Mesa 25.1.0 Release Notes / 2025-05-07

Mesa 25.1.0 is a new development release. People who are concerned with stability and reliability should stick with a previous release or wait for Mesa 25.1.1.

Mesa 25.1.0 implements the OpenGL 4.6 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 4.6. OpenGL 4.6 is only available if requested at context creation. Compatibility contexts may report a lower version depending on each driver.

Mesa 25.1.0 implements the Vulkan 1.4 API, but the version reported by the apiVersion property of the VkPhysicalDeviceProperties struct depends on the particular driver being used.

SHA checksums

SHA256: b1c45888969ee5df997e2542654f735ab1b772924b442f3016d2293414c99c14  mesa-25.1.0.tar.xz
SHA512: 40298370727fa1ad6d59be8692dfef01c42e41780816e9b951a05f779c7acca127162d2d3cedfccb7bfe3834b9e917c2b9bc6cb76887488d919cb61741a1da1a  mesa-25.1.0.tar.xz

New features

  • cl_khr_spirv_linkonce_odr in rusticl

  • storagePushConstant16 on panvk

  • storageInputOutput16 on panvk

  • VK_KHR_depth_stencil_resolve on panvk

  • VK_KHR_separate_depth_stencil_layouts on panvk

  • VK_EXT_separate_stencil_usage on panvk

  • VK_KHR_sampler_ycbcr_conversion on panvk/v10+

  • VK_EXT_ycbcr_2plane_444_formats on panvk/v10+

  • VK_EXT_ycbcr_image_arrays on panvk/v10+

  • VK_KHR_imageless_framebuffer on panvk

  • VK_KHR_uniform_buffer_standard_layout on panvk

  • VK_EXT_border_color_swizzle on panvk

  • VK_MESA_image_alignment_control on NVK

  • shaderFloat16 on panvk

  • VK_KHR_shader_subgroup_uniform_control_flow on panvk/v10+

  • VK_KHR_shader_maximal_reconvergence on panvk/v10+

  • VK_EXT_device_memory_report on RADV

  • VK_KHR_shader_subgroup_extended_types on panvk/v10+

  • shaderStorageImageExtendedFormats on panvk

  • VK_KHR_display on panvk

  • VK_EXT_display_control on panvk

  • EXT_shader_framebuffer_image_fetch on v3d

  • EXT_shader_framebuffer_image_fetch_coherent on v3d

  • KHR_blend_equation_advanced on v3d

  • KHR_blend_equation_advanced_coherent on v3d

  • KHR_partial_update on etnaviv

  • VK_KHR_line_rasterization on panvk

  • VK_EXT_line_rasterization on panvk

  • shaderImageGatherExtended on panvk

  • textureCompressionBC on panvk

  • VK_EXT_sample_locations on RADV for GFX10+

  • GL_ARB_shader_clock on panvk

  • VK_KHR_shader_float_controls on panvk

  • VK_KHR_shader_float_controls2 on panvk/v10+

  • storageBuffer8BitAccess on panvk

  • storagePushConstant8 on panvk

  • uniformAndStorageBuffer8BitAccess on panvk

  • MSAA with 8 and 16 sample counts on panvk

  • VK_KHR_spirv_1_4 on panvk/v10+

  • Mali G720 and G725 on Panfrost and panvk

  • dualSrcBlend on panvk

  • VK_KHR_dynamic_rendering_local_read on panvk

  • VK_EXT_subgroup_size_control on panvk/v10+

  • VK_KHR_format_feature_flags2 on panvk

  • shaderStorageImageReadWithoutFormat on panvk

  • shaderStorageImageWriteWithoutFormat on panvk

  • VK_EXT_direct_mode_display on panvk

  • Vulkan 1.2 on panvk/v10+

Bug fixes

  • RADV:RX 9070:Mesa-25.0.5 GTA 5 Enhanced GPU HANG

  • [ANV/DG2] nvpro-samples/vk_raytracing_tutorial_KHR/ray_tracing_reflections renders not all reflections

  • radv: Flickering in Kingdom Come: Deliverance II

  • RADV regression causes severe glitches in Hunt Showdown 1896 on Polaris

  • Z-Fighting in Tomb Raider IV - VI Remastered Linux

  • [anv] VK_ERROR_DEVICE_LOST on Linux 6.13.8 while playing Dota 2 on Intel Graphics

  • Variable Rate Shading (VRS) produces very aliased results on RADV with an AMD gpu

  • WWE 2k23 small “artifacts”

  • RadeonSI - ACO does not spill enough SGPRs despite low VGPR pressure in some scenarios

  • RADV: Performance Regression (~18%) on Vega 64 (GFX9) in Cyberpunk 2077 caused by commit 08918f08805f (MR 34361)

  • radeonsi: CL conformance test `vector_swizzle` fails since 177427877bb50ad7ba24abfa13e55a2684d804df

  • [macOS] glapi_gentable.c:46:10: fatal error: ‘glapitable.h’ file not found

  • commit 90faadae regression - Warhammer 40,000: Space Marine 2 crashing

  • RADV: GPU hangs always at certain places in the game “Sniper Elite: Resistance”

  • Random mesa crashes in kwin_wayland on a 6600XT

  • Patch to fix clinfo on rusticl

  • radv/aco: Ghost of Tsushima hangs and causes gpu resets on RDNA 3 GPU

  • mesa-vulkan-driver-git.x86_64 causes strange colored rectangle artifacts in Final Fantasy XIV

  • NVK - Out of video memory error when starting Marvel Rivals

  • NVK - Out of video memory error when starting Marvel Rivals

  • RADV: Performance regression in Elden Ring on GFX8/Polaris

  • RADV: Performance regression in Elden Ring on GFX8/Polaris

  • Vulkan issues after sleeping on 9070 XT

  • ring gfx_0.0.0 timeout after waking from sleep - RX 9070

  • Confidential issue #12324

  • Confidential issue #12946

  • svga: how do I test this driver?

  • brw: new Xe2 CTS failures

  • [NVK] NAK assert in The Last of Us Part 2 shader

  • anv: S.T.A.L.K.E.R. 2: Heart of Chornobyl crashes after starting a new game

  • [ANV][LNL] - Lost Records: Bloom & Rage (1902960) - Title hangs on launch and subsequently crashes to desktop.

  • [BMG] Intel b580 battlemage: Fort Solis (Unreal Engine game) boots to menu, hangs while loading after hitting continue from the main menu

  • [ANV][LNL] - NINJA GAIDEN 2 Black (3287520) - Environment assets are incorrectly rendered or missing.

  • [ANV][LNL] - The Headliners (3059070) - Title hangs a few minutes after launch.

  • anv, regression: Invisibly blinking cliffs & rocks in Satisfactory DX12 on BMG

  • [Feature request] Add HK support for VK_EXT_queue_family_foreign vulkan extension

  • vk/overlay: output_file option failing

  • The Last of Us Part I GPU hang on gfx1201

  • glsl: IR validation fails on some shaders after lower_precision (breaks GLSL-to-NIR)

  • [bisected, LNL] brw: 341e5117ecbc (“brw/nir: Treat load_const as convergent”) regresses arb_gpu_shader5-interpolateAtOffset on LNL

  • vulkan regression mesa 24.3.4 to 25.0.0.rc3 with broadcom

  • AMD: broken Minetest trace on Vega iGPU

  • All OpenGL applications segfault

  • radv: nir_opt_varyings.c:2766: deduplicate_outputs: Assertion `list_index == 0’ failed.

