1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd"> 2<html lang="en"> 3<head> 4 <meta http-equiv="content-type" content="text/html; charset=utf-8"> 5 <title>Mesa Release Notes</title> 6 <link rel="stylesheet" type="text/css" href="../mesa.css"> 7</head> 8<body> 9 10<div class="header"> 11 <h1>The Mesa 3D Graphics Library</h1> 12</div> 13 14<iframe src="../contents.html"></iframe> 15<div class="content"> 16 17<h1>Mesa 19.0.0 Release Notes / TBD</h1> 18 19<p> 20Mesa 19.0.0 is a new development release. People who are concerned 21with stability and reliability should stick with a previous release or 22wait for Mesa 19.0.1. 23</p> 24<p> 25Mesa 19.0.0 implements the OpenGL 4.5 API, but the version reported by 26glGetString(GL_VERSION) or glGetIntegerv(GL_MAJOR_VERSION) / 27glGetIntegerv(GL_MINOR_VERSION) depends on the particular driver being used. 28Some drivers don't support all the features required in OpenGL 4.5. OpenGL 294.5 is <strong>only</strong> available if requested at context creation. 30Compatibility contexts may report a lower version depending on each driver. 31</p> 32 33<h2>SHA256 checksums</h2> 34<pre> 35 4c5b9c5227d37c1f6bdc786a6fa7ee7fbce40b2e8a87340c7d3234534ece3304 mesa-19.0.0.tar.gz 36 5a549dfb40ec31e5c36c47aadac04554cb2e2a8d144a046a378fc16da57e38f8 mesa-19.0.0.tar.xz 37</pre> 38 39 40<h2>New features</h2> 41 42<ul> 43<li>GL_AMD_texture_texture4 on all GL 4.0 drivers.</li> 44<li>GL_EXT_shader_implicit_conversions on all drivers (ES extension).</li> 45<li>GL_EXT_texture_compression_bptc on all GL 4.0 drivers (ES extension).</li> 46<li>GL_EXT_texture_compression_rgtc on all GL 3.0 drivers (ES extension).</li> 47<li>GL_EXT_render_snorm on gallium drivers (ES extension).</li> 48<li>GL_EXT_texture_view on drivers supporting texture views (ES extension).</li> 49<li>GL_OES_texture_view on drivers supporting texture views (ES extension).</li> 50<li>GL_NV_shader_atomic_float on nvc0 (Fermi/Kepler only).</li> 51<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> 52<li>VK_ANDROID_external_memory_android_hardware_buffer on Intel</li> 53<li>Fixed and re-exposed VK_EXT_pci_bus_info on Intel and RADV</li> 54<li>VK_EXT_scalar_block_layout on Intel and RADV</li> 55<li>VK_KHR_depth_stencil_resolve on Intel</li> 56<li>VK_KHR_draw_indirect_count on Intel</li> 57<li>VK_EXT_conditional_rendering on Intel</li> 58<li>VK_EXT_memory_budget on RADV</li> 59</ul> 60 61<h2>Bug fixes</h2> 62 63<ul> 64 65<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> 66 67<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=102349">Bug 102349</a> - nv4x crashing with plasmashell - gdb log included</li> 68 69<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> 70 71<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> 72 73<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> 74 75<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=105301">Bug 105301</a> - The big SKQP bug</li> 76 77<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106577">Bug 106577</a> - broken rendering with nine and nouveau (GM107)</li> 78 79<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106595">Bug 106595</a> - [RADV] Rendering distortions only when MSAA is enabled</li> 80 81<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> 82 83<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> 84 85<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> 86 87<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=107728">Bug 107728</a> - Wrong background in Sascha Willem's Multisampling Demo</li> 88 89<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> 90 91<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> 92 93<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=108114">Bug 108114</a> - [vulkancts] new VK_KHR_16bit_storage tests fail.</li> 94 95<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=108116">Bug 108116</a> - [vulkancts] stencil partial clear tests fail.</li> 96 97<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> 98 99<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=108311">Bug 108311</a> - Query buffer object support is broken on r600.</li> 100 101<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> 102 103<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=108560">Bug 108560</a> - Mesa 32 is built without sse</li> 104 105<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=108624">Bug 108624</a> - [regression][bisected] "nir: Copy propagation between blocks" regression</li> 106 107<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> 108 109<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=108635">Bug 108635</a> - Mesa master commit 68dc591af16ebb36814e4c187e4998948103c99c causes XWayland to segfault</li> 110 111<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> 112 113<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=108713">Bug 108713</a> - Gallium: use after free with transform feedback</li> 114 115<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> 116 117<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=108805">Bug 108805</a> - i965 regressions from EXT_texture_sRGB_R8</li> 118 119<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=108829">Bug 108829</a> - [meson] libglapi exports internal API</li> 120 121<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=108868">Bug 108868</a> - [BYT IVB] Tesselation test regressions</li> 122 123<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> 124 125<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=108894">Bug 108894</a> - [anv] vkCmdCopyBuffer() and vkCmdCopyQueryPoolResults() write-after-write hazard</li> 126 127<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> 128 129<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=108910">Bug 108910</a> - Vkd3d test failure test_multisample_array_texture()</li> 130 131<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=108911">Bug 108911</a> - Vkd3d test failure test_clear_render_target_view()</li> 132 133<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> 134 135<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> 136 137<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> 138 139<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=108943">Bug 108943</a> - Build fails on ppc64le with meson</li> 140 141<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=108961">Bug 108961</a> - make check test_replace_src_bitsize failure</li> 142 143<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=108974">Bug 108974</a> - make check DispatchSanity_test regression</li> 144 145<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> 146 147<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> 148 149<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=109072">Bug 109072</a> - GPU hang in blender 2.80</li> 150 151<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> 152 153<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> 154 155<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=109086">Bug 109086</a> - Crash software mesa with gl_select render mode</li> 156 157<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> 158 159<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> 160 161<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> 162 163<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> 164 165<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> 166 167<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> 168 169<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=109229">Bug 109229</a> - glLinkProgram locks up for ~30 seconds</li> 170 171<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> 172 173<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=109242">Bug 109242</a> - [RADV] The Witcher 3 system freeze</li> 174 175<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=109304">Bug 109304</a> - GfxBench AztecRuins Vulkan version Segfault</li> 176 177<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> 178 179<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> 180 181<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> 182 183<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=109401">Bug 109401</a> - [DXVK] Project Cars rendering problems</li> 184 185<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=109404">Bug 109404</a> - [ANV] The Witcher 3 shadows flickering</li> 186 187<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> 188 189<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> 190 191<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=109449">Bug 109449</a> - [snb] quakespasm triggers a segmentation fault.</li> 192 193<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> 194 195<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> 196 197<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> 198 199<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> 200 201<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> 202 203<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> 204 205<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> 206 207<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=109597">Bug 109597</a> - wreckfest issues with transparent objects & skybox</li> 208 209<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> 210 211<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> 212 213<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=109698">Bug 109698</a> - dri.pc contents invalid when built with meson</li> 214 215<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=109717">Bug 109717</a> - [regression] Cull distance tests asserting</li> 216 217<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=109735">Bug 109735</a> - [Regression] broken font with mesa_vulkan_overlay</li> 218 219<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> 220 221</ul> 222 223<h2>Changes</h2> 224 225<p>Adam Jackson (4):</p> 226<ul> 227 <li>glx: Demand success from CreateContext requests (v2)</li> 228 <li>specs: Remove GLES profile interaction text from GLX_MESA_query_renderer</li> 229 <li>specs: Remove GLX_RENDERER_ID_MESA from GLX_MESA_query_renderer</li> 230 <li>specs: Bump GLX_MESA_query_renderer to version 9</li> 231</ul> 232 233<p>Aditya Swarup (1):</p> 234<ul> 235 <li>i965: Lift restriction in external textures for EGLImage support</li> 236</ul> 237 238<p>Alejandro Piñeiro (3):</p> 239<ul> 240 <li>nir: remove unused variable</li> 241 <li>nir/xfb: don't assert when xfb_buffer/stride is present but not xfb_offset</li> 242 <li>nir/xfb: distinguish array of structs vs array of blocks</li> 243</ul> 244 245<p>Alex Deucher (3):</p> 246<ul> 247 <li>pci_ids: add new vega10 pci ids</li> 248 <li>pci_ids: add new vega20 pci id</li> 249 <li>pci_ids: add new VegaM pci id</li> 250</ul> 251 252<p>Alex Smith (1):</p> 253<ul> 254 <li>radv: Flush before vkCmdWriteTimestamp() if needed</li> 255</ul> 256 257<p>Alexander von Gluck IV (1):</p> 258<ul> 259 <li>egl/haiku: Fix reference to disp vs dpy</li> 260</ul> 261 262<p>Alok Hota (8):</p> 263<ul> 264 <li>swr/rast: Use gfxptr_t value in JitGatherVertices</li> 265 <li>swr/rast: Add annotator to interleave isa text</li> 266 <li>swr/rast: partial support for Tiled Resources</li> 267 <li>swr/rast: Unaligned and translations in gathers</li> 268 <li>swr/rast: Scope MEM_CLIENT enum for mem usages</li> 269 <li>swr/rast: New execution engine per JIT</li> 270 <li>swr/rast: Store cached files in multiple subdirs</li> 271 <li>swr/rast: bypass size limit for non-sampled textures</li> 272</ul> 273 274<p>Alyssa Rosenzweig (1):</p> 275<ul> 276 <li>util: Fix warning in u_cpu_detect on non-x86</li> 277</ul> 278 279<p>Andre Heider (4):</p> 280<ul> 281 <li>st/nine: fix stack corruption due to ABI mismatch</li> 282 <li>st/nine: plug thread related leaks</li> 283 <li>st/nine: clean up thead shutdown sequence a bit</li> 284 <li>d3dadapter9: use snprintf(..., "%s", ...) instead of strncpy</li> 285</ul> 286 287<p>Andres Gomez (8):</p> 288<ul> 289 <li>glsl/linker: complete documentation for assign_attribute_or_color_locations</li> 290 <li>docs: update 18.3 and add 19.x cycles for the release calendar</li> 291 <li>glsl: correct typo in GLSL compilation error message</li> 292 <li>editorconfig: Add max_line_length property</li> 293 <li>glsl/linker: specify proper direction in location aliasing error</li> 294 <li>docs: complete the calendar and release schedule documentation</li> 295 <li>bin/get-pick-list.sh: fix the oneline printing</li> 296 <li>bin/get-pick-list.sh: fix redirection in sh</li> 297</ul> 298 299<p>Andrii Simiklit (9):</p> 300<ul> 301 <li>intel/tools: avoid 'unused variable' warnings</li> 302 <li>compiler: avoid 'unused variable' warnings</li> 303 <li>i965: avoid 'unused variable' warnings</li> 304 <li>i965/batch: avoid reverting batch buffer if saved state is an empty</li> 305 <li>intel/tools: make sure the binary file is properly read</li> 306 <li>anv/pipeline: remove unnecessary null-pointer check</li> 307 <li>intel/batch-decoder: fix vertex buffer size calculation for gen<8</li> 308 <li>intel/batch-decoder: fix a vb end address calculation</li> 309 <li>i965: re-emit index buffer state on a reset option change.</li> 310</ul> 311 312<p>Anuj Phogat (7):</p> 313<ul> 314 <li>i965/icl: Set Error Detection Behavior Control Bit in L3CNTLREG</li> 315 <li>anv/icl: Set Error Detection Behavior Control Bit in L3CNTLREG</li> 316 <li>anv/icl: Disable prefetching of sampler state entries</li> 317 <li>i965/icl: Fix L3 configurations</li> 318 <li>i965/icl: Set use full ways in L3CNTLREG</li> 319 <li>intel/icl: Set way_size_per_bank to 4</li> 320 <li>anv/icl: Set use full ways in L3CNTLREG</li> 321</ul> 322 323<p>Axel Davy (12):</p> 324<ul> 325 <li>st/nine: Allow 'triple buffering' with thread_submit</li> 326 <li>st/nine: Remove thread_submit warning</li> 327 <li>st/nine: Use helper to release swapchain buffers later</li> 328 <li>st/nine: Switch to presentation buffer if resize is detected</li> 329 <li>st/nine: Fix volumetexture dtor on ctor failure</li> 330 <li>st/nine: Bind src not dst in nine_context_box_upload</li> 331 <li>st/nine: Add src reference to nine_context_range_upload</li> 332 <li>st/nine: Increase the limit of cached ff shaders</li> 333 <li>st/nine: Immediately upload user provided textures</li> 334 <li>st/nine: Enable debug info if NDEBUG is not set</li> 335 <li>st/nine: Ignore window size if error</li> 336 <li>st/nine: Ignore multisample quality level if no ms</li> 337</ul> 338 339<p>Bart Oldeman (1):</p> 340<ul> 341 <li>gallium-xlib: query MIT-SHM before using it.</li> 342</ul> 343 344<p>Bas Nieuwenhuizen (41):</p> 345<ul> 346 <li>radv: Use structured intrinsics instead of indexing workaround for GFX9.</li> 347 <li>vulkan: Allow storage images in the WSI.</li> 348 <li>radv: Fix opaque metadata descriptor last layer.</li> 349 <li>radv: Clamp gfx9 image view extents to the allocated image extents.</li> 350 <li>radv: Align large buffers to the fragment size.</li> 351 <li>radv/android: Mark android WSI image as shareable.</li> 352 <li>radv/android: Use buffer metadata to determine scanout compat.</li> 353 <li>radv: Check for shareable images in central place.</li> 354 <li>radv: Remove redundant format check.</li> 355 <li>radv: Fix multiview depth clears</li> 356 <li>radv: Work around non-renderable 128bpp compressed 3d textures on GFX9.</li> 357 <li>radv: Fix wrongly positioned paren.</li> 358 <li>radv: Do a cache flush if needed before reading predicates.</li> 359 <li>radv: Implement buffer stores with less than 4 components.</li> 360 <li>anv/android: Do not reject storage images.</li> 361 <li>radv: Remove device path.</li> 362 <li>radv: Remove unused variable.</li> 363 <li>amd/common: Add some parentheses to silence warning.</li> 364 <li>radv: Fix rasterization precision bits.</li> 365 <li>spirv: Fix matrix parameters in function calls.</li> 366 <li>freedreno: Move register constant files to src/freedreno.</li> 367 <li>radv: Only use 32 KiB per threadgroup on Stoney.</li> 368 <li>radv: Set partial_vs_wave for pipelines with just GS, not tess.</li> 369 <li>nir: Account for atomics in copy propagation.</li> 370 <li>radv: Remove unused variable.</li> 371 <li>radv/winsys: Set winsys bo priority on creation.</li> 372 <li>radv/winsys: Add priority handling during submit.</li> 373 <li>radv: Enable VK_EXT_memory_priority.</li> 374 <li>radv: Fix the shader info pass for not having the variable.</li> 375 <li>amd/common: Fix stores to derefs with unknown variable.</li> 376 <li>amd/common: Add gep helper for pointer increment.</li> 377 <li>amd/common: Handle nir_deref_type_ptr_as_array for shared memory.</li> 378 <li>amd/common: handle nir_deref_cast for shared memory from integers.</li> 379 <li>radv: Only look at pImmutableSamples if the descriptor has a sampler.