polynomial approximation for 16-bit asin()< / li >
< li > anv/pipeline_cache: fix incorrect guards for NIR cache< / li >
< li > anv/pipeline_cache: free NIR shader cache< / li >
< li > anv/device: fix maximum number of images supported< / li >
< / ul >
< p > Ian Romanick (28):< / p >
< ul >
< li > glsl: Add warning tests for identifiers with __< / li >
< li > glsl: Add pragma to disable all warnings< / li >
< li > glsl: prevent qualifiers modification of predeclared variables< / li >
< li > glsl: Omit redundant qualifier checks on redeclarations< / li >
< li > glsl: Refactor type checking for redeclarations< / li >
< li > nir: Add a saturated unsigned integer add opcode< / li >
< li > i965/fs: Implement nir_op_uadd_sat< / li >
< li > nir/phi_builder: Internal users should use nir_phi_builder_value_set_block_def too< / li >
< li > util/slab: Rename slab_mempool typed parameters to mempool< / li >
< li > util/hash_table: Add _mesa_hash_table_init function< / li >
< li > nir/phi_builder: Use per-value hash table to store [block] -> def mapping< / li >
< li > nir: Fix holes in nir_instr< / li >
< li > nir: Release per-block metadata in nir_sweep< / li >
< li > i965/vec4: Silence unused parameter warnings in vec4 compiler tests< / li >
< li > i965/vec4/dce: Don't narrow the write mask if the flags are used< / li >
< li > i965/fs: Eliminate unary op on operand of compare-with-zero< / li >
< li > i965/vec4: Propagate conditional modifiers from more compares to other compares< / li >
< li > nir/opt_peephole_select: Don't try to remove flow control around indirect loads< / li >
< li > intel/compiler: More peephole select< / li >
< li > nir/opt_peephole_select: Don't peephole_select expensive math instructions< / li >
< li > intel/compiler: More peephole_select for pre-Gen6< / li >
< li > Revert "nir/lower_indirect: Bail early if modes == 0"< / li >
< li > nir/algebraic: Don't put quotes around floating point literals< / li >
< li > glsl: Add utility to convert text files to C strings< / li >
< li > nir: Silence zillions of unused parameter warnings in release builds< / li >
< li > spirv: Add missing break< / li >
< li > intel/fs: nir_op_extract_i8 extracts a byte, not a word< / li >
< li > intel/fs: Fix extract_u8 of an odd byte from a 64-bit integer< / li >
< / ul >
< p > Ilia Mirkin (37):< / p >
< ul >
< li > nv50/ir: delete MINMAX instruction that is no longer in the BB< / li >
< li > nv50/ir/ra: improve condition for short regs, unify with cond for 16-bit< / li >
< li > nv50/ir/ra: enforce max register requirement, and change spill order< / li >
< li > nv50/ir: remove dnz flag when converting MAD to ADD due to optimizations< / li >
< li > nv50: always keep TSC slot 0 bound< / li >
< li > nv50,nvc0: add explicit handling of PIPE_CAP_MAX_VERTEX_ELEMENT_SRC_OFFSET< / li >
< li > nouveau: set texture upload budget< / li >
< li > nvc0: replace use of explicit default_tsc with entry 0< / li >
< li > nvc0: always keep TSC slot 0 bound to fix TXF< / li >
< li > st/mesa: remove sampler associated with buffer texture in pbo logic< / li >
< li > st/mesa: allow glDrawElements to work with GL_SELECT feedback< / li >
< li > tgsi: add ATOMFADD operation< / li >
< li > gallium: add PIPE_CAP_TGSI_ATOMFADD to indicate support< / li >
< li > st/mesa: select ATOMFADD when source type is float< / li >
< li > st/mesa: expose GL_NV_shader_atomic_float when ATOMFADD is supported< / li >
< li > nv50/ir: add support for converting ATOMFADD to proper ir< / li >
< li > nvc0: enable GL_NV_shader_atomic_float on pre-Maxwell< / li >
< li > nv50,nvc0: add missing CAPs for unsupported features< / li >
< li > nv30: avoid setting user_priv without setting cur_ctx< / li >
< li > nv30: fix rare issue with fp unbinding not finding the bufctx< / li >
< li > nv30: add support for multi-layer transfers< / li >
< li > nv30: use correct helper to get blocks in y direction< / li >
< li > nv30: fix some s3tc layout issues< / li >
< li > nv30: disable rendering to 3D textures< / li >
< li > docs: fix gallium screen cap docs< / li >
< li > nv50,nvc0: mark textures dirty on fb update< / li >
< li > nvc0: don't put text segment into bufctx< / li >
< li > nvc0/ir: fix second tex argument after levelZero optimization< / li >
< li > nv50,nvc0: add explicit settings for recent caps< / li >
< li > nvc0: add support for handling indirect draws with attrib conversion< / li >
< li > nvc0/ir: always use CG mode for loads from atomic-only buffers< / li >
< li > nvc0: fix 3d images on kepler< / li >
< li > nv50,nvc0: use condition for occlusion queries when already complete< / li >
< li > nvc0: stick zero values for the compute invocation counts< / li >
< li > nvc0: we have 16k-sized framebuffers, fix default scissors< / li >
< li > swr: set PIPE_CAP_MAX_VARYINGS correctly< / li >
< li > glsl: fix recording of variables for XFB in TCS shaders< / li >
< / ul >
< p > Indrajit Das (1):< / p >
< ul >
< li > st/va: Return correct status from vlVaQuerySurfaceStatus< / li >
< / ul >
< p > Jakob Bornecrantz (1):< / p >
< ul >
< li > virgl/vtest: Use default socket name from protocol header< / li >
< / ul >
< p > Jan Vesely (2):< / p >
< ul >
< li > amd: Make vgpr-spilling depend on llvm version< / li >
< li > clover: Fix build after clang r348827< / li >
< / ul >
< p > Jason Ekstrand (207):< / p >
< ul >
< li > vulkan: Update the XML and headers to 1.1.91< / li >
< li > intel/fs,vec4: Clean up a repeated pattern with SSBOs< / li >
< li > intel/fs: Use the new nir_src_is_const and friends< / li >
< li > nir: Add a read_mask helper for ALU instructions< / li >
< li > intel/vec4: Use the new nir_src_is_const and friends< / li >
< li > intel/analyze_ubo_ranges: Use nir_src_is_const and friends< / li >
< li > anv: Use nir_src_is_const and friends in lowering code< / li >
< li > intel/fs: Add an assert to optimize_frontfacing_ternary< / li >
< li > nir/lower_alu_to_scalar: Don't try to lower unpack_32_2x16< / li >
< li > nir/builder: Assert that intN_t immediates fit< / li >
< li > nir/builder: Add iadd_imm and imul_imm helpers< / li >
< li > nir/builder: Add a nir_pack/unpack/bitcast helpers< / li >
< li > nir/spirv: Force 32-bit for UBO and SSBO Booleans< / li >
< li > nir/glsl: Force 32-bit for UBO and SSBO Booleans< / li >
< li > nir/lower_io: Add shared to get_io_offset_src< / li >
< li > nir: Add alignment parameters to SSBO, UBO, and shared access< / li >
< li > intel/compiler: Lower SSBO and shared loads/stores in NIR< / li >
< li > intel,nir: Move gl_LocalInvocationID lowering to nir_lower_system_values< / li >
< li > intel/fs,vec4: Fix a compiler warning< / li >
< li > vulkan: Update the XML and headers to 1.1.93< / li >
< li > anv: Expose VK_EXT_scalar_block_layout< / li >
< li > anv: Put robust buffer access in the pipeline hash< / li >
< li > anv/nir: Rework arguments to apply_pipeline_layout< / li >
< li > nir/derefs: Add a nir_derefs_do_not_alias enum value< / li >
< li > vulkan: Update the XML and headers to 1.1.95< / li >
< li > nir/opcodes: Pull in the type helpers from constant_expressions< / li >
< li > nir/opcodes: Rename tbool to tbool32< / li >
< li > nir/algebraic: Clean up some __str__ cruft< / li >
< li > nir/algebraic: Refactor codegen a bit< / li >
< li > nir/algebraic: Add support for unsized conversion opcodes< / li >
< li > nir/opt_algebraic: Simplify an optimization using the new search ops< / li >
< li > nir/opt_algebraic: Drop bit-size suffixes from conversions< / li >
< li > nir/opt_algebraic: Add 32-bit specifiers to a bunch of booleans< / li >
< li > nir: Make boolean conversions sized just like the others< / li >
< li > anv,radv: Disable VK_EXT_pci_bus_info< / li >
< li > intel/ir: Don't allow allocating zero registers< / li >
< li > spirv: Add support for MinLod< / li >
< li > nir/lower_tex: Simplify lower_gradient logic< / li >
< li > nir/lower_tex: Modify txd instructions instead of replacing them< / li >
< li > nir/lower_tex: Add lowering for some min_lod cases< / li >
< li > intel/fs: Support min_lod parameters on texture instructions< / li >
< li > anv: Advertise support for MinLod on Skylake+< / li >
< li > anv/pipeline: Set the correct binding count for compute shaders< / li >
< li > intel/blorp: Assert that we don't re-layout a compressed surface< / li >
< li > nir: Document the function inlining process< / li >
< li > nir: Allow [iu]mul_high on non-32-bit types< / li >
< li > nir/lower_int64: Add support for [iu]mul_high< / li >
< li > nir: Add a pass for lowering integer division by constants< / li >
< li > i965/vec4: Implement nir_op_uadd_sat< / li >
< li > i965: Enable nir_opt_idiv_const for 32 and 64-bit integers< / li >
< li > nir/lower_idiv: Use ilt instead of bit twiddling< / li >
< li > nir/tgsi: Use nir_bany in ttn_kill_if< / li >
< li > nir/constant_folding: Fix source bit size logic< / li >
< li > nir/algebraic: Optimize x2b(xneg(a)) -> a< / li >
< li > nir: Drop support for lower_b2f< / li >
< li > nir/algebraic: Make an optimization more specific< / li >
< li > nir: Rename Boolean-related opcodes to include 32 in the name< / li >
< li > nir/constant_expressions: Rework Boolean handling< / li >
< li > nir: Add support for 1-bit data types< / li >
< li > nir/large_constants: Properly handle 1-bit bools< / li >
< li > nir/algebraic: Generalize an optimization< / li >
< li > nir: Add 1-bit Boolean opcodes< / li >
< li > nir: Add a bool to int32 lowering pass< / li >
< li > nir: Switch to using 1-bit Booleans for almost everything< / li >
< li > nir/algebraic: Optimize 1-bit Booleans< / li >
< li > nir/algebraic: Add some optimizations for D3D-style Booleans< / li >
< li > radv: Fix a stupid if in gather_intrinsic_info< / li >
