2019-05-07 17:07:16 +01:00
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
< html lang = "en" >
< head >
< meta http-equiv = "content-type" content = "text/html; charset=utf-8" >
< title > Mesa Release Notes< / title >
< link rel = "stylesheet" type = "text/css" href = "../mesa.css" >
< / head >
< body >
< div class = "header" >
< h1 > The Mesa 3D Graphics Library< / h1 >
< / div >
< iframe src = "../contents.html" > < / iframe >
< div class = "content" >
2019-09-25 18:40:00 +01:00
< h1 > Mesa 19.2.0 Release Notes / 2019.09.25< / h1 >
2019-05-07 17:07:16 +01:00
< p >
Mesa 19.2.0 is a new development release. People who are concerned
with stability and reliability should stick with a previous release or
wait for Mesa 19.2.1.
< / p >
< p >
Mesa 19.2.0 implements the OpenGL 4.5 API, but the version reported by
glGetString(GL_VERSION) or glGetIntegerv(GL_MAJOR_VERSION) /
glGetIntegerv(GL_MINOR_VERSION) depends on the particular driver being used.
Some drivers don't support all the features required in OpenGL 4.5. OpenGL
4.5 is < strong > only< / strong > available if requested at context creation.
Compatibility contexts may report a lower version depending on each driver.
< / p >
2019-09-18 10:44:13 +01:00
< p >
Mesa 19.2.0 implements the Vulkan 1.1 API, but the version reported by
the apiVersion property of the VkPhysicalDeviceProperties struct
depends on the particular driver being used.
< / p >
2019-05-07 17:07:16 +01:00
< h2 > SHA256 checksums< / h2 >
< pre >
2019-09-25 18:39:51 +01:00
b060caa2a00f856431160ff7377d0e8f58f2aa48c16ee5a9e265ebdccb10852a mesa-19.2.0.tar.xz
2019-05-07 17:07:16 +01:00
< / pre >
< h2 > New features< / h2 >
< ul >
2019-07-05 22:30:08 +01:00
< li > GL_ARB_post_depth_coverage on radeonsi (Navi)< / li >
2019-08-14 08:37:35 +01:00
< li > GL_ARB_seamless_cubemap_per_texture on etnaviv (if GPU supports SEAMLESS_CUBE_MAP)< / li >
2019-08-01 09:17:26 +01:00
< li > GL_EXT_shader_image_load_store on radeonsi (with LLVM >= 10)< / li >
2019-07-31 23:47:34 +01:00
< li > GL_EXT_shader_samples_identical on iris and radeonsi (if using NIR)< / li >
2019-08-07 01:09:46 +01:00
< li > GL_EXT_texture_shadow_lod on i965, iris< / li >
2019-05-16 18:01:40 +01:00
< li > EGL_EXT_platform_device< / li >
2019-08-02 22:05:43 +01:00
< li > VK_AMD_buffer_marker on radv< / li >
< li > VK_EXT_index_type_uint8 on radv< / li >
< li > VK_EXT_post_depth_coverage on radv< / li >
2019-08-02 22:06:37 +01:00
< li > VK_EXT_queue_family_foreign on radv< / li >
2019-08-02 22:05:43 +01:00
< li > VK_EXT_sample_locations on radv< / li >
2019-06-08 01:37:38 +01:00
< li > VK_EXT_shader_demote_to_helper_invocation on Intel.< / li >
2019-08-02 22:05:43 +01:00
< li > VK_KHR_depth_stencil_resolve on radv< / li >
2019-08-02 22:06:37 +01:00
< li > VK_KHR_imageless_framebuffer on radv< / li >
< li > VK_KHR_shader_atomic_int64 on radv< / li >
2019-08-02 22:05:43 +01:00
< li > VK_KHR_uniform_buffer_standard_layout on radv< / li >
2019-05-07 17:07:16 +01:00
< / ul >
< h2 > Bug fixes< / h2 >
< ul >
2019-09-25 17:55:33 +01:00
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=103674" > Bug 103674< / a > - u_queue.c:173:7: error: implicit declaration of function 'timespec_get' is invalid in C99< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=104395" > Bug 104395< / a > - [CTS] GTF-GL46.gtf32.GL3Tests.packed_pixels.packed_pixels tests fail on 32bit Mesa< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=110765" > Bug 110765< / a > - ANV regression: Assertion `pass-> attachment_count == framebuffer-> attachment_count' failed< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=110814" > Bug 110814< / a > - KWin compositor crashes on launch< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=111069" > Bug 111069< / a > - Assertion fails in nir_opt_remove_phis.c during compilation of SPIR-V shader< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=111213" > Bug 111213< / a > - VA-API