  • vulkan/wsi: memory leak from wsi_CreateSwapchainKHR

  • [RADV][RDNA3][Phoenix3][APU] NARAKA: BLADEPOINT (1203220) gpu hang reproducible (ice/water regression mesa 24.1 bisected SAMPLE_MASK_TRACKER_WATERMARK=15) random (maybe other apps/games)

  • GPU hangs running Octopath Traveler II with 780M

  • GPU crash on Radeon 780M with Tales of Arise

  • The Last of Us - shadows flickering on gfx1201 without nohiz flag

  • brw: Hit unreachable nir_op_fsign case that brw_nir_lower_fsign missed

  • Panvk:Add Support for BCn (BC1–BC7) Texture Compression

  • anv: Dark pattern overlayed on objects in Eve Online DX11 mode on BMG

  • radv/aco: Strobing artifacts in Pacific Drive

  • aco: Distorted light halos, Star Citizen

  • Mesa 25 removes VA-API encoding for R9 390

  • debian-android build is huge

  • isl: CPCBCompressionEnable is now LosslessCompressionEnable

  • Video stuttering / anv: extend implicit fencing support

  • anv, bmg: Visual issues in AC Origins, Odyssey and Fenyx Rising when dxvk doesn’t export PointSize

  • [ANV][LNL] - A Game About Digging A Hole (3244220) - Title throws an assertion failure on launch.

  • [ANV][BMG] - A Plague Tale: Requiem - Severe shadow flicker

  • [ANV][LNL] - Hogwarts Legacy (990080) - Flickering artifacts visible on magical barrier.

  • brw: regression on Gfx9 dEQP-VK

  • HTML Docs fail to build from source with Sphinx 8.2.1

  • libglapi.dll ain’t built anymore under Windows

  • anv/video: Timestamps are exposed in video encode queue, but it crashes

  • Getting a crash with manually built llvmpipe (OpenGL)

  • [RadeonSI] Blender assetshelf icons are borken in mesa >= 25.0.0

  • radeonsi regression after 24.3.4

  • misc OpenGL CTS failures

  • glBindVertexBuffer regression due to ID reuse

  • RADV: logic used to avoid running on CDNA is faulty

  • [ANV][LNL] - Company of Heroes 3 (1677280) - Circular banding is present on screen during gameplay.

  • video hardware acceleration issue with “VA”

  • [LNL/BMG] Assassin’s Creed Valhalla trace replay hang

  • Vulkan conformanceVersion is reported as 0.0.0.0 in Mesa 25.0.0

  • X11 + Zink on NVK flickers older frames in Firefox based browsers

  • VRAM Abnormal use on mesa 25.0

  • [radv][regression] Multiple games detect the wrong amount of vram

  • Resident Evil 2 Remake flickers

  • OpConstantNull not supported for OpTypeCooperativeMatrixKHR

  • v3dv: vkcube-wayland crashes on raspberry pi 5 kernel 6.12 and latest mesa

  • GMSH Visualization Fails with radeonsi:can’t compile a main shader part, Fedora 41 AMD 7900xt

  • AMD VDPAU deinterlacing SIGSEGV

  • radv: vkd3d-proton test_primitive_restart_list_topology_stream_output randomly fails on NAVI2X

  • Mesa 24.1 introduced a Vulkan problem with DOOM 2016 on AMD 780M GPU

  • radeonsi: Firefox fails assertion requiring binding vertex elements before vertex_buffers

  • nouveau & zink+nvk: Flashing in Firefox and Thunderbird on Hyprland

  • Zink: Kopper’s present thread causes Wayland protocol races

  • GLmatrix needs aligned malloc

  • Lavapipe crashes if no Position is output in mesh shader

  • radeonsi/video: Allocate video buffers with modifiers 7f7206f1a9d brake ‘mplayer -vo vdpau…’ on GFX8 (Polaris 20)

  • [RADV/aco][regression][bisected] - Avowed (2457220) - GPU hangs near Watermill outside of Dawnshore

  • radv/sqtt: assertion “layout transition marker should be only emitted inside a barrier marker”

  • mesa_shader_cache directory is created even if mesa_shader_cache_db is used.

  • nak: Fold i2b(b2i(x))

  • [radv] Glitchy ground geometry regression in Total War Warhammer III on RX 7600

  • NVK: Implement shaderSharedInt64Atomics

  • nvk: Implement VK_MESA_image_alignment_control

  • radeonsi: regression with running DaVinci Resolve under rusticl since 666a6eb871d5dec79362bdc5d16f15915eb52f96

  • [ANV][LNL] - Black Myth: Wukong (2358720) - Corruption is visible near the edge of water.

  • [ANV][LNL] - Hogwarts Legacy (990080) - Pixelated corruption is visible when looking out at the water.

  • radv/video/h265: pps.flags.transform_skip_enabled_flag = 1 randomly hangs GPU

  • turnip falls with `assertion “errno == ETIME” failed`

  • [ANV][LNL] - Steel Rats (619700) - Game crashes after opening logos play before reaching main menu

  • nvk: Implement host-only descriptors

  • Gnome-shell Wayland fails to start with segfault at modifier-less driver

  • [ANV][LNL] - DYNASTY WARRIORS: ORIGINS (2384580) - Dithered transparency has vertical bands.

  • AMD Radeon R9 270 randomly causes video playback applications to crash with “amdgpu: The CS has been rejected”

  • ci: a306 jobs use a307 files

  • Rendering issues on GravityMark with RadeonSI ACO

  • Expose low latency encoding for radv vulkan video encode

  • i915: multiple tests assert with tgsi_ureg.h:893: ureg_swizzle: Assertion `reg.File != TGSI_FILE_NULL’ failed.

  • ci: debian-build-testing fails in nightly

  • shaders/closed/steam/deus-ex-mankind-divided/260.shader_test fails NIR validation

  • shaders/closed/steam/deus-ex-mankind-divided/260.shader_test fails NIR validation

  • panvk : vk_pipeline_cache_object_deserialize: Assertion `reader.current == reader.end && !reader.overrun’ failed.

  • 46a8d5e7ef61735416d0c54886a7a9930621ae2c causes a permission denied spam

  • [BUILD] Build Failure: Implicit Function Declaration ‘timespec_sub_saturate’ (loader_wayland_helper.c)

  • anv: `MESA: warning: INTEL_HWCONFIG_MIN_GS_URB_ENTRIES (2) != devinfo->urb.min_entries[MESA_SHADER_GEOMETRY] (0)`

  • intel genX_acceleration_structure: missing dependency to bvh/header.spv.h

  • ci: what’s going on with zink-venus-lavapipe ?