< li > st/nir: Use nir_src_as_uint for tokens< / li >
< li > vulkan: Update the XML and headers to 1.1.96< / li >
< li > anv,radv: Re-enable VK_EXT_pci_bus_info< / li >
< li > anv: Bump the patch version to 96< / li >
< li > nir/propagate_invariant: Skip unknown vars< / li >
< li > nir/linking_helpers: Look at derefs for modes< / li >
< li > nir/lower_io_arrays_to_elements: Look at derefs for modes< / li >
< li > nir/lower_io_to_scalar: Look at derefs for modes< / li >
< li > nir/lower_wpos_center: Look at derefs for modes< / li >
< li > nir/copy_prop_vars: Get modes directly from derefs< / li >
< li > nir/dead_write_vars: Get modes directly from derefs< / li >
< li > radv/query: Add a nir_test_flag helper< / li >
< li > radv/query: Use 1-bit booleans in query shaders< / li >
< li > intel/blorp: Be more conservative about copying clear colors< / li >
< li > vulkan: Update the XML and headers to 1.1.97< / li >
< li > glsl_type: Support serializing 8 and 16-bit types< / li >
< li > spirv: Handle any bit size in vector_insert/extract< / li >
< li > anv/apply_pipeline_layout: Set the cursor in lower_res_reindex_intrinsic< / li >
< li > spirv: Sign-extend array indices< / li >
< li > spirv: Emit NIR deref instructions on-the-fly< / li >
< li > nir/builder: Add nir_i2i and nir_u2u helpers which take a bit size< / li >
< li > spirv: Handle arbitrary bit sizes for deref array indices< / li >
< li > nir/validate: Require array indices to match the deref bit size< / li >
< li > nir: Allow storing to shader_storage< / li >
< li > nir: Distinguish between normal uniforms and UBOs< / li >
< li > glsl_type: Drop the glsl_get_array_instance C helper< / li >
< li > glsl_type: Add a C wrapper to get struct field offsets< / li >
< li > glsl_type: Simplify glsl_channel_type< / li >
< li > glsl_type: Add support for explicitly laid out matrices and arrays< / li >
< li > spirv: Propagate layout decorations to created glsl_types< / li >
< li > nir: Move propagation of cast derefs to a new nir_opt_deref pass< / li >
< li > nir: Add a ptr_as_array deref type< / li >
< li > nir/validate: Don't allow derefs in if conditions< / li >
< li > nir/opt_deref: Properly optimize ptr_as_array derefs< / li >
< li > nir/deref: Support casts and ptr_as_array in comparisons< / li >
< li > nir/deref: Skip over casts in fixup_deref_modes< / li >
< li > nir/remove_dead_variables: Properly handle deref casts< / li >
< li > nir/validate: Allow derefs in phi nodes< / li >
< li > nir/intrinsics: Allow deref sources to consume anything< / li >
< li > nir/intrinsics: Add access flags to load/store_deref< / li >
< li > nir/validate: Allow array derefs on vectors in more modes< / li >
< li > nir/lower_io: Add "explicit" IO lowering< / li >
< li > nir/vulkan: Add a descriptor type to vulkan resource intrinsics< / li >
< li > spirv: Add error checking for Block and BufferBlock decorations< / li >
< li > spirv: Choose atomic deref type with pointer_uses_ssa_offset< / li >
< li > spirv: Add explicit pointer types< / li >
< li > spirv: Make better use of vtn_pointer_uses_ssa_offset< / li >
< li > spirv: Add support for using derefs for UBO/SSBO access< / li >
< li > anv: Enable the new deref-based UBO/SSBO path< / li >
< li > spirv: Sort supported capabilities< / li >
< li > anv: Sort properties and features switch statements< / li >
< li > nir: Add some more int64 lowering helpers< / li >
< li > anv/pipeline: Constant fold after apply_pipeline_layout< / li >
< li > anv/pipeline: Move wpos and input attachment lowering to lower_nir< / li >
< li > compiler/types: Serialize/deserialize subpass input types correctly< / li >
< li > anv/pipeline: Hash shader modules and spec constants separately< / li >
< li > anv/pipeline_cache: Add support for caching NIR< / li >
< li > anv/pipeline: Cache the pre-lowered NIR< / li >
< li > intel/peephole_ffma: Fix swizzle propagation< / li >
< li > spirv: Whack sampler/image pointers to uniform< / li >
< li > spirv: Contain the GLSLang issue #179 workaround to old GLSLang< / li >
< li > intel/nir: Call nir_opt_deref in brw_nir_optimize< / li >
< li > nir/gcm: Support deref instructions< / li >
< li > spirv: Emit switch conditions on-the-fly< / li >
< li > intel/blorp: Add two more filter modes< / li >
< li > anv: Rename has_resolve to has_color_resolve< / li >
< li > anv/blorp: Refactor MSAA resolves into an exportable helper function< / li >
< li > anv: Move resolve_subpass to genX_cmd_buffer.c< / li >
< li > anv: Implement VK_KHR_depth_stencil_resolve< / li >
< li > nir: Add a bool to float32 lowering pass< / li >
< li > intel/eu: Stop overriding exec sizes in send_indirect_message< / li >
< li > intel/fs: Don't touch accumulator destination while applying regioning alignment rule< / li >
< li > anv: Re-sort the extensions list< / li >
< li > anv: Only parse pImmutableSamplers if the descriptor has samplers< / li >
< li > relnotes: Add newly added Vulkan extensions< / li >
< li > anv/pipeline: Add a pdevice helper variable< / li >
< li > nir: Mark deref UBO and SSBO access as non-scalar< / li >
< li > spirv: Update the JSON and headers from Khronos master< / li >
< li > anv: Always emit at least one vertex element< / li >
< li > spirv: Initialize struct member offsets to -1< / li >
< li > spirv: Only split blocks< / li >
< li > spirv: Only set interface_type on blocks< / li >
< li > nir: Preserve offsets in lower_io_to_scalar_early< / li >
< li > nir/xfb: Fix offset accounting for dvec3/4< / li >
< li > nir/xfb: Properly handle arrays of blocks< / li >
< li > anv: Add but do not enable VK_EXT_transform_feedback< / li >
< li > anv: Add pipeline cache support for xfb_info< / li >
< li > anv: Implement the basic form of VK_EXT_transform_feedback< / li >
< li > anv: Implement vkCmdDrawIndirectByteCountEXT< / li >
< li > anv: Implement CmdBegin/EndQueryIndexed< / li >
< li > genxml: Add SO_PRIM_STORAGE_NEEDED and SO_NUM_PRIMS_WRITTEN< / li >
< li > anv: Implement transform feedback queries< / li >
< li > nir: Add load/store/atomic global intrinsics< / li >
< li > nir/lower_io: Add a 32 and 64-bit global address formats< / li >
< li > nir/lower_io: Add support for nir_var_mem_global< / li >
< li > nir/validate: Allow array derefs of vectors for nir_var_mem_global< / li >
< li > nir: Allow SSBOs and global to alias< / li >
< li > spirv: Drop a bogus assert< / li >
< li > spirv: Handle OpTypeForwardPointer< / li >
< li > spirv: Implement OpConvertPtrToU and OpConvertUToPtr< / li >
< li > spirv: Add support for SPV_EXT_physical_storage_buffer< / li >
< li > intel/fs: Get rid of fs_inst::equals< / li >
< li > intel/defines: Explicitly cast to uint32_t in SET_FIELD and SET_BITS< / li >
< li > intel/fs: Handle IMAGE_SIZE in size_read() and is_send_from_grf()< / li >
< li > intel/fs: Take an explicit exec size in brw_surface_payload_size()< / li >
< li > intel/eu: Add has_simd4x2 bools to surface_write functions< / li >
< li > intel/eu: Rework surface descriptor helpers< / li >
< li > intel/fs: Add a generic SEND opcode< / li >
< li > intel/fs: Use SHADER_OPCODE_SEND for surface messages< / li >
< li > intel/fs: Use a logical opcode for IMAGE_SIZE< / li >
< li > intel/fs: Use SHADER_OPCODE_SEND for texturing on gen7+< / li >
< li > intel/fs: Use SHADER_OPCODE_SEND for varying UBO pulls on gen7+< / li >
< li > intel/eu: Use GET_BITS in brw_inst_set_send_ex_desc< / li >
< li > intel/eu/validate: SEND restrictions also apply to SENDC< / li >
< li > intel/eu: Add more message descriptor helpers< / li >
< li > intel/disasm: Rework SEND decoding to use descriptors< / li >
< li > intel/inst: Fix the ia16_addr_imm helpers< / li >
< li > intel/inst: Indent some code< / li >
< li > intel/eu: Add support for the SENDS[C] messages< / li >
< li > intel/disasm: Properly disassemble split sends< / li >
< li > intel/fs: Support SENDS in SHADER_OPCODE_SEND< / li >
< li > intel/fs: Add interference between SENDS sources< / li >
< li > intel/fs: Use split sends for surface writes on gen9+< / li >
< li > intel/fs: Do the grf127 hack on SIMD8 instructions in SIMD16 mode< / li >
< li > nir/deref: Rematerialize parents in rematerialize_derefs_in_use_blocks< / li >
< li > intel/fs: Bail in optimize_extract_to_float if we have modifiers< / li >
< li > compiler/types: Add a contains_64bit helper< / li >
< li > nir/xfb: Properly align 64-bit values< / li >
< li > nir: Rewrite lower_clip_cull_distance_arrays to do a lot less lowering< / li >
< li > nir/xfb: Work in terms of components rather than slots< / li >
< li > nir/xfb: Handle compact arrays in gather_xfb_info< / li >
< li > nir/lower_clip_cull: Fix an incorrect assert< / li >
< li > anv: Count surfaces for non-YCbCr images in GetDescriptorSetLayoutSupport< / li >
< li > spirv: OpImageQueryLod requires a sampler< / li >
< li > intel,nir: Lower TXD with min_lod when the sampler index is not < 16< / li >
< li > spirv: Pull offset/stride from the pointer for OpArrayLength< / li >
< li > anv: Refactor descriptor pushing a bit< / li >
< li > anv: Take references to push descriptor set layouts< / li >
< li > nir: Add a pass for lowering IO back to vector when possible< / li >
< li > intel/nir: Vectorize all IO< / li >
< / ul >
< p > Jiang, Sonny (1):< / p >
< ul >
< li > radeonsi: add compute_last_block to configure the partial block fields< / li >
< / ul >
< p > Jon Turney (3):< / p >
< ul >
< li > glx: Fix compilation with GLX_USE_WINDOWSGL< / li >
< li > appveyor: put build steps in a script, rather than inline in appveyor.yml< / li >
< li > appveyor: Add a Cygwin build script< / li >
< / ul >
< p > Jonathan Marek (42):< / p >
< ul >
< li > nir: add fceil lowering< / li >
< li > freedreno: a2xx: fd2_draw update< / li >