</li> 380 <li>amd/common: Use correct writemask for shared memory stores.</li> 381 <li>radv: Sync ETC2 whitelisted devices.</li> 382 <li>radv: Fix float16 interpolation set up.</li> 383 <li>radv: Allow interpolation on non-float types.</li> 384 <li>radv: Handle clip+cull distances more generally as compact arrays.</li> 385 <li>radv: Fix rebase issue in 19.0 for float16 fix.</li> 386 <li>radv: Interpolate less aggressively.</li> 387</ul> 388 389<p>Boyan Ding (3):</p> 390<ul> 391 <li>gk110/ir: Add rcp f64 implementation</li> 392 <li>gk110/ir: Add rsq f64 implementation</li> 393 <li>gk110/ir: Use the new rcp/rsq in library</li> 394</ul> 395 396<p>Brian Paul (3):</p> 397<ul> 398 <li>svga: add new gallium formats to the format conversion table</li> 399 <li>mesa: fix display list corner case assertion</li> 400 <li>svga: remove SVGA_RELOC_READ flag in SVGA3D_BindGBSurface()</li> 401</ul> 402 403<p>Bruce Cherniak (1):</p> 404<ul> 405 <li>gallium/swr: Fix multi-context sync fence deadlock.</li> 406</ul> 407 408<p>Caio Marcelo de Oliveira Filho (10):</p> 409<ul> 410 <li>nir: properly clear the entry sources in copy_prop_vars</li> 411 <li>nir: properly find the entry to keep in copy_prop_vars</li> 412 <li>nir: add a way to print the deref chain</li> 413 <li>nir: remove dead code from copy_prop_vars</li> 414 <li>nir: fix warning in nir_lower_io.c</li> 415 <li>util: Helper to create sets and hashes with pointer keys</li> 416 <li>src/compiler: use new hash table and set creation helpers</li> 417 <li>src/intel: use new hash table and set creation helpers</li> 418 <li>nir: check NIR_SKIP to skip passes by name</li> 419 <li>gallium: Add PIPE_CAP_GLSL_TESS_LEVELS_AS_INPUTS</li> 420</ul> 421 422<p>Carlos Garnacho (1):</p> 423<ul> 424 <li>wayland/egl: Ensure EGL surface is resized on DRI update_buffers()</li> 425</ul> 426 427<p>Carsten Haitzler (Rasterman) (2):</p> 428<ul> 429 <li>vc4: Use named parameters for the NEON inline asm.</li> 430 <li>vc4: Declare the cpu pointers as being modified in NEON asm.</li> 431</ul> 432 433<p>Chad Versace (1):</p> 434<ul> 435 <li>i965: Fix -Wswitch on INTEL_COPY_STREAMING_LOAD</li> 436</ul> 437 438<p>Chia-I Wu (2):</p> 439<ul> 440 <li>meson: fix EGL/X11 build without GLX</li> 441 <li>freedreno/drm: sync uapi again</li> 442</ul> 443 444<p>Christian Gmeiner (6):</p> 445<ul> 446 <li>nir: add lowering for ffloor</li> 447 <li>etnaviv: drop redundant ctx function parameter</li> 448 <li>meson: add etnaviv to the tools option</li> 449 <li>etnaviv: extend etna_resource with an addressing mode</li> 450 <li>etnaviv: update headers from rnndb</li> 451 <li>etnaviv: add linear sampling support</li> 452</ul> 453 454<p>Connor Abbott (4):</p> 455<ul> 456 <li>Revert "radv: disable VK_SUBGROUP_FEATURE_VOTE_BIT"</li> 457 <li>nir/algebraic: Rewrite bit-size inference</li> 458 <li>nir/algebraic: Add unit tests for bitsize validation</li> 459 <li>nir: Fixup algebraic test for variable-sized conversions</li> 460</ul> 461 462<p>Daniel Stone (1):</p> 463<ul> 464 <li>gbm: Clarify acceptable formats for gbm_bo</li> 465</ul> 466 467<p>Danylo Piliaiev (9):</p> 468<ul> 469 <li>i965: Fix calculation of layers array length for isl_view</li> 470 <li>nir: add if opt opt_if_loop_last_continue()</li> 471 <li>glsl/linker: Fix unmatched TCS outputs being reduced to local variable</li> 472 <li>glsl: Make invariant outputs in ES fragment shader not to cause error</li> 473 <li>glsl: Fix copying function's out to temp if dereferenced by array</li> 474 <li>anv: Implement VK_KHR_draw_indirect_count for gen 7+</li> 475 <li>anv: Implement VK_EXT_conditional_rendering for gen 7.5+</li> 476 <li>anv: Fix VK_EXT_transform_feedback working with varyings packed in PSIZ</li> 477 <li>anv: Fix destroying descriptor sets when pool gets reset</li> 478</ul> 479 480<p>Dave Airlie (19):</p> 481<ul> 482 <li>radv: apply xfb buffer offset at buffer binding time not later. (v2)</li> 483 <li>radv: fix begin/end transform feedback with 0 counter buffers.</li> 484 <li>virgl: fix vtest regression since fencing changes.</li> 485 <li>spirv/vtn: handle variable pointers without offset lowering</li> 486 <li>nir: move getting deref from var after we check deref type.</li> 487 <li>nir: handle shared pointers in lowering indirect derefs.</li> 488 <li>ac: avoid casting pointers on bcsel and stores</li> 489 <li>radv: handle loading from shared pointers</li> 490 <li>ac: handle cast derefs</li> 491 <li>r600: make suballocator 256-bytes align</li> 492 <li>virgl: fix undefined shift to use unsigned.</li> 493 <li>virgl: fix const warning on debug flags.</li> 494 <li>radv: use 3d shader for gfx9 copies if dst is 3d</li> 495 <li>radv/xfb: fix counter buffer bounds checks.</li> 496 <li>virgl/vtest: fix front buffer flush with protocol version 0.</li> 497 <li>virgl: use primconvert provoking vertex properly</li> 498 <li>dri_interface: add put shm image2 (v2)</li> 499 <li>glx: add support for putimageshm2 path (v2)</li> 500 <li>gallium: use put image shm2 path (v2)</li> 501</ul> 502 503<p>David Shao (1):</p> 504<ul> 505 <li>meson: ensure that xmlpool_options.h is generated for gallium targets that need it</li> 506</ul> 507 508<p>Dieter Nützel (1):</p> 509<ul> 510 <li>docs/features: Delete double nv50 entry and wrong enumeration</li> 511</ul> 512 513<p>Dylan Baker (48):</p> 514<ul> 515 <li>meson: link gallium nine with pthreads</li> 516 <li>meson: Don't set -Wall</li> 517 <li>meson: fix libatomic tests</li> 518 <li>meson: Add tests to suites</li> 519 <li>util: promote u_memory to src/util</li> 520 <li>meson: Add nir_algebraic_parser_test to suites</li> 521 <li>meson: Fix ppc64 little endian detection</li> 522 <li>meson: remove duplicate definition</li> 523 <li>meson: Add support for gnu hurd</li> 524 <li>meson: Add toggle for glx-direct</li> 525 <li>docs/meson: Recommend not using CFLAGS and friends</li> 526 <li>travis: meson: use native files to override llvm-config</li> 527 <li>travis: Don't try to read libdrm out of configure.ac</li> 528 <li>travis: meson: enable unit tests</li> 529 <li>docs: add note about using backticks for rbs in gitlab</li> 530 <li>docs/install: Add meson to the main install page</li> 531 <li>docs/meson: Update LLVM section with information about native files</li> 532 <li>docs/install: Update python dependency section</li> 533 <li>docs/autoconf: Mark autoconf as being replaced</li> 534 <li>meson: Override C++ standard to gnu++11 when building with altivec on ppc64</li> 535 <li>meson: Error out if building nouveau and using LLVM without rtti</li> 536 <li>autotools: Remove tegra vdpau driver</li> 537 <li>meson: Add a script to extract the cmd line used for meson</li> 538 <li>meson: allow building dri driver without window system if osmesa is classic</li> 539 <li>bin/meson-cmd-extract: Also handle cross and native files</li> 540 <li>meson: fix swr KNL build</li> 541 <li>meson: Fix compiler checks for SWR with ICC</li> 542 <li>meson: Add warnings and errors when using ICC</li> 543 <li>automake: Fix path to generated source</li> 544 <li>automake: Add float64.glsl to dist tarball</li> 545 <li>automake: Add include dir for nir src directory</li> 546 <li>configure: Bump SWR LLVM requirement to 7</li> 547 <li>automake: Add --enable-autotools to distcheck flags</li> 548 <li>android,autotools,i965: Fix location of float64_glsl.h</li> 549 <li>VERSION: bump to 19.0.0-rc1</li> 550 <li>Version: Bump for rc2</li> 551 <li>cherry-ignore: Add some patches</li> 552 <li>Revert "intel/compiler: More peephole_select for pre-Gen6"</li> 553 <li>Revert "nir/opt_peephole_select: Don't peephole_select expensive math instructions"</li> 554 <li>Revert "intel/compiler: More peephole select"</li> 555 <li>Bump version for 19.0-rc3</li> 556 <li>version: bump for 19.0-rc4</li> 557 <li>get-pick-list: Add --pretty=medium to the arguments for Cc patches</li> 558 <li>meson: Add dependency on genxml to anvil</li> 559 <li>Version: update to 19.0-rc5</li> 560 <li>Bump version for rc6</li> 561 <li>VERSION: bump version for rc7</li> 562 <li>cherry-ignore: Update the cherry-ignore file</li> 563</ul> 564 565<p>Eduardo Lima Mitev (2):</p> 566<ul> 567 <li>freedreno/ir3: Make imageStore use num components from image format</li> 568 <li>freedreno/ir3: Handle GL_NONE in get_num_components_for_glformat()</li> 569</ul> 570 571<p>Eleni Maria Stea (1):</p> 572<ul> 573 <li>i965: fixed clamping in set_scissor_bits when the y is flipped</li> 574</ul> 575 576<p>Elie Tournier (17):</p> 577<ul> 578 <li>glsl: Add "built-in" function to do abs(fp64)</li> 579 <li>glsl: Add "built-in" functions to do neg(fp64)</li> 580 <li>glsl: Add "built-in" function to do sign(fp64)</li> 581 <li>glsl: Add "built-in" functions to do eq/ne(fp64, fp64)</li> 582 <li>glsl: Add utility function to extract 64-bit sign</li> 583 <li>glsl: Add "built-in" functions to do lt(fp64, fp64)</li> 584 <li>glsl: Add "built-in" functions to do add(fp64, fp64)</li> 585 <li>glsl: Add "built-in" functions to do mul(fp64, fp64)</li> 586 <li>glsl: Add "built-in" functions to do fp64_to_uint(fp64)</li> 587 <li>glsl: Add "built-in" functions to do uint_to_fp64(uint)</li> 588 <li>glsl: Add "built-in" functions to do fp64_to_int(fp64)</li> 589 <li>glsl: Add "built-in" functions to do int_to_fp64(int)</li> 590 <li>glsl: Add "built-in" functions to do fp64_to_fp32(fp64)</li> 591 <li>glsl: Add "built-in" functions to do fp32_to_fp64(fp32)</li> 592 <li>glsl: Add "built-in" functions to do sqrt(fp64)</li> 593 <li>glsl: Add "built-in" functions to do trunc(fp64)</li> 594 <li>glsl: Add "built-in" functions to do round(fp64)</li> 595</ul> 596 597<p>Emil Velikov (81):</p> 598<ul> 599 <li>mesa: bump version to 19.1.0-devel</li> 600 <li>docs: add 19.0.0-devel release notes template</li> 601 <li>docs: mention EXT_shader_implicit_conversions</li> 602 <li>egl: add EGL_EXT_device_base entrypoints</li> 603 <li>egl/glvnd: correctly report errors when vendor cannot be found</li> 604 <li>docs/releasing.html: polish cherry-picking/testing text</li> 605 <li>docs/submittingpatches.html: correctly handle the <p> tag</li> 606 <li>docs: document the staging branch and add reference to it</li> 607 <li>bin/get-pick-list.sh: simplify git oneline printing</li> 608 <li>bin/get-pick-list.sh: prefix output with "[stable] "</li> 609 <li>bin/get-pick-list.sh: handle "typod" usecase.</li> 610 <li>bin/get-pick-list.sh: handle the fixes tag</li> 611 <li>bin/get-pick-list.sh: tweak the commit sha matching pattern</li> 612 <li>bin/get-pick-list.sh: flesh out is_sha_nomination</li> 613 <li>bin/get-pick-list.sh: handle fixes tag with missing colon</li> 614 <li>bin/get-pick-list.sh: handle unofficial "broken by" tag</li> 615 <li>bin/get-pick-list.sh: use test instead of [ ]</li> 616 <li>bin/get-pick-list.sh: handle reverts prior to the branchpoint</li> 617 <li>travis: drop unneeded x11proto-xf86vidmode-dev</li> 618 <li>glx: make xf86vidmode mandatory for direct rendering</li> 619 <li>travis: adding missing x11-xcb for meson+vulkan</li> 620 <li>egl/wayland: bail out when drmGetMagic fails</li> 621 <li>egl/wayland: plug memory leak in drm_handle_device()</li> 622 <li>docs: update 18.3.0 release notes</li> 623 <li>docs: add sha256 checksums for 18.3.0</li> 624 <li>docs: update calendar, add news item and link release notes for 18.3.0</li> 625 <li>freedreno: drop duplicate MKDIR_GEN declaration</li> 626 <li>freedreno: add the missing _la in libfreedreno_ir3_la</li> 627 <li>amd/addrlib: drop si_ci_vi_merged_enum.h from the list</li> 628 <li>docs: add release notes for 18.3.1</li> 629 <li>docs: add sha256 checksums for 18.3.1</li> 630 <li>docs: update calendar, add news item and link release notes for 18.3.1</li> 631 <li>glx: mandate xf86vidmode only for "drm" dri platforms</li> 632 <li>bin/get-pick-list.sh: rework handing of sha nominations</li> 633 <li>bin/get-pick-list.sh: warn when commit lists invalid sha</li> 634 <li>meson: don't require glx/egl/gbm with gallium drivers</li> 635 <li>pipe-loader: meson: reference correct library</li> 636 <li>TODO: glx: meson: build dri based glx tests, only with -Dglx=dri</li> 637 <li>glx: meson: drop includes from a link-only library</li> 638 <li>glx: meson: wire up the dispatch-index-check test</li> 639 <li>glx/test: meson: assorted include fixes</li> 640 <li>configure: add CXX11_CXXFLAGS to LLVM_CXXFLAGS</li> 641 <li>travis: flip to distro xenial, drop sudo false</li> 642 <li>travis: meson: print the configured state</li> 643 <li>travis: printout llvm-config --version</li> 644 <li>travis: meson: use FOO_DRIVERS directly</li> 645 <li>travis: meson: add unwind handling</li> 646 <li>travis: meson: explicitly control the DRI loaders</li> 647 <li>travis: meson: add explicit handling to gallium ST</li> 648 <li>travis: meson: port gallium build combinations over</li> 649 <li>docs: add release notes for 18.3.2</li> 650 <li>docs: add sha256 checksums for 18.3.2</li> 651 <li>docs: update calendar, add news item and link release notes for 18.3.2</li> 652 <li>freedreno: automake: ship ir3_nir_trig.py in the tarball</li> 653 <li>mesa: correctly use os.path.join in our python scripts</li> 654 <li>Revert "mesa/main: remove ARB suffix from glGetnTexImage"</li> 655 <li>mapi: sort static entrypoints numerically</li> 656 <li>mapi: add all _glapi_table entrypoints to static_data.py</li> 657 <li>genCommon.py: Fix typo in _LIBRARY_FEATURE_NAMES.</li> 658 <li>mapi: move genCommon.py to src/mapi/new</li> 659 <li>mapi/new: import mapi scripts from glvnd</li> 660 <li>mapi/new: sort by slot number</li> 661 <li>mapi/new: use the static_data offsets in the new generator</li> 662 <li>mapi/new: reinstate _NO_HIDDEN suffixes in the new generator</li> 663 <li>mapi/new: split out public_entries handling</li> 664 <li>mapi/new: don't print info we don't need for ES1/ES2</li> 665 <li>mapi/new: fixup the GLDEBUGPROCKHR typedef to the non KHR one</li> 666 <li>mapi/new: remove duplicate GLvoid/void substitution</li> 667 <li>autotools: wire the new generator for es1 and es2</li> 668 <li>meson: wire the new generator for es1 and es2</li> 669 <li>scons: wire the new generator for es1 and es2</li> 670 <li>Revert "mapi/new: sort by slot number"</li> 671 <li>mapi/es*api: remove GL_OES_EGL_image entrypoints</li> 672 <li>mapi/es*api: remove GL_EXT_multi_draw_arrays entrypoints</li> 673 <li>mapi/es2api: remove no longer present entrypoints</li> 674 <li>mapi: remove old, unused ES* generator code</li> 675 <li>mapi: remove machinery handling CSV files</li> 676 <li>mapi: print function declarations for shared glapi</li> 677 <li>vc4: Declare the last cpu pointer as being modified in NEON asm.</li> 678 <li>anv: wire up the state_pool_padding test</li> 679 <li>meson: egl: correctly manage loader/xmlconfig</li> 680</ul> 681 682<p>Eric Anholt (171):</p> 683<ul> 684 <li>v3d: Fix a copy-and-paste comment in the simulator code.</li> 685 <li>v3d: Fix a typo in a comment in job handling.</li> 686 <li>v3d: Drop #if 0-ed out v3d_dump_to_file().</li> 687 <li>v3d: Respect user-passed strides for BO imports.</li> 688 <li>v3d: Take advantage of _mesa_hash_table_remove_key() in the simulator.</li> 689 <li>v3d: Use the TLB R/B swapping instead of recompiles when available.</li> 690 <li>v3d: Update the TLB config for depth writes on V3D 4.2.</li> 691 <li>vc4: Drop the winsys_stride relayout in the simluator</li> 692 <li>v3d: Maintain a mapping of the GEM buffer in the simulator.</li> 693 <li>v3d: Remove the special path for simulaton of the submit ioctl.</li> 694 <li>vc4: Take advantage of _mesa_hash_table_remove_key() in the simulator.</li> 695 <li>vc4: Maintain a separate GEM mapping of BOs in the simulator.</li> 696 <li>vc4: Use the normal simulator ioctl path for CL submit as well.</li> 697 <li>gbm: Move gbm_format_canonicalize() to the core.</li> 698 <li>gbm: Introduce a helper function for printing GBM format names.</li> 699 <li>egl: Improve the debugging of gbm format matching in DRI configs.</li> 700 <li>v3d: Fix double-swapping of R/B on V3D 4.1</li> 701 <li>v3d: Don't try to set PF flags on a LDTMU operation</li> 702 <li>vc4: Make sure we make ro scanout resources for create_with_modifiers.</li> 703 <li>vc4: Don't return a vc4 BO handle on a renderonly screen.</li> 704 <li>glx: Remove an old DEFAULT_DRIVER_DIR default.</li> 705 <li>glx: Move DRI extensions pointer loading to driOpenDriver().</li> 706 <li>egl: Move loader_set_logger() up to egl_dri2.c.</li> 707 <li>loader: Stop using a local definition for an in-tree header</li> 708 <li>loader: Factor out the common driver opening logic from each loader.</li> 709 <li>egl: Print the actual message to the console from _eglError().</li> 710 <li>gallium: Fix uninitialized variable warning in compute test.</li> 711 <li>gallium: Remove unused variable in u_tests.