Changes

Aaron Ruby (17):

  • gfxstream: Replace pre-processor (LINUX && !ANDROID) checks with LINUX_GUEST_BUILD

  • gfxstream: Make the virtgpu device discovery for LinuxVirtGpu more robust

  • gfxstream: Change “mesaOnly” nomenclature to be “guestOnly”

  • gfxstream: Add common interfaces in the VirtGpuDevice to query DrmInfo and PciBusInfo

  • gfxstream: Clean up the gfxstream_vk device and instance init

  • gfxstream: Fix precedence and ownership issues on Linux for imported FD and the VirtGpuResource

  • gfxstream: Add reference counting for GEM handles in LinuxVirtGpuBlob

  • gfxstream: Downgrade log severity when enabling params in LinuxVirtGpu

  • gfxstream: Remove vkGetImageSubresourceLayout ResourceTracker entry

  • gfxstream: Full emulation support for VK_EXT_image_drm_format_modifier

  • gfxstream: Resolve/clean-up inconsistencies with advertising emulated extensions

  • gfxstream: No VIRGL_BIND_LINEAR for ColorBuffers

  • gfxstream: Move virtgpu_gfxstream_protocols.h to the common location for house protocols

  • egl: Remove FallbackZink config item entirely

  • gbm/dri: Match zink autoloading from egl

  • drm-uapi: Sync virtgpu header

  • virtio: Remove virglrenderer_hw.h entirely

Adam Jackson (8):

  • zink: Enable KHR_shader_subgroup

  • lvp: set subgroupQuadOperationsInAllStages to true

  • glx: Make #undef GLX_INDIRECT_RENDERING do something

  • mapi/glx: Remove xserver code generation

  • glx: Remove (almost) all usage of _X_HIDDEN / _X_INTERNAL

  • mapi/glx: Remove FASTCALL/PURE

  • loader: Stop looking in ${libdir}/tls/

  • meson: Simplify the power8 optimization logic

Aditya Kumar (1):

  • gfxstream: Fix compiling gfxstream for musl libs

Alejandro Piñeiro (1):

  • nir: aliasing checks should be also done with index != 0

Aleksi Sapon (3):

  • draw: fix gl_PrimitiveID in tessellation

  • llvmpipe: improve aniso filtering

  • lp: fix gnu-empty-initializer warning

Alyssa Rosenzweig (226):

  • meson: factor out with_driver_using_cl

  • nir: add nir_function_intrinsics_pass

  • nir/lower_scratch_to_var: handle KERNELs

  • nir/lower_scratch_to_var: handle multi-function shaders

  • nir/print: extract nir_print_function_body

  • nir/serialize: add specialized function serialization

  • nir: introduce bindgen_return

  • nir: add nir_call_serialized helper

  • vtn: add vtn_bindgen2 tool

  • libcl: add MIN3/MAX3 macros like on the host

  • libcl: define GLSL-style compute built-ins

  • libagx: port to glsl-style compute builtins

  • asahi: port to vtn_bindgen2

  • nir: include __LINE__ in NIR_PASS validation results

  • libcl: add unreachable() macro

  • nir: add image_min_lod_agx

  • nir: add lod_bias_min_agx tex src

  • agx: lower min LOD for txf

  • agx: switch to nir_tex_src_lod_bias_min_agx

  • hk: pack has_border with clamp_0_sampler_index

  • hk: rearrange sampler image desc

  • hk: emulate EXT_image_view_min_lod

  • hk: advertise EXT_image_view_min_lod

  • Revert “hk: Stop using strings or common key types for meta keys”