< li > freedreno/a2xx: fix POINT_MINMAX_MAX overflow< / li >
< li > freedreno: add missing a20x ids< / li >
< li > freedreno/a2xx: set VIZ_QUERY_ID on a20x< / li >
< li > freedreno/a2xx: Compute depth base in gmem correctly< / li >
< li > freedreno: a2xx texture update< / li >
< li > freedreno: use GENERIC instead of TEXCOORD for blit program< / li >
< li > freedreno: use MSM_BO_SCANOUT with scanout buffers< / li >
< li > glsl/nir: int constants as float for native_integers=false< / li >
< li > glsl/nir: ftrunc for native_integers=false float to int cast< / li >
< li > glsl/nir: keep bool types when native_integers=false< / li >
< li > freedreno: a2xx: cleanup init_shader_const< / li >
< li > freedreno: a2xx: cleanup REG_A2XX_PA_CL_VTE_CNTL< / li >
< li > freedreno: a2xx: fix gmem2mem viewport< / li >
< li > freedreno: a2xx: fix VERTEX_REUSE/DEALLOC on a20x< / li >
< li > freedreno: a2xx: fix non-zero texture base offsets< / li >
< li > freedreno: a2xx: sysmem rendering< / li >
< li > freedreno: a2xx: NIR backend< / li >
< li > freedreno: a2xx: insert scalar MOV to allow 2 source scalar< / li >
< li > freedreno: a2xx: add ir2 copy propagation< / li >
< li > freedreno: a2xx: add partial lower_scalar pass for ir2< / li >
< li > freedreno: add renderonly scanout< / li >
< li > freedreno: a2xx: ir2 cleanup< / li >
< li > freedreno: a2xx: enable early-Z testing< / li >
< li > freedreno: update a2xx registers< / li >
< li > freedreno: a2xx: a20x hw binning< / li >
< li > freedreno: a2xx: clear fixes and fast clear path< / li >
< li > freedreno: a2xx: minor solid_vertexbuf fixups< / li >
< li > freedreno: a2xx: add perfcntrs< / li >
< li > kmsro: Add freedreno renderonly support< / li >
< li > st/dri: invalidate_resource depth/stencil before flush_resource< / li >
< li > mesa/st: wire up DiscardFramebuffer< / li >
< li > freedreno: fix invalidate logic< / li >
< li > freedreno: fix depth usage logic< / li >
< li > freedreno: fix sysmem rendering being used when clear is used< / li >
< li > freedreno: a2xx: fix fast clear< / li >
< li > freedreno: a2xx: don't write 4th vertex in mem2gmem< / li >
< li > freedreno: a2xx: add use_hw_binning function< / li >
< li > freedreno: a2xx: fix fast clear for some gmem configurations< / li >
< li > freedreno: a2xx: fix mipmapping for NPOT textures< / li >
< li > freedreno: use renderonly path for buffers allocated with modifiers< / li >
< / ul >
< p > Jordan Justen (3):< / p >
< ul >
< li > docs: Document GitLab merge request process (email alternative)< / li >
< li > i965/genX_state: Add register access functions< / li >
< li > i965/compute: Emit GPGPU_WALKER in genX_state_upload< / li >
< / ul >
< p > Jose Maria Casanova Crespo (1):< / p >
< ul >
< li > glsl: TCS outputs can not be transform feedback candidates on GLES< / li >
< / ul >
< p > José Fonseca (2):< / p >
< ul >
< li > appveyor: Revert commits adding Cygwin support.< / li >
< li > scons: Workaround failures with MSVC when using SCons 3.0.[2-4].< / li >
< / ul >
< p > Juan A. Suarez Romero (17):< / p >
< ul >
< li > docs: add release notes for 18.2.5< / li >
< li > docs: add sha256 checksums for 18.2.5< / li >
< li > docs: update calendar, add news item and link release notes for 18.2.5< / li >
< li > docs: add release notes for 18.2.6< / li >
< li > docs: add sha256 checksums for 18.2.6< / li >
< li > docs: update calendar, add news item and link release notes for 18.2.6< / li >
< li > docs: extends 18.2 lifecycle< / li >
< li > docs: add release notes for 18.2.7< / li >
< li > docs: add sha256 checksums for 18.2.7< / li >
< li > docs: update calendar, add news item and link release notes for 18.2.7< / li >
< li > docs: add release notes for 18.2.8< / li >
< li > docs: add sha256 checksums for 18.2.8< / li >
< li > docs: update calendar, add news item and link release notes for 18.2.8< / li >
< li > anv/cmd_buffer: check for NULL framebuffer< / li >
< li > genxml: add missing field values for 3DSTATE_SF< / li >
< li > anv: advertise 8 subpixel precision bits< / li >
< li > anv: destroy descriptor sets when pool gets reset< / li >
< / ul >
< p > Józef Kucia (1):< / p >
< ul >
< li > nir: Fix assert in print_intrinsic_instr().< / li >
< / ul >
< p > Karol Herbst (35):< / p >
< ul >
< li > nv50/ir: print color masks of tex instructions< / li >
< li > nv50/ra: add condenseDef overloads for partial condenses< / li >
< li > nv50/ir: add scalar field to TexInstructions< / li >
< li > gm107/ir: use scalar tex instructions where possible< / li >
< li > gm107/ir: fix compile time warning in getTEXSMask< / li >
< li > nir: add const_index parameters to system value builder function< / li >
< li > nir: replace nir_load_system_value calls with appropiate builder functions< / li >
< li > nir/spirv: cast shift operand to u32< / li >
< li > nv50,nvc0: Fix gallium nine regression regarding sampler bindings< / li >
< li > nv50/ir: initialize relDegree staticly< / li >
< li > nouveau: use atomic operations for driver statistics< / li >
< li > nv50/ir: fix use-after-free in ConstantFolding::visit< / li >
< li > nir: rename global/local to private/function memory< / li >
< li > nv50/ir: disable tryCollapseChainedMULs in ConstantFolding for precise instructions< / li >
< li > gm107/ir: disable TEXS for tex with derivAll set< / li >
< li > nir: rename nir_var_private to nir_var_shader_temp< / li >
< li > nir: rename nir_var_function to nir_var_function_temp< / li >
< li > nir: rename nir_var_ubo to nir_var_mem_ubo< / li >
< li > nir: rename nir_var_ssbo to nir_var_mem_ssbo< / li >
< li > nir: rename nir_var_shared to nir_var_mem_shared< / li >
< li > nir/spirv: handle SpvStorageClassCrossWorkgroup< / li >
< li > glsl/lower_output_reads: set invariant and precise flags on temporaries< / li >
< li > nir: replace more nir_load_system_value calls with builder functions< / li >
< li > nir/validate: allow to check against a bitmask of bit_sizes< / li >
< li > nir: add legal bit_sizes to intrinsics< / li >
< li > nir: add bit_size parameter to system values with multiple allowed bit sizes< / li >
< li > mesa: add MESA_SHADER_KERNEL< / li >
< li > vtn: handle SpvExecutionModelKernel< / li >
< li > nir/spirv: handle ContractionOff execution mode< / li >
< li > gk104/ir: Use the new rcp/rsq in library< / li >
< li > gm107/ir: add fp64 rcp< / li >
< li > gm107/ir: add fp64 rsq< / li >
< li > gallium: add PIPE_CAP_MAX_VARYINGS< / li >
< li > st/mesa: require RGBA2, RGB4, and RGBA4 to be renderable< / li >
< li > nir/opt_if: don't mark progress if nothing changes< / li >
< / ul >
< p > Kenneth Graunke (41):< / p >
< ul >
< li > intel: Use a URB start offset of 0 for disabled stages.< / li >
< li > st/mesa: Pull nir_lower_wpos_ytransform work into a helper function.< / li >
< li > st/nir: Drop unused parameter from st_nir_assign_uniform_locations().< / li >
< li > st/mesa: Don't record garbage streamout information in the non-SSO case.< / li >
< li > i915: Delete swizzling detection logic.< / li >
< li > nir: Use nir_shader_get_entrypoint in nir_lower_clip_vs().< / li >
< li > nir: Inline lower_clip_vs() into nir_lower_clip_vs().< / li >
< li > nir: Save nir_variable pointers in nir_lower_clip_vs rather than locs.< / li >
< li > nir: Make nir_lower_clip_vs optionally work with variables.< / li >
< li > i965: Allow only one slot of clip distances to be set on Gen4-5.< / li >
< li > i965: Use a 'nir' temporary rather than poking at brw_program< / li >
< li > i965: Do NIR shader cloning in the caller.< / li >
< li > intel/compiler: Use nir's info when checking uses_streams.< / li >
< li > intel/blorp: Expand blorp_address::offset to be 64 bits.< / li >
< li > i965: Delete dead brw_meta_resolve_color prototype.< / li >
< li > i965: Flip arguments to load_register_reg helpers.< / li >
< li > genxml: Consistently use a numeric "MOCS" field< / li >
< li > i965: Don't override subslice count to 4 on Gen11.< / li >
< li > st/mesa: Drop dead 'passthrough_fs' field.< / li >
< li > st/mesa: Drop !passColor optimization in drawpixels shaders.< / li >
< li > st/mesa: Don't open code the drawpixels vertex shader.< / li >
< li > st/mesa: Combine the DrawPixels and Bitmap passthrough VS programs.< / li >
< li > st/nir: Gather info after applying lowering FS variant features< / li >
< li > st/nir: Drop unused gl_program parameter in VS input handling helper.< / li >
< li > nir: Fix gl_nir_lower_samplers_as_deref's structure type handling.< / li >
< li > nir: Make gl_nir_lower_samplers use gl_nir_lower_samplers_as_deref< / li >
< li > blorp: Add blorp_get_surface_address to the driver interface.< / li >
< li > blorp: Pass the batch to lookup/upload_shader instead of context< / li >
< li > nir: Allow a non-existent sampler deref in nir_lower_samplers_as_deref< / li >
< li > st/nir: Lower TES gl_PatchVerticesIn to a constant if linked with a TCS.< / li >
< li > i965: Drop mark_surface_used mechanism.< / li >
< li > st/mesa: Make an enum for pipeline statistics query result indices.< / li >
< li > st/mesa: Rearrange PIPE_QUERY_PIPELINE_STATISTICS result fetching.< / li >
< li > gallium: Add the ability to query a single pipeline statistics counter< / li >
< li > st/mesa: Optionally override RGB/RGBX dst alpha blend factors< / li >
< li > gallium: Add forgotten docs for PIPE_CAP_GLSL_TESS_LEVELS_AS_INPUTS.< / li >
< li > st/mesa: Limit GL_MAX_[NATIVE_]PROGRAM_PARAMETERS_ARB to 2048< / li >
< li > anv: Put MOCS in the correct location< / li >
< li > nir: Don't reassociate add/mul chains containing only constants< / li >
< li > compiler: Mark clip/cull distance arrays as compact before lowering.< / li >
< li > spirv: Eliminate dead input/output variables after translation.< / li >