nouveau SIGSEGV and asserts< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=111241" > Bug 111241< / a > - Shadertoy shader causing hang< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=111248" > Bug 111248< / a > - Navi10 Font rendering issue in Overwatch< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=111271" > Bug 111271< / a > - Crash in eglMakeCurrent< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=111308" > Bug 111308< / a > - [Regression, NIR, bisected] Black squares in Unigine Heaven via DXVK< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=111401" > Bug 111401< / a > - Vulkan overlay layer - async compute not supported, making overlay disappear in Doom< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=111405" > Bug 111405< / a > - Some infinite 'do{}while' loops lead mesa to an infinite compilation< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=111411" > Bug 111411< / a > - SPIR-V shader leads to GPU hang, sometimes making machine unstable< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=111414" > Bug 111414< / a > - [REGRESSION] [BISECTED] Segmentation fault in si_bind_blend_state after removal of the blend state NULL check< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=111467" > Bug 111467< / a > - WOLF RPG Editor + Gallium Nine Standalone: Rendering issue when using Iris driver< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=111490" > Bug 111490< / a > - [REGRESSION] [BISECTED] Shadow Tactics: Blades of the Shogun - problems rendering water< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=111493" > Bug 111493< / a > - In the game The Surge (378540) - textures disappear then appear again when I change the camera angle view< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=111509" > Bug 111509< / a > - [regression][bisected] piglit.spec.ext_image_dma_buf_import.ext_image_dma_buf_import-export fails on iris< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=111522" > Bug 111522< / a > - [bisected] Supraland no longer start< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=111529" > Bug 111529< / a > - EGL_PLATFORM=drm doesn't expose MESA_query_driver extension< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=111552" > Bug 111552< / a > - Geekbench 5.0 Vulkan compute benchmark fails on Anvil< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=111566" > Bug 111566< / a > - [REGRESSION] [BISECTED] Large CS workgroup sizes broken in combination with FP64 on Intel.< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=111576" > Bug 111576< / a > - [bisected] Performance regression in X4:Foundations in 19.2< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=111676" > Bug 111676< / a > - Tropico 6 apitrace throws error into logs< / li >
< li > < a href = "https://bugs.freedesktop.org/show_bug.cgi?id=111734" > Bug 111734< / a > - Geometry shader with double interpolators fails in LLVM< / li >
2019-05-07 17:07:16 +01:00
< / ul >
< h2 > Changes< / h2 >
2019-09-25 17:55:33 +01:00
< p > Adam Jackson (1):< / p >
< ul >
< li > docs: Update bug report URLs for the gitlab migration< / li >
< / ul >
< p > Alex Smith (1):< / p >
< ul >
< li > radv: Change memory type order for GPUs without dedicated VRAM< / li >
< / ul >
< p > Alyssa Rosenzweig (1):< / p >
< ul >
< li > pan/midgard: Fix writeout combining< / li >
< / ul >
< p > Andres Gomez (1):< / p >
< ul >
< li > docs: Add the maximum implemented Vulkan API version in 19.2 rel notes< / li >
< / ul >
< p > Andres Rodriguez (1):< / p >
< ul >
< li > radv: additional query fixes< / li >
< / ul >
< p > Arcady Goldmints-Orlov (1):< / p >
< ul >
< li > anv: fix descriptor limits on gen8< / li >
< / ul >
< p > Bas Nieuwenhuizen (6):< / p >
< ul >
< li > radv: Use correct vgpr_comp_cnt for VS if both prim_id and instance_id are needed.< / li >
< li > radv: Emit VGT_GS_ONCHIP_CNTL for tess on GFX10.< / li >
< li > radv: Disable NGG for geometry shaders.< / li >