  • intel: drop nir_lower_printf calls

  • intel/nir_lower_printf: modernize nir

  • intel: port to u_printf context + singleton

  • nir/lower_printf: hash format strings in nir_printf_fmt

  • nir: drop printf_base_identifier

  • nir: add nir_intrinsic_has_semantic helper

  • nir: switch intrinsic semantics to BIT

  • nir: mark subgroup/quadgroup ops

  • nir/gather_info: use subgroup/quadgroup flags

  • nir/opt_move_discards_to_top: use semantic

  • nir: mark a few more subgroup ops

  • libcl/vk: add common query copy write routine

  • nir: default-initialize next_stage

  • asahi: clang-format

  • asahi: fix libwrap.dylib

  • asahi: fix cull distance with GS

  • hk: implement calibrated timestamps

  • hk: fix increment CS invs

  • hk: do not incorrectly offset host-image-copy sources

  • asahi: bind zero-page

  • libagx: use zero page

  • asahi: use zero sink for vbuf

  • hk: use zero sink for null index buffer

  • hk: don’t allocate zero sink

  • libagx: add missing null pointer check

  • asahi: perf debug indirect tess

  • asahi: use NIR_PASS to validate more

  • asahi: switch tib lower to intrinsic pass

  • libagx: fix subgroup id confusion

  • libagx: fix wraparound issue with robust draw kernel

  • libagx: use assert instead of 0xdeadbeef writes

  • asahi: add more alignment asserts

  • hk: reject non-2D modifiers

  • hk: unstub UnmapMemory2KHR

  • hk: unbind VAs

  • hk: fix buffer binding

  • hk: enable bufferDeviceAddressCaptureReplay

  • docs/asahi: update varying info

  • docs/asahi: add some section headers

  • docs/asahi: describe sparse page tables

  • asahi: rename Null layout

  • asahi: identify image mode enum

  • asahi: add sparse block XML

  • ail: model sparse page tables

  • ail: unit test sparse_table_size_B

  • ail: report mip_tail_first_lod for sparse

  • ail: report miptail stride

  • ail: move helpers to layout.h for sharing

  • ail: extract a blocksize helper

  • ail: expose ail_get_max_tile_size

  • ail: add ail_bytes_to_pages helper

  • ail: add ail_get_twiddled_block_B helper

  • ail: unit test ail_get_twiddled_block_B

  • asahi: assert page alignment in vm_bind

  • asahi: support unbinding VA in agx_va_free

  • agx: implement sparse residency queries

  • hk: ban sparse host-image-copy

  • hk: implement sparse

  • asahi: drop cargo culted disk cache disable

  • agx: fix uniform atomic opts

  • hk: ensure space with the dispatch

  • hk: pass cmdbuf, not control stream, into precomp dispatch

  • hk: assert more

  • hk: accelerate buffer copies with CL

  • hk: pass cmdbuf to perf_debug

  • libagx: fix ia_primitives with tessellation

  • libagx: vectorize tess level loads

  • hk: perf debug sparse binding

  • asahi,libagx,hk: don’t set custom_target names

  • asahi: drop trivial depends

  • agx: assert shift bounds

  • agx: fix botched address fuses

  • hk: do not clamp txf for copy shaders

  • libagx: use 64-bit multiply for image atomic calculation

  • asahi: fix null deref in error path

  • hk: fake min/max filtering for proton

  • agx: handle rgb32 residency queries

  • hk: ban sparse RGB32

  • agx: handle sparse buffer images

  • asahi: support sparse in virtgpu

  • nir/builder: add nir_shader_tex_pass

  • treewide: use nir_shader_tex_pass

  • radv/nir_lower_viewport_to_zero: intrinsic pass

  • radv/nir_lower_view_index: intrinsic pass

  • radv/nir_lower_fs_barycentric: intrinsic pass

  • radv/nir_lower_intrinsics_early: intrinsic pass

  • radv/nir_lower_fs_intrinsics: intrinsic pass

  • nvk: rewrite query copy shader in CL C

  • docs/asahi: clarify twiddled vs GPU-tiled

  • ail: rename twiddled -> gpu tiled

  • ail: split compression up

  • asahi: rename wip modifier

  • asahi: drop silly

  • hk: drop silly

  • nir: add nir_progress/nir_no_progress helpers

  • asahi: clang-format

  • nir: clang-format

  • treewide: Switch to nir_progress

  • treewide: port remaining nir_metadata_preserve users

  • nir,nak: update comments referencing nir_metadata_preserve

  • nir: eliminate nir_metadata_preserve

  • nir: clean up progress

  • ir3: clean up progress manually

  • asahi: clang-format

  • vulkan: add common VK_PRINT_STR/VK_COPY_STR macros

  • hk: switch to common VK_COPY/PRINT_STR

  • anv,hasvk: switch to common VK_COPY/PRINT_STR

  • v3dv: switch to common VK_COPY/PRINT_STR

  • radv: use VK_COPY_STR

  • panfrost,panvk: fix clang warnings

  • vulkan: add helpers to work with executable statistics

  • util: add shader statistic framework

  • asahi: port to common stats framework

  • v3dv: fix clang warning

  • meson: make CL args common

  • meson: set NDEBUG appropriately for OpenCL

  • intel: use common CL args

  • meson,clc: set CL standard from meson

  • panfrost: clean up CL meson rules

  • panfrost: port to common stats framework

  • agx: call nir_lower_is_helper_invocation

  • pan/mdg: call nir_lower_is_helper_invocation

  • nir/lower_helper_writes: fix stores after discard

  • pan/lower_helper_invocation: clean up

  • bin: add script for applying review trailers

  • bin: add list of Mesa contributors

  • libcl: add u_foreach_bit

  • hk: fix cull distance confusion

  • asahi: integrate printf/abort support

  • libagx: do not use prefix sums for GS index buffer

  • asahi/gs: drop non-XFB prefix sums

  • asahi/gs: factor out output info

  • asahi/gs: avoid recalculating

  • asahi/gs: only prefix sum with XFB

  • libagx,asahi: hoist GS draw generation

  • asahi/gs: drop unused params

  • asahi/gs: report whether xfb is needed

  • asahi: do not dispatch count/pre-GS unless needed

  • hk: do not dispatch count/pre-GS unless needed

  • libagx: use indirect draw struct

  • libagx: clean up

  • tu: use the stats framework

  • v3dv: use the stats framework

  • drm-uapi: update drm_fourcc.h

  • ail: drop fake modifiers

  • libagx: reduce # of variants of unroll restart

  • asahi: add assert for max variant count

  • agx: fix ctz of zero with address calcs

  • agx: fix tg4 offset residency

  • hk: advertise semaphore extensions

  • hk: advertise bufferDeviceAddressCaptureReplayEXT

  • hk: fix unaligned copies

  • libagx: clean up query copy; bug fix

  • libagx: assert alignment for copies

  • libagx: use common heap allocs

  • asahi: fix printf without result buffer

  • asahi: fix depth buffer feedback loops

  • asahi: clang-format

  • glsl_to_nir: upcast array indices

  • nir/lower_blend: refactor logicop variables

  • nir/lower_blend: disable logic ops for unsupported formats

  • nir/lower_tex: use vector_insert_imm

  • vk/meta: generalize 3D blit code

  • asahi,hk: disentangle logicop_enable

  • agx: optimize nonuniform offset

  • hk: drop dead todo

  • hk: fix last VGT output component limits

  • hk: support colour <–> z/s copies

  • hk: advertise VK_EXT_queue_family_foreign

  • hk: advertise maintenance7

  • hk: advertise maintenance8

  • bin: add rebase mode

  • drm-uapi: add asahi uapi

  • asahi: disable virtio gpu for upstream

  • asahi: port to stable uAPI

  • asahi: remove unstable uapi header

  • gallium: wire up asahi driver

  • asahi: fix i/a queries with tess

  • panfrost: invert and rename no_ubo_to_push flag

  • panfrost: do not push “true” UBOs

  • asahi: shrink VA space for sparse emulation

  • asahi: add sparse emu helpers

  • asahi: fix zero bo leak

  • hk: bind for sparse emulation

  • hk: use ro maps

  • hk: advertise sparseResidencyBuffer

  • hk: drop soft fault assumption in hk_buffer_addr_range

  • util/simple_mtx: fix duplicate definition

  • nir: unvendor lod_bias(_agx)

  • nir: handle mismatched bias/lod bitsizes

  • nir: add sampler LOD bias lowering

  • hk: handle lod/min_lod size mismatch

  • agx: switch to common LOD bias lowering

  • hk: fix underbinding scratch

  • hk: fix tessellation + clipper queries

  • hk: fix null FS corner cases

  • agx: model sources as late-kill in demand calcs

  • agx: late-kill sources

  • agx: early-kill sources only if it won’t shuffle

  • hk: fix patch count = 0 handling

  • asahi: fix possible null deref

  • asahi: do not use “Null” layout

Andrew Wolfers (1):

  • vulkan: Add BGRA format support

Antonino Maniscalco (2):

  • freedreno/crashdec: Avoid crashing on some traces

  • freedreno/crashdec: Add check for `CP_BV_SQE_STAT_ADDR`

Antonio Ospite (28):

  • ci/android: fix building deqp on Android

  • ci/android: add an android-angle-lavapipe job

  • ci/android: don’t do unnecessary cleanup

  • ci/android: don’t set HOME globally in cuttlefish-runner.sh

  • ci/android: disable audio in cuttlefish, it is not needed

  • ci/android: disable modem simulator in cuttlefish, it is not needed

  • ci/android: stop cuttlefish before copying the logs, to log everything

  • ci/android: increase the waiting time to stop the cuttlefish launcher

  • ci/android: bump CUTTLEFISH_BUILD_NUMBER

  • ci/android: use an x86_64_only cuttlefish image

  • ci/android: push /deqp-gles/mustpass/egl-main.txt.zst

  • ci/android: pass -vsock_guest_cid to launch_cvd

  • ci/android: move all dEQP handling in one place in cuttlefish-runner.sh

  • ci/android: add an android-deqp-runner.sh script

  • ci/android: add a job using android-cts instead of deqp-runner

  • ci/android: don’t set EGL_PLATFORM on the host before launching cuttlefish

  • ci/android: increase the waiting time to stop the cuttlefish launcher

  • ci/android: factor out a generic android-runner.sh from cuttlefish-runner.sh

  • ci/android: remove some unnecessary adb commands from android-runner.sh

  • ci/android: check for ANGLE_TAG in android-deqp-runner.sh

  • ci/android: remove old mesa and ANGLE libraries before pushing new ones

  • ci/android: always push ANGLE libraries

  • ci/android: push also the intel vulkan driver

  • ci/android: handle ANGLE being installed under /system on Android 15+

  • ci/android: factor out GLES runtime version retrieval to a function

  • ci/android: only ship x86_64 artifacts in debian-android build job

  • ci/android: strip the artifacts of the debian-android job

  • meson: pass –no-pager to `meson configure` command

Ashley Smith (2):

  • panfrost: Reset syncobj after use to avoid kernel warnings

  • pan/bi: Enable ARB_shader_clock extension support

Assadian, Navid (1):

  • amd/vpelib: More parameters to the segmentation process and introduce validation hook

Autumn Ashton (5):

  • radv: Fix compute resolve rounding

  • radv: Enable fragmentShadingRateWithCustomSampleLocations

  • radv: Expose EXT_sample_locations everywhere

  • radv: Expose VK_SAMPLE_COUNT_1_BIT for sample position on GFX10+

  • ci/zink+radv: Add sample locations fails

Bas Nieuwenhuizen (1):

  • radv: Move support check out of winsys.