< / ul >
< p > Kirill Burtsev (1):< / p >
< ul >
< li > loader: free error state, when checking the drawable type< / li >
< / ul >
< p > Kristian H. Kristensen (14):< / p >
< ul >
< li > freedreno/a6xx: Clear z32 and separate stencil with blitter< / li >
< li > freedreno/a6xx: Move restore blits to IB< / li >
< li > freedreno/a6xx: Move resolve blits to an IB< / li >
< li > freedreno/a6xx: Clear gmem buffers at flush time< / li >
< li > gallium: Android build fixes< / li >
< li > mesa: Add core support for EXT_multisampled_render_to_texture{,2}< / li >
< li > gallium: Add new PIPE_CAP_SURFACE_SAMPLE_COUNT< / li >
< li > st/mesa: Add support for EXT_multisampled_render_to_texture< / li >
< li > freedreno: Add support for EXT_multisampled_render_to_texture< / li >
< li > freedreno: Fix the Makefile.am fix< / li >
< li > glapi: fixup EXT_multisampled_render_to_texture dispatch< / li >
< li > freedreno: Synchronize batch and flush for staging resource< / li >
< li > freedreno/a6xx: Turn on texture tiling by default< / li >
< li > freedreno/a6xx: Emit blitter dst with OUT_RELOCW< / li >
< / ul >
< p > Leo Liu (2):< / p >
< ul >
< li > st/va: fix the incorrect max profiles report< / li >
< li > st/va/vp9: set max reference as default of VP9 reference number< / li >
< / ul >
< p > Lionel Landwerlin (47):< / p >
< ul >
< li > intel/dump_gpu: add missing gdb option< / li >
< li > intel/sanitize_gpu: add help/gdb options to wrapper< / li >
< li > intel/sanitize_gpu: deal with non page multiple buffer sizes< / li >
< li > intel/sanitize_gpu: add debug message on mmap fail< / li >
< li > intel/decoders: fix instruction base address parsing< / li >
< li > anv: stub internal android code< / li >
< li > anv/android: mark gralloc allocated BOs as external< / li >
< li > intel/dump_gpu: move output option together< / li >
< li > intel/dump_gpu: add platform option< / li >
< li > intel/aub_read: remove useless breaks< / li >
< li > nir/lower_tex: add alpha channel parameter for yuv lowering< / li >
< li > nir/lower_tex: Add AYUV lowering support< / li >
< li > dri: add AYUV format< / li >
< li > i965: add support for sampling from AYUV< / li >
< li > anv: simplify internal address offset< / li >
< li > anv/image: remove unused parameter< / li >
< li > anv/lower_ycbcr: make sure to set 0s on all components< / li >
< li > anv: associate vulkan formats with aspects< / li >
< li > anv: use image aspects rather than computed ones< / li >
< li > anv: move helper function internally< / li >
< li > egl/dri: fix error value with unknown drm format< / li >
< li > intel/decoders: read ring buffer length< / li >
< li > intel/aubinator: fix ring buffer pointer< / li >
< li > intel/aub_viewer: fix dynamic state printing< / li >
< li > intel/aub_viewer: Print blend states properly< / li >
< li > anv: flush pipeline before query result copies< / li >
< li > anv/query: flush render target before copying results< / li >
< li > anv: don't do partial resolve on layer > 0< / li >
< li > intel/aub_viewer: fix shader get_bo< / li >
< li > intel/aub_viewer: fixup 0x address prefix< / li >
< li > intel/aub_viewer: print address of missing shader< / li >
< li > intel/aub_viewer: fix shader view< / li >
< li > intel/aub_viewer: fold binding/sampler table items< / li >
< li > intel/aub_viewer: highlight true booleans< / li >
< li > i965: limit VF caching workaround to gen8/9/10< / li >
< li > intel/blorp: emit VF caching workaround before 3DSTATE_VERTEX_BUFFERS< / li >
< li > i965: include draw_params/derived_draw_params for VF cache workaround< / li >
< li > i965: add CS stall on VF invalidation workaround< / li >
< li > anv: explictly specify format for blorp ccs/mcs op< / li >
< li > anv: flush fast clear colors into compressed surfaces< / li >
< li > anv: fix invalid binding table index computation< / li >
< li > anv: narrow flushing of the render target to buffer writes< / li >
< li > anv: document cache flushes & invalidations< / li >
< li > intel/genxml: add missing MI_PREDICATE compare operations< / li >
< li > vulkan: make generated enum to strings helpers available from c++< / li >
< li > intel: fix urb size for CFL GT1< / li >
< li > intel/compiler: use correct swizzle for replacement< / li >
< / ul >
< p > Lucas Stach (6):< / p >
< ul >
< li > etnaviv: use dummy RT buffer when rendering without color buffer< / li >
< li > etnaviv: use surface format directly< / li >
< li > st/dri: allow both render and sampler compatible dma-buf formats< / li >
< li > st/dri: replace format conversion functions with single mapping table< / li >
< li > etnaviv: enable full overwrite in a few more cases< / li >
< li > etnaviv: annotate variables only used in debug build< / li >
< / ul >
< p > Marek Olšák (94):< / p >
< ul >
< li > st/va: fix incorrect use of resource_destroy< / li >
< li > ac/surface: remove the overallocation workaround for Vega12< / li >
< li > radeonsi: use better DCC clear codes< / li >
< li > radeonsi: don't set the CB clear color registers for 0/1 clear colors on Raven2< / li >
< li > gallium: add PIPE_CONTEXT_LOSE_CONTEXT_ON_RESET< / li >
< li > radeonsi: stop command submission with PIPE_CONTEXT_LOSE_CONTEXT_ON_RESET only< / li >
< li > st/mesa: disable L3 thread pinning< / li >
< li > mesa: mark GL_SR8_EXT non-renderable on GLES< / li >
< li > radeonsi: fix video APIs on Raven2< / li >
< li > gallium/u_tests: add a compute shader test that clears an image< / li >
< li > gallium/u_tests: fix MSVC build by using old-style zero initializers< / li >
< li > mesa/glthread: pass the function name to _mesa_glthread_restore_dispatch< / li >
< li > mesa/glthread: enable immediate mode< / li >
< li > drirc: enable glthread for Talos Principle< / li >
< li > st/mesa: regularly re-pin driver threads to the CCX where the app thread is< / li >
< li > st/mesa: pin driver threads to a fixed CCX when glthread is enabled< / li >
< li > radeonsi: don't send data after write-confirm with BOTTOM_OF_PIPE_TS< / li >
< li > radeonsi: go back to using bottom-of-pipe for beginning of TIME_ELAPSED< / li >
< li > winsys/amdgpu: fix a buffer leak in amdgpu_bo_from_handle< / li >
< li > winsys/amdgpu: fix a device handle leak in amdgpu_winsys_create< / li >
< li > radeonsi: clean up primitive binning enablement< / li >
< li > radeonsi: use structured buffer intrinsics for image views< / li >
< li > radeonsi: fix is_oneway_access_only for image stores< / li >
< li > radeonsi: small cleanup for memory opcodes< / li >
< li > tgsi/scan: add more information about bindless usage< / li >
< li > radeonsi/nir: parse more information about bindless usage< / li >
< li > radeonsi: fix is_oneway_access_only for bindless images< / li >
< li > winsys/amdgpu: always reclaim/release slabs if there is not enough memory< / li >
< li > radeonsi: generalize the slab allocator code to allow layered slab allocators< / li >
< li > winsys/amdgpu: optimize slab allocation for 2 MB amdgpu page tables< / li >
< li > winsys/amdgpu: clean up code around BO VM alignment< / li >
< li > winsys/amdgpu: use > = instead of > for VM address alignment< / li >
< li > winsys/amdgpu: increase the VM alignment to the MSB of the size for Gfx9< / li >
< li > winsys/amdgpu: overallocate buffers for faster address translation on Gfx9< / li >
< li > winsys/amdgpu,radeon: pass vm_alignment to buffer_from_handle< / li >
< li > winsys/amdgpu: use optimal VM alignment for imported buffers< / li >
< li > winsys/amdgpu: use optimal VM alignment for CPU allocations< / li >
< li > radeonsi: allow si_cp_dma_clear_buffer to clear GDS from any IB< / li >
< li > winsys/amdgpu: add support for allocating GDS and OA resources< / li >
< li > radeonsi: add memory management stress tests for GDS< / li >
< li > Revert "winsys/amdgpu: overallocate buffers for faster address translation on Gfx9"< / li >
< li > st/mesa: expose GL_OES_texture_view< / li >
< li > mesa: expose GL_EXT_texture_view as an alias of GL_OES_texture_view< / li >
< li > mesa: expose EXT_texture_compression_rgtc on GLES< / li >
< li > mesa: expose EXT_texture_compression_bptc in GLES< / li >
< li > mesa: expose AMD_texture_texture4< / li >
< li > st/mesa: expose EXT_render_snorm on GLES< / li >
< li > radeonsi: don't emit redundant PKT3_NUM_INSTANCES packets< / li >
< li > radeonsi: call si_fix_resource_usage for the GS copy shader as well< / li >
< li > radeonsi: make si_cp_wait_mem more configurable< / li >
< li > radeonsi: use u_decomposed_prims_for_vertices instead of u_prims_for_vertices< / li >
< li > radeonsi: remove unused variables in si_insert_input_ptr< / li >
< li > radeonsi: always unmap texture CPU mappings on 32-bit CPU architectures< / li >
< li > ac: remove unused variable from ac_build_ddxy< / li >
< li > st/mesa: unify window-system renderbuffer initialization< / li >
< li > st/mesa: don't reference pipe_surface locally in PBO code< / li >
< li > st/mesa: don't leak pipe_surface if pipe_context is not current< / li >
< li > st/dri: fix dri2_format_table for argb1555 and rgb565< / li >
< li > radeonsi: also apply the GS hang workaround to draws without tessellation< / li >
< li > winsys/amdgpu: fix whitespace< / li >
< li > winsys/amdgpu: use the new BO list API< / li >
< li > radeonsi: fix a u_blitter crash after a shader with FBFETCH< / li >
< li > radeonsi: fix rendering to tiny viewports where the viewport center is > 8K< / li >
< li > radeonsi: use buffer_store_format_x & xy< / li >
< li > radeonsi: remove redundant call to emit_cache_flush in compute clear/copy< / li >
< li > radeonsi: compile clear and copy buffer compute shaders on demand< / li >
< li > radeonsi: correct WRITE_DATA.DST_SEL definitions< / li >