</li> 712 <li>v3d: Add renderonly support.</li> 713 <li>v3d: Add support for RGBA_SRGB along with BGRA_SRGB.</li> 714 <li>v3d: Add missing OES_half_float_linear support.</li> 715 <li>v3d: Use combined input/output segments.</li> 716 <li>v3d: Add the V3D TFU submit interface to the simulator.</li> 717 <li>v3d: Use the TFU to do generatemipmap.</li> 718 <li>v3d: Update simulator cache flushing code to match the kernel better.</li> 719 <li>v3d: Create a state uploader for packing our shaders together.</li> 720 <li>v3d: Put default vertex attribute values into the state uploader as well.</li> 721 <li>v3d: Re-use the wrap mode uniform on V3D 3.3.</li> 722 <li>v3d: Make an array for frag/vert texture state in the context.</li> 723 <li>v3d: Don't forget to flush writes to UBOs.</li> 724 <li>v3d: Convert to using nir_src_as_uint() from const_value derefs.</li> 725 <li>v3d: Fix a comment typo</li> 726 <li>v3d: Return the right gl_SampleMaskIn[] value.</li> 727 <li>v3d: Fix handling of texture first_layer offsets for 3D textures.</li> 728 <li>v3d: Avoid confusing auto-indenting in TEXTURE_SHADER_STATE packing</li> 729 <li>v3d: Split most of TEXTURE_SHADER_STATE setup out of sampler views.</li> 730 <li>v3d: Garbage collect unused uniforms code.</li> 731 <li>v3d: Simplify VIR uniform dumping using a temporary.</li> 732 <li>v3d: Add VIR dumping of TMU config p0/p1.</li> 733 <li>v3d: Fix a leak of the transfer helper on screen destroy.</li> 734 <li>vc4: Fix a leak of the transfer helper on screen destroy.</li> 735 <li>v3d: Fix a leak of the disassembled instruction string during debug dumps.</li> 736 <li>tfu</li> 737 <li>shader-packing</li> 738 <li>nir: Add some more consts to the nir_format_convert.h helpers.</li> 739 <li>nir: Pull some of intel's image load/store format conversion to nir_format.h</li> 740 <li>intel: Simplify the half-float packing in image load/store lowering.</li> 741 <li>mesa/st: Expose compute shaders when NIR support is advertised.</li> 742 <li>nir: Print the format of image variables.</li> 743 <li>Revert "intel: Simplify the half-float packing in image load/store lowering."</li> 744 <li>nir: Move intel's half-float image store lowering to to nir_format.h.</li> 745 <li>v3d: Don't forget to wait for our TFU job before rendering from it.</li> 746 <li>v3d: Set up the right stride for raster TFU.</li> 747 <li>v3d: Don't forget to bump the number of writes when doing TFU ops.</li> 748 <li>v3d: Add support for using the TFU to do some blits.</li> 749 <li>v3d: Add support for texturing from linear.</li> 750 <li>v3d: Add safety checks for resource_create().</li> 751 <li>v3d: Make sure that a thrsw doesn't split a multop from its umul24.</li> 752 <li>v3d: Add missing flagging of SYNCB as a TSY op.</li> 753 <li>v3d: Add support for draw indirect for GLES3.1.</li> 754 <li>v3d: Avoid assertion failures when removing end-of-shader instructions.</li> 755 <li>v3d: Move uinfo->data[] dereference to the top of v3d_write_uniforms().</li> 756 <li>v3d: Move uniform pretty-printing to its own helper function.</li> 757 <li>v3d: Use the uniform pretty-printer in v3d_write_uniforms()'s debug code.</li> 758 <li>v3d: Do uniform pretty-printing in the QPU dump.</li> 759 <li>v3d: Drop in a bunch of notes about performance improvement opportunities.</li> 760 <li>vc4: Use the original bit size when scalarizing uniform loads.</li> 761 <li>v3d: Use the original bit size when scalarizing uniform loads.</li> 762 <li>vc4: Reuse nir_format_convert.h in our blend lowering.</li> 763 <li>v3d: Fix the argument type for vir_BRANCH().</li> 764 <li>nir: Fix clamping of uints for image store lowering.</li> 765 <li>v3d: Put the dst bo first in the list of BOs for TFU calls.</li> 766 <li>v3d: Fix check for TFU job completion in the simulator.</li> 767 <li>v3d: Don't try to create shadow tiled temporaries for 1D textures.</li> 768 <li>v3d: Remove dead prototypes for load/store utile functions.</li> 769 <li>v3d: Implement texture_subdata to reduce teximage upload copies.</li> 770 <li>vc4: Move the utile load/store functions to a header for reuse by v3d.</li> 771 <li>v3d: Add a fallthrough path for utile load/store of 32 byte lines.</li> 772 <li>v3d: Load and store aligned utiles all at once.</li> 773 <li>docs: Add a note that MRs should still include any r-b or a-b tags.</li> 774 <li>docs: Add an encouraging note about providing reviews and acks.</li> 775 <li>v3d: Fix simulator mode on i915 render nodes.</li> 776 <li>v3d: Drop shadow comparison state from shader variant key.</li> 777 <li>v3d: Hook up perf_debug() output to GL_ARB_debug output as well.</li> 778 <li>vc4: Hook up perf_debug() output to GL_ARB_debug_output as well.</li> 779 <li>gallium/ttn: Fix setup of outputs_written.</li> 780 <li>v3d: Fix uniform pretty printing assertion failure with branches.</li> 781 <li>v3d: Add a "precompile" debug flag for shader-db.</li> 782 <li>v3d: Hook up some shader-db output to GL_ARB_debug_output.</li> 783 <li>v3d: Drop unused count_nir_instrs() helper.</li> 784 <li>v3d: Drop incorrect dependency for flpop.</li> 785 <li>v3d: Move "does this instruction have flags" from sched to generic helpers.</li> 786 <li>v3d: Don't generate temps for comparisons.</li> 787 <li>v3d: Dead-code eliminate unused flags updates.</li> 788 <li>v3d: Add a note for a potential performance win on multop/umul24.</li> 789 <li>v3d: Force sampling from base level for tg4.</li> 790 <li>v3d: Add support for non-constant texture offsets.</li> 791 <li>v3d: Add support for requesting the sample offsets.</li> 792 <li>v3d: Add support for textureSize() on MSAA textures.</li> 793 <li>v3d: Add support for gl_HelperInvocation.</li> 794 <li>v3d: Fix segfault when failing to compile a program.</li> 795 <li>v3d: Don't forget to include RT writes in precompiles.</li> 796 <li>v3d: Simplify the emission of comparisons for the bcsel optimization.</li> 797 <li>v3d: Move the "Find the ALU instruction generating our bool" out of bcsel.</li> 798 <li>v3d: Don't try to fold non-SSA-src comparisons into bcsels.</li> 799 <li>v3d: Fold comparisons for IF conditions into the flags for the IF.</li> 800 <li>v3d: Handle dynamically uniform IF statements with uniform control flow.</li> 801 <li>v3d: Refactor compiler entrypoints.</li> 802 <li>v3d: Reinstate the new shader-db output after v3d_compile() refactor.</li> 803 <li>v3d: Fix up VS output setup during precompiles.</li> 804 <li>v3d: Remove dead switch cases and comments from v3d_nir_lower_io.</li> 805 <li>v3d: Do UBO loads a vector at a time.</li> 806 <li>v3d: Stop scalarizing our uniform loads.</li> 807 <li>nir: Allow nir_format_unpack_int/sint to unpack larger values.</li> 808 <li>nir: Add nir_lower_tex options to lower sampler return formats.</li> 809 <li>v3d: Use the core tex lowering.</li> 810 <li>nir: Add nir_lower_tex support for Broadcom's swizzled TG4 results.</li> 811 <li>v3d: Enable GL_ARB_texture_gather on V3D 4.x.</li> 812 <li>nir: Make nir_deref_instr_build/get_const_offset actually use size_align.</li> 813 <li>glsl: Fix buffer overflow with an atomic buffer binding out of range.</li> 814 <li>v3d: Add support for flushing dirty TMU data at job end.</li> 815 <li>v3d: Add support for the early_fragment_tests flag.</li> 816 <li>v3d: Add support for GL_ARB_framebuffer_no_attachments.</li> 817 <li>v3d: Fix txf_ms 2D_ARRAY array index.</li> 818 <li>v3d: Add an isr to the simulator to catch GMP violations.</li> 819 <li>v3d: Add support for matrix inputs to the FS.</li> 820 <li>v3d: Drop the GLSL version level.</li> 821 <li>v3d: Add SSBO/atomic counters support.</li> 822 <li>v3d: Add support for shader_image_load_store.</li> 823 <li>v3d: Add support for CS workgroup/invocation id intrinsics.</li> 824 <li>v3d: Add support for CS shared variable load/store/atomics.</li> 825 <li>v3d: Add support for CS barrier() intrinsics.</li> 826 <li>v3d: SHARED but not necessarily SCANOUT buffers on RO must be linear.</li> 827 <li>v3d: If the modifier is not known on BO import, default to linear for RO.</li> 828 <li>v3d: Restructure RO allocations using resource_from_handle.</li> 829 <li>v3d: Don't leak the GPU fd for renderonly usage.</li> 830 <li>vc4: Don't leak the GPU fd for renderonly usage.</li> 831 <li>gallium: Enable unit tests as actual meson unit tests.</li> 832 <li>gallium: Fix comment about possible colorspaces.</li> 833 <li>gallium: Make sure we return is_unorm/is_snorm for compressed formats.</li> 834 <li>v3d: Rename gallium-local limits defines from VC5 to V3D.</li> 835 <li>v3d: Fix overly-large vattr_sizes structs.</li> 836 <li>v3d: Avoid duplicating limits defines between gallium and v3d core.</li> 837 <li>v3d: Drop maximum number of texture units down to 16.</li> 838 <li>v3d: Fix BO stats accounting for imported buffers.</li> 839 <li>v3d: Flush blit jobs immediately after generating them.</li> 840 <li>v3d: Fix release-build warning about utile_h.</li> 841 <li>v3d: Fix stencil sampling from packed depth/stencil.</li> 842 <li>v3d: Fix stencil sampling from a separate-stencil buffer.</li> 843 <li>v3d: Use the symbolic names for wrap modes from the XML.</li> 844 <li>v3d: Move the sampler state to the long-lived state uploader.</li> 845 <li>v3d: Create separate sampler states for the various blend formats.</li> 846 <li>pl111: Rename the pl111 driver to "kmsro".</li> 847 <li>kmsro: Extend to include hx8357d.</li> 848 <li>vc4: Enable NEON asm on meson cross-builds.</li> 849 <li>v3d: Fix the autotools build.</li> 850 <li>mesa: Skip partial InvalidateFramebuffer of packed depth/stencil.</li> 851 <li>v3d: Fix image_load_store clamping of signed integer stores.</li> 852 <li>v3d: Use the early_fragment_tests flag for the shader's disable-EZ field.</li> 853 <li>v3d: Fix the check for "is the last thrsw inside control flow"</li> 854 <li>st/dri: Set the PIPE_BIND_SHARED flag on create_image_with_modifiers.</li> 855</ul> 856 857<p>Eric Engestrom (47):</p> 858<ul> 859 <li>wsi/wayland: use proper VkResult type</li> 860 <li>wsi/wayland: only finish() a successfully init()ed display</li> 861 <li>REVIEWERS: add include path for EGL</li> 862 <li>REVIEWERS: add Emil as EGL reviewer</li> 863 <li>REVIEWERS: add Vulkan reviewer group</li> 864 <li>xmlpool: update translation po files</li> 865 <li>meson: only run vulkan's meson.build when building vulkan</li> 866 <li>gbm: remove unnecessary meson include</li> 867 <li>meson: fix wayland-less builds</li> 868 <li>gbm: add new entrypoint to symbols check</li> 869 <li>egl: add missing glvnd entrypoint for EGL_ANDROID_blob_cache</li> 870 <li>egl: fix bad rebase</li> 871 <li>gbm: add missing comma between strings</li> 872 <li>glapi: add missing visibility args</li> 873 <li>anv: correctly use vulkan 1.0 by default</li> 874 <li>vulkan/utils: s/VERSION/PACKAGE_VERSION/</li> 875 <li>build: stop defining unused VERSION</li> 876 <li>wsi/display: fix mem leak when freeing swapchains</li> 877 <li>vulkan/wsi: fix s/,/;/ typo</li> 878 <li>meson: skip asm check when asm is disabled</li> 879 <li>anv: add unreachable() for VK_EXT_fragment_density_map</li> 880 <li>mesa: drop unused & deprecated lib</li> 881 <li>loader: deduplicate logger function declaration</li> 882 <li>docs: add meson cross compilation instructions</li> 883 <li>docs: format code blocks a bit nicely</li> 884 <li>docs: fix the meson aarch64 cross-file</li> 885 <li>docs: advertise distro-provided meson cross-files</li> 886 <li>anv: drop unneeded KHR suffix</li> 887 <li>wsi: drop unneeded KHR suffix</li> 888 <li>radv: remove a few more unnecessary KHR suffixes</li> 889 <li>egl: add missing includes</li> 890 <li>egl: remove unused include</li> 891 <li>travis: avoid using unset llvm-config</li> 892 <li>egl: fix python lib deprecation warning</li> 893 <li>docs: explain how to see what meson options exist</li> 894 <li>travis: fix autotools build after --enable-autotools switch addition</li> 895 <li>configure: EGL requirements only apply if EGL is built</li> 896 <li>egl: finalize EGL_MESA_query_driver</li> 897 <li>egl: update headers from Khronos</li> 898 <li>egl: add glvnd entrypoints for EGL_MESA_query_driver</li> 899 <li>travis: bump libdrm to 2.4.97</li> 900 <li>egl/glvnd: sync egl.xml from Khronos</li> 901 <li>anv: drop always-successful VkResult</li> 902 <li>meson/vdpau: add missing soversion</li> 903 <li>xvmc: fix string comparison</li> 904 <li>xvmc: fix string comparison</li> 905 <li>egl: fix libdrm-less builds</li> 906</ul> 907 908<p>Erik Faye-Lund (70):</p> 909<ul> 910 <li>glsl: add has_implicit_conversions()-helper</li> 911 <li>glsl: add has_implicit_uint_to_int_conversion()-helper</li> 912 <li>glsl: fall back to inexact function-match</li> 913 <li>mesa/glsl: add support for EXT_shader_implicit_conversions</li> 914 <li>glsl: do not allow implicit casts of unsized array initializers</li> 915 <li>mesa: expose NV_conditional_render on GLES</li> 916 <li>mesa/main: fixup make check after NV_conditional_render for gles</li> 917 <li>Revert "mesa/main: fixup make check after NV_conditional_render for gles"</li> 918 <li>Revert "mesa: expose NV_conditional_render on GLES"</li> 919 <li>mesa/main: correct requirement for EXT_occlusion_query_boolean</li> 920 <li>mesa/main: correct year for EXT_occlusion_query_boolean</li> 921 <li>mesa/main: use non-prefixed enums for consistency</li> 922 <li>mesa/main: simplify pipeline-statistics query validation</li> 923 <li>mesa/main: fix validation of GL_SAMPLES_PASSED</li> 924 <li>mesa/main: fix validation of GL_ANY_SAMPLES_PASSED</li> 925 <li>mesa/main: fix validation of GL_ANY_SAMPLES_PASSED_CONSERVATIVE</li> 926 <li>mesa/main: fix validation of GL_TIME_ELAPSED</li> 927 <li>mesa/main: fix validation of transform-feedback queries</li> 928 <li>mesa/main: fix validation of transform-feedback overflow queries</li> 929 <li>mesa/main: fix validation of ARB_query_buffer_object</li> 930 <li>mesa/main: fix validation of GL_TIMESTAMP</li> 931 <li>mesa/main: remove overly strict query-validation</li> 932 <li>mesa/main: remove ARB suffix from glGetnTexImage</li> 933 <li>mesa/main: remove bogus error for zero-sized images</li> 934 <li>mesa/main: factor out tex-image error-checking</li> 935 <li>mesa/main: factor out common error-checking</li> 936 <li>mesa/main: check cube-completeness in common code</li> 937 <li>mesa/main: fix incorrect depth-error</li> 938 <li>mesa/main: fixup requirements for GL_PRIMITIVES_GENERATED</li> 939 <li>mesa/main: make _mesa_has_tessellation return bool</li> 940 <li>mesa/main: rename format-check function</li> 941 <li>mesa/main: clean up S3_s3tc check</li> 942 <li>mesa/main: clean up OES_texture_float_linear check</li> 943 <li>mesa/main: clean up ES2_compatibility check</li> 944 <li>mesa/main: clean up integer texture check</li> 945 <li>mesa/main: use _mesa_has_FOO_bar for compressed format checks</li> 946 <li>mesa/main: do not allow s3tc enums on gles1</li> 947 <li>mesa/main: do not allow etc2 enums on gles1</li> 948 <li>mesa/main: do not allow astc enums on gles1</li> 949 <li>mesa/main: do not allow depth-texture enums on gles1</li> 950 <li>mesa/main: do not allow stencil-texture enums on gles1</li> 951 <li>mesa/main: do not allow ARB_texture_rgb10_a2ui enums before gles3</li> 952 <li>mesa/main: do not allow integer-texture enums before gles3</li> 953 <li>mesa/main: do not allow ARB_depth_buffer_float enums before gles3</li> 954 <li>mesa/main: do not allow EXT_packed_float enums before gles3</li> 955 <li>mesa/main: do not allow rg-textures enums before gles3</li> 956 <li>mesa/main: do not allow EXT_texture_shared_exponent enums before gles3</li> 957 <li>mesa/main: do not allow MESA_ycbcr_texture enums on gles</li> 958 <li>mesa/main: do not allow type_2_10_10_10_REV enums before gles3</li> 959 <li>mesa/main: do not allow floating-point texture enums on gles1</li> 960 <li>mesa/main: do not allow snorm-texture enums before gles3</li> 961 <li>mesa/main: do not allow sRGB texture enums before gles3</li> 962 <li>mesa/main: do not allow EXT_texture_sRGB_R8 enums before gles3</li> 963 <li>mesa/main: split float-texture support checking in two</li> 964 <li>mesa/main: require EXT_texture_type_2_10_10_10_REV for gles3</li> 965 <li>mesa/main: require EXT_texture_sRGB for gles3</li> 966 <li>mesa/st: do not probe for the same texture-formats twice</li> 967 <li>mesa/main: do not require float-texture filtering for es3</li> 968 <li>mesa/main: correct validation for GL_RGB565</li> 969 <li>mesa/main: fix up _mesa_has_rg_textures for gles2</li> 970 <li>virgl: force linear texturing support</li> 971 <li>virgl: simplify virgl_hw_set_vertex_buffers</li> 972 <li>virgl: simplify virgl_hw_set_index_buffer</li> 973 <li>virgl: wrap vertex element state in a struct</li> 974 <li>virgl: work around bad assumptions in virglrenderer</li> 975 <li>anv/meson: make sure tests link with -msse2</li> 976 <li>anv/autotools: make sure tests link with -msse2</li> 977 <li>docs: add note about sending merge-requests from forks</li> 978 <li>mapi: drop unneeded gl_dispatch_stub declarations</li> 979 <li>virgl: remove unused variable</li> 980</ul> 981 982<p>Ernestas Kulik (2):</p> 983<ul> 984 <li>vc4: Fix leak in HW queries error path</li> 985 <li>v3d: Fix leak in resource setup error path</li> 986</ul> 987 988<p>Francisco Jerez (14):</p> 989<ul> 990 <li>intel/fs: Prevent emission of IR instructions not aligned to their own execution size.