< li > Revert "ac/nir: Lower large indirect variables to scratch"< / li >
< li > tu: Set up glsl types.< / li >
< li > radv: Add workaround for hang in The Surge 2.< / li >
< / ul >
< p > Caio Marcelo de Oliveira Filho (2):< / p >
< ul >
< li > nir/lower_explicit_io: Handle 1 bit loads and stores< / li >
< li > glsl/nir: Avoid overflow when setting max_uniform_location< / li >
< / ul >
< p > Connor Abbott (1):< / p >
< ul >
< li > radv: Call nir_propagate_invariant()< / li >
< / ul >
< p > Danylo Piliaiev (3):< / p >
< ul >
< li > nir/loop_unroll: Prepare loop for unrolling in wrapper_unroll< / li >
< li > nir/loop_analyze: Treat do{}while(false) loops as 0 iterations< / li >
< li > tgsi_to_nir: Translate TGSI_INTERPOLATE_COLOR as INTERP_MODE_NONE< / li >
< / ul >
< p > Dave Airlie (2):< / p >
< ul >
< li > virgl: fix format conversion for recent gallium changes.< / li >
< li > gallivm: fix atomic compare-and-swap< / li >
< / ul >
< p > Dave Stevenson (1):< / p >
< ul >
< li > broadcom/v3d: Allow importing linear BOs with arbitrary offset/stride.< / li >
< / ul >
< p > Dylan Baker (9):< / p >
< ul >
< li > bump version to 19.2-rc2< / li >
< li > nir: Add is_not_negative helper function< / li >
< li > Bump version for rc3< / li >
< li > meson: don't generate file into subdirs< / li >
< li > add patches to be ignored< / li >
< li > Bump version for 19.2.0-rc4< / li >
< li > cherry-ignore: Add patches< / li >
< li > rehardcode from origin/master to upstream/master< / li >
< li > bin/get-pick-list: use --oneline=pretty instead of --oneline< / li >
< / ul >
< p > Emil Velikov (1):< / p >
< ul >
< li > Update version to 19.2.0-rc1< / li >
< / ul >
< p > Eric Engestrom (14):< / p >
< ul >
< li > ttn: fix 64-bit shift on 32-bit `1`< / li >
< li > egl: fix deadlock in malloc error path< / li >
< li > util/os_file: fix double-close()< / li >
< li > anv: fix format string in error message< / li >
< li > freedreno/drm-shim: fix mem leak< / li >
< li > nir: fix memleak in error path< / li >
< li > anv: add support for driconf< / li >
< li > wsi: add minImageCount override< / li >
< li > anv: add support for vk_x11_override_min_image_count< / li >
< li > amd: move adaptive sync to performance section, as it is defined in xmlpool< / li >
< li > radv: add support for vk_x11_override_min_image_count< / li >
< li > drirc: override minImageCount=2 for gfxbench< / li >
< li > gl: drop incorrect pkg-config file for glvnd< / li >
< li > meson: re-add incorrect pkg-config files with GLVND for backward compatibility< / li >
< / ul >
< p > Erik Faye-Lund (2):< / p >
< ul >
< li > gallium/auxiliary/indices: consistently apply start only to input< / li >
< li > util: fix SSE-version needed for double opcodes< / li >
< / ul >
< p > Haihao Xiang (1):< / p >
< ul >
< li > i965: support AYUV/XYUV for external import only< / li >
< / ul >
< p > Hal Gentz (2):< / p >
< ul >
< li > glx: Fix SEGV due to dereferencing a NULL ptr from XCB-GLX.< / li >
< li > gallium/osmesa: Fix the inability to set no context as current.< / li >
< / ul >
< p > Iago Toral Quiroga (1):< / p >
< ul >
< li > v3d: make sure we have enough space in the CL for the primitive counts packet< / li >
< / ul >
< p > Ian Romanick (8):< / p >
< ul >
< li > nir/algrbraic: Don't optimize open-coded bitfield reverse when lowering is enabled< / li >
< li > intel/compiler: Request bitfield_reverse lowering on pre-Gen7 hardware< / li >
< li > nir/algebraic: Mark some value range analysis-based optimizations imprecise< / li >
< li > nir/range-analysis: Adjust result range of exp2 to account for flush-to-zero< / li >
< li > nir/range-analysis: Adjust result range of multiplication to account for flush-to-zero< / li >
< li > nir/range-analysis: Fix incorrect fadd range result for (ne_zero, ne_zero)< / li >
< li > nir/range-analysis: Handle constants in nir_op_mov just like nir_op_bcsel< / li >