< li > radeonsi: fix the top-of-pipe fence on SI< / li >
< li > radeonsi: don't use WRITE_DATA.DST_SEL == MEM_GRBM on > = CIK< / li >
< li > radeonsi: move PKT3_WRITE_DATA generation into a helper function< / li >
< li > gallium: add SINT formats to have exact counterparts to SNORM formats< / li >
< li > gallium/util: add util_format_snorm8_to_sint8 (from radeonsi)< / li >
< li > radeonsi: disable render cond & pipeline stats for internal compute dispatches< / li >
< li > radeonsi: rename rscreen -> sscreen< / li >
< li > radeonsi: rename rview -> sview< / li >
< li > winsys/amdgpu: rename rfence, rsrc, rdst -> afence, asrc, adst< / li >
< li > radeonsi: remove r600 from comments< / li >
< li > radeonsi: rename r600_resource -> si_resource< / li >
< li > radeonsi: rename rquery -> squery< / li >
< li > radeonsi: rename rsrc -> ssrc, rdst -> sdst< / li >
< li > radeonsi: rename rbo, rbuffer to buf or buffer< / li >
< li > radeonsi: rename rfence -> sfence< / li >
< li > st/mesa: purge framebuffers when unbinding a context< / li >
< li > st/mesa: fix PRIMITIVES_GENERATED query after the "pipeline stat single" changes< / li >
< li > ac: use the correct LLVM processor name on Raven2< / li >
< li > radeonsi: fix crashing performance counters (division by zero)< / li >
< li > meson: drop the xcb-xrandr version requirement< / li >
< li > gallium/u_threaded: fix EXPLICIT_FLUSH for flush offsets > 0< / li >
< li > radeonsi: fix EXPLICIT_FLUSH for flush offsets > 0< / li >
< li > winsys/amdgpu: don't drop manually added fence dependencies< / li >
< li > radeonsi: add driconf option radeonsi_enable_nir< / li >
< li > radeonsi: always enable NIR for Civilization 6 to fix corruption< / li >
< li > driconf: add Civ6Sub executable for Civilization 6< / li >
< li > tgsi: don't set tgsi_info::uses_bindless_images for constbufs and hw atomics< / li >
< / ul >
< p > Mario Kleiner (4):< / p >
< ul >
< li > radeonsi: Fix use of 1- or 2- component GL_DOUBLE vbo's.< / li >
< li > egl/wayland: Allow client-> server format conversion for PRIME offload. (v2)< / li >
< li > egl/wayland-drm: Only announce formats via wl_drm which the driver supports.< / li >
< li > drirc: Add sddm-greeter to adaptive_sync blacklist.< / li >
< / ul >
< p > Mark Janes (3):< / p >
< ul >
< li > Revert "i965/batch: avoid reverting batch buffer if saved state is an empty"< / li >
< li > Revert "Implementation of egl dri2 drivers for MESA_query_driver"< / li >
< li > Revert "Implement EGL API for MESA_query_driver"< / li >
< / ul >
< p > Mathias Fröhlich (17):< / p >
< ul >
< li > mesa: Remove needless indirection in some draw functions.< / li >
< li > mesa: Rename gl_vertex_array_object::_Enabled -> Enabled.< / li >
< li > mesa: Use the gl_vertex_array_object::Enabled bitfield.< / li >
< li > mesa: Use gl_vertex_array_object::Enabled for glGet.< / li >
< li > mesa: Remove gl_array_attributes::Enabled.< / li >
< li > mesa: Work with bitmasks when en/dis-abling VAO arrays.< / li >
< li > mesa: Unify glEdgeFlagPointer data type.< / li >
< li > nouveau: Use gl_array_attribute::_ElementSize.< / li >
< li > tnl: Use gl_array_attribute::_ElementSize.< / li >
< li > mesa: Factor out struct gl_vertex_format.< / li >
< li > mesa: Remove unneeded bitfield widths from the VAO.< / li >
< li > mesa/st: Only care about the uploader if it was used.< / li >
< li > mesa/st: Only unmap the uploader that was actually used.< / li >
< li > mesa/st: Factor out array and buffer setup from st_atom_array.c.< / li >
< li > mesa/st: Avoid extra references in the feedback draw function scope.< / li >
< li > mesa/st: Use binding information from the VAO in feedback rendering.< / li >
< li > mesa/st: Make st_pipe_vertex_format static.< / li >
< / ul >
< p > Matt Turner (41):< / p >
< ul >
< li > util/ralloc: Switch from DEBUG to NDEBUG< / li >
< li > util/ralloc: Make sizeof(linear_header) a multiple of 8< / li >
< li > nir: Call fflush() at the end of nir_print_shader()< / li >
< li > glsl: Remove unused member variable< / li >
< li > gallivm: Use nextafterf(0.5, 0.0) as rounding constant< / li >
< li > mesa: Revert INTEL_fragment_shader_ordering support< / li >
< li > Revert "st/mesa: silenced unhanded enum warning in st_glsl_to_tgsi.cpp"< / li >
< li > i965/fs: Handle V/UV immediates in dump_instructions()< / li >
< li > glsl: Add function support to glsl_to_nir< / li >
< li > glsl: Create file to contain software fp64 functions< / li >
< li > glsl: Add "built-in" functions to do ffma(fp64)< / li >
< li > glsl: Add "built-in" functions to do fmin/fmax(fp64)< / li >
< li > glsl: Add "built-in" function to do ffloor(fp64)< / li >
< li > glsl: Add "built-in" functions to do ffract(fp64)< / li >
< li > glsl: Add "built-in" functions to convert bool to double< / li >
< li > nir: Rework nir_lower_constant_initializers() to handle functions< / li >
< li > nir: Tag entrypoint for easy recognition by nir_shader_get_entrypoint()< / li >
< li > nir: Wire up int64 lowering functions< / li >
< li > nir: Implement lowering of 64-bit shift operations< / li >
< li > nir: Add and set info::uses_64bit< / li >
< li > nir: Create nir_builder in nir_lower_doubles_impl()< / li >
< li > nir: Add lowering support for 64-bit operations to software< / li >
< li > nir: Unset metadata debug bit if no progress made< / li >
< li > intel/compiler: Lower 64-bit MOV/SEL operations< / li >
< li > intel/compiler: Split 64-bit MOV-indirects if needed< / li >
< li > intel/compiler: Avoid false positive assertions< / li >
< li > intel/compiler: Rearrange code to avoid future problems< / li >
< li > intel/compiler: Prevent warnings in the following patch< / li >
< li > intel/compiler: Expand size of the 'nr' field< / li >
< li > intel/compiler: Heap-allocate temporary storage< / li >
< li > i965: Compile fp64 software routines and lower double-ops< / li >
< li > i965: Enable 64-bit GLSL extensions< / li >
< li > i965: Compile fp64 funcs only if we do not have 64-bit hardware support< / li >
< li > intel/compiler: Reset default flag register in brw_find_live_channel()< / li >
< li > gallium: Enable ASIMD/NEON on aarch64.< / li >
< li > gallivm: Return true from arch_rounding_available() if NEON is available< / li >
< li > intel/compiler: Add a file-level description of brw_eu_validate.c< / li >
< li > i965: Always compile fp64 funcs when needed< / li >
< li > nir: Optimize double-precision lower_round_even()< / li >
< li > intel/compiler: Avoid propagating inequality cmods if types are different< / li >
< li > intel/compiler/test: Add unit test for mismatched signedness comparison< / li >
< / ul >
< p > Mauro Rossi (6):< / p >
< ul >
< li > android: gallium/auxiliary: add include to get u_debug.h header< / li >
< li > android: radv: add libmesa_git_sha1 static dependency< / li >
< li > android: amd/addrlib: update Mesa's copy of addrlib< / li >
< li > android: st/mesa: fix building error due to sched_getcpu()< / li >
< li > android: anv: fix generated files depedencies (v2)< / li >
< li > android: anv: fix libexpat shared dependency< / li >
< / ul >
< p > Maya Rashish (2):< / p >
< ul >
< li > radeon: fix printf format specifier.< / li >
< li > configure: fix test portability< / li >
< / ul >
< p > Michal Srb (2):< / p >
< ul >
< li > gallium: Constify drisw_loader_funcs struct< / li >
< li > drisw: Use separate drisw_loader_funcs for shm< / li >
< / ul >
< p > Michel Dänzer (4):< / p >
< ul >
< li > winsys/amdgpu: Stop using amdgpu_bo_handle_type_kms_noimport< / li >
< li > winsys/amdgpu: Pull in LLVM CFLAGS< / li >
< li > amd/common: Restore v4i32 suffix for llvm.SI.load.const intrinsic< / li >
< li > loader/dri3: Use strlen instead of sizeof for creating VRR property atom< / li >
< / ul >
< p > Neha Bhende (1):< / p >
< ul >
< li > st/mesa: Fix topogun-1.06-orc-84k-resize.trace crash< / li >
< / ul >
< p > Neil Roberts (4):< / p >
< ul >
< li > freedreno: Add .dir-locals to the common directory< / li >
< li > spirv/nir: handle location decorations on block interface members< / li >
< li > glsl_types: Rename parameter of glsl_count_attribute_slots< / li >
< li > spirv: Don't use special semantics when counting vertex attribute size< / li >
< / ul >
< p > Nicholas Kazlauskas (5):< / p >
< ul >
< li > util: Get program name based on path when possible< / li >
< li > util: Add adaptive_sync driconf option< / li >
< li > drirc: Initial blacklist for adaptive sync< / li >
< li > loader/dri3: Enable adaptive_sync via _VARIABLE_REFRESH property< / li >
< li > radeonsi: Enable adaptive_sync by default for radeon< / li >
< / ul >
< p > Nicolai Hähnle (37):< / p >
< ul >
< li > radv: include LLVM IR in the VK_AMD_shader_info "disassembly"< / li >
< li > radeonsi: fix an out-of-bounds read reported by ASAN< / li >
< li > winsys/amdgpu: add amdgpu_winsys_bo::lock< / li >
< li > winsys/amdgpu: explicitly declare whether buffer_map is permanent or not< / li >
< li > egl/wayland: rather obvious build fix< / li >
< li > radv: remove dependency on addrlib gfx9_enum.h< / li >
< li > ac/surface/gfx9: let addrlib choose the preferred swizzle kind< / li >
< li > amd/addrlib: update Mesa's copy of addrlib< / li >
< li > meson: link LLVM 'native' component when LLVM is available< / li >
< li > ddebug: simplify watchdog loop and fix crash in the no-timeout case< / li >
< li > ddebug: always flush when requested, even when hang detection is disabled< / li >
< li > r600: remove redundant semicolon< / li >
< li > amd/sid_tables: add additional python3 compatibility imports< / li >
< li > amd/common: whitespace fixes< / li >
< li > amd/common: add ac_build_ifcc< / li >
< li > amd/common: scan/reduce across waves of a workgroup< / li >
< li > amd/common: add i1 special case to ac_build_{inclusive,exclusive}_scan< / li >