Benjamin Cheng (1):

  • d3d12/video_enc: Fix AV1 tile_info() coding

Benjamin Lee (41):

  • panfrost: minor refactors in preparation for panvk 16-bit IO

  • panfrost: move handling for bifrost mediump lowering to pan_collect_varyings

  • panfrost: support 16-bit vertex attributes

  • panfrost: support 16-bit varyings

  • panvk: advertise storagePushConstant16 and storageInputOutput16

  • panvk: enable VK_KHR_depth_stencil_resolve

  • panvk: enable KHR_separate_depth_stencil_layouts

  • panvk: document missed extensions in new_features

  • panvk: implement VK_EXT_separate_stencil_usage

  • panfrost: remove NIR_PASS_V usage for noperspective lowering

  • meson: update wayland-protocols source_hash

  • panfrost: fix large int32->float16 conversions

  • panfrost: reorder lower_bit_size pass

  • panfrost: fix condition in bi_nir_is_replicated

  • panfrost/va: remove swizzle mod from LDEXP

  • panfrost: implement 16-bit ldexp

  • panvk: advertise shaderFloat16

  • panfrost: implement 16-bit pack/unpack intrinsics

  • panvk/csf: fix uninitialized read in utrace_clone_init_builder

  • panfrost/pps: fix omitting several counters

  • panfrost: fix libpan_v4 header include

  • panfrost/va: remove dead code for packing BRANCHZI.i16 lane mod

  • driconf: add uint64 type

  • panvk: add core mask driconf options

  • panfrost: add core mask driconf options

  • pan/va: add roundmode modifier to additional instructions

  • panfrost: implement float controls rounding mode

  • pan/va: preserve signed zero in f32->f16 conversions

  • pan/bi: refactor bi_instr_ftz to allow dontcare FTZ states

  • pan/bi: ignore ftz mode when scheduling int instructions

  • pan/bi: implement denorm behavior float controls

  • panvk/va: don’t advertise independent denorm behavior

  • panvk: advertise VK_KHR_shader_float_controls

  • panvk: advertise VK_KHR_float_controls2

  • panfrost: define bi_swizzle alias values

  • panfrost: add bi_swizzle values for unused swizzles

  • panfrost/va: use ‘lane’ modifier for MKVEC.v2i8

  • panfrost: use bifrost instruction table for bi_lower_swizzle

  • panfrost/va: valhall-specific swizzle lowering

  • panvk: enable uniformAndStorageBuffer8BitAccess and storagePushConstant8

  • panvk: don’t advertise VK_KHR_float_controls2 on bifrost

Benjamin Otte (1):

  • lavapipe: Don’t advertise support for multiplane drm formats

Benjamin ROBIN (2):

  • util/disk_cache: Do not try to delete old cache if cache is disabled

  • docs: Update MESA_SHADER_CACHE_DIR env variable behavior

Bo Hu (2):

  • gfxstream: Adding support for VK_KHR_global_priority extension

  • gfxstream: update code generator for simplified dep graph

Boris Brezillon (29):

  • panvk: Don’t clobber registers if the render pass was suspended

  • pan/decode: Fix the blend_count mask

  • panvk/csf: Don’t free the resources twice when init_render_desc_ringbuf() fails

  • panvk: Initialize device virtual address space after the VM creation

  • vulkan/state: Fix input attachment map state initialization/copy

  • vk/pass: Add input attachment location info

  • vulkan/state: Fix default input attachment map values

  • panvk/jm: Don’t force a preload if the previous batch didn’t have draws

  • pan/bi: Allow depth/stencil tile buffer access using LD_TILE

  • pan/bi: Pass load_converted_output_pan target through a source

  • pan/bi: Pass an explicit sampleid to load_converted_output_pan

  • panvk/csf: Set invalidate_inherited_ctx only if the render pass is inherited

  • panvk: Re-order things in panvk_deserialize_shader()

  • panvk: Isolate CS specific bits in panvk_shader

  • panvk: Generate the earlyzs LUT at shader creation time

  • pan/earlyzs: Support the shader ZS read-only case and its optimization on v10+

  • panvk/jm: Move cmd_prepare_draw_sysvals() out of the layer loop

  • panvk: Support color attachment remapping

  • pan/bi: Introduce two intrinsics to support input attachment remapping

  • panvk: Optimize input attachment loads when we can

  • panvk: Skip BY_REGION barriers if we’re in a render pass

  • panvk: Advertise KHR_dynamic_rendering_local_read support

  • pan/va: Support relaxed waits on read-only render targets

  • panvk/csf: Optimize read-only tile buffer access

  • panfrost: also consider z/s for tile-size

  • pan/earlyzs: Fix the read-only ZS optimization

  • panvk/csf: Pass less arguments to emit_vs_attrib()

  • panvk/csf: Fix instance attribute offseting

  • pan/format: Disable image storage on A8_UNORM

Caio Oliveira (112):

  • intel/brw: Move fs_inst implementation code together

  • intel/brw: Rename fs_inst to brw_inst

  • intel/brw: Rename fs_inst_box to brw_inst_box

  • intel/brw: Move some larger functions from brw_inst.h to brw_inst.cpp

  • intel/brw: Remove brw_gs_compile struct

  • intel/brw: Rename file brw_fs_nir.cpp to brw_from_nir.cpp

  • intel/brw: Remove ‘fs’ prefix from brw_from_nir functions

  • anv: Add statistic for ‘GRF registers’ for Xe3+

  • intel/brw: Add brw_analysis.h

  • intel/brw: Use brw_analysis prefix for liveness analysis files

  • intel/brw: Use brw_analysis prefix for def analysis file

  • intel/brw: Use brw_analysis prefix for performance analysis files

  • intel/brw: Move idom_tree declaration to brw_analysis.h

  • intel/brw: Move analysis passes without own file to brw_analysis.cpp

  • intel/brw: Merge brw_ir_analysis.h into brw_analysis.h

  • intel/brw: Use brw prefix instead of namespace for dep analysis enum

  • intel/brw: Use brw prefix instead of namespace for analysis implementations

  • intel/brw: Remove ‘using namespace brw’ directives

  • intel/brw: Remove offsets and total_size from VGRF allocator

  • intel/brw: Pass fs_visitor around instead of the simple_allocator

  • intel/brw: Add functions to allocate VGRF space

  • intel/brw: Fold simple_allocator into the shader

  • intel/brw: Rename and move thread_payload types to own header

  • intel/brw: Merge brw_fs_visitor.cpp into brw_fs.cpp

  • intel/brw: Rename files brw_fs.cpp/h to brw_shader.cpp/h

  • intel/brw: Rename fs_visitor to brw_shader

  • intel/brw: Rename fs_copy_prop_dataflow to brw_copy_prop_dataflow

  • intel/brw: Rename a few remaining functions to remove fs prefix

  • intel/brw: Update outdated comments

  • intel/compiler: Use -Werror=vla

  • intel/brw: Use the builder DO() function in all places

  • intel/brw: Always have a (non-DO) block after a DO in the CFG

  • intel/brw: Don’t need to repair CFG in brw_opt_combine_constants

  • brw: Reformat brw_gram.y and brw_lex.l

  • brw: Fix size in assembler when compacting

  • brw: Always verify EU compaction in debug mode

  • brw: Remove unused function

  • brw: Add block pointer in brw_inst

  • brw: Use brw_inst::block in Def analysis

  • brw: Use brw_inst::block in Combine Constants

  • brw: Use brw_inst::block in CSE

  • brw: Remove block parameter from brw_inst::remove()

  • brw: Simplify brw_builder “insert before inst” constructor

  • brw: Add explicit way to get an empty brw_builder

  • brw: Embed at_end() inside brw_builder(brw_shader *) constructor

  • brw: Remove bblock_t parameters from various passes

  • brw: Make assembler strict about JIP and UIP order

  • brw: Add support for GOTO/JOIN in the assembler

  • brw: Rework label tracking in assembler

  • brw: Remove extra interface in brw_cfg types

  • brw: Remove dead code from control flow

  • brw: Add missing dependency classes to various passes

  • brw: Get the reference to brw_def_analysis only once in saturate propagation

  • brw: Move defs analysis back to its place in saturate propagation

  • brw: Simplify the test code for brw passes

  • brw: Add assembler support for DPAS

  • brw: Remove extra SHADER_OPCODE_FLOW emitted during NIR conversion

  • spirv: Update headers and metadata from latest Khronos commit

  • vulkan: Update XML and headers to 1.4.311

  • spirv: In SpecConstantOp handling don’t adjust convert ops bit_size twice

  • brw: Fix decoding of 3-src destination stride in EU validation

  • brw: Allow generating destination with stride 2 in 3-src instructions

  • brw: Remove prefix gfx10 from enum types

  • brw: Make some integer check more explicit

  • brw: Add BRW_TYPE_BF for bfloat16

  • brw: Add BRW_TYPE_BF validation

  • brw: Consider bfloat16 in scoreboard

  • brw: Add EU assembler support for bfloat16

  • intel/executor: Add small example with bfloat

  • brw: Expand EU validation for DPAS

  • brw: Fix disassembler trying to decode 3src_hstride in Gfx9

  • brw: Remove brw_cfg::dump()