< li > nir/algebraic: Do not apply late DPH optimization in vertex processing stages< / li >
< / ul >
< p > Ilia Mirkin (1):< / p >
< ul >
< li > gallium/vl: use compute preference for all multimedia, not just blit< / li >
< / ul >
< p > Jason Ekstrand (9):< / p >
< ul >
< li > anv: Bump maxComputeWorkgroupSize< / li >
< li > nir: Handle complex derefs in nir_split_array_vars< / li >
< li > nir: Don't infinitely recurse in lower_ssa_defs_to_regs_block< / li >
< li > nir: Add a block_is_unreachable helper< / li >
< li > nir/repair_ssa: Repair dominance for unreachable blocks< / li >
< li > nir/repair_ssa: Insert deref casts when needed< / li >
< li > nir/dead_cf: Repair SSA if the pass makes progress< / li >
< li > intel/fs: Handle UNDEF in split_virtual_grfs< / li >
< li > nir/repair_ssa: Replace the unreachable check with the phi builder< / li >
< / ul >
< p > Jonathan Marek (1):< / p >
< ul >
< li > freedreno/a2xx: ir2: fix lowering of instructions after float lowering< / li >
< / ul >
< p > Jose Maria Casanova Crespo (1):< / p >
< ul >
< li > mesa: recover target_check before get_current_tex_objects< / li >
< / ul >
< p > Juan A. Suarez Romero (1):< / p >
< ul >
< li > bin/get-pick-list.sh: sha1 commits can be smaller than 8 chars< / li >
< / ul >
< p > Kenneth Graunke (20):< / p >
< ul >
< li > gallium/ddebug: Wrap resource_get_param if available< / li >
< li > gallium/trace: Wrap resource_get_param if available< / li >
< li > gallium/rbug: Wrap resource_get_param if available< / li >
< li > gallium/noop: Implement resource_get_param< / li >
< li > iris: Replace devinfo-> gen with GEN_GEN< / li >
< li > iris: Fix broken aux.possible/sampler_usages bitmask handling< / li >
< li > iris: Update fast clear colors on Gen9 with direct immediate writes.< / li >
< li > iris: Drop copy format hacks from copy region based transfer path.< / li >
< li > iris: Avoid unnecessary resolves on transfer maps< / li >
< li > iris: Fix large timeout handling in rel2abs()< / li >
< li > isl: Drop UnormPathInColorPipe for buffer surfaces.< / li >
< li > isl: Don't set UnormPathInColorPipe for integer surfaces.< / li >
< li > util: Add a _mesa_i64roundevenf() helper.< / li >
< li > mesa: Fix _mesa_float_to_unorm() on 32-bit systems.< / li >
< li > iris: Fix partial fast clear checks to account for miplevel.< / li >
< li > iris: Report correct number of planes for planar images< / li >
< li > iris: Fix constant buffer sizes for non-UBOs< / li >
< li > gallium: Fix util_format_get_depth_only< / li >
< li > iris: Initialize ice-> state.prim_mode to an invalid value< / li >
< li > intel: Increase Gen11 compute shader scratch IDs to 64.< / li >
< / ul >
< p > Lepton Wu (1):< / p >
< ul >
< li > virgl: Fix pipe_resource leaks under multi-sample.< / li >
< / ul >
< p > Lionel Landwerlin (9):< / p >
< ul >
< li > util/timespec: use unsigned 64 bit integers for nsec values< / li >
< li > util: fix compilation on macos< / li >
< li > egl: fix platform selection< / li >
< li > vulkan/overlay: bounce image back to present layout< / li >
< li > radv: store engine name< / li >
< li > driconfig: add a new engine name/version parameter< / li >
< li > vulkan: add vk_x11_strict_image_count option< / li >
< li > util/xmlconfig: fix regexp compile failure check< / li >
< li > drirc: include unreal engine version 0 to 23< / li >
< / ul >
< p > Marek Olšák (23):< / p >
< ul >
< li > radeonsi/gfx10: fix the legacy pipeline by storing as_ngg in the shader cache< / li >
< li > radeonsi: move some global shader cache flags to per-binary flags< / li >
< li > radeonsi/gfx10: fix tessellation for the legacy pipeline< / li >
< li > radeonsi/gfx10: fix the PRIMITIVES_GENERATED query if using legacy streamout< / li >
< li > radeonsi/gfx10: create the GS copy shader if using legacy streamout< / li >
< li > radeonsi/gfx10: add as_ngg variant for VS as ES to select Wave32/64< / li >
< li > radeonsi/gfx10: fix InstanceID for legacy VS+GS< / li >