< li > ac/surface: 3D and cube surfaces are never displayable< / li >
< li > radeonsi: move SI_FORCE_FAMILY functionality to winsys< / li >
< li > radeonsi: extract declare_vs_blit_inputs< / li >
< li > radeonsi: add si_init_draw_functions and make some functions static< / li >
< li > radeonsi/gfx9: use SET_UCONFIG_REG_INDEX packets when available< / li >
< li > radeonsi: don't set RAW_WAIT for CP DMA clears< / li >
< li > radeonsi: rename SI_RESOURCE_FLAG_FORCE_TILING to clarify its purpose< / li >
< li > radeonsi: const-ify si_set_tesseval_regs< / li >
< li > radeonsi: show the fixed function TCS in debug dumps< / li >
< li > radeonsi: avoid using hard-coded SI_NUM_RW_BUFFERS< / li >
< li > radeonsi: add an si_set_rw_shader_buffer convenience function< / li >
< li > radeonsi: use si_set_rw_shader_buffer for setting streamout buffers< / li >
< li > radeonsi: track constant buffer bind history in si_pipe_set_constant_buffer< / li >
< li > radeonsi: move remaining perfcounter code into si_perfcounter.c< / li >
< li > radeonsi: move query suspend logic into the top-level si_query struct< / li >
< li > radeonsi: factor si_query_buffer logic out of si_query_hw< / li >
< li > radeonsi: split perfcounter queries from si_query_hw< / li >
< li > radeonsi: const-ify the si_query_ops< / li >
< li > amd/common: use llvm.amdgcn.s.buffer.load for LLVM 8.0< / li >
< li > amd/common/vi+: enable SMEM loads with GLC=1< / li >
< / ul >
< p > Niklas Haas (3):< / p >
< ul >
< li > glsl: fix block member alignment validation for vec3< / li >
< li > radv: correctly use vulkan 1.0 by default< / li >
< li > radv: add device-> instance extension dependencies< / li >
< / ul >
< p > Olivier Fourdan (1):< / p >
< ul >
< li > wayland/egl: Resize EGL surface on update buffer for swrast< / li >
< / ul >
< p > Oscar Blumberg (1):< / p >
< ul >
< li > radeonsi: Fix guardband computation for large render targets< / li >
< / ul >
< p > Pierre Moreau (2):< / p >
< ul >
< li > clover/meson: Ignore 'svn' suffix when computing CLANG_RESOURCE_DIR< / li >
< li > meson: Fix with_gallium_icd to with_opencl_icd< / li >
< / ul >
< p > Plamena Manolova (1):< / p >
< ul >
< li > nir: Don't lower the local work group size if it's variable.< / li >
< / ul >
< p > Rafael Antognolli (24):< / p >
< ul >
< li > intel/genxml: Add register for object preemption.< / li >
< li > i965/gen10+: Enable object level preemption.< / li >
< li > i965/gen9: Add workarounds for object preemption.< / li >
< li > anv/tests: Fix block_pool_no_free test.< / li >
< li > anv/allocator: Add anv_state_table.< / li >
< li > anv/allocator: Add getter for anv_block_pool.< / li >
< li > anv/allocator: Add helper to push states back to the state table.< / li >
< li > anv/allocator: Use anv_state_table on anv_state_pool_alloc.< / li >
< li > anv/allocator: Use anv_state_table on back_alloc too.< / li >
< li > anv/allocator: Remove anv_free_list.< / li >
< li > anv/allocator: Rename anv_free_list2 to anv_free_list.< / li >
< li > anv/allocator: Remove pool-> map.< / li >
< li > anv: Update usage of block_pool-> bo.< / li >
< li > anv/allocator: Add support for a list of BOs in block pool.< / li >
< li > anv: Split code to add BO dependencies to execbuf.< / li >
< li > anv: Validate the list of BOs from the block pool.< / li >
< li > anv: Remove some asserts.< / li >
< li > anv/allocator: Rework chunk return to the state pool.< / li >
< li > anv/allocator: Add padding information.< / li >
< li > anv/allocator: Enable snooping on block pool and anv_bo_pool BOs.< / li >
< li > anv: Remove state flush.< / li >
< li > anv/allocator: Add support for non-userptr.< / li >
< li > anv/tests: Adding test for the state_pool padding.< / li >
< li > anv/allocator: Avoid race condition in anv_block_pool_map.< / li >
< / ul >
< p > Ray Zhang (1):< / p >
< ul >
< li > glx: fix shared memory leak in X11< / li >
< / ul >
< p > Rhys Kidd (5):< / p >
< ul >
< li > travis: radeonsi and radv require LLVM 7.0< / li >
< li > meson: libfreedreno depends upon libdrm (for fence support)< / li >
< li > v3d: Wire up core pipe_debug_callback< / li >
< li > vc4: Wire up core pipe_debug_callback< / li >
< li > nv50,nvc0: add missing CAPs for unsupported features< / li >
< / ul >
< p > Rhys Perry (14):< / p >
< ul >
< li > nir: fix constness in nir_intrinsic_align()< / li >
< li > ac: refactor visit_load_buffer< / li >
< li > ac: split 16-bit ssbo loads that may not be dword aligned< / li >
< li > radv: don't set surf_index for stencil-only images< / li >
< li > radv: switch from nir_bcsel to nir_b32csel< / li >
< li > ac/nir,radv,radeonsi/nir: use correct indices for interpolation intrinsics< / li >
< li > nir: fix copy-paste error in nir_lower_constant_initializers< / li >
< li > radv: use dithered alpha-to-coverage< / li >
< li > radv: pass radv_draw_info to radv_emit_draw_registers()< / li >
< li > radv: add missed situations for scissor bug workaround< / li >
< li > radv: avoid context rolls when binding graphics pipelines< / li >
< li > radv: prevent dirtying of dynamic state when it does not change< / li >
< li > radv: bitcast 16-bit outputs to integers< / li >
< li > radv: ensure export arguments are always float< / li >
< / ul >
< p > Rob Clark (79):< / p >
< ul >
< li > freedreno: update generated headers< / li >
< li > freedreno/a6xx: fix VSC bug with larger # of tiles< / li >
< li > freedreno/drm: fix unused 'entry' warnings< / li >
< li > freedreno/drm: remove dependency on gallium driver< / li >
< li > freedreno: move drm to common location< / li >
< li > freedreno/ir3: standalone compiler updates< / li >
< li > freedreno: shader_t -> gl_shader_stage< / li >
< li > freedreno: remove shader_stage_name()< / li >
< li > freedreno: FD_SHADER_DEBUG -> IR3_SHADER_DEBUG< / li >
< li > freedreno/ir3: move disasm and optmsgs debug flags< / li >
< li > util: env_var_as_unsigned() helper< / li >
< li > freedreno/ir3: use env_var_as_unsigned()< / li >
< li > freedreno/ir3: some header file cleanup< / li >
< li > freedreno/ir3: remove pipe_stream_output_info dependency< / li >
< li > freedreno/ir3: split up ir3_shader< / li >
< li > freedreno/ir3: remove u_inlines usage< / li >
< li > freedreno: move ir3 to common location< / li >
< li > mesa/st: swap order of clear() and clear_with_quad()< / li >
< li > mesa/st: better colormask check for clear fallback< / li >
< li > freedreno/a6xx: disable LRZ for z32< / li >
< li > freedreno/a6xx: set guardband clip< / li >
< li > freedreno: update generated headers< / li >
< li > freedreno/a3xx: also set FSSUPERTHREADENABLE< / li >
< li > freedreno/a6xx: MSAA< / li >
< li > freedreno: remove unused fd_surface fields< / li >
< li > gallium: fix typo< / li >
< li > freedreno/a5xx+a6xx: remove unused fs/vs pvt mem< / li >
< li > freedreno/drm: fix relocs in nested stateobjs< / li >
< li > freedreno: update generated headers< / li >
< li > freedreno/a6xx: blitter fixes< / li >
< li > freedreno/ir3: don't fetch unused tex components< / li >
< li > freedreno/ir3: sync instr/disasm< / li >
< li > freedreno/ir3: code-motion< / li >
< li > freedreno/ir3: track max flow control depth for a5xx/a6xx< / li >
< li > freedreno/drm: fix memory leak< / li >
< li > nir: fix spelling typo< / li >
< li > mesa/st/nir: fix missing nir_compact_varyings< / li >
< li > freedreno/drm: sync uapi and enable softpin< / li >
< li > freedreno: debug GEM obj names< / li >
< li > freedreno: also set DUMP flag on shaders< / li >
< li > freedreno/ir3: fix crash< / li >
< li > freedreno/ir3: don't remove unused input components< / li >
< li > freedreno/a6xx: fix blitter crash< / li >
< li > gallium/aux: add is_unorm() helper< / li >
< li > freedreno: update generated headers< / li >
< li > freedreno/a6xx: more blitter fixes< / li >
< li > freedreno: move fd_resource_copy_region()< / li >
< li > freedreno/a6xx: fix resource_copy_region()< / li >
< li > freedreno/a6xx: fix corrupted uniforms< / li >
< li > freedreno/ir3: fix fallout of extra assert< / li >
< li > freedreno/ir3: don't treat all inputs/outputs as vec4< / li >
< li > freedreno: combine fd_resource_layer_offset()/fd_resource_offset()< / li >
< li > freedreno/a6xx: simplify special case for 3d layout< / li >
< li > freedreno/a6xx: improve setup_slices() debug msgs< / li >
< li > freedreno: update generated headers< / li >
< li > freedreno/a6xx: fix 3d texture layout< / li >
< li > freedreno: skip depth resolve if not written< / li >
< li > freedreno: rework blit API< / li >
< li > freedreno: try blitter for fd_resource_copy_region()< / li >
< li > freedreno/a6xx: rework blitter API< / li >
< li > freedreno: remove blit_via_copy_region()< / li >
< li > freedreno: fix staging resource size for arrays< / li >
< li > freedreno: make cmdstream bo's read-only to GPU< / li >
< li > freedreno/a6xx: separate stencil restore/resolve fixes< / li >
< li > freedreno/a6xx: move tile_mode to sampler-view CSO< / li >
< li > freedreno/a6xx: fix 3d+tiled layout< / li >
< li > nir/vtn: add caps for some cl related capabilities< / li >
< li > loader: fix the no-modifiers case< / li >
< li > freedreno: core buffer modifier support< / li >
< li > freedreno: set modifier when exporting buffer< / li >
< li > freedreno: limit tiling to PIPE_BIND_SAMPLER_VIEW< / li >
< li > freedreno/a2xx: fix unused variable warning< / li >
< li > freedreno/a5xx: fix blitter nr_samples check< / li >
< li > freedreno/a6xx: fix blitter nr_samples check< / li >
< li > freedreno: stop frob'ing pipe_resource::nr_samples< / li >
< li > freedreno: minor cleanups< / li >
< li > mesa: wire up InvalidateFramebuffer< / li >
< li > freedreno: fix release tarball< / li >