</li> 991 <li>intel/fs: Handle source modifiers in lower_integer_multiplication().</li> 992 <li>intel/fs: Implement quad swizzles on ICL+.</li> 993 <li>intel/fs: Fix bug in lower_simd_width while splitting an instruction which was already split.</li> 994 <li>intel/eu/gen7: Fix brw_MOV() with DF destination and strided source.</li> 995 <li>intel/fs: Respect CHV/BXT regioning restrictions in copy propagation pass.</li> 996 <li>intel/fs: Constify fs_inst::can_do_source_mods().</li> 997 <li>intel/fs: Introduce regioning lowering pass.</li> 998 <li>intel/fs: Remove existing lower_conversions pass.</li> 999 <li>intel/fs: Remove nasty open-coded CHV/BXT 64-bit workarounds.</li> 1000 <li>intel/fs: Remove FS_OPCODE_UNPACK_HALF_2x16_SPLIT opcodes.</li> 1001 <li>intel/fs: Promote execution type to 32-bit when any half-float conversion is needed.</li> 1002 <li>intel/fs: Exclude control sources from execution type and region alignment calculations.</li> 1003 <li>intel/fs: Implement extended strides greater than 4 for IR source regions.</li> 1004</ul> 1005 1006<p>Fritz Koenig (2):</p> 1007<ul> 1008 <li>freedreno: drm_fourcc.h header include</li> 1009 <li>freedreno: add query for dmabuf modifiers</li> 1010</ul> 1011 1012<p>Gert Wollny (30):</p> 1013<ul> 1014 <li>mesa/core: Add definitions and translations for EXT_texture_sRGB_R8</li> 1015 <li>Gallium: Add format PIPE_FORMAT_R8_SRGB</li> 1016 <li>mesa/st: Add support for EXT_texture_sRGB_R8</li> 1017 <li>virgl/vtest-winsys: Use virgl version of bind flags</li> 1018 <li>r600: Add support for EXT_texture_sRGB_R8</li> 1019 <li>mesa: Reference count shaders that are used by transform feedback objects</li> 1020 <li>virgl: Add command and flags to initiate debugging on the host (v2)</li> 1021 <li>nir: Allow to skip integer ops in nir_lower_to_source_mods</li> 1022 <li>i965: Correct L8_UNORM_SRGB table entry</li> 1023 <li>i965: be more specific about FBO completeness errors</li> 1024 <li>i965: Force zero swizzles for unused components in GL_RED and GL_RG</li> 1025 <li>i965: Add support for and expose EXT_texture_sRGB_R8</li> 1026 <li>virgl: Use file descriptor instead of un-allocated object</li> 1027 <li>i965:use FRAMEBUFFER_UNSUPPORTED instead of FRAMEBUFFER_INCOMPLETE_DIMENSIONS</li> 1028 <li>r600: Only set context streamout strides info from the shader that has outputs</li> 1029 <li>r600: clean up the GS ring buffers when the context is destroyed</li> 1030 <li>glsl: free or reuse memory allocated for TF varying</li> 1031 <li>virgl,vtest: Initialize return value</li> 1032 <li>virgl: Don't try handling server fences when they are not supported</li> 1033 <li>i965: Explicitely handle swizzles for MESA_FORMAT_R_SRGB8</li> 1034 <li>i965: Set the FBO error state INCOMPLETE_ATTACHMENT only for SRGB_R8</li> 1035 <li>autotools: Deprecate the use of autotools</li> 1036 <li>Gallium: Add new CAPS to indicate whether a driver can switch SRGB write</li> 1037 <li>virgl: Set sRGB write control CAP based on host capabilities</li> 1038 <li>mesa:main: Add flag for EXT_sRGB to gl_extensions</li> 1039 <li>i965: Set flag for EXT_sRGB</li> 1040 <li>mesa/st: rework support for sRGB framebuffer attachements</li> 1041 <li>mesa/main: Use flag for EXT_sRGB instead of EXT_framebuffer_sRGB where possible</li> 1042 <li>mesa/main/version: Lower the requirements for GLES 3.0</li> 1043 <li>mesa/main: Expose EXT_sRGB_write_control</li> 1044</ul> 1045 1046<p>Guido Günther (2):</p> 1047<ul> 1048 <li>etnaviv: Make sure rs alignment checks match</li> 1049 <li>etnaviv: fix typo in cflush_all description</li> 1050</ul> 1051 1052<p>Gurchetan Singh (18):</p> 1053<ul> 1054 <li>egl: add missing #include <stddef.h> in egldevice.h</li> 1055 <li>virgl: quadruple command buffer size</li> 1056 <li>virgl: avoid large inline transfers</li> 1057 <li>virgl: don't mark buffers as unclean after a write</li> 1058 <li>virgl: texture_transfer_pool --> transfer_pool</li> 1059 <li>virgl: remove unnessecary code</li> 1060 <li>virgl: move texture metadata to common code</li> 1061 <li>virgl: move virgl_resource_layout to common code</li> 1062 <li>virgl: move vrend_get_tex_image_offset to common code</li> 1063 <li>virgl: store layer_stride in metadata</li> 1064 <li>virgl: consolidate transfer code</li> 1065 <li>virgl: make transfer code with PIPE_BUFFER targets</li> 1066 <li>virgl: make virgl_buffers use resource helpers</li> 1067 <li>virgl: modify how we handle GL_MAP_FLUSH_EXPLICIT_BIT</li> 1068 <li>virgl: move resource metadata into base resource</li> 1069 <li>virgl: move resource creation / import / destruction to common code</li> 1070 <li>virgl: don't flush an empty range</li> 1071 <li>virgl: remove empty file</li> 1072</ul> 1073 1074<p>Hanno Böck (1):</p> 1075<ul> 1076 <li>glsl/test: Fix use after free in test_optpass.</li> 1077</ul> 1078 1079<p>Hyunjun Ko (1):</p> 1080<ul> 1081 <li>freedreno: implements get_sample_position</li> 1082</ul> 1083 1084<p>Iago Toral Quiroga (22):</p> 1085<ul> 1086 <li>intel/compiler: fix node interference of simd16 instructions</li> 1087 <li>nir/constant_folding: fix incorrect bit-size check</li> 1088 <li>nir/from_ssa: fix bit-size of temporary register</li> 1089 <li>Revert "nir/builder: Assert that intN_t immediates fit"</li> 1090 <li>intel/compiler: fix indentation style in opt_algebraic()</li> 1091 <li>intel/compiler: fix register allocation in opt_peephole_sel</li> 1092 <li>intel/compiler: do not copy-propagate strided regions to ddx/ddy arguments</li> 1093 <li>intel/compiler: move nir_lower_bool_to_int32 before nir_lower_locals_to_regs</li> 1094 <li>compiler/nir: add a nir_b2f() helper</li> 1095 <li>compiler/nir: add nir_fadd_imm() and nir_fmul_imm() helpers</li> 1096 <li>compiler/spirv: handle 16-bit float in radians() and degrees()</li> 1097 <li>compiler/spirv: implement 16-bit asin</li> 1098 <li>compiler/spirv: implement 16-bit acos</li> 1099 <li>compiler/spirv: implement 16-bit atan</li> 1100 <li>compiler/spirv: implement 16-bit atan2</li> 1101 <li>compiler/spirv: implement 16-bit exp and log</li> 1102 <li>compiler/spirv: implement 16-bit hyperbolic trigonometric functions</li> 1103 <li>compiler/spirv: implement 16-bit frexp</li> 1104 <li>compiler/spirv: use 32-bit polynomial approximation for 16-bit asin()</li> 1105 <li>anv/pipeline_cache: fix incorrect guards for NIR cache</li> 1106 <li>anv/pipeline_cache: free NIR shader cache</li> 1107 <li>anv/device: fix maximum number of images supported</li> 1108</ul> 1109 1110<p>Ian Romanick (28):</p> 1111<ul> 1112 <li>glsl: Add warning tests for identifiers with __</li> 1113 <li>glsl: Add pragma to disable all warnings</li> 1114 <li>glsl: prevent qualifiers modification of predeclared variables</li> 1115 <li>glsl: Omit redundant qualifier checks on redeclarations</li> 1116 <li>glsl: Refactor type checking for redeclarations</li> 1117 <li>nir: Add a saturated unsigned integer add opcode</li> 1118 <li>i965/fs: Implement nir_op_uadd_sat</li> 1119 <li>nir/phi_builder: Internal users should use nir_phi_builder_value_set_block_def too</li> 1120 <li>util/slab: Rename slab_mempool typed parameters to mempool</li> 1121 <li>util/hash_table: Add _mesa_hash_table_init function</li> 1122 <li>nir/phi_builder: Use per-value hash table to store [block] -> def mapping</li> 1123 <li>nir: Fix holes in nir_instr</li> 1124 <li>nir: Release per-block metadata in nir_sweep</li> 1125 <li>i965/vec4: Silence unused parameter warnings in vec4 compiler tests</li> 1126 <li>i965/vec4/dce: Don't narrow the write mask if the flags are used</li> 1127 <li>i965/fs: Eliminate unary op on operand of compare-with-zero</li> 1128 <li>i965/vec4: Propagate conditional modifiers from more compares to other compares</li> 1129 <li>nir/opt_peephole_select: Don't try to remove flow control around indirect loads</li> 1130 <li>intel/compiler: More peephole select</li> 1131 <li>nir/opt_peephole_select: Don't peephole_select expensive math instructions</li> 1132 <li>intel/compiler: More peephole_select for pre-Gen6</li> 1133 <li>Revert "nir/lower_indirect: Bail early if modes == 0"</li> 1134 <li>nir/algebraic: Don't put quotes around floating point literals</li> 1135 <li>glsl: Add utility to convert text files to C strings</li> 1136 <li>nir: Silence zillions of unused parameter warnings in release builds</li> 1137 <li>spirv: Add missing break</li> 1138 <li>intel/fs: nir_op_extract_i8 extracts a byte, not a word</li> 1139 <li>intel/fs: Fix extract_u8 of an odd byte from a 64-bit integer</li> 1140</ul> 1141 1142<p>Ilia Mirkin (37):</p> 1143<ul> 1144 <li>nv50/ir: delete MINMAX instruction that is no longer in the BB</li> 1145 <li>nv50/ir/ra: improve condition for short regs, unify with cond for 16-bit</li> 1146 <li>nv50/ir/ra: enforce max register requirement, and change spill order</li> 1147 <li>nv50/ir: remove dnz flag when converting MAD to ADD due to optimizations</li> 1148 <li>nv50: always keep TSC slot 0 bound</li> 1149 <li>nv50,nvc0: add explicit handling of PIPE_CAP_MAX_VERTEX_ELEMENT_SRC_OFFSET</li> 1150 <li>nouveau: set texture upload budget</li> 1151 <li>nvc0: replace use of explicit default_tsc with entry 0</li> 1152 <li>nvc0: always keep TSC slot 0 bound to fix TXF</li> 1153 <li>st/mesa: remove sampler associated with buffer texture in pbo logic</li> 1154 <li>st/mesa: allow glDrawElements to work with GL_SELECT feedback</li> 1155 <li>tgsi: add ATOMFADD operation</li> 1156 <li>gallium: add PIPE_CAP_TGSI_ATOMFADD to indicate support</li> 1157 <li>st/mesa: select ATOMFADD when source type is float</li> 1158 <li>st/mesa: expose GL_NV_shader_atomic_float when ATOMFADD is supported</li> 1159 <li>nv50/ir: add support for converting ATOMFADD to proper ir</li> 1160 <li>nvc0: enable GL_NV_shader_atomic_float on pre-Maxwell</li> 1161 <li>nv50,nvc0: add missing CAPs for unsupported features</li> 1162 <li>nv30: avoid setting user_priv without setting cur_ctx</li> 1163 <li>nv30: fix rare issue with fp unbinding not finding the bufctx</li> 1164 <li>nv30: add support for multi-layer transfers</li> 1165 <li>nv30: use correct helper to get blocks in y direction</li> 1166 <li>nv30: fix some s3tc layout issues</li> 1167 <li>nv30: disable rendering to 3D textures</li> 1168 <li>docs: fix gallium screen cap docs</li> 1169 <li>nv50,nvc0: mark textures dirty on fb update</li> 1170 <li>nvc0: don't put text segment into bufctx</li> 1171 <li>nvc0/ir: fix second tex argument after levelZero optimization</li> 1172 <li>nv50,nvc0: add explicit settings for recent caps</li> 1173 <li>nvc0: add support for handling indirect draws with attrib conversion</li> 1174 <li>nvc0/ir: always use CG mode for loads from atomic-only buffers</li> 1175 <li>nvc0: fix 3d images on kepler</li> 1176 <li>nv50,nvc0: use condition for occlusion queries when already complete</li> 1177 <li>nvc0: stick zero values for the compute invocation counts</li> 1178 <li>nvc0: we have 16k-sized framebuffers, fix default scissors</li> 1179 <li>swr: set PIPE_CAP_MAX_VARYINGS correctly</li> 1180 <li>glsl: fix recording of variables for XFB in TCS shaders</li> 1181</ul> 1182 1183<p>Indrajit Das (1):</p> 1184<ul> 1185 <li>st/va: Return correct status from vlVaQuerySurfaceStatus</li> 1186</ul> 1187 1188<p>Jakob Bornecrantz (1):</p> 1189<ul> 1190 <li>virgl/vtest: Use default socket name from protocol header</li> 1191</ul> 1192 1193<p>Jan Vesely (2):</p> 1194<ul> 1195 <li>amd: Make vgpr-spilling depend on llvm version</li> 1196 <li>clover: Fix build after clang r348827</li> 1197</ul> 1198 1199<p>Jason Ekstrand (207):</p> 1200<ul> 1201 <li>vulkan: Update the XML and headers to 1.1.91</li> 1202 <li>intel/fs,vec4: Clean up a repeated pattern with SSBOs</li> 1203 <li>intel/fs: Use the new nir_src_is_const and friends</li> 1204 <li>nir: Add a read_mask helper for ALU instructions</li> 1205 <li>intel/vec4: Use the new nir_src_is_const and friends</li> 1206 <li>intel/analyze_ubo_ranges: Use nir_src_is_const and friends</li> 1207 <li>anv: Use nir_src_is_const and friends in lowering code</li> 1208 <li>intel/fs: Add an assert to optimize_frontfacing_ternary</li> 1209 <li>nir/lower_alu_to_scalar: Don't try to lower unpack_32_2x16</li> 1210 <li>nir/builder: Assert that intN_t immediates fit</li> 1211 <li>nir/builder: Add iadd_imm and imul_imm helpers</li> 1212 <li>nir/builder: Add a nir_pack/unpack/bitcast helpers</li> 1213 <li>nir/spirv: Force 32-bit for UBO and SSBO Booleans</li> 1214 <li>nir/glsl: Force 32-bit for UBO and SSBO Booleans</li> 1215 <li>nir/lower_io: Add shared to get_io_offset_src</li> 1216 <li>nir: Add alignment parameters to SSBO, UBO, and shared access</li> 1217 <li>intel/compiler: Lower SSBO and shared loads/stores in NIR</li> 1218 <li>intel,nir: Move gl_LocalInvocationID lowering to nir_lower_system_values</li> 1219 <li>intel/fs,vec4: Fix a compiler warning</li> 1220 <li>vulkan: Update the XML and headers to 1.1.93</li> 1221 <li>anv: Expose VK_EXT_scalar_block_layout</li> 1222 <li>anv: Put robust buffer access in the pipeline hash</li> 1223 <li>anv/nir: Rework arguments to apply_pipeline_layout</li> 1224 <li>nir/derefs: Add a nir_derefs_do_not_alias enum value</li> 1225 <li>vulkan: Update the XML and headers to 1.1.95</li> 1226 <li>nir/opcodes: Pull in the type helpers from constant_expressions</li> 1227 <li>nir/opcodes: Rename tbool to tbool32</li> 1228 <li>nir/algebraic: Clean up some __str__ cruft</li> 1229 <li>nir/algebraic: Refactor codegen a bit</li> 1230 <li>nir/algebraic: Add support for unsized conversion opcodes</li> 1231 <li>nir/opt_algebraic: Simplify an optimization using the new search ops</li> 1232 <li>nir/opt_algebraic: Drop bit-size suffixes from conversions</li> 1233 <li>nir/opt_algebraic: Add 32-bit specifiers to a bunch of booleans</li> 1234 <li>nir: Make boolean conversions sized just like the others</li> 1235 <li>anv,radv: Disable VK_EXT_pci_bus_info</li> 1236 <li>intel/ir: Don't allow allocating zero registers</li> 1237 <li>spirv: Add support for MinLod</li> 1238 <li>nir/lower_tex: Simplify lower_gradient logic</li> 1239 <li>nir/lower_tex: Modify txd instructions instead of replacing them</li> 1240 <li>nir/lower_tex: Add lowering for some min_lod cases</li> 1241 <li>intel/fs: Support min_lod parameters on texture instructions</li> 1242 <li>anv: Advertise support for MinLod on Skylake+</li> 1243 <li>anv/pipeline: Set the correct binding count for compute shaders</li> 1244 <li>intel/blorp: Assert that we don't re-layout a compressed surface</li> 1245 <li>nir: Document the function inlining process</li> 1246 <li>nir: Allow [iu]mul_high on non-32-bit types</li> 1247 <li>nir/lower_int64: Add support for [iu]mul_high</li> 1248 <li>nir: Add a pass for lowering integer division by constants</li> 1249 <li>i965/vec4: Implement nir_op_uadd_sat</li> 1250 <li>i965: Enable nir_opt_idiv_const for 32 and 64-bit integers</li> 1251 <li>nir/lower_idiv: Use ilt instead of bit twiddling</li> 1252 <li>nir/tgsi: Use nir_bany in ttn_kill_if</li> 1253 <li>nir/constant_folding: Fix source bit size logic</li> 1254 <li>nir/algebraic: Optimize x2b(xneg(a)) -> a</li> 1255 <li>nir: Drop support for lower_b2f</li> 1256 <li>nir/algebraic: Make an optimization more specific</li> 1257 <li>nir: Rename Boolean-related opcodes to include 32 in the name</li> 1258 <li>nir/constant_expressions: Rework Boolean handling</li> 1259 <li>nir: Add support for 1-bit data types</li> 1260 <li>nir/large_constants: Properly handle 1-bit bools</li> 1261 <li>nir/algebraic: Generalize an optimization</li> 1262 <li>nir: Add 1-bit Boolean opcodes</li> 1263 <li>nir: Add a bool to int32 lowering pass</li> 1264 <li>nir: Switch to using 1-bit Booleans for almost everything</li> 1265 <li>nir/algebraic: Optimize 1-bit Booleans</li> 1266 <li>nir/algebraic: Add some optimizations for D3D-style Booleans</li> 1267 <li>radv: Fix a stupid if in gather_intrinsic_info</li> 1268 <li>st/nir: Use nir_src_as_uint for tokens</li> 1269 <li>vulkan: Update the XML and headers to 1.1.96</li> 1270 <li>anv,radv: Re-enable VK_EXT_pci_bus_info</li> 1271 <li>anv: Bump the patch version to 96</li> 1272 <li>nir/propagate_invariant: Skip unknown vars</li> 1273 <li>nir/linking_helpers: Look at derefs for modes</li> 1274 <li>nir/lower_io_arrays_to_elements: Look at derefs for modes</li> 1275 <li>nir/lower_io_to_scalar: Look at derefs for modes</li> 1276 <li>nir/lower_wpos_center: Look at derefs for modes</li> 1277 <li>nir/copy_prop_vars: Get modes directly from derefs</li> 1278 <li>nir/dead_write_vars: Get modes directly from derefs</li> 1279 <li>radv/query: Add a nir_test_flag helper</li> 1280 <li>radv/query: Use 1-bit booleans in query shaders</li> 1281 <li>intel/blorp: Be more conservative about copying clear colors</li> 1282 <li>vulkan: Update the XML and headers to 1.1.97</li> 1283 <li>glsl_type: Support serializing 8 and 16-bit types</li> 1284 <li>spirv: Handle any bit size in vector_insert/extract</li> 1285 <li>anv/apply_pipeline_layout: Set the cursor in lower_res_reindex_intrinsic</li> 1286 <li>spirv: Sign-extend array indices</li> 1287 <li>spirv: Emit NIR deref instructions on-the-fly</li> 1288 <li>nir/builder: Add nir_i2i and nir_u2u helpers which take a bit size</li> 1289 <li>spirv: Handle arbitrary bit sizes for deref array indices</li> 1290 <li>nir/validate: Require array indices to match the deref bit size</li> 1291 <li>nir: Allow storing to shader_storage</li> 1292 <li>nir: Distinguish between normal uniforms and UBOs</li> 1293 <li>glsl_type: Drop the glsl_get_array_instance C helper</li> 1294 <li>glsl_type: Add a C wrapper to get struct field offsets</li> 1295 <li>glsl_type: Simplify glsl_channel_type</li> 1296 <li>glsl_type: Add support for explicitly laid out matrices and arrays</li> 1297 <li>spirv: Propagate layout decorations to created glsl_types</li> 1298 <li>nir: Move propagation of cast derefs to a new nir_opt_deref pass</li> 1299 <li>nir: Add a ptr_as_array deref type</li> 1300 <li>nir/validate: Don't allow derefs in if conditions</li> 1301 <li>nir/opt_deref: Properly optimize ptr_as_array derefs</li> 1302 <li>nir/deref: Support casts and ptr_as_array in comparisons</li> 1303 <li>nir/deref: Skip over casts in fixup_deref_modes</li> 1304 <li>nir/remove_dead_variables: Properly handle deref casts</li> 1305 <li>nir/validate: Allow derefs in phi nodes</li> 1306 <li>nir/intrinsics: Allow deref sources to consume anything</li> 1307 <li>nir/intrinsics: Add access flags to load/store_deref</li> 1308 <li>nir/validate: Allow array derefs on vectors in more modes</li> 1309 <li>nir/lower_io: Add "explicit" IO lowering</li> 1310 <li>nir/vulkan: Add a descriptor type to vulkan resource intrinsics</li> 1311 <li>spirv: Add error checking for Block and BufferBlock decorations</li> 1312 <li>spirv: Choose atomic deref type with pointer_uses_ssa_offset</li> 1313 <li>spirv: Add explicit pointer types</li> 1314 <li>spirv: Make better use of vtn_pointer_uses_ssa_offset</li> 1315 <li>spirv: Add support for using derefs for UBO/SSBO access</li> 1316 <li>anv: Enable the new deref-based UBO/SSBO path</li> 1317 <li>spirv: Sort supported capabilities</li> 1318 <li>anv: Sort properties and features switch statements</li> 1319 <li>nir: Add some more int64 lowering helpers</li> 1320 <li>anv/pipeline: Constant fold after apply_pipeline_layout</li> 1321 <li>anv/pipeline: Move wpos and input attachment lowering to lower_nir</li> 1322 <li>compiler/types: Serialize/deserialize subpass input types correctly</li> 1323 <li>anv/pipeline: Hash shader modules and spec constants separately</li> 1324 <li>anv/pipeline_cache: Add support for caching NIR</li> 1325 <li>anv/pipeline: Cache the pre-lowered NIR</li> 1326 <li>intel/peephole_ffma: Fix swizzle propagation</li> 1327 <li>spirv: Whack sampler/image pointers to uniform</li> 1328 <li>spirv: Contain the GLSLang issue #179 workaround to old GLSLang</li> 1329 <li>intel/nir: Call nir_opt_deref in brw_nir_optimize</li> 1330 <li>nir/gcm: Support deref instructions</li> 1331 <li>spirv: Emit switch conditions on-the-fly</li> 1332 <li>intel/blorp: Add two more filter modes</li> 1333 <li>anv: Rename has_resolve to has_color_resolve</li> 1334 <li>anv/blorp: Refactor MSAA resolves into an exportable helper function</li> 1335 <li>anv: Move resolve_subpass to genX_cmd_buffer.c</li> 1336 <li>anv: Implement VK_KHR_depth_stencil_resolve</li> 1337 <li>nir: Add a bool to float32 lowering pass</li> 1338 <li>intel/eu: Stop overriding exec sizes in send_indirect_message</li> 1339 <li>intel/fs: Don't touch accumulator destination while applying regioning alignment rule</li> 1340 <li>anv: Re-sort the extensions list</li> 1341 <li>anv: Only parse pImmutableSamplers if the descriptor has samplers</li> 1342 <li>relnotes: Add newly added Vulkan extensions</li> 1343 <li>anv/pipeline: Add a pdevice helper variable</li> 1344 <li>nir: Mark deref UBO and SSBO access as non-scalar</li> 1345 <li>spirv: Update the JSON and headers from Khronos master</li> 1346 <li>anv: Always emit at least one vertex element</li> 1347 <li>spirv: Initialize struct member offsets to -1</li> 1348 <li>spirv: Only split blocks</li> 1349 <li>spirv: Only set interface_type on blocks</li> 1350 <li>nir: Preserve offsets in lower_io_to_scalar_early</li> 1351 <li>nir/xfb: Fix offset accounting for dvec3/4</li> 1352 <li>nir/xfb: Properly handle arrays of blocks</li> 1353 <li>anv: Add but do not enable VK_EXT_transform_feedback</li> 1354 <li>anv: Add pipeline cache support for xfb_info</li> 1355 <li>anv: Implement the basic form of VK_EXT_transform_feedback</li> 1356 <li>anv: Implement vkCmdDrawIndirectByteCountEXT</li> 1357 <li>anv: Implement CmdBegin/EndQueryIndexed</li> 1358 <li>genxml: Add SO_PRIM_STORAGE_NEEDED and SO_NUM_PRIMS_WRITTEN</li> 1359 <li>anv: Implement transform feedback queries</li> 1360 <li>nir: Add load/store/atomic global intrinsics</li> 1361 <li>nir/lower_io: Add a 32 and 64-bit global address formats</li> 1362 <li>nir/lower_io: Add support for nir_var_mem_global</li> 1363 <li>nir/validate: Allow array derefs of vectors for nir_var_mem_global</li> 1364 <li>nir: Allow SSBOs and global to alias</li> 1365 <li>spirv: Drop a bogus assert</li> 1366 <li>spirv: Handle OpTypeForwardPointer</li> 1367 <li>spirv: Implement OpConvertPtrToU and OpConvertUToPtr</li> 1368 <li>spirv: Add support for SPV_EXT_physical_storage_buffer</li> 1369 <li>intel/fs: Get rid of fs_inst::equals</li> 1370 <li>intel/defines: Explicitly cast to uint32_t in SET_FIELD and SET_BITS</li> 1371 <li>intel/fs: Handle IMAGE_SIZE in size_read() and is_send_from_grf()</li> 1372 <li>intel/fs: Take an explicit exec size in brw_surface_payload_size()</li> 1373 <li>intel/eu: Add has_simd4x2 bools to surface_write functions</li> 1374 <li>intel/eu: Rework surface descriptor helpers</li> 1375 <li>intel/fs: Add a generic SEND opcode</li> 1376 <li>intel/fs: Use SHADER_OPCODE_SEND for surface messages</li> 1377 <li>intel/fs: Use a logical opcode for IMAGE_SIZE</li> 1378 <li>intel/fs: Use SHADER_OPCODE_SEND for texturing on gen7+</li> 1379 <li>intel/fs: Use SHADER_OPCODE_SEND for varying UBO pulls on gen7+</li> 1380 <li>intel/eu: Use GET_BITS in brw_inst_set_send_ex_desc</li> 1381 <li>intel/eu/validate: SEND restrictions also apply to SENDC</li> 1382 <li>intel/eu: Add more message descriptor helpers</li> 1383 <li>intel/disasm: Rework SEND decoding to use descriptors</li> 1384 <li>intel/inst: Fix the ia16_addr_imm helpers</li> 1385 <li>intel/inst: Indent some code</li> 1386 <li>intel/eu: Add support for the SENDS[C] messages</li> 1387 <li>intel/disasm: Properly disassemble split sends</li> 1388 <li>intel/fs: Support SENDS in SHADER_OPCODE_SEND</li> 1389 <li>intel/fs: Add interference between SENDS sources</li> 1390 <li>intel/fs: Use split sends for surface writes on gen9+</li> 1391 <li>intel/fs: Do the grf127 hack on SIMD8 instructions in SIMD16 mode</li> 1392 <li>nir/deref: Rematerialize parents in rematerialize_derefs_in_use_blocks</li> 1393 <li>intel/fs: Bail in optimize_extract_to_float if we have modifiers</li> 1394 <li>compiler/types: Add a contains_64bit helper</li> 1395 <li>nir/xfb: Properly align 64-bit values</li> 1396 <li>nir: Rewrite lower_clip_cull_distance_arrays to do a lot less lowering</li> 1397 <li>nir/xfb: Work in terms of components rather than slots</li> 1398 <li>nir/xfb: Handle compact arrays in gather_xfb_info</li> 1399 <li>nir/lower_clip_cull: Fix an incorrect assert</li> 1400 <li>anv: Count surfaces for non-YCbCr images in GetDescriptorSetLayoutSupport</li> 1401 <li>spirv: OpImageQueryLod requires a sampler</li> 1402 <li>intel,nir: Lower TXD with min_lod when the sampler index is not < 16</li> 1403 <li>spirv: Pull offset/stride from the pointer for OpArrayLength</li> 1404 <li>anv: Refactor descriptor pushing a bit</li> 1405 <li>anv: Take references to push descriptor set layouts</li> 1406 <li>nir: Add a pass for lowering IO back to vector when possible</li> 1407 <li>intel/nir: Vectorize all IO</li> 1408</ul> 1409 1410<p>Jiang, Sonny (1):</p> 1411<ul> 1412 <li>radeonsi: add compute_last_block to configure the partial block fields</li> 1413</ul> 1414 1415<p>Jon Turney (3):</p> 1416<ul> 1417 <li>glx: Fix compilation with GLX_USE_WINDOWSGL</li> 1418 <li>appveyor: put build steps in a script, rather than inline in appveyor.yml</li> 1419 <li>appveyor: Add a Cygwin build script</li> 1420</ul> 1421 1422<p>Jonathan Marek (42):</p> 1423<ul> 1424 <li>nir: add fceil lowering</li> 1425 <li>freedreno: a2xx: fd2_draw update</li> 1426 <li>freedreno/a2xx: fix POINT_MINMAX_MAX overflow</li> 1427 <li>freedreno: add missing a20x ids</li> 1428 <li>freedreno/a2xx: set VIZ_QUERY_ID on a20x</li> 1429 <li>freedreno/a2xx: Compute depth base in gmem correctly</li> 1430 <li>freedreno: a2xx texture update</li> 1431 <li>freedreno: use GENERIC instead of TEXCOORD for blit program</li> 1432 <li>freedreno: use MSM_BO_SCANOUT with scanout buffers</li> 1433 <li>glsl/nir: int constants as float for native_integers=false</li> 1434 <li>glsl/nir: ftrunc for native_integers=false float to int cast</li> 1435 <li>glsl/nir: keep bool types when native_integers=false</li> 1436 <li>freedreno: a2xx: cleanup init_shader_const</li> 1437 <li>freedreno: a2xx: cleanup REG_A2XX_PA_CL_VTE_CNTL</li> 1438 <li>freedreno: a2xx: fix gmem2mem viewport</li> 1439 <li>freedreno: a2xx: fix VERTEX_REUSE/DEALLOC on a20x</li> 1440 <li>freedreno: a2xx: fix non-zero texture base offsets</li> 1441 <li>freedreno: a2xx: sysmem rendering</li> 1442 <li>freedreno: a2xx: NIR backend</li> 1443 <li>freedreno: a2xx: insert scalar MOV to allow 2 source scalar</li> 1444 <li>freedreno: a2xx: add ir2 copy propagation</li> 1445 <li>freedreno: a2xx: add partial lower_scalar pass for ir2</li> 1446 <li>freedreno: add renderonly scanout</li> 1447 <li>freedreno: a2xx: ir2 cleanup</li> 1448 <li>freedreno: a2xx: enable early-Z testing</li> 1449 <li>freedreno: update a2xx registers</li> 1450 <li>freedreno: a2xx: a20x hw binning</li> 1451 <li>freedreno: a2xx: clear fixes and fast clear path</li> 1452 <li>freedreno: a2xx: minor solid_vertexbuf fixups</li> 1453 <li>freedreno: a2xx: add perfcntrs</li> 1454 <li>kmsro: Add freedreno renderonly support</li> 1455 <li>st/dri: invalidate_resource depth/stencil before flush_resource</li> 1456 <li>mesa/st: wire up DiscardFramebuffer</li> 1457 <li>freedreno: fix invalidate logic</li> 1458 <li>freedreno: fix depth usage logic</li> 1459 <li>freedreno: fix sysmem rendering being used when clear is used</li> 1460 <li>freedreno: a2xx: fix fast clear</li> 1461 <li>freedreno: a2xx: don't write 4th vertex in mem2gmem</li> 1462 <li>freedreno: a2xx: add use_hw_binning function</li> 1463 <li>freedreno: a2xx: fix fast clear for some gmem configurations</li> 1464 <li>freedreno: a2xx: fix mipmapping for NPOT textures</li> 1465 <li>freedreno: use renderonly path for buffers allocated with modifiers</li> 1466</ul> 1467 1468<p>Jordan Justen (3):</p> 1469<ul> 1470 <li>docs: Document GitLab merge request process (email alternative)</li> 1471 <li>i965/genX_state: Add register access functions</li> 1472 <li>i965/compute: Emit GPGPU_WALKER in genX_state_upload</li> 1473</ul> 1474 1475<p>Jose Maria Casanova Crespo (1):</p> 1476<ul> 1477 <li>glsl: TCS outputs can not be transform feedback candidates on GLES</li> 1478</ul> 1479 1480<p>José Fonseca (2):</p> 1481<ul> 1482 <li>appveyor: Revert commits adding Cygwin support.</li> 1483 <li>scons: Workaround failures with MSVC when using SCons 3.0.[2-4].</li> 1484</ul> 1485 1486<p>Juan A. Suarez Romero (17):</p> 1487<ul> 1488 <li>docs: add release notes for 18.2.5</li> 1489 <li>docs: add sha256 checksums for 18.2.5</li> 1490 <li>docs: update calendar, add news item and link release notes for 18.2.5</li> 1491 <li>docs: add release notes for 18.2.6</li> 1492 <li>docs: add sha256 checksums for 18.2.6</li> 1493 <li>docs: update calendar, add news item and link release notes for 18.2.6</li> 1494 <li>docs: extends 18.2 lifecycle</li> 1495 <li>docs: add release notes for 18.2.7</li> 1496 <li>docs: add sha256 checksums for 18.2.7</li> 1497 <li>docs: update calendar, add news item and link release notes for 18.2.7</li> 1498 <li>docs: add release notes for 18.2.8</li> 1499 <li>docs: add sha256 checksums for 18.2.8</li> 1500 <li>docs: update calendar, add news item and link release notes for 18.2.8</li> 1501 <li>anv/cmd_buffer: check for NULL framebuffer</li> 1502 <li>genxml: add missing field values for 3DSTATE_SF</li> 1503 <li>anv: advertise 8 subpixel precision bits</li> 1504 <li>anv: destroy descriptor sets when pool gets reset</li> 1505</ul> 1506 1507<p>Józef Kucia (1):</p> 1508<ul> 1509 <li>nir: Fix assert in print_intrinsic_instr().</li> 1510</ul> 1511 1512<p>Karol Herbst (35):</p> 1513<ul> 1514 <li>nv50/ir: print color masks of tex instructions</li> 1515 <li>nv50/ra: add condenseDef overloads for partial condenses</li> 1516 <li>nv50/ir: add scalar field to TexInstructions</li> 1517 <li>gm107/ir: use scalar tex instructions where possible</li> 1518 <li>gm107/ir: fix compile time warning in getTEXSMask</li> 1519 <li>nir: add const_index parameters to system value builder function</li> 1520 <li>nir: replace nir_load_system_value calls with appropiate builder functions</li> 1521 <li>nir/spirv: cast shift operand to u32</li> 1522 <li>nv50,nvc0: Fix gallium nine regression regarding sampler bindings</li> 1523 <li>nv50/ir: initialize relDegree staticly</li> 1524 <li>nouveau: use atomic operations for driver statistics</li> 1525 <li>nv50/ir: fix use-after-free in ConstantFolding::visit</li> 1526 <li>nir: rename global/local to private/function memory</li> 1527 <li>nv50/ir: disable tryCollapseChainedMULs in ConstantFolding for precise instructions</li> 1528 <li>gm107/ir: disable TEXS for tex with derivAll set</li> 1529 <li>nir: rename nir_var_private to nir_var_shader_temp</li> 1530 <li>nir: rename nir_var_function to nir_var_function_temp</li> 1531 <li>nir: rename nir_var_ubo to nir_var_mem_ubo</li> 1532 <li>nir: rename nir_var_ssbo to nir_var_mem_ssbo</li> 1533 <li>nir: rename nir_var_shared to nir_var_mem_shared</li> 1534 <li>nir/spirv: handle SpvStorageClassCrossWorkgroup</li> 1535 <li>glsl/lower_output_reads: set invariant and precise flags on temporaries</li> 1536 <li>nir: replace more nir_load_system_value calls with builder functions</li> 1537 <li>nir/validate: allow to check against a bitmask of bit_sizes</li> 1538 <li>nir: add legal bit_sizes to intrinsics</li> 1539 <li>nir: add bit_size parameter to system values with multiple allowed bit sizes</li> 1540 <li>mesa: add MESA_SHADER_KERNEL</li> 1541 <li>vtn: handle SpvExecutionModelKernel</li> 1542 <li>nir/spirv: handle ContractionOff execution mode</li> 1543 <li>gk104/ir: Use the new rcp/rsq in library</li> 1544 <li>gm107/ir: add fp64 rcp</li> 1545 <li>gm107/ir: add fp64 rsq</li> 1546 <li>gallium: add PIPE_CAP_MAX_VARYINGS</li> 1547 <li>st/mesa: require RGBA2, RGB4, and RGBA4 to be renderable</li> 1548 <li>nir/opt_if: don't mark progress if nothing changes</li> 1549</ul> 1550 1551<p>Kenneth Graunke (41):</p> 1552<ul> 1553 <li>intel: Use a URB start offset of 0 for disabled stages.</li> 1554 <li>st/mesa: Pull nir_lower_wpos_ytransform work into a helper function.</li> 1555 <li>st/nir: Drop unused parameter from st_nir_assign_uniform_locations().</li> 1556 <li>st/mesa: Don't record garbage streamout information in the non-SSO case.</li> 1557 <li>i915: Delete swizzling detection logic.</li> 1558 <li>nir: Use nir_shader_get_entrypoint in nir_lower_clip_vs().</li> 1559 <li>nir: Inline lower_clip_vs() into nir_lower_clip_vs().</li> 1560 <li>nir: Save nir_variable pointers in nir_lower_clip_vs rather than locs.</li> 1561 <li>nir: Make nir_lower_clip_vs optionally work with variables.</li> 1562 <li>i965: Allow only one slot of clip distances to be set on Gen4-5.</li> 1563 <li>i965: Use a 'nir' temporary rather than poking at brw_program</li> 1564 <li>i965: Do NIR shader cloning in the caller.</li> 1565 <li>intel/compiler: Use nir's info when checking uses_streams.