< li > radeonsi/gfx10: don't initialize VGT_INSTANCE_STEP_RATE_0< / li >
< li > radeonsi/gfx10: always use the legacy pipeline for streamout< / li >
< li > radeonsi/gfx10: finish up Navi14, add PCI ID< / li >
< li > radeonsi/gfx10: add AMD_DEBUG=nongg< / li >
< li > winsys/amdgpu+radeon: process AMD_DEBUG in addition to R600_DEBUG< / li >
< li > radeonsi: add PKT3_CONTEXT_REG_RMW< / li >
< li > radeonsi/gfx10: remove incorrect ngg/pos_writes_edgeflag variables< / li >
< li > radeonsi/gfx10: set PA_CL_VS_OUT_CNTL with CONTEXT_REG_RMW to fix edge flags< / li >
< li > radeonsi: consolidate determining VGPR_COMP_CNT for API VS< / li >
< li > radeonsi: unbind blend/DSA/rasterizer state correctly in delete functions< / li >
< li > radeonsi: fix scratch buffer WAVESIZE setting leading to corruption< / li >
< li > radeonsi/gfx10: don't call gfx10_destroy_query with compute-only contexts< / li >
< li > radeonsi/gfx10: fix wave occupancy computations< / li >
< li > radeonsi: add Navi12 PCI ID< / li >
< li > amd: add more PCI IDs for Navi14< / li >
< li > ac/addrlib: fix chip identification for Vega10, Arcturus, Raven2, Renoir< / li >
< / ul >
< p > Mauro Rossi (2):< / p >
< ul >
< li > android: mesa: revert "Enable asm unconditionally"< / li >
< li > android: anv: libmesa_vulkan_common: add libmesa_util static dependency< / li >
< / ul >
< p > Paulo Zanoni (2):< / p >
< ul >
< li > intel/fs: grab fail_msg from v32 instead of v16 when v32-> run_cs fails< / li >
< li > intel/fs: fix SHADER_OPCODE_CLUSTER_BROADCAST for SIMD32< / li >
< / ul >
< p > Pierre-Eric Pelloux-Prayer (1):< / p >
< ul >
< li > glsl: replace 'x + (-x)' with constant 0< / li >
< / ul >
< p > Rafael Antognolli (1):< / p >
< ul >
< li > anv: Only re-emit non-dynamic state that has changed.< / li >
< / ul >
< p > Rhys Perry (1):< / p >
< ul >
< li > radv: always emit a position export in gs copy shaders< / li >
< / ul >
< p > Samuel Iglesias Gonsálvez (1):< / p >
< ul >
< li > intel/nir: do not apply the fsin and fcos trig workarounds for consts< / li >
< / ul >
< p > Samuel Pitoiset (11):< / p >
< ul >
< li > radv: allow to enable VK_AMD_shader_ballot only on GFX8+< / li >
< li > radv: add a new debug option called RADV_DEBUG=noshaderballot< / li >
< li > radv: force enable VK_AMD_shader_ballot for Wolfenstein Youngblood< / li >
< li > ac: fix exclusive scans on GFX8-GFX9< / li >
< li > radv/gfx10: don't initialize VGT_INSTANCE_STEP_RATE_0< / li >
< li > radv/gfx10: do not use NGG with NAVI14< / li >
< li > radv: fix getting the index type size for uint8_t< / li >
< li > nir: do not assume that the result of fexp2(a) is always an integral< / li >
< li > radv: fix allocating number of user sgprs if streamout is used< / li >
< li > radv: fix loading 64-bit GS inputs< / li >
< li > radv/gfx10: fix VK_KHR_pipeline_executable_properties with NGG GS< / li >
< / ul >
< p > Sergii Romantsov (2):< / p >
< ul >
< li > intel/dri: finish proper glthread< / li >
< li > nir/large_constants: more careful data copying< / li >
< / ul >
< p > Tapani Pälli (5):< / p >
< ul >
< li > util: fix os_create_anonymous_file on android< / li >
< li > iris/android: fix build and link with libmesa_intel_perf< / li >
< li > egl: reset blob cache set/get functions on terminate< / li >
< li > iris: close screen fd on iris_destroy_screen< / li >
< li > egl: check for NULL value like eglGetSyncAttribKHR does< / li >
< / ul >
< p > Thong Thai (1):< / p >
< ul >
< li > Revert "radeonsi: don't emit PKT3_CONTEXT_CONTROL on amdgpu"< / li >
< / ul >
< p > Timur Kristóf (1):< / p >
< ul >
< li > st/nine: Properly initialize GLSL types for NIR shaders.< / li >
< / ul >
< p > Vinson Lee (2):< / p >
2019-05-07 17:07:16 +01:00
< ul >
2019-09-25 17:55:33 +01:00
< li > swr: Fix build with llvm-9.0 again.< / li >
< li > travis: Fail build if any command in if statement fails.< / li >
2019-05-07 17:07:16 +01:00
< / ul >
< / div >
< / body >
< / html >