< li > freedreno: more fixing release tarball< / li >
< / ul >
< p > Rob Herring (3):< / p >
< ul >
< li > pipe-loader: Fallback to kmsro driver when no matching driver name found< / li >
< li > kmsro: Add etnaviv renderonly support< / li >
< li > Switch imx to kmsro and remove the imx winsys< / li >
< / ul >
< p > Robert Foss (3):< / p >
< ul >
< li > virgl: native fence fd support< / li >
< li > virgl: Clean up fences commit< / li >
< li > virgl: add assert and missing function parameter< / li >
< / ul >
< p > Rodrigo Vivi (1):< / p >
< ul >
< li > intel: Add more PCI Device IDs for Coffee Lake and Ice Lake.< / li >
< / ul >
< p > Roland Scheidegger (7):< / p >
< ul >
< li > gallivm: fix improper clamping of vertex index when fetching gs inputs< / li >
< li > draw: fix infinite loop in line stippling< / li >
< li > gallivm: remove unused float coord wrapping for aos sampling< / li >
< li > gallivm: use llvm jit code for decoding s3tc< / li >
< li > gallivm: don't use pavg.b intrinsic on llvm > = 6.0< / li >
< li > gallivm: abort when trying to use non-existing intrinsic< / li >
< li > Revert "llvmpipe: Always return some fence in flush (v2)"< / li >
< / ul >
< p > Sagar Ghuge (14):< / p >
< ul >
< li > intel/compiler: Disassemble GEN6_SFID_DATAPORT_SAMPLER_CACHE as dp_sampler< / li >
< li > intel/compiler: Set swizzle to BRW_SWIZZLE_XXXX for scalar region< / li >
< li > intel/compiler: Always print flag subregister number< / li >
< li > nir: Add a new lowering option to lower 3D surfaces from txd to txl.< / li >
< li > glsl: Add "built-in" functions to do uint64_to_fp64(uint64_t)< / li >
< li > glsl: Add "built-in" functions to do int64_to_fp64(int64_t)< / li >
< li > glsl: Add "built-in" functions to do uint64_to_fp32(uint64_t)< / li >
< li > glsl: Add "built-in" functions to do int64_to_fp32(int64_t)< / li >
< li > glsl: Add utility function to round and pack uint64_t value< / li >
< li > glsl: Add "built-in" functions to do fp64_to_uint64(fp64)< / li >
< li > glsl: Add utility function to round and pack int64_t value< / li >
< li > glsl: Add "built-in" functions to do fp64_to_int64(fp64)< / li >
< li > glsl: Add "built-in" functions to do fp32_to_uint64(fp32)< / li >
< li > glsl: Add "built-in" functions to do fp32_to_int64(fp32)< / li >
< / ul >
< p > Samuel Pitoiset (103):< / p >
< ul >
< li > radv: remove useless sync after copying query results with compute< / li >
< li > radv: add missing TFB queries support to CmdCopyQueryPoolsResults()< / li >
< li > radv: replace si_emit_wait_fence() with radv_cp_wait_mem()< / li >
< li > radv: more use of radv_cp_wait_mem()< / li >
< li > radv: allocate enough space in CS when copying query results with compute< / li >
< li > radv: disable conditional rendering for vkCmdCopyQueryPoolResults()< / li >
< li > radv: only expose VK_SUBGROUP_FEATURE_ARITHMETIC_BIT for VI+< / li >
< li > radv: use LOAD_CONTEXT_REG when loading fast clear values< / li >
< li > radv: fix GPU hangs when loading depth/stencil clear values on SI/CIK< / li >
< li > radv: cleanup and document a Hawaii bug with offchip buffers< / li >
< li > radv: clean up setting partial_es_wave for distributed tess on VI< / li >
< li > radv: make use of num_good_cu_per_sh in si_emit_graphics() too< / li >
< li > radv: binding streamout buffers doesn't change context regs< / li >
< li > radv: set PA.SC_CONSERVATIVE_RASTERIZATION.NULL_SQUAD_AA_MASK_ENABLE< / li >
< li > radv: set optimal OVERWRITE_COMBINER_WATERMARK on GFX9< / li >
< li > radv: add a debug option for disabling primitive binning< / li >
< li > radv: enable primitive binning by default< / li >
< li > radv: tidy up radv_set_dcc_need_cmask_elim_pred()< / li >
< li > radv: always clear the FCE predicate after DCC/FMASK/CMASK decompressions< / li >
< li > radv/winsys: remove the max IBs per submit limit for the fallback path< / li >
< li > radv/winsys: remove the max IBs per submit limit for the sysmem path< / li >
< li > radv: remove unnecessary goto in the fast clear paths< / li >
< li > radv: add radv_get_htile_fast_clear_value() helper< / li >
< li > radv: add radv_is_fast_clear_{depth,stencil}_allowed() helpers< / li >
< li > radv: check allowed fast HTILE clears a bit earlier< / li >
< li > radv: rewrite the condition that checks allowed depth/stencil values< / li >
< li > radv: implement fast HTILE clears for depth or stencil only on GFX9< / li >
< li > ac/nir: fix intrinsic name string size in visit_image_atomic()< / li >
< li > radv: ignore subpass self-dependencies< / li >
< li > radv: only sync CP DMA for transfer operations or bottom pipe< / li >
< li > radv: remove useless sync after CmdClear{Color,DepthStencil}Image()< / li >
< li > radv: remove useless sync before CmdClear{Color,DepthStencil}Image()< / li >
< li > radv: ignore subpass self-dependencies for CreateRenderPass() too< / li >
< li > radv: remove useless check in emit_fast_color_clear()< / li >
< li > radv: add radv_image_can_fast_clear() helper< / li >
< li > radv: add radv_image_view_can_fast_clear() helper< / li >
< li > radv: add radv_can_fast_clear_{color,depth}() helpers< / li >
< li > radv: simplify a check in emit_fast_color_clear()< / li >
< li > radv: refactor the fast clear path for better re-use< / li >
< li > radv: optimize CmdClear{Color,DepthStencil}Image() for layered textures< / li >
< li > radv: remove unused pending_clears param in the transition path< / li >
< li > radv: drop few useless state changes when doing color/depth decompressions< / li >
< li > radv: rework the TC-compat HTILE hardware bug with COND_EXEC< / li >
< li > radv: reset pending_reset_query when flushing caches< / li >
< li > radv: wait on the high 32 bits of timestamp queries< / li >
< li > spirv: add SpvCapabilityInt64Atomics< / li >
< li > radv: expose VK_EXT_scalar_block_layout< / li >
< li > amd: remove support for LLVM 6.0< / li >
< li > gallium: add missing PIPE_CAP_SURFACE_SAMPLE_COUNT default value< / li >
< li > radv: bump reported version to 1.1.90< / li >
< li > radv: add a predicate for reflecting DCC decompression state< / li >
< li > radv: allow to skip DCC decompressions with the new predicate< / li >
< li > radv: switch on EOP when primitive restart is enabled with triangle strips< / li >
< li > radv: check if addrlib enabled HTILE in radv_image_can_enable_htile()< / li >
< li > radv: don't check if format is depth in radv_image_can_enable_hile()< / li >
< li > radv: report Vulkan version 1.1.90 for real< / li >
< li > ac/nir: remove the bitfield_extract workaround for LLVM 8< / li >
< li > radv: drop the amdgpu-skip-threshold=1 workaround for LLVM 8< / li >
< li > radv: fix subpass image transitions with multiviews< / li >
< li > radv: compute optimal VM alignment for imported buffers< / li >
< li > spirv: add support for SpvCapabilityStorageImageMultisample< / li >
< li > ac/nir: restrict fmask lookup to image load intrinsics< / li >
< li > radv: initialize FMASK for images in fully expanded mode< / li >
< li > radv: add support for FMASK expand< / li >
< li > radv: enable shaderStorageImageMultisample feature on GFX8+< / li >
< li > radv: get rid of bunch of KHR suffixes< / li >
< li > radv: enable variable pointers< / li >
< li > radv: skip draws with instance_count == 0< / li >
< li > ac/nir: add get_cache_policy() helper and use it< / li >
< li > ac/nir: set cache policy when loading/storing buffer images< / li >
< li > ac: add missing 16-bit types to glsl_base_to_llvm_type()< / li >
< li > radv: remove unnecessary returns in GetPhysicalDevice*Properties()< / li >
< li > radv: add two small helpers for getting VRAM and visible VRAM sizes< / li >
< li > radv: add support for VK_EXT_memory_budget< / li >
< li > ac/nir: don't trash L1 caches for store operations with writeonly memory< / li >
< li > radv: drop unused code related to 16 sample locations< / li >
< li > radv: reduce size of the per-queue descriptor BO< / li >
< li > radv: do not write unused descriptors to the per-queue BO< / li >
< li > radv: initialize the per-queue descriptor BO only once< / li >
< li > nir: do not remove varyings used for transform feedback< / li >
< li > nir: fix lowering arrays to elements for XFB outputs< / li >
< li > radv: improve gathering of load_push_constants with dynamic bindings< / li >
< li > radv: remove old_fence parameter from si_cs_emit_write_event_eop()< / li >
< li > radv: only allocate the GFX9 fence and EOP BOs for the gfx queue< / li >
< li > radv: compute the GFX9 fence VA at allocation time< / li >
< li > radv: always pass the GFX9 fence data to si_cs_emit_cache_flush()< / li >
< li > radv: fix computing number of user SGPRs for streamout buffers< / li >
< li > radv: remove radv_userdata_info::indirect field< / li >
< li > radv: simplify allocating user SGPRS for descriptor sets< / li >
< li > radv: set noalias/dereferenceable LLVM attributes based on param types< / li >
< li > radv: re-enable fast depth clears for 16-bit surfaces on VI< / li >
< li > radv/winsys: fix hash when adding internal buffers< / li >
< li > radv: fix compiler issues with GCC 9< / li >
< li > radv: fix using LOAD_CONTEXT_REG with old GFX ME firmwares on GFX8< / li >
< li > radv/winsys: fix BO list creation when RADV_DEBUG=allbos is set< / li >
< li > radv: always export gl_SampleMask when the fragment shader uses it< / li >
< li > radv: write the alpha channel of MRT0 when alpha coverage is enabled< / li >
< li > radv: fix writing the alpha channel of MRT0 when alpha coverage is enabled< / li >
< li > radv: fix out-of-bounds access when copying descriptors BO list< / li >
< li > radv: don't copy buffer descriptors list for samplers< / li >
< li > radv: fix clearing attachments in secondary command buffers< / li >