  • brw: Track num_instructions in a block

  • brw: Use block’s num_instructions in scoreboard tests

  • brw: Track total_instructions in a shader

  • brw: Add analysis for block IP ranges

  • brw: Use brw_ip_ranges in passes

  • brw: Use brw_ip_ranges in scheduling / regalloc

  • brw: Use brw_ip_ranges in liveness analysis

  • brw: Remove adjust_block_ips and brw_inst::remove() with defer

  • intel/executor: Use getopt for command line arguments

  • intel/executor: Allow selecting a device to use

  • brw: Use control flow helpers in scoreboard tests

  • brw: Use SIMD16 shaders in scoreboard tests for Xe2+

  • brw: Stop setting SFID in scoreboard tests

  • brw: Return actual progress in brw_lower_scoreboard

  • brw: Add parser of SWSB annotations to use in tests

  • brw: Use new bld/exp style in scoreboard tests

  • brw: Remove HSW specific code from brw_compile_cs.cpp

  • brw: Add brw_builder::uniform()

  • brw: Fix invalid memory access in scoreboard test

  • brw: Fix memory leak in EU validation tests

  • intel/executor: Fix check for open() failure

  • brw: Properly handle cooperative matrices created with constants

  • spirv: Take a separate data_type when creating a new vtn_variable

  • brw: Add a few basic tests for register coalesce

  • brw: Clean up saturate propagation after non-defs version removal

  • brw: Add brw_range struct

  • brw: Use brw_range in IP ranges analysis

  • brw: Use brw_range when operating with live ranges

  • brw: Use brw_range to store VGRF ranges

  • brw: Use brw_range to store Vars ranges

  • brw: Use brw_range::last() to explicit get the last valid IP

  • brw: Make brw_range use half-open ranges

  • broadcom/ci: Skip test due to timeout

  • nir/load_store_vectorize: Skip new bit-sizes that are unaligned with high_offset

  • intel: Add intel_device_info::has_systolic

  • intel: Disable has_bfloat16 for MTL

  • brw: Allow DPAS with BF on Gfx125

  • brw: Update EU validation to allow packed BF mixed with packed F

  • intel/executor: Update bfloat example

  • intel/executor: Fix bfloat example for converting F to packed BF

Caleb Callaway (3):

  • anv: add INTEL_DEBUG=rt_notrace

  • intel/compiler: fix lingering i965 references

  • intel/tools: fix 32b build for EU stall tool

Casey Bowman (2):

  • vulkan/screenshot-layer: Add RGBA surface format support

  • vulkan/screenshot-layer: Correct queueFamilyIndex source

Caterina Shablia (11):

  • panvk: enable imagelessFramebuffer

  • panvk: enable uniformBufferStandardLayout

  • nir: teach nir_lower_bit_size to handle ballot and ballot_relaxed

  • pan/bi: ensure src bit sizes of ballot{,_relaxed} and read_invocation

  • pan/bi: lower non-32-bit ballot{,_relaxed}

  • pan/bi: lower subgroups before lowering int64

  • panvk: enable subgroupExtendedTypes

  • panfrost: don’t overwrite push uniforms and sysvals UBO with user’s UBO

  • panfrost: update nr_uniform_buffers before dispatching XFB

  • panfrost: require buffer_count and pushed_words to be passed to panfrost_emit_const_buf

  • panfrost: move the comment closer to what it’s about

Charmaine Lee (1):

  • svga: remove tgsi semantic in shader compiler key

Christian Gmeiner (20):

  • docs: Update perfetto with the latest status

  • docs: Update the list of drivers with CPU tracepoints

  • docs: Add perfetto driver specifics for V3D

  • etnaviv: isa: Add swizzle instruction

  • nir: Add bool return value to nir_lower_clip_halfz(..)

  • nir: Add bool return value to nir_lower_texcoord_replace(..)

  • etnaviv: nir: Return progress for etna_lower_alu(..)

  • etnaviv: nir: Return progress for etna_lower_io(..)

  • etnaviv: nir: Don’t use deprecated NIR_PASS_V macro anymore.

  • zink/ntv: Only emit GeometryStreams cap if multiple streams are used

  • etnaviv/ci: Bring back GC7000

  • etnaviv/ci: Start using the revision number for GPU_VERSION

  • etnaviv/ci: Update flakes for gc7000-r6214

  • etnaviv/ci: Add gles2 run for GC3000

  • etnaviv/ci: Add missing rev to GC3000

  • etnaviv: rs: Factor out box alignment function

  • etnaviv: rs: Change param etna_get_rs_alignment_mask(..)

  • etnaviv: Add support for KHR_partial_update

  • mesa/formats: Add MESA_FORMAT_NV12

  • etnaviv: Add multi-planar YUV support

Colin Marc (1):

  • vulkan/wsi: implement the Wayland color management protocol

Collabora’s Gfx CI Team (7):

  • Uprev Piglit to fc8179d319046f45346bcbcc5aaeabebdf151f03

  • Uprev Piglit to 04d901e49de6b650f9dceaf73220371273d87f73

  • Uprev Piglit to 708a9e365b18fdd881af989f75e1a6c1409cae8c

  • Uprev ANGLE to 1b34d2a18af12cc55a3bc74dd679c2937d10cc5c

  • Uprev ANGLE to 3818d37d5e94317f01810053b8f28c1f1e8b98e6

  • Uprev Piglit to 0ecdebb0f592

  • Uprev ANGLE to a3f2545f6bb3

Connor Abbott (52):

  • tu: Constify frag_areas argument of tu_fdm_bin_apply_t

  • tu: Refactor fragment density map sampling

  • freedreno: Add VSC mask parameter to CP_SET_BIN_DATA5

  • tu: Implement bin merging for fragment density map

  • freedreno: Add a750+ “absolute” VSC bin mask

  • tu: Implement “absolute” bin mask on a750

  • tu: Make sure tiles being merged are adjacent

  • tu: Fix vertical tile merging check

  • tu: Fix static blend bandwidth calculation

  • tu: Remove useless prim_order state merging

  • ir3: Fix const allocation when parsing assembly

  • freedreno/decode: Push hostaddr->gpuaddr conversion into highlight_gpuaddr()