</li> 1566 <li>intel/blorp: Expand blorp_address::offset to be 64 bits.</li> 1567 <li>i965: Delete dead brw_meta_resolve_color prototype.</li> 1568 <li>i965: Flip arguments to load_register_reg helpers.</li> 1569 <li>genxml: Consistently use a numeric "MOCS" field</li> 1570 <li>i965: Don't override subslice count to 4 on Gen11.</li> 1571 <li>st/mesa: Drop dead 'passthrough_fs' field.</li> 1572 <li>st/mesa: Drop !passColor optimization in drawpixels shaders.</li> 1573 <li>st/mesa: Don't open code the drawpixels vertex shader.</li> 1574 <li>st/mesa: Combine the DrawPixels and Bitmap passthrough VS programs.</li> 1575 <li>st/nir: Gather info after applying lowering FS variant features</li> 1576 <li>st/nir: Drop unused gl_program parameter in VS input handling helper.</li> 1577 <li>nir: Fix gl_nir_lower_samplers_as_deref's structure type handling.</li> 1578 <li>nir: Make gl_nir_lower_samplers use gl_nir_lower_samplers_as_deref</li> 1579 <li>blorp: Add blorp_get_surface_address to the driver interface.</li> 1580 <li>blorp: Pass the batch to lookup/upload_shader instead of context</li> 1581 <li>nir: Allow a non-existent sampler deref in nir_lower_samplers_as_deref</li> 1582 <li>st/nir: Lower TES gl_PatchVerticesIn to a constant if linked with a TCS.</li> 1583 <li>i965: Drop mark_surface_used mechanism.</li> 1584 <li>st/mesa: Make an enum for pipeline statistics query result indices.</li> 1585 <li>st/mesa: Rearrange PIPE_QUERY_PIPELINE_STATISTICS result fetching.</li> 1586 <li>gallium: Add the ability to query a single pipeline statistics counter</li> 1587 <li>st/mesa: Optionally override RGB/RGBX dst alpha blend factors</li> 1588 <li>gallium: Add forgotten docs for PIPE_CAP_GLSL_TESS_LEVELS_AS_INPUTS.</li> 1589 <li>st/mesa: Limit GL_MAX_[NATIVE_]PROGRAM_PARAMETERS_ARB to 2048</li> 1590 <li>anv: Put MOCS in the correct location</li> 1591 <li>nir: Don't reassociate add/mul chains containing only constants</li> 1592 <li>compiler: Mark clip/cull distance arrays as compact before lowering.</li> 1593 <li>spirv: Eliminate dead input/output variables after translation.</li> 1594</ul> 1595 1596<p>Kirill Burtsev (1):</p> 1597<ul> 1598 <li>loader: free error state, when checking the drawable type</li> 1599</ul> 1600 1601<p>Kristian H. Kristensen (14):</p> 1602<ul> 1603 <li>freedreno/a6xx: Clear z32 and separate stencil with blitter</li> 1604 <li>freedreno/a6xx: Move restore blits to IB</li> 1605 <li>freedreno/a6xx: Move resolve blits to an IB</li> 1606 <li>freedreno/a6xx: Clear gmem buffers at flush time</li> 1607 <li>gallium: Android build fixes</li> 1608 <li>mesa: Add core support for EXT_multisampled_render_to_texture{,2}</li> 1609 <li>gallium: Add new PIPE_CAP_SURFACE_SAMPLE_COUNT</li> 1610 <li>st/mesa: Add support for EXT_multisampled_render_to_texture</li> 1611 <li>freedreno: Add support for EXT_multisampled_render_to_texture</li> 1612 <li>freedreno: Fix the Makefile.am fix</li> 1613 <li>glapi: fixup EXT_multisampled_render_to_texture dispatch</li> 1614 <li>freedreno: Synchronize batch and flush for staging resource</li> 1615 <li>freedreno/a6xx: Turn on texture tiling by default</li> 1616 <li>freedreno/a6xx: Emit blitter dst with OUT_RELOCW</li> 1617</ul> 1618 1619<p>Leo Liu (2):</p> 1620<ul> 1621 <li>st/va: fix the incorrect max profiles report</li> 1622 <li>st/va/vp9: set max reference as default of VP9 reference number</li> 1623</ul> 1624 1625<p>Lionel Landwerlin (47):</p> 1626<ul> 1627 <li>intel/dump_gpu: add missing gdb option</li> 1628 <li>intel/sanitize_gpu: add help/gdb options to wrapper</li> 1629 <li>intel/sanitize_gpu: deal with non page multiple buffer sizes</li> 1630 <li>intel/sanitize_gpu: add debug message on mmap fail</li> 1631 <li>intel/decoders: fix instruction base address parsing</li> 1632 <li>anv: stub internal android code</li> 1633 <li>anv/android: mark gralloc allocated BOs as external</li> 1634 <li>intel/dump_gpu: move output option together</li> 1635 <li>intel/dump_gpu: add platform option</li> 1636 <li>intel/aub_read: remove useless breaks</li> 1637 <li>nir/lower_tex: add alpha channel parameter for yuv lowering</li> 1638 <li>nir/lower_tex: Add AYUV lowering support</li> 1639 <li>dri: add AYUV format</li> 1640 <li>i965: add support for sampling from AYUV</li> 1641 <li>anv: simplify internal address offset</li> 1642 <li>anv/image: remove unused parameter</li> 1643 <li>anv/lower_ycbcr: make sure to set 0s on all components</li> 1644 <li>anv: associate vulkan formats with aspects</li> 1645 <li>anv: use image aspects rather than computed ones</li> 1646 <li>anv: move helper function internally</li> 1647 <li>egl/dri: fix error value with unknown drm format</li> 1648 <li>intel/decoders: read ring buffer length</li> 1649 <li>intel/aubinator: fix ring buffer pointer</li> 1650 <li>intel/aub_viewer: fix dynamic state printing</li> 1651 <li>intel/aub_viewer: Print blend states properly</li> 1652 <li>anv: flush pipeline before query result copies</li> 1653 <li>anv/query: flush render target before copying results</li> 1654 <li>anv: don't do partial resolve on layer > 0</li> 1655 <li>intel/aub_viewer: fix shader get_bo</li> 1656 <li>intel/aub_viewer: fixup 0x address prefix</li> 1657 <li>intel/aub_viewer: print address of missing shader</li> 1658 <li>intel/aub_viewer: fix shader view</li> 1659 <li>intel/aub_viewer: fold binding/sampler table items</li> 1660 <li>intel/aub_viewer: highlight true booleans</li> 1661 <li>i965: limit VF caching workaround to gen8/9/10</li> 1662 <li>intel/blorp: emit VF caching workaround before 3DSTATE_VERTEX_BUFFERS</li> 1663 <li>i965: include draw_params/derived_draw_params for VF cache workaround</li> 1664 <li>i965: add CS stall on VF invalidation workaround</li> 1665 <li>anv: explictly specify format for blorp ccs/mcs op</li> 1666 <li>anv: flush fast clear colors into compressed surfaces</li> 1667 <li>anv: fix invalid binding table index computation</li> 1668 <li>anv: narrow flushing of the render target to buffer writes</li> 1669 <li>anv: document cache flushes & invalidations</li> 1670 <li>intel/genxml: add missing MI_PREDICATE compare operations</li> 1671 <li>vulkan: make generated enum to strings helpers available from c++</li> 1672 <li>intel: fix urb size for CFL GT1</li> 1673 <li>intel/compiler: use correct swizzle for replacement</li> 1674</ul> 1675 1676<p>Lucas Stach (6):</p> 1677<ul> 1678 <li>etnaviv: use dummy RT buffer when rendering without color buffer</li> 1679 <li>etnaviv: use surface format directly</li> 1680 <li>st/dri: allow both render and sampler compatible dma-buf formats</li> 1681 <li>st/dri: replace format conversion functions with single mapping table</li> 1682 <li>etnaviv: enable full overwrite in a few more cases</li> 1683 <li>etnaviv: annotate variables only used in debug build</li> 1684</ul> 1685 1686<p>Marek Olšák (94):</p> 1687<ul> 1688 <li>st/va: fix incorrect use of resource_destroy</li> 1689 <li>ac/surface: remove the overallocation workaround for Vega12</li> 1690 <li>radeonsi: use better DCC clear codes</li> 1691 <li>radeonsi: don't set the CB clear color registers for 0/1 clear colors on Raven2</li> 1692 <li>gallium: add PIPE_CONTEXT_LOSE_CONTEXT_ON_RESET</li> 1693 <li>radeonsi: stop command submission with PIPE_CONTEXT_LOSE_CONTEXT_ON_RESET only</li> 1694 <li>st/mesa: disable L3 thread pinning</li> 1695 <li>mesa: mark GL_SR8_EXT non-renderable on GLES</li> 1696 <li>radeonsi: fix video APIs on Raven2</li> 1697 <li>gallium/u_tests: add a compute shader test that clears an image</li> 1698 <li>gallium/u_tests: fix MSVC build by using old-style zero initializers</li> 1699 <li>mesa/glthread: pass the function name to _mesa_glthread_restore_dispatch</li> 1700 <li>mesa/glthread: enable immediate mode</li> 1701 <li>drirc: enable glthread for Talos Principle</li> 1702 <li>st/mesa: regularly re-pin driver threads to the CCX where the app thread is</li> 1703 <li>st/mesa: pin driver threads to a fixed CCX when glthread is enabled</li> 1704 <li>radeonsi: don't send data after write-confirm with BOTTOM_OF_PIPE_TS</li> 1705 <li>radeonsi: go back to using bottom-of-pipe for beginning of TIME_ELAPSED</li> 1706 <li>winsys/amdgpu: fix a buffer leak in amdgpu_bo_from_handle</li> 1707 <li>winsys/amdgpu: fix a device handle leak in amdgpu_winsys_create</li> 1708 <li>radeonsi: clean up primitive binning enablement</li> 1709 <li>radeonsi: use structured buffer intrinsics for image views</li> 1710 <li>radeonsi: fix is_oneway_access_only for image stores</li> 1711 <li>radeonsi: small cleanup for memory opcodes</li> 1712 <li>tgsi/scan: add more information about bindless usage</li> 1713 <li>radeonsi/nir: parse more information about bindless usage</li> 1714 <li>radeonsi: fix is_oneway_access_only for bindless images</li> 1715 <li>winsys/amdgpu: always reclaim/release slabs if there is not enough memory</li> 1716 <li>radeonsi: generalize the slab allocator code to allow layered slab allocators</li> 1717 <li>winsys/amdgpu: optimize slab allocation for 2 MB amdgpu page tables</li> 1718 <li>winsys/amdgpu: clean up code around BO VM alignment</li> 1719 <li>winsys/amdgpu: use >= instead of > for VM address alignment</li> 1720 <li>winsys/amdgpu: increase the VM alignment to the MSB of the size for Gfx9</li> 1721 <li>winsys/amdgpu: overallocate buffers for faster address translation on Gfx9</li> 1722 <li>winsys/amdgpu,radeon: pass vm_alignment to buffer_from_handle</li> 1723 <li>winsys/amdgpu: use optimal VM alignment for imported buffers</li> 1724 <li>winsys/amdgpu: use optimal VM alignment for CPU allocations</li> 1725 <li>radeonsi: allow si_cp_dma_clear_buffer to clear GDS from any IB</li> 1726 <li>winsys/amdgpu: add support for allocating GDS and OA resources</li> 1727 <li>radeonsi: add memory management stress tests for GDS</li> 1728 <li>Revert "winsys/amdgpu: overallocate buffers for faster address translation on Gfx9"</li> 1729 <li>st/mesa: expose GL_OES_texture_view</li> 1730 <li>mesa: expose GL_EXT_texture_view as an alias of GL_OES_texture_view</li> 1731 <li>mesa: expose EXT_texture_compression_rgtc on GLES</li> 1732 <li>mesa: expose EXT_texture_compression_bptc in GLES</li> 1733 <li>mesa: expose AMD_texture_texture4</li> 1734 <li>st/mesa: expose EXT_render_snorm on GLES</li> 1735 <li>radeonsi: don't emit redundant PKT3_NUM_INSTANCES packets</li> 1736 <li>radeonsi: call si_fix_resource_usage for the GS copy shader as well</li> 1737 <li>radeonsi: make si_cp_wait_mem more configurable</li> 1738 <li>radeonsi: use u_decomposed_prims_for_vertices instead of u_prims_for_vertices</li> 1739 <li>radeonsi: remove unused variables in si_insert_input_ptr</li> 1740 <li>radeonsi: always unmap texture CPU mappings on 32-bit CPU architectures</li> 1741 <li>ac: remove unused variable from ac_build_ddxy</li> 1742 <li>st/mesa: unify window-system renderbuffer initialization</li> 1743 <li>st/mesa: don't reference pipe_surface locally in PBO code</li> 1744 <li>st/mesa: don't leak pipe_surface if pipe_context is not current</li> 1745 <li>st/dri: fix dri2_format_table for argb1555 and rgb565</li> 1746 <li>radeonsi: also apply the GS hang workaround to draws without tessellation</li> 1747 <li>winsys/amdgpu: fix whitespace</li> 1748 <li>winsys/amdgpu: use the new BO list API</li> 1749 <li>radeonsi: fix a u_blitter crash after a shader with FBFETCH</li> 1750 <li>radeonsi: fix rendering to tiny viewports where the viewport center is > 8K</li> 1751 <li>radeonsi: use buffer_store_format_x & xy</li> 1752 <li>radeonsi: remove redundant call to emit_cache_flush in compute clear/copy</li> 1753 <li>radeonsi: compile clear and copy buffer compute shaders on demand</li> 1754 <li>radeonsi: correct WRITE_DATA.DST_SEL definitions</li> 1755 <li>radeonsi: fix the top-of-pipe fence on SI</li> 1756 <li>radeonsi: don't use WRITE_DATA.DST_SEL == MEM_GRBM on >= CIK</li> 1757 <li>radeonsi: move PKT3_WRITE_DATA generation into a helper function</li> 1758 <li>gallium: add SINT formats to have exact counterparts to SNORM formats</li> 1759 <li>gallium/util: add util_format_snorm8_to_sint8 (from radeonsi)</li> 1760 <li>radeonsi: disable render cond & pipeline stats for internal compute dispatches</li> 1761 <li>radeonsi: rename rscreen -> sscreen</li> 1762 <li>radeonsi: rename rview -> sview</li> 1763 <li>winsys/amdgpu: rename rfence, rsrc, rdst -> afence, asrc, adst</li> 1764 <li>radeonsi: remove r600 from comments</li> 1765 <li>radeonsi: rename r600_resource -> si_resource</li> 1766 <li>radeonsi: rename rquery -> squery</li> 1767 <li>radeonsi: rename rsrc -> ssrc, rdst -> sdst</li> 1768 <li>radeonsi: rename rbo, rbuffer to buf or buffer</li> 1769 <li>radeonsi: rename rfence -> sfence</li> 1770 <li>st/mesa: purge framebuffers when unbinding a context</li> 1771 <li>st/mesa: fix PRIMITIVES_GENERATED query after the "pipeline stat single" changes</li> 1772 <li>ac: use the correct LLVM processor name on Raven2</li> 1773 <li>radeonsi: fix crashing performance counters (division by zero)</li> 1774 <li>meson: drop the xcb-xrandr version requirement</li> 1775 <li>gallium/u_threaded: fix EXPLICIT_FLUSH for flush offsets > 0</li> 1776 <li>radeonsi: fix EXPLICIT_FLUSH for flush offsets > 0</li> 1777 <li>winsys/amdgpu: don't drop manually added fence dependencies</li> 1778 <li>radeonsi: add driconf option radeonsi_enable_nir</li> 1779 <li>radeonsi: always enable NIR for Civilization 6 to fix corruption</li> 1780 <li>driconf: add Civ6Sub executable for Civilization 6</li> 1781 <li>tgsi: don't set tgsi_info::uses_bindless_images for constbufs and hw atomics</li> 1782</ul> 1783 1784<p>Mario Kleiner (4):</p> 1785<ul> 1786 <li>radeonsi: Fix use of 1- or 2- component GL_DOUBLE vbo's.</li> 1787 <li>egl/wayland: Allow client->server format conversion for PRIME offload. (v2)</li> 1788 <li>egl/wayland-drm: Only announce formats via wl_drm which the driver supports.</li> 1789 <li>drirc: Add sddm-greeter to adaptive_sync blacklist.</li> 1790</ul> 1791 1792<p>Mark Janes (3):</p> 1793<ul> 1794 <li>Revert "i965/batch: avoid reverting batch buffer if saved state is an empty"</li> 1795 <li>Revert "Implementation of egl dri2 drivers for MESA_query_driver"</li> 1796 <li>Revert "Implement EGL API for MESA_query_driver"</li> 1797</ul> 1798 1799<p>Mathias Fröhlich (17):</p> 1800<ul> 1801 <li>mesa: Remove needless indirection in some draw functions.</li> 1802 <li>mesa: Rename gl_vertex_array_object::_Enabled -> Enabled.</li> 1803 <li>mesa: Use the gl_vertex_array_object::Enabled bitfield.</li> 1804 <li>mesa: Use gl_vertex_array_object::Enabled for glGet.</li> 1805 <li>mesa: Remove gl_array_attributes::Enabled.</li> 1806 <li>mesa: Work with bitmasks when en/dis-abling VAO arrays.</li> 1807 <li>mesa: Unify glEdgeFlagPointer data type.</li> 1808 <li>nouveau: Use gl_array_attribute::_ElementSize.</li> 1809 <li>tnl: Use gl_array_attribute::_ElementSize.</li> 1810 <li>mesa: Factor out struct gl_vertex_format.</li> 1811 <li>mesa: Remove unneeded bitfield widths from the VAO.</li> 1812 <li>mesa/st: Only care about the uploader if it was used.</li> 1813 <li>mesa/st: Only unmap the uploader that was actually used.</li> 1814 <li>mesa/st: Factor out array and buffer setup from st_atom_array.c.</li> 1815 <li>mesa/st: Avoid extra references in the feedback draw function scope.</li> 1816 <li>mesa/st: Use binding information from the VAO in feedback rendering.</li> 1817 <li>mesa/st: Make st_pipe_vertex_format static.</li> 1818</ul> 1819 1820<p>Matt Turner (41):</p> 1821<ul> 1822 <li>util/ralloc: Switch from DEBUG to NDEBUG</li> 1823 <li>util/ralloc: Make sizeof(linear_header) a multiple of 8</li> 1824 <li>nir: Call fflush() at the end of nir_print_shader()</li> 1825 <li>glsl: Remove unused member variable</li> 1826 <li>gallivm: Use nextafterf(0.5, 0.0) as rounding constant</li> 1827 <li>mesa: Revert INTEL_fragment_shader_ordering support</li> 1828 <li>Revert "st/mesa: silenced unhanded enum warning in st_glsl_to_tgsi.cpp"</li> 1829 <li>i965/fs: Handle V/UV immediates in dump_instructions()</li> 1830 <li>glsl: Add function support to glsl_to_nir</li> 1831 <li>glsl: Create file to contain software fp64 functions</li> 1832 <li>glsl: Add "built-in" functions to do ffma(fp64)</li> 1833 <li>glsl: Add "built-in" functions to do fmin/fmax(fp64)</li> 1834 <li>glsl: Add "built-in" function to do ffloor(fp64)</li> 1835 <li>glsl: Add "built-in" functions to do ffract(fp64)</li> 1836 <li>glsl: Add "built-in" functions to convert bool to double</li> 1837 <li>nir: Rework nir_lower_constant_initializers() to handle functions</li> 1838 <li>nir: Tag entrypoint for easy recognition by nir_shader_get_entrypoint()</li> 1839 <li>nir: Wire up int64 lowering functions</li> 1840 <li>nir: Implement lowering of 64-bit shift operations</li> 1841 <li>nir: Add and set info::uses_64bit</li> 1842 <li>nir: Create nir_builder in nir_lower_doubles_impl()</li> 1843 <li>nir: Add lowering support for 64-bit operations to software</li> 1844 <li>nir: Unset metadata debug bit if no progress made</li> 1845 <li>intel/compiler: Lower 64-bit MOV/SEL operations</li> 1846 <li>intel/compiler: Split 64-bit MOV-indirects if needed</li> 1847 <li>intel/compiler: Avoid false positive assertions</li> 1848 <li>intel/compiler: Rearrange code to avoid future problems</li> 1849 <li>intel/compiler: Prevent warnings in the following patch</li> 1850 <li>intel/compiler: Expand size of the 'nr' field</li> 1851 <li>intel/compiler: Heap-allocate temporary storage</li> 1852 <li>i965: Compile fp64 software routines and lower double-ops</li> 1853 <li>i965: Enable 64-bit GLSL extensions</li> 1854 <li>i965: Compile fp64 funcs only if we do not have 64-bit hardware support</li> 1855 <li>intel/compiler: Reset default flag register in brw_find_live_channel()</li> 1856 <li>gallium: Enable ASIMD/NEON on aarch64.