< li > radv: properly align the fence and EOP bug VA on GFX9< / li >
< li > radv: fix pointSizeRange limits< / li >
< / ul >
< p > Sergii Romantsov (4):< / p >
< ul >
< li > autotools: library-dependency when no sse and 32-bit< / li >
< li > i965/batch/debug: Allow log be dumped before assert< / li >
< li > nir: Length of boolean vtn_value now is 1< / li >
< li > dri: meson: do not prefix user provided dri-drivers-path< / li >
< / ul >
< p > Sonny Jiang (1):< / p >
< ul >
< li > radeonsi: use compute for resource_copy_region when possible< / li >
< / ul >
< p > Tapani Pälli (27):< / p >
< ul >
< li > anv: allow exporting an imported SYNC_FD semaphore type< / li >
< li > anv: add create_flags as part of anv_image< / li >
< li > anv: refactor make_surface to use data from anv_image< / li >
< li > anv: make anv_get_image_format_features public< / li >
< li > anv: add from/to helpers with android and vulkan formats< / li >
< li > anv/android: add GetAndroidHardwareBufferPropertiesANDROID< / li >
< li > anv: add anv_ahw_usage_from_vk_usage helper function< / li >
< li > anv: refactor, remove else block in AllocateMemory< / li >
< li > anv/android: support import/export of AHardwareBuffer objects< / li >
< li > anv/android: add ahardwarebuffer external memory properties< / li >
< li > anv/android: support creating images from external format< / li >
< li > anv: support VkExternalFormatANDROID in vkCreateSamplerYcbcrConversion< / li >
< li > anv: add VkFormat field as part of anv_format< / li >
< li > anv: support VkSamplerYcbcrConversionInfo in vkCreateImageView< / li >
< li > anv: ignore VkSamplerYcbcrConversion on non-yuv formats< / li >
< li > anv/android: turn on VK_ANDROID_external_memory_android_hardware_buffer< / li >
< li > dri3: initialize adaptive_sync as false before configQueryb< / li >
< li > intel/isl: move tiled_memcpy static libs from i965 to isl< / li >
< li > anv: do not advertise AHW support if extension not enabled< / li >
< li > nir: cleanup glsl_get_struct_field_offset, glsl_get_explicit_stride< / li >
< li > android: fix build issues with libmesa_anv_gen* libraries< / li >
< li > mesa: return NULL if we exceed MaxColorAttachments in get_fb_attachment< / li >
< li > nir: initialize value in copy_prop_vars_block< / li >
< li > anv: retain the is_array state in create_plane_tex_instr_implicit< / li >
< li > anv: release memory allocated by glsl types during spirv_to_nir< / li >
< li > anv: revert "anv: release memory allocated by glsl types during spirv_to_nir"< / li >
< li > anv: destroy descriptor sets when pool gets destroyed< / li >
< / ul >
< p > Thomas Hellstrom (9):< / p >
< ul >
< li > st/xa: Render update. Better support for solid pictures< / li >
< li > st/xa: Support higher color precision for solid pictures< / li >
< li > st/xa: Support a couple of new formats< / li >
< li > st/xa: Fix transformations when we have both source and mask samplers< / li >
< li > st/xa: Minor renderer cleanups< / li >
< li > st/xa: Support Component Alpha with trivial blending< / li >
< li > st/xa: Bump minor< / li >
< li > st/xa: Fix a memory leak< / li >
< li > winsys/svga: Fix a memory leak< / li >
< / ul >
< p > Timothy Arceri (56):< / p >
< ul >
< li > nir: allow propagation of if evaluation for bcsel< / li >
< li > nir: fix condition propagation when src has a swizzle< / li >
< li > ac/nir_to_llvm: fix b2f for f64< / li >
< li > nir: add new linking opt nir_link_constant_varyings()< / li >
< li > st/mesa: make use of nir_link_constant_varyings()< / li >
< li > nir: add glsl_type_is_integer() helper< / li >
< li > nir: don't pack varyings ints with floats unless flat< / li >
< li > anv/i965: make use of nir_link_constant_varyings()< / li >
< li > nir: add support for removing redundant stores to copy prop var< / li >
< li > radv: make use of nir_move_out_const_to_consumer()< / li >
< li > nir: small tidy ups for nir_loop_analyze()< / li >
< li > nir: clarify some nit_loop_info member names< / li >
< li > nir: add a new nir_cf_list_clone_and_reinsert() helper< / li >
< li > nir: make use of new nir_cf_list_clone_and_reinsert() helper< / li >
< li > nir: factor out some of the complex loop unroll code to a helper< / li >
< li > nir: rework force_unroll_array_access()< / li >
< li > nir: in loop analysis track actual control flow type< / li >
< li > nir: reword code comment< / li >
< li > nir: detect more induction variables< / li >
< li > nir: fix opt_if_loop_last_continue()< / li >
< li > tgsi/scan: fix loop exit point in tgsi_scan_tess_ctrl()< / li >
< li > tgsi/scan: correctly walk instructions in tgsi_scan_tess_ctrl()< / li >
< li > radeonsi: remove unrequired param in si_nir_scan_tess_ctrl()< / li >
< li > ac/nir_to_llvm: add ac_are_tessfactors_def_in_all_invocs()< / li >
< li > radeonsi: make use of ac_are_tessfactors_def_in_all_invocs()< / li >
< li > st/glsl_to_nir: call nir_lower_load_const_to_scalar() in the st< / li >
< li > nir: rename nir_link_constant_varyings() nir_link_opt_varyings()< / li >
< li > nir: add can_replace_varying() helper< / li >
< li > nir: rework nir_link_opt_varyings()< / li >
< li > nir: link time opt duplicate varyings< / li >
< li > nir: make nir_opt_remove_phis_impl() static< / li >
< li > nir: make use of does_varying_match() helper< / li >
< li > nir: simplify does_varying_match()< / li >
< li > nir: add rewrite_phi_predecessor_blocks() helper< / li >
< li > nir: merge some basic consecutive ifs< / li >
< li > st/glsl: refactor st_link_nir()< / li >
< li > nir: avoid uninitialized variable warning< / li >
< li > glsl: Copy function out to temp if we don't directly ref a variable< / li >
< li > ac/nir_to_llvm: fix type handling in image code< / li >
< li > radeonsi/nir: get correct type for images inside structs< / li >
< li > ac/nir_to_llvm: fix regression in bindless support< / li >
< li > ac/nir_to_llvm: add support for structs to get_sampler_desc()< / li >
< li > glsl: don't skip GLSL IR opts on first-time compiles< / li >
< li > glsl: be much more aggressive when skipping shader compilation< / li >
< li > Revert "glsl: be much more aggressive when skipping shader compilation"< / li >
< li > ac/nir_to_llvm: fix interpolateAt* for arrays< / li >
< li > glsl: be much more aggressive when skipping shader compilation< / li >
< li > radeonsi/nir: add missing piece for bindless image support< / li >
< li > ac/nir_to_llvm: add bindless support for uniform handles< / li >
< li > ac/nir_to_llvm: fix interpolateAt* for structs< / li >
< li > ac/nir_to_llvm: fix clamp shadow reference for more hardware< / li >
< li > tgsi: remove culldist semantic from docs< / li >
< li > radv/ac: fix some fp16 handling< / li >
< li > glsl: use remap location when serialising uniform program resource data< / li >
< li > radeonsi: fix query buffer allocation< / li >
< li > glsl: fix shader cache for packed param list< / li >
< / ul >
< p > Tobias Klausmann (1):< / p >
< ul >
< li > amd/vulkan: meson build - use radv_deps for libvulkan_radeon< / li >
< / ul >
< p > Tomasz Figa (1):< / p >
< ul >
< li > llvmpipe: Always return some fence in flush (v2)< / li >
< / ul >
< p > Tomeu Vizoso (1):< / p >
< ul >
< li > etnaviv: Consolidate buffer references from framebuffers< / li >
< / ul >
< p > Toni Lönnberg (14):< / p >
< ul >
< li > intel/decoder: Engine parameter for instructions< / li >
< li > intel/decoder: tools: gen_engine to drm_i915_gem_engine_class< / li >
< li > intel/decoder: tools: Use engine for decoding batch instructions< / li >
< li > intel/genxml: Add engine definition to render engine instructions (gen4)< / li >
< li > intel/genxml: Add engine definition to render engine instructions (gen45)< / li >
< li > intel/genxml: Add engine definition to render engine instructions (gen5)< / li >
< li > intel/genxml: Add engine definition to render engine instructions (gen6)< / li >
< li > intel/genxml: Add engine definition to render engine instructions (gen7)< / li >
< li > intel/genxml: Add engine definition to render engine instructions (gen75)< / li >
< li > intel/genxml: Add engine definition to render engine instructions (gen8)< / li >
< li > intel/genxml: Add engine definition to render engine instructions (gen9)< / li >
< li > intel/genxml: Add engine definition to render engine instructions (gen10)< / li >
< li > intel/genxml: Add engine definition to render engine instructions (gen11)< / li >
< li > intel/aubinator_error_decode: Get rid of warning for missing switch case< / li >
< / ul >
< p > Topi Pohjolainen (1):< / p >
< ul >
< li > i965/icl: Disable prefetching of sampler state entries< / li >
< / ul >
< p > Veluri Mithun (5):< / p >
< ul >
< li > Add extension doc for MESA_query_driver< / li >
< li > Implement EGL API for MESA_query_driver< / li >
< li > Implementation of egl dri2 drivers for MESA_query_driver< / li >
< li > egl: Implement EGL API for MESA_query_driver< / li >
< li > egl: Implementation of egl dri2 drivers for MESA_query_driver< / li >
< / ul >
< p > Vinson Lee (7):< / p >
< ul >
< li > r600/sb: Fix constant logical operand in assert.< / li >
< li > freedreno: Fix autotools build.< / li >
< li > st/xvmc: Add X11 include path.< / li >
< li > nir/algebraic: Make algebraic_parser_test.sh executable.< / li >
< li > meson: Fix typo.< / li >
< li > meson: Fix libsensors detection.< / li >
< li > meson: Fix typo.< / li >
< / ul >
< p > Yevhenii Kolesnikov (1):< / p >
< ul >
< li > i965: Fix allow_higher_compat_version workaround limited by OpenGL 3.0< / li >
< / ul >
< p > pal1000 (1):< / p >
< ul >
< li > scons: Compatibility with Scons development version string< / li >
< / ul >
2018-11-01 18:56:54 +00:00
< / div >
< / body >
< / html >