  • freedreno/crashdec: Use register for RB rptr

  • freedreno/crashdec: Handle hangs where the SQE is in RB

  • freedreno/crashdec: Fix and extend control reg dumping on a750

  • freedreno/crashdec: Dump CP_BV_SQE_UCODE_DBG

  • freedreno: Misc control registers updates

  • tu: Implement VK_KHR_maintenance7

  • tu: Plumb through VkMemoryBarrierAccessFlags3KHR

  • tu: Handle D32S8 -> R32 and R32 -> D32S8 copies

  • tu: Handle R8->D24S8 and D24S8->R8 copies

  • tu: Handle a pipelineStageCreationFeedbackCount of 0

  • ir3: Remove ir3_info::data

  • tu: Expose VK_KHR_maintenance8

  • compiler/shader_info: Better document require_full_quads

  • nir, compiler: Rename needs_quad_helper_invocations

  • compiler,nir: Gather needs_full_quad_helper_invocations info

  • ir3: Use needs_full_helper_invocations

  • tu: Fix binning_possible detection with bin merging

  • tu: Fix GMEM offset for multisample layered separate stencil

  • tu: Enable GMEM with layered rendering

  • tu: Fix size of frag_size_ir3 and frag_offset_ir3 driver params

  • tu: Fix reported FDM fragment size with multiview

  • tu: Fix layer_count with dynamic rendering + multiview

  • ir3: Split mad with scalar ALU

  • tu: Only allow power-of-two fragment areas

  • tu: Split out part of tiling config to vsc config

  • tu: Fix CmdClearAttachments with fragment density map

  • tu/fdm: Skip some patchpoints when binning

  • tu: Implement VK_QCOM_fragment_density_map_offset

  • vk/runtime: Use vk_command_buffer in renderpass wrappers

  • vk/runtime: Add common CmdEndRendering

  • tu: Implement VK_EXT_fragment_density_map_offset

  • ir3: Pass through access flags when lowering global accesses

  • nir/opt_preamble: Handle load_global_ir3

  • ir3: Move load/store vectorization to finalize

  • ir3: Vectorize shared memory loads/stores

  • tu: Fix flushing when using a staging buffer for copies

  • freedreno: Add compute_lb_size device info

  • freedreno/a6xx: Define CONSTANTRAMMODE

  • freedreno/a6xx, turnip: Set CONSTANTRAMMODE correctly

  • ir3: Take LB restriction on constlen into account on a7xx

Corentin Noël (12):

  • venus/ci: Skip flaky test due to intermittent timeouts

  • ci: Update CrosVM and Virglrenderer

  • lavapipe: Remove doubly initialized features

  • lavapipe: Change lvp_cmd_type to anonymous enum

  • ci: Update CrosVM and Virglrenderer

  • virgl: nir: Don’t use deprecated NIR_PASS_V macro anymore

  • nir: Add bool return value to nir_legacy_trivialize(..)

  • ntt: nir: Don’t use deprecated NIR_PASS_V macro anymore

  • tnn: nir: Don’t use deprecated NIR_PASS_V macro anymore

  • ci: Uprev virglrenderer to latest version on time

  • virgl: Close handle on resource info failure

  • virgl: Use drmCloseBufferHandle instead of calling dmIoctl directly

Daniel Schürmann (43):

  • aco/isel: fix empty exec tracking for uniform branches

  • aco/isel: move cf_info into separate struct cf_context

  • aco/isel: rename cf_context::has_divergent_branch

  • aco/isel: use cf_context in if_context to restore cf information

  • aco/isel: use cf_context in loop_context to restore cf information

  • aco/isel: add and use exec_info::empty() helper

  • aco/isel: fix assumptions about potential empty exec mask in nested control flow

  • aco/isel: remove loop nest information from exec_info

  • aco/isel: track control flow divergence in loops more accurately

  • nir: make divergence information metadata

  • nir: require nir_metadata_divergence if needed

  • amd: switch to nir_metadata_divergence

  • intel: switch to nir_metadata_divergence

  • nouveau: switch to nir_metadata_divergence

  • nir: only print divergence information if metadata is valid

  • nir/serialize: don’t serialize divergence information

  • nir/validate: validate divergence metadata

  • nir/print: print phi sources sorted by predecessor blocks

  • aco/scheduler: always respect min_waves on GFX10+

  • aco/scheduler: stop rounding down the target number of waves on GFX10+

  • aco: unify get_addr_sgpr_from_waves() and get_addr_vgpr_from_waves() into one function

  • aco/scheduler: keep track of RegisterDemand at UpwardsCursor::insert_idx

  • aco/scheduler: keep track of RegisterDemand at DownwardsCursor::insert_idx{_clause}

  • aco/scheduler: remove unused include of unordered_set

  • aco/insert_exec_mask: Don’t immediately set exec to zero in break/continue blocks

  • aco/insert_exec_mask: don’t restore exec in continue_or_break blocks

  • aco/lower_branches: allow other instructions after s_andn2 in break blocks

  • aco/ssa_elimination: insert parallelcopies for p_phi immediately before branch

  • aco/ssa_elimination: refactor scratch_sgpr handling

  • aco/ssa_elimination: create a single parallelcopy instruction for linear and logical phis

  • aco/assembler: Fix short jumps over chained branches

  • aco/assembler: Don’t insert chained branches into otherwise empty blocks

  • aco/tests: add more tests for chained branches

  • zink: lower {demote|terminate}_if to CF after lowering point smooth

  • nir: remove nir_lower_discard_if_to_cf option

  • nir: set SYSTEM_VALUE_HELPER_INVOCATION read for nir_intrinsic_is_helper_invocation

  • zink: clean up HelperInvocation code emission

  • zink: enable CapabilityDemoteToHelperInvocation and extension only if required

  • aco/lower_branches: properly consider exec mask needs of branch targets

  • aco: don’t assume that demote doesn’t cause an empty exec mask

  • aco/insert_exec_mask: if applicable, use s_wqm to restore exec after divergent CF

  • aco/insert_exec_mask: don’t disable dead quads on demote in divergent CF

  • aco: Remove empty exec skipping after demote

Daniel Stone (29):

  • ci/vvl: Use appropriate level of parallelism

  • ci: Move common testing packages to test-base

  • ci: Move apitrace to test-base container

  • ci: Add debian/arm32_test-* images

  • ci: Build libdrm for Android host builds

  • ci/angle: Use native toolchain for builds

  • ci/zink: Expand flake definition on radv

  • ci: Update kernel for Panthor scheduling fixes

  • ci/panfrost: Move G610 testing to pre-merge

  • ci: Re-enable Collabora CI

  • docs/ci: Fix nginx caching instructions

  • ci/lava: Fail faster when getting results

  • etnaviv: Add missing build dependency on generated header

  • ci/amd: Add new radeonsi fail seen in the wild

  • ci/zink: Flake out sparse tests

  • ci: Add daniels to restricted-trace users

  • ci: Re-enable trace jobs with updated Piglit

  • ci: Move softpipe issue from fail to flake

  • ci/amd: Disable radv-fossils

  • ci/windows: Don’t copy non-existent libglapi.dll

  • ci/panfrost: Split inherit definitions into -inc

  • ci: Reduce build job timeouts

  • ci/softpipe: Add timeouts to softpipe jobs

  • ci/llvmpipe: Add llvmpipe and lavapipe timeouts

  • ci/virgl: Add timeout to software jobs

  • ci/docs: Add timeout to doc build

  • ci/microsoft: Add job timeouts and pin Piglit to GSt

  • ci: Make all job timeouts explicit

  • doc/ci: Update nginx caching snippets

Danylo Piliaiev (28):