</li> 1857 <li>gallivm: Return true from arch_rounding_available() if NEON is available</li> 1858 <li>intel/compiler: Add a file-level description of brw_eu_validate.c</li> 1859 <li>i965: Always compile fp64 funcs when needed</li> 1860 <li>nir: Optimize double-precision lower_round_even()</li> 1861 <li>intel/compiler: Avoid propagating inequality cmods if types are different</li> 1862 <li>intel/compiler/test: Add unit test for mismatched signedness comparison</li> 1863</ul> 1864 1865<p>Mauro Rossi (6):</p> 1866<ul> 1867 <li>android: gallium/auxiliary: add include to get u_debug.h header</li> 1868 <li>android: radv: add libmesa_git_sha1 static dependency</li> 1869 <li>android: amd/addrlib: update Mesa's copy of addrlib</li> 1870 <li>android: st/mesa: fix building error due to sched_getcpu()</li> 1871 <li>android: anv: fix generated files depedencies (v2)</li> 1872 <li>android: anv: fix libexpat shared dependency</li> 1873</ul> 1874 1875<p>Maya Rashish (2):</p> 1876<ul> 1877 <li>radeon: fix printf format specifier.</li> 1878 <li>configure: fix test portability</li> 1879</ul> 1880 1881<p>Michal Srb (2):</p> 1882<ul> 1883 <li>gallium: Constify drisw_loader_funcs struct</li> 1884 <li>drisw: Use separate drisw_loader_funcs for shm</li> 1885</ul> 1886 1887<p>Michel Dänzer (4):</p> 1888<ul> 1889 <li>winsys/amdgpu: Stop using amdgpu_bo_handle_type_kms_noimport</li> 1890 <li>winsys/amdgpu: Pull in LLVM CFLAGS</li> 1891 <li>amd/common: Restore v4i32 suffix for llvm.SI.load.const intrinsic</li> 1892 <li>loader/dri3: Use strlen instead of sizeof for creating VRR property atom</li> 1893</ul> 1894 1895<p>Neha Bhende (1):</p> 1896<ul> 1897 <li>st/mesa: Fix topogun-1.06-orc-84k-resize.trace crash</li> 1898</ul> 1899 1900<p>Neil Roberts (4):</p> 1901<ul> 1902 <li>freedreno: Add .dir-locals to the common directory</li> 1903 <li>spirv/nir: handle location decorations on block interface members</li> 1904 <li>glsl_types: Rename parameter of glsl_count_attribute_slots</li> 1905 <li>spirv: Don't use special semantics when counting vertex attribute size</li> 1906</ul> 1907 1908<p>Nicholas Kazlauskas (5):</p> 1909<ul> 1910 <li>util: Get program name based on path when possible</li> 1911 <li>util: Add adaptive_sync driconf option</li> 1912 <li>drirc: Initial blacklist for adaptive sync</li> 1913 <li>loader/dri3: Enable adaptive_sync via _VARIABLE_REFRESH property</li> 1914 <li>radeonsi: Enable adaptive_sync by default for radeon</li> 1915</ul> 1916 1917<p>Nicolai Hähnle (37):</p> 1918<ul> 1919 <li>radv: include LLVM IR in the VK_AMD_shader_info "disassembly"</li> 1920 <li>radeonsi: fix an out-of-bounds read reported by ASAN</li> 1921 <li>winsys/amdgpu: add amdgpu_winsys_bo::lock</li> 1922 <li>winsys/amdgpu: explicitly declare whether buffer_map is permanent or not</li> 1923 <li>egl/wayland: rather obvious build fix</li> 1924 <li>radv: remove dependency on addrlib gfx9_enum.h</li> 1925 <li>ac/surface/gfx9: let addrlib choose the preferred swizzle kind</li> 1926 <li>amd/addrlib: update Mesa's copy of addrlib</li> 1927 <li>meson: link LLVM 'native' component when LLVM is available</li> 1928 <li>ddebug: simplify watchdog loop and fix crash in the no-timeout case</li> 1929 <li>ddebug: always flush when requested, even when hang detection is disabled</li> 1930 <li>r600: remove redundant semicolon</li> 1931 <li>amd/sid_tables: add additional python3 compatibility imports</li> 1932 <li>amd/common: whitespace fixes</li> 1933 <li>amd/common: add ac_build_ifcc</li> 1934 <li>amd/common: scan/reduce across waves of a workgroup</li> 1935 <li>amd/common: add i1 special case to ac_build_{inclusive,exclusive}_scan</li> 1936 <li>ac/surface: 3D and cube surfaces are never displayable</li> 1937 <li>radeonsi: move SI_FORCE_FAMILY functionality to winsys</li> 1938 <li>radeonsi: extract declare_vs_blit_inputs</li> 1939 <li>radeonsi: add si_init_draw_functions and make some functions static</li> 1940 <li>radeonsi/gfx9: use SET_UCONFIG_REG_INDEX packets when available</li> 1941 <li>radeonsi: don't set RAW_WAIT for CP DMA clears</li> 1942 <li>radeonsi: rename SI_RESOURCE_FLAG_FORCE_TILING to clarify its purpose</li> 1943 <li>radeonsi: const-ify si_set_tesseval_regs</li> 1944 <li>radeonsi: show the fixed function TCS in debug dumps</li> 1945 <li>radeonsi: avoid using hard-coded SI_NUM_RW_BUFFERS</li> 1946 <li>radeonsi: add an si_set_rw_shader_buffer convenience function</li> 1947 <li>radeonsi: use si_set_rw_shader_buffer for setting streamout buffers</li> 1948 <li>radeonsi: track constant buffer bind history in si_pipe_set_constant_buffer</li> 1949 <li>radeonsi: move remaining perfcounter code into si_perfcounter.c</li> 1950 <li>radeonsi: move query suspend logic into the top-level si_query struct</li> 1951 <li>radeonsi: factor si_query_buffer logic out of si_query_hw</li> 1952 <li>radeonsi: split perfcounter queries from si_query_hw</li> 1953 <li>radeonsi: const-ify the si_query_ops</li> 1954 <li>amd/common: use llvm.amdgcn.s.buffer.load for LLVM 8.0</li> 1955 <li>amd/common/vi+: enable SMEM loads with GLC=1</li> 1956</ul> 1957 1958<p>Niklas Haas (3):</p> 1959<ul> 1960 <li>glsl: fix block member alignment validation for vec3</li> 1961 <li>radv: correctly use vulkan 1.0 by default</li> 1962 <li>radv: add device->instance extension dependencies</li> 1963</ul> 1964 1965<p>Olivier Fourdan (1):</p> 1966<ul> 1967 <li>wayland/egl: Resize EGL surface on update buffer for swrast</li> 1968</ul> 1969 1970<p>Oscar Blumberg (1):</p> 1971<ul> 1972 <li>radeonsi: Fix guardband computation for large render targets</li> 1973</ul> 1974 1975<p>Pierre Moreau (2):</p> 1976<ul> 1977 <li>clover/meson: Ignore 'svn' suffix when computing CLANG_RESOURCE_DIR</li> 1978 <li>meson: Fix with_gallium_icd to with_opencl_icd</li> 1979</ul> 1980 1981<p>Plamena Manolova (1):</p> 1982<ul> 1983 <li>nir: Don't lower the local work group size if it's variable.</li> 1984</ul> 1985 1986<p>Rafael Antognolli (24):</p> 1987<ul> 1988 <li>intel/genxml: Add register for object preemption.</li> 1989 <li>i965/gen10+: Enable object level preemption.</li> 1990 <li>i965/gen9: Add workarounds for object preemption.</li> 1991 <li>anv/tests: Fix block_pool_no_free test.</li> 1992 <li>anv/allocator: Add anv_state_table.</li> 1993 <li>anv/allocator: Add getter for anv_block_pool.</li> 1994 <li>anv/allocator: Add helper to push states back to the state table.</li> 1995 <li>anv/allocator: Use anv_state_table on anv_state_pool_alloc.</li> 1996 <li>anv/allocator: Use anv_state_table on back_alloc too.</li> 1997 <li>anv/allocator: Remove anv_free_list.</li> 1998 <li>anv/allocator: Rename anv_free_list2 to anv_free_list.</li> 1999 <li>anv/allocator: Remove pool->map.</li> 2000 <li>anv: Update usage of block_pool->bo.</li> 2001 <li>anv/allocator: Add support for a list of BOs in block pool.</li> 2002 <li>anv: Split code to add BO dependencies to execbuf.</li> 2003 <li>anv: Validate the list of BOs from the block pool.</li> 2004 <li>anv: Remove some asserts.</li> 2005 <li>anv/allocator: Rework chunk return to the state pool.</li> 2006 <li>anv/allocator: Add padding information.</li> 2007 <li>anv/allocator: Enable snooping on block pool and anv_bo_pool BOs.</li> 2008 <li>anv: Remove state flush.</li> 2009 <li>anv/allocator: Add support for non-userptr.</li> 2010 <li>anv/tests: Adding test for the state_pool padding.</li> 2011 <li>anv/allocator: Avoid race condition in anv_block_pool_map.</li> 2012</ul> 2013 2014<p>Ray Zhang (1):</p> 2015<ul> 2016 <li>glx: fix shared memory leak in X11</li> 2017</ul> 2018 2019<p>Rhys Kidd (5):</p> 2020<ul> 2021 <li>travis: radeonsi and radv require LLVM 7.0</li> 2022 <li>meson: libfreedreno depends upon libdrm (for fence support)</li> 2023 <li>v3d: Wire up core pipe_debug_callback</li> 2024 <li>vc4: Wire up core pipe_debug_callback</li> 2025 <li>nv50,nvc0: add missing CAPs for unsupported features</li> 2026</ul> 2027 2028<p>Rhys Perry (14):</p> 2029<ul> 2030 <li>nir: fix constness in nir_intrinsic_align()</li> 2031 <li>ac: refactor visit_load_buffer</li> 2032 <li>ac: split 16-bit ssbo loads that may not be dword aligned</li> 2033 <li>radv: don't set surf_index for stencil-only images</li> 2034 <li>radv: switch from nir_bcsel to nir_b32csel</li> 2035 <li>ac/nir,radv,radeonsi/nir: use correct indices for interpolation intrinsics</li> 2036 <li>nir: fix copy-paste error in nir_lower_constant_initializers</li> 2037 <li>radv: use dithered alpha-to-coverage</li> 2038 <li>radv: pass radv_draw_info to radv_emit_draw_registers()</li> 2039 <li>radv: add missed situations for scissor bug workaround</li> 2040 <li>radv: avoid context rolls when binding graphics pipelines</li> 2041 <li>radv: prevent dirtying of dynamic state when it does not change</li> 2042 <li>radv: bitcast 16-bit outputs to integers</li> 2043 <li>radv: ensure export arguments are always float</li> 2044</ul> 2045 2046<p>Rob Clark (79):</p> 2047<ul> 2048 <li>freedreno: update generated headers</li> 2049 <li>freedreno/a6xx: fix VSC bug with larger # of tiles</li> 2050 <li>freedreno/drm: fix unused 'entry' warnings</li> 2051 <li>freedreno/drm: remove dependency on gallium driver</li> 2052 <li>freedreno: move drm to common location</li> 2053 <li>freedreno/ir3: standalone compiler updates</li> 2054 <li>freedreno: shader_t -> gl_shader_stage</li> 2055 <li>freedreno: remove shader_stage_name()</li> 2056 <li>freedreno: FD_SHADER_DEBUG -> IR3_SHADER_DEBUG</li> 2057 <li>freedreno/ir3: move disasm and optmsgs debug flags</li> 2058 <li>util: env_var_as_unsigned() helper</li> 2059 <li>freedreno/ir3: use env_var_as_unsigned()</li> 2060 <li>freedreno/ir3: some header file cleanup</li> 2061 <li>freedreno/ir3: remove pipe_stream_output_info dependency</li> 2062 <li>freedreno/ir3: split up ir3_shader</li> 2063 <li>freedreno/ir3: remove u_inlines usage</li> 2064 <li>freedreno: move ir3 to common location</li> 2065 <li>mesa/st: swap order of clear() and clear_with_quad()</li> 2066 <li>mesa/st: better colormask check for clear fallback</li> 2067 <li>freedreno/a6xx: disable LRZ for z32</li> 2068 <li>freedreno/a6xx: set guardband clip</li> 2069 <li>freedreno: update generated headers</li> 2070 <li>freedreno/a3xx: also set FSSUPERTHREADENABLE</li> 2071 <li>freedreno/a6xx: MSAA</li> 2072 <li>freedreno: remove unused fd_surface fields</li> 2073 <li>gallium: fix typo</li> 2074 <li>freedreno/a5xx+a6xx: remove unused fs/vs pvt mem</li> 2075 <li>freedreno/drm: fix relocs in nested stateobjs</li> 2076 <li>freedreno: update generated headers</li> 2077 <li>freedreno/a6xx: blitter fixes</li> 2078 <li>freedreno/ir3: don't fetch unused tex components</li> 2079 <li>freedreno/ir3: sync instr/disasm</li> 2080 <li>freedreno/ir3: code-motion</li> 2081 <li>freedreno/ir3: track max flow control depth for a5xx/a6xx</li> 2082 <li>freedreno/drm: fix memory leak</li> 2083 <li>nir: fix spelling typo</li> 2084 <li>mesa/st/nir: fix missing nir_compact_varyings</li> 2085 <li>freedreno/drm: sync uapi and enable softpin</li> 2086 <li>freedreno: debug GEM obj names</li> 2087 <li>freedreno: also set DUMP flag on shaders</li> 2088 <li>freedreno/ir3: fix crash</li> 2089 <li>freedreno/ir3: don't remove unused input components</li> 2090 <li>freedreno/a6xx: fix blitter crash</li> 2091 <li>gallium/aux: add is_unorm() helper</li> 2092 <li>freedreno: update generated headers</li> 2093 <li>freedreno/a6xx: more blitter fixes</li> 2094 <li>freedreno: move fd_resource_copy_region()</li> 2095 <li>freedreno/a6xx: fix resource_copy_region()</li> 2096 <li>freedreno/a6xx: fix corrupted uniforms</li> 2097 <li>freedreno/ir3: fix fallout of extra assert</li> 2098 <li>freedreno/ir3: don't treat all inputs/outputs as vec4</li> 2099 <li>freedreno: combine fd_resource_layer_offset()/fd_resource_offset()</li> 2100 <li>freedreno/a6xx: simplify special case for 3d layout</li> 2101 <li>freedreno/a6xx: improve setup_slices() debug msgs</li> 2102 <li>freedreno: update generated headers</li> 2103 <li>freedreno/a6xx: fix 3d texture layout</li> 2104 <li>freedreno: skip depth resolve if not written</li> 2105 <li>freedreno: rework blit API</li> 2106 <li>freedreno: try blitter for fd_resource_copy_region()</li> 2107 <li>freedreno/a6xx: rework blitter API</li> 2108 <li>freedreno: remove blit_via_copy_region()</li> 2109 <li>freedreno: fix staging resource size for arrays</li> 2110 <li>freedreno: make cmdstream bo's read-only to GPU</li> 2111 <li>freedreno/a6xx: separate stencil restore/resolve fixes</li> 2112 <li>freedreno/a6xx: move tile_mode to sampler-view CSO</li> 2113 <li>freedreno/a6xx: fix 3d+tiled layout</li> 2114 <li>nir/vtn: add caps for some cl related capabilities</li> 2115 <li>loader: fix the no-modifiers case</li> 2116 <li>freedreno: core buffer modifier support</li> 2117 <li>freedreno: set modifier when exporting buffer</li> 2118 <li>freedreno: limit tiling to PIPE_BIND_SAMPLER_VIEW</li> 2119 <li>freedreno/a2xx: fix unused variable warning</li> 2120 <li>freedreno/a5xx: fix blitter nr_samples check</li> 2121 <li>freedreno/a6xx: fix blitter nr_samples check</li> 2122 <li>freedreno: stop frob'ing pipe_resource::nr_samples</li> 2123 <li>freedreno: minor cleanups</li> 2124 <li>mesa: wire up InvalidateFramebuffer</li> 2125 <li>freedreno: fix release tarball</li> 2126 <li>freedreno: more fixing release tarball</li> 2127</ul> 2128 2129<p>Rob Herring (3):</p> 2130<ul> 2131 <li>pipe-loader: Fallback to kmsro driver when no matching driver name found</li> 2132 <li>kmsro: Add etnaviv renderonly support</li> 2133 <li>Switch imx to kmsro and remove the imx winsys</li> 2134</ul> 2135 2136<p>Robert Foss (3):</p> 2137<ul> 2138 <li>virgl: native fence fd support</li> 2139 <li>virgl: Clean up fences commit</li> 2140 <li>virgl: add assert and missing function parameter</li> 2141</ul> 2142 2143<p>Rodrigo Vivi (1):</p> 2144<ul> 2145 <li>intel: Add more PCI Device IDs for Coffee Lake and Ice Lake.</li> 2146</ul> 2147 2148<p>Roland Scheidegger (7):</p> 2149<ul> 2150 <li>gallivm: fix improper clamping of vertex index when fetching gs inputs</li> 2151 <li>draw: fix infinite loop in line stippling</li> 2152 <li>gallivm: remove unused float coord wrapping for aos sampling</li> 2153 <li>gallivm: use llvm jit code for decoding s3tc</li> 2154 <li>gallivm: don't use pavg.b intrinsic on llvm >= 6.0</li> 2155 <li>gallivm: abort when trying to use non-existing intrinsic</li> 2156 <li>Revert "llvmpipe: Always return some fence in flush (v2)"</li> 2157</ul> 2158 2159<p>Sagar Ghuge (14):</p> 2160<ul> 2161 <li>intel/compiler: Disassemble GEN6_SFID_DATAPORT_SAMPLER_CACHE as dp_sampler</li> 2162 <li>intel/compiler: Set swizzle to BRW_SWIZZLE_XXXX for scalar region</li> 2163 <li>intel/compiler: Always print flag subregister number</li> 2164 <li>nir: Add a new lowering option to lower 3D surfaces from txd to txl.</li> 2165 <li>glsl: Add "built-in" functions to do uint64_to_fp64(uint64_t)</li> 2166 <li>glsl: Add "built-in" functions to do int64_to_fp64(int64_t)</li> 2167 <li>glsl: Add "built-in" functions to do uint64_to_fp32(uint64_t)</li> 2168 <li>glsl: Add "built-in" functions to do int64_to_fp32(int64_t)</li> 2169 <li>glsl: Add utility function to round and pack uint64_t value</li> 2170 <li>glsl: Add "built-in" functions to do fp64_to_uint64(fp64)</li> 2171 <li>glsl: Add utility function to round and pack int64_t value</li> 2172 <li>glsl: Add "built-in" functions to do fp64_to_int64(fp64)</li> 2173 <li>glsl: Add "built-in" functions to do fp32_to_uint64(fp32)</li> 2174 <li>glsl: Add "built-in" functions to do fp32_to_int64(fp32)</li> 2175</ul> 2176 2177<p>Samuel Pitoiset (103):</p> 2178<ul> 2179 <li>radv: remove useless sync after copying query results with compute</li> 2180 <li>radv: add missing TFB queries support to CmdCopyQueryPoolsResults()</li> 2181 <li>radv: replace si_emit_wait_fence() with radv_cp_wait_mem()</li> 2182 <li>radv: more use of radv_cp_wait_mem()</li> 2183 <li>radv: allocate enough space in CS when copying query results with compute</li> 2184 <li>radv: disable conditional rendering for vkCmdCopyQueryPoolResults()</li> 2185 <li>radv: only expose VK_SUBGROUP_FEATURE_ARITHMETIC_BIT for VI+</li> 2186 <li>radv: use LOAD_CONTEXT_REG when loading fast clear values</li> 2187 <li>radv: fix GPU hangs when loading depth/stencil clear values on SI/CIK</li> 2188 <li>radv: cleanup and document a Hawaii bug with offchip buffers</li> 2189 <li>radv: clean up setting partial_es_wave for distributed tess on VI</li> 2190 <li>radv: make use of num_good_cu_per_sh in si_emit_graphics() too</li> 2191 <li>radv: binding streamout buffers doesn't change context regs</li> 2192 <li>radv: set PA.SC_CONSERVATIVE_RASTERIZATION.NULL_SQUAD_AA_MASK_ENABLE</li> 2193 <li>radv: set optimal OVERWRITE_COMBINER_WATERMARK on GFX9</li> 2194 <li>radv: add a debug option for disabling primitive binning</li> 2195 <li>radv: enable primitive binning by default</li> 2196 <li>radv: tidy up radv_set_dcc_need_cmask_elim_pred()</li> 2197 <li>radv: always clear the FCE predicate after DCC/FMASK/CMASK decompressions</li> 2198 <li>radv/winsys: remove the max IBs per submit limit for the fallback path</li> 2199 <li>radv/winsys: remove the max IBs per submit limit for the sysmem path</li> 2200 <li>radv: remove unnecessary goto in the fast clear paths</li> 2201 <li>radv: add radv_get_htile_fast_clear_value() helper</li> 2202 <li>radv: add radv_is_fast_clear_{depth,stencil}_allowed() helpers</li> 2203 <li>radv: check allowed fast HTILE clears a bit earlier</li> 2204 <li>radv: rewrite the condition that checks allowed depth/stencil values</li> 2205 <li>radv: implement fast HTILE clears for depth or stencil only on GFX9</li> 2206 <li>ac/nir: fix intrinsic name string size in visit_image_atomic()</li> 2207 <li>radv: ignore subpass self-dependencies</li> 2208 <li>radv: only sync CP DMA for transfer operations or bottom pipe</li> 2209 <li>radv: remove useless sync after CmdClear{Color,DepthStencil}Image()</li> 2210 <li>radv: remove useless sync before CmdClear{Color,DepthStencil}Image()</li> 2211 <li>radv: ignore subpass self-dependencies for CreateRenderPass() too</li> 2212 <li>radv: remove useless check in emit_fast_color_clear()</li> 2213 <li>radv: add radv_image_can_fast_clear() helper</li> 2214 <li>radv: add radv_image_view_can_fast_clear() helper</li> 2215 <li>radv: add radv_can_fast_clear_{color,depth}() helpers</li> 2216 <li>radv: simplify a check in emit_fast_color_clear()</li> 2217 <li>radv: refactor the fast clear path for better re-use</li> 2218 <li>radv: optimize CmdClear{Color,DepthStencil}Image() for layered textures</li> 2219 <li>radv: remove unused pending_clears param in the transition path</li> 2220 <li>radv: drop few useless state changes when doing color/depth decompressions</li> 2221 <li>radv: rework the TC-compat HTILE hardware bug with COND_EXEC</li> 2222 <li>radv: reset pending_reset_query when flushing caches</li> 2223 <li>radv: wait on the high 32 bits of timestamp queries</li> 2224 <li>spirv: add SpvCapabilityInt64Atomics</li> 2225 <li>radv: expose VK_EXT_scalar_block_layout</li> 2226 <li>amd: remove support for LLVM 6.0</li> 2227 <li>gallium: add missing PIPE_CAP_SURFACE_SAMPLE_COUNT default value</li> 2228 <li>radv: bump reported version to 1.1.90</li> 2229 <li>radv: add a predicate for reflecting DCC decompression state</li> 2230 <li>radv: allow to skip DCC decompressions with the new predicate</li> 2231 <li>radv: switch on EOP when primitive restart is enabled with triangle strips</li> 2232 <li>radv: check if addrlib enabled HTILE in radv_image_can_enable_htile()</li> 2233 <li>radv: don't check if format is depth in radv_image_can_enable_hile()</li> 2234 <li>radv: report Vulkan version 1.1.90 for real</li> 2235 <li>ac/nir: remove the bitfield_extract workaround for LLVM 8</li> 2236 <li>radv: drop the amdgpu-skip-threshold=1 workaround for LLVM 8</li> 2237 <li>radv: fix subpass image transitions with multiviews</li> 2238 <li>radv: compute optimal VM alignment for imported buffers</li> 2239 <li>spirv: add support for SpvCapabilityStorageImageMultisample</li> 2240 <li>ac/nir: restrict fmask lookup to image load intrinsics</li> 2241 <li>radv: initialize FMASK for images in fully expanded mode</li> 2242 <li>radv: add support for FMASK expand</li> 2243 <li>radv: enable shaderStorageImageMultisample feature on GFX8+</li> 2244 <li>radv: get rid of bunch of KHR suffixes</li> 2245 <li>radv: enable variable pointers</li> 2246 <li>radv: skip draws with instance_count == 0</li> 2247 <li>ac/nir: add get_cache_policy() helper and use it</li> 2248 <li>ac/nir: set cache policy when loading/storing buffer images</li> 2249 <li>ac: add missing 16-bit types to glsl_base_to_llvm_type()</li> 2250 <li>radv: remove unnecessary returns in GetPhysicalDevice*Properties()</li> 2251 <li>radv: add two small helpers for getting VRAM and visible VRAM sizes</li> 2252 <li>radv: add support for VK_EXT_memory_budget</li> 2253 <li>ac/nir: don't trash L1 caches for store operations with writeonly memory</li> 2254 <li>radv: drop unused code related to 16 sample locations</li> 2255 <li>radv: reduce size of the per-queue descriptor BO</li> 2256 <li>radv: do not write unused descriptors to the per-queue BO</li> 2257 <li>radv: initialize the per-queue descriptor BO only once</li> 2258 <li>nir: do not remove varyings used for transform feedback</li> 2259 <li>nir: fix lowering arrays to elements for XFB outputs</li> 2260 <li>radv: improve gathering of load_push_constants with dynamic bindings</li> 2261 <li>radv: remove old_fence parameter from si_cs_emit_write_event_eop()</li> 2262 <li>radv: only allocate the GFX9 fence and EOP BOs for the gfx queue</li> 2263 <li>radv: compute the GFX9 fence VA at allocation time</li> 2264 <li>radv: always pass the GFX9 fence data to si_cs_emit_cache_flush()</li> 2265 <li>radv: fix computing number of user SGPRs for streamout buffers</li> 2266 <li>radv: remove radv_userdata_info::indirect field</li> 2267 <li>radv: simplify allocating user SGPRS for descriptor sets</li> 2268 <li>radv: set noalias/dereferenceable LLVM attributes based on param types</li> 2269 <li>radv: re-enable fast depth clears for 16-bit surfaces on VI</li> 2270 <li>radv/winsys: fix hash when adding internal buffers</li> 2271 <li>radv: fix compiler issues with GCC 9</li> 2272 <li>radv: fix using LOAD_CONTEXT_REG with old GFX ME firmwares on GFX8</li> 2273 <li>radv/winsys: fix BO list creation when RADV_DEBUG=allbos is set</li> 2274 <li>radv: always export gl_SampleMask when the fragment shader uses it</li> 2275 <li>radv: write the alpha channel of MRT0 when alpha coverage is enabled</li> 2276 <li>radv: fix writing the alpha channel of MRT0 when alpha coverage is enabled</li> 2277 <li>radv: fix out-of-bounds access when copying descriptors BO list</li> 2278 <li>radv: don't copy buffer descriptors list for samplers</li> 2279 <li>radv: fix clearing attachments in secondary command buffers</li> 2280 <li>radv: properly align the fence and EOP bug VA on GFX9</li> 2281 <li>radv: fix pointSizeRange limits</li> 2282</ul> 2283 2284<p>Sergii Romantsov (4):</p> 2285<ul> 2286 <li>autotools: library-dependency when no sse and 32-bit</li> 2287 <li>i965/batch/debug: Allow log be dumped before assert</li> 2288 <li>nir: Length of boolean vtn_value now is 1</li> 2289 <li>dri: meson: do not prefix user provided dri-drivers-path</li> 2290</ul> 2291 2292<p>Sonny Jiang (1):</p> 2293<ul> 2294 <li>radeonsi: use compute for resource_copy_region when possible</li> 2295</ul> 2296 2297<p>Tapani Pälli (27):</p> 2298<ul> 2299 <li>anv: allow exporting an imported SYNC_FD semaphore type</li> 2300 <li>anv: add create_flags as part of anv_image</li> 2301 <li>anv: refactor make_surface to use data from anv_image</li> 2302 <li>anv: make anv_get_image_format_features public</li> 2303 <li>anv: add from/to helpers with android and vulkan formats</li> 2304 <li>anv/android: add GetAndroidHardwareBufferPropertiesANDROID</li> 2305 <li>anv: add anv_ahw_usage_from_vk_usage helper function</li> 2306 <li>anv: refactor, remove else block in AllocateMemory</li> 2307 <li>anv/android: support import/export of AHardwareBuffer objects</li> 2308 <li>anv/android: add ahardwarebuffer external memory properties</li> 2309 <li>anv/android: support creating images from external format</li> 2310 <li>anv: support VkExternalFormatANDROID in vkCreateSamplerYcbcrConversion</li> 2311 <li>anv: add VkFormat field as part of anv_format</li> 2312 <li>anv: support VkSamplerYcbcrConversionInfo in vkCreateImageView</li> 2313 <li>anv: ignore VkSamplerYcbcrConversion on non-yuv formats</li> 2314 <li>anv/android: turn on VK_ANDROID_external_memory_android_hardware_buffer</li> 2315 <li>dri3: initialize adaptive_sync as false before configQueryb</li> 2316 <li>intel/isl: move tiled_memcpy static libs from i965 to isl</li> 2317 <li>anv: do not advertise AHW support if extension not enabled</li> 2318 <li>nir: cleanup glsl_get_struct_field_offset, glsl_get_explicit_stride</li> 2319 <li>android: fix build issues with libmesa_anv_gen* libraries</li> 2320 <li>mesa: return NULL if we exceed MaxColorAttachments in get_fb_attachment</li> 2321 <li>nir: initialize value in copy_prop_vars_block</li> 2322 <li>anv: retain the is_array state in create_plane_tex_instr_implicit</li> 2323 <li>anv: release memory allocated by glsl types during spirv_to_nir</li> 2324 <li>anv: revert "anv: release memory allocated by glsl types during spirv_to_nir"</li> 2325 <li>anv: destroy descriptor sets when pool gets destroyed</li> 2326</ul> 2327 2328<p>Thomas Hellstrom (9):</p> 2329<ul> 2330 <li>st/xa: Render update. Better support for solid pictures</li> 2331 <li>st/xa: Support higher color precision for solid pictures</li> 2332 <li>st/xa: Support a couple of new formats</li> 2333 <li>st/xa: Fix transformations when we have both source and mask samplers</li> 2334 <li>st/xa: Minor renderer cleanups</li> 2335 <li>st/xa: Support Component Alpha with trivial blending</li> 2336 <li>st/xa: Bump minor</li> 2337 <li>st/xa: Fix a memory leak</li> 2338 <li>winsys/svga: Fix a memory leak</li> 2339</ul> 2340 2341<p>Timothy Arceri (56):</p> 2342<ul> 2343 <li>nir: allow propagation of if evaluation for bcsel</li> 2344 <li>nir: fix condition propagation when src has a swizzle</li> 2345 <li>ac/nir_to_llvm: fix b2f for f64</li> 2346 <li>nir: add new linking opt nir_link_constant_varyings()</li> 2347 <li>st/mesa: make use of nir_link_constant_varyings()</li> 2348 <li>nir: add glsl_type_is_integer() helper</li> 2349 <li>nir: don't pack varyings ints with floats unless flat</li> 2350 <li>anv/i965: make use of nir_link_constant_varyings()</li> 2351 <li>nir: add support for removing redundant stores to copy prop var</li> 2352 <li>radv: make use of nir_move_out_const_to_consumer()</li> 2353 <li>nir: small tidy ups for nir_loop_analyze()</li> 2354 <li>nir: clarify some nit_loop_info member names</li> 2355 <li>nir: add a new nir_cf_list_clone_and_reinsert() helper</li> 2356 <li>nir: make use of new nir_cf_list_clone_and_reinsert() helper</li> 2357 <li>nir: factor out some of the complex loop unroll code to a helper</li> 2358 <li>nir: rework force_unroll_array_access()</li> 2359 <li>nir: in loop analysis track actual control flow type</li> 2360 <li>nir: reword code comment</li> 2361 <li>nir: detect more induction variables</li> 2362 <li>nir: fix opt_if_loop_last_continue()</li> 2363 <li>tgsi/scan: fix loop exit point in tgsi_scan_tess_ctrl()</li> 2364 <li>tgsi/scan: correctly walk instructions in tgsi_scan_tess_ctrl()</li> 2365 <li>radeonsi: remove unrequired param in si_nir_scan_tess_ctrl()</li> 2366 <li>ac/nir_to_llvm: add ac_are_tessfactors_def_in_all_invocs()</li> 2367 <li>radeonsi: make use of ac_are_tessfactors_def_in_all_invocs()</li> 2368 <li>st/glsl_to_nir: call nir_lower_load_const_to_scalar() in the st</li> 2369 <li>nir: rename nir_link_constant_varyings() nir_link_opt_varyings()</li> 2370 <li>nir: add can_replace_varying() helper</li> 2371 <li>nir: rework nir_link_opt_varyings()</li> 2372 <li>nir: link time opt duplicate varyings</li> 2373 <li>nir: make nir_opt_remove_phis_impl() static</li> 2374 <li>nir: make use of does_varying_match() helper</li> 2375 <li>nir: simplify does_varying_match()</li> 2376 <li>nir: add rewrite_phi_predecessor_blocks() helper</li> 2377 <li>nir: merge some basic consecutive ifs</li> 2378 <li>st/glsl: refactor st_link_nir()</li> 2379 <li>nir: avoid uninitialized variable warning</li> 2380 <li>glsl: Copy function out to temp if we don't directly ref a variable</li> 2381 <li>ac/nir_to_llvm: fix type handling in image code</li> 2382 <li>radeonsi/nir: get correct type for images inside structs</li> 2383 <li>ac/nir_to_llvm: fix regression in bindless support</li> 2384 <li>ac/nir_to_llvm: add support for structs to get_sampler_desc()</li> 2385 <li>glsl: don't skip GLSL IR opts on first-time compiles</li> 2386 <li>glsl: be much more aggressive when skipping shader compilation</li> 2387 <li>Revert "glsl: be much more aggressive when skipping shader compilation"</li> 2388 <li>ac/nir_to_llvm: fix interpolateAt* for arrays</li> 2389 <li>glsl: be much more aggressive when skipping shader compilation</li> 2390 <li>radeonsi/nir: add missing piece for bindless image support</li> 2391 <li>ac/nir_to_llvm: add bindless support for uniform handles</li> 2392 <li>ac/nir_to_llvm: fix interpolateAt* for structs</li> 2393 <li>ac/nir_to_llvm: fix clamp shadow reference for more hardware</li> 2394 <li>tgsi: remove culldist semantic from docs</li> 2395 <li>radv/ac: fix some fp16 handling</li> 2396 <li>glsl: use remap location when serialising uniform program resource data</li> 2397 <li>radeonsi: fix query buffer allocation</li> 2398 <li>glsl: fix shader cache for packed param list</li> 2399</ul> 2400 2401<p>Tobias Klausmann (1):</p> 2402<ul> 2403 <li>amd/vulkan: meson build - use radv_deps for libvulkan_radeon</li> 2404</ul> 2405 2406<p>Tomasz Figa (1):</p> 2407<ul> 2408 <li>llvmpipe: Always return some fence in flush (v2)</li> 2409</ul> 2410 2411<p>Tomeu Vizoso (1):</p> 2412<ul> 2413 <li>etnaviv: Consolidate buffer references from framebuffers</li> 2414</ul> 2415 2416<p>Toni Lönnberg (14):</p> 2417<ul> 2418 <li>intel/decoder: Engine parameter for instructions</li> 2419 <li>intel/decoder: tools: gen_engine to drm_i915_gem_engine_class</li> 2420 <li>intel/decoder: tools: Use engine for decoding batch instructions</li> 2421 <li>intel/genxml: Add engine definition to render engine instructions (gen4)</li> 2422 <li>intel/genxml: Add engine definition to render engine instructions (gen45)</li> 2423 <li>intel/genxml: Add engine definition to render engine instructions (gen5)</li> 2424 <li>intel/genxml: Add engine definition to render engine instructions (gen6)</li> 2425 <li>intel/genxml: Add engine definition to render engine instructions (gen7)</li> 2426 <li>intel/genxml: Add engine definition to render engine instructions (gen75)</li> 2427 <li>intel/genxml: Add engine definition to render engine instructions (gen8)</li> 2428 <li>intel/genxml: Add engine definition to render engine instructions (gen9)</li> 2429 <li>intel/genxml: Add engine definition to render engine instructions (gen10)</li> 2430 <li>intel/genxml: Add engine definition to render engine instructions (gen11)</li> 2431 <li>intel/aubinator_error_decode: Get rid of warning for missing switch case</li> 2432</ul> 2433 2434<p>Topi Pohjolainen (1):</p> 2435<ul> 2436 <li>i965/icl: Disable prefetching of sampler state entries</li> 2437</ul> 2438 2439<p>Veluri Mithun (5):</p> 2440<ul> 2441 <li>Add extension doc for MESA_query_driver</li> 2442 <li>Implement EGL API for MESA_query_driver</li> 2443 <li>Implementation of egl dri2 drivers for MESA_query_driver</li> 2444 <li>egl: Implement EGL API for MESA_query_driver</li> 2445 <li>egl: Implementation of egl dri2 drivers for MESA_query_driver</li> 2446</ul> 2447 2448<p>Vinson Lee (7):</p> 2449<ul> 2450 <li>r600/sb: Fix constant logical operand in assert.</li> 2451 <li>freedreno: Fix autotools build.</li> 2452 <li>st/xvmc: Add X11 include path.</li> 2453 <li>nir/algebraic: Make algebraic_parser_test.sh executable.</li> 2454 <li>meson: Fix typo.</li> 2455 <li>meson: Fix libsensors detection.</li> 2456 <li>meson: Fix typo.</li> 2457</ul> 2458 2459<p>Yevhenii Kolesnikov (1):</p> 2460<ul> 2461 <li>i965: Fix allow_higher_compat_version workaround limited by OpenGL 3.0</li> 2462</ul> 2463 2464<p>pal1000 (1):</p> 2465<ul> 2466 <li>scons: Compatibility with Scons development version string</li> 2467</ul> 2468 2469</div> 2470</body> 2471</html> 2472