  • u_trace: print tracepoint params in csv output

  • util: Make debug_dump_flags thread safe

  • util: Add dump_debug_control_string to dump debug_control

  • tu: Add gmem disable reason to renderpass tracepoint

  • tu: Add info about debug options to command buffer tracepoint

  • tu: Get correct src view when storing gmem attachment

  • freedreno/fdl: Log mutability when dumping layout

  • tu: Handle mismatched mutability when resolving from GMEM

  • tu/a6xx: Emit VSC addresses for each bin to restore after preemption

  • nir/nir_lower_multiview: Don’t assert if load_deref doesn’t have var

  • ir3: Set need_full_quad depending on info.fs.require_full_quads

  • tu: Be more granular in calculating whether blend state reads color

  • tu: Fix NULL deref in trace_end_render_pass

  • freedreno/regs: Rename BINNING bit to FS_DISABLE in a few regs

  • ir3: Detect empty fragment shaders

  • tu/a7xx: Bypass invoking empty FS for D/S-only draw calls

  • ir3: Detect FS that write only color without other side effects

  • tu: Disable FS in certain cases even if FS is not empty

  • tu/lrz: Improve LRZ around stencil tests and reads_dest cases

  • tu: Use EARLY_Z also for stencil tests

  • ir3: Fix shaders that write only color classified as empty

  • freedreno: Bump kernel uapi (linux 6.14)

  • freedreno,tu: Read and pass to compiler uche_trap_base

  • tu: Implement VK_KHR_shader_clock

  • freedreno/a6xx: Implement ARB_shader_clock

  • tu,freedreno: Don’t fallback to LINEAR with DRM_FORMAT_MOD_QCOM_COMPRESSED

  • ir3: VARYING_SLOT_LAYER output is used for binning

  • tu: Fix disable_fs state update condition

Dave Airlie (24):

  • radv/video: move encoder to using a buffer instead of an image

  • radv/video: calculate colloc buffer size for h264 B frames.

  • radv/video: add h264 b frame encoding support.

  • vulkan/wsi/x11: don’t use update_region for damage if not created

  • nak: adjust latencies on fp16/64 instructions on Turing

  • nvk: enable float16 on turing.

  • loader/nouveau: load zink as the GL driver for turing and above.

  • radv/video: don’t try and send events on UVD devices.

  • vulkan/video: add simple parameter retrieval wrappers

  • radv/video: convert to using common parameter wrappers.

  • anv/video: convert to common parameters retrieval code

  • anv/video: don’t write to params if not set.

  • vulkan/video: add support for inline session paramters.

  • radv: expose VK_KHR_video_mainteance2

  • anv: expose VK_KHR_video_maintenance2

  • gallivm: check for avx512vbmi and tell LLVM the correct answer.

  • nak: add reads after setting writes

  • nak: Add an a_has_pred parameter to waw_latency

  • nak: Add Turing latency information

  • nak: Add Ampere and Ada latency information

  • nvk: update nvidia class header files.

  • nvk: add ADA compute class to nv_push_dump

  • nvk: add hopper support to nv_push_dump

  • nak: add F2FP to sm75 instr latencies

David Rosca (64):

  • ac/vcn_dec: Fix AV1 film grain on VCN5

  • radeonsi/video: Avoid stream handle duplicates in PID namespace

  • frontends/va: Don’t try to switch to protected buffer in EndPicture

  • frontends/va: Add CreateContext flag to enable protected context

  • frontends/va: Require protected context for VAProtectedSliceDataBuffer

  • frontends/va: Switch to graphics context when creating protected surface

  • radeonsi/vcn: Use texture instead of video buffer for DPB buffers

  • radeonsi/video: Fix creating video buffers with AMD_DEBUG=tmz

  • frontends/vdpau: Set H264 chroma_format_idc

  • radeonsi/vcn: Set correct chroma format for H264 decode

  • radeonsi/uvd: Set correct chroma format for H264 decode

  • ci/amd: Remove VAAPI skips

  • frontends/va: Use transfer stride and offset in DeriveImage

  • radv/video: Fix setting balanced preset for HEVC encode with SAO enabled

  • radv/video: Move IB header from begin/end to encode_video

  • radeonsi: Use gfx for TMZ buffer clears

  • winsys/amdgpu: Add assert for secure submissions on compute ring

  • frontends/vdpau: Use extra reference buffer for AV1 film grain

  • ac/surface: Only allow linear modifier for subsampled 422 formats

  • ac/surface: Allow DCC for multi-plane formats on GFX12

  • radeonsi/vcn: Fix chroma pitch for JPEG decode

  • radeonsi/video: Allocate video buffers with modifiers

  • radeonsi/vcn: Add UDT support for VCN5

  • radeonsi/vcn: Rework decode ref handling

  • radeonsi/video: Fix crash when creating buffers without modifiers support

  • frontends/va: Set AV1 max_width/height to surface size

  • frontends/vdpau: Set AV1 max_width/height to surface size

  • Revert “radeonsi/vcn: Limit size to target size in AV1 decode”

  • pipe: Remove PIPE_AV1_ENC_FRAME_TYPE_SHOW_EXISTING

  • radeonsi/vcn: Set all pic params for H264 encode references

  • radeonsi/vcn: Add radeon_enc_av1_picture_type

  • radeonsi/vcn: Support H264 encode weighted_bipred_idc

  • radeonsi/video: Remove mpeg12 shader decoder support

  • gallium/vl: Fix video buffer supported format check

  • Revert “frontends/vdpau: Alloc interlaced surface for interlaced pics”

  • frontends/vdpau: Fix creating deinterlace filter for interleaved buffers

  • frontends/va: Support A8R8G8B8 format for processing

  • frontends/va: Use ARGB as default fourcc for RGB32 RT format

  • frontends/va: Don’t filter supported formats according to config RT format

  • gallium/vl: Return YUV plane order for single plane formats

  • radeonsi/video: Only allow 64K_S swizzle mode for VCN < 2.2

  • radeonsi/vce: Support old VCE firmware

  • radeonsi/video: Allow DCC 256B block size with drm minor >= 63

  • gallium/vl: Fix rotation with scaling for compute shaders

  • gallium/vl: Fix mirror with rotation for compute shaders

  • frontends/va: Don’t ignore rotation and mirror for conversions to RGB

  • ac/parse_ib: Fix parsing output format on VCN5

  • ac/parse_ib: Parse VCN DYNAMIC_REFLIST_BUFFER

  • radv: Use radv_format_to_pipe_format instead of vk_format_to_pipe_format

  • radv: Add radv_format_description to remap 10/12bit formats to 16bit

  • radeonsi/vcn: Disable AV1 unidir compound with rate control

  • egl/x11: Fix swap interval setup

  • radv/video: Fix msg header total size

  • radv/video: Fix encode session info for VCN3+

  • radv/video: Use ac_vcn_enc_init_cmds

  • radv/video: Always enable B pictures for H264 encode

  • radeonsi/vpe: Don’t try to flush cs from buffer_map

  • radeonsi/vcn: Don’t try to flush cs from buffer_map

  • radeonsi/uvd,vce: Don’t try to flush cs from buffer_map

  • radeonsi/video: Remove cs argument from si_video_resize_buffer

  • radeonsi/vpe: Use float division to get scaling ratio

  • radeonsi/vpe: Fix process_frame return value

  • radeonsi/vpe: Use studio range for YUV and full for RGB by default

  • radeonsi/vcn: Fix decode target index for H264 interlaced streams

David Tobolik (1):

  • rusticl/feat: LinkOnce ODR

Dmitry Baryshkov (4):

  • freedreno/registers: allow skipping the validation

  • meson: add freedreno (turnip) Vulkan to arm64 defaults

  • meson: disable SIMD blake optimisations on x32 host

  • mesa-clc: add an option to force inclusion of OpenCL headers

Dmitry Osipenko (1):

  • virtio/vpipe: Correct vdrm_vpipe_connect() definition

Dudemanguy (1):

  • treewide: remove unneeded executable bit in non-scripting files

Dylan Baker (6):

  • intel: output a depfile with mesa_clc

  • iris: Correctly set NOS for geometry shader state changes

  • iris: fix handling of GL_*_VERTEX_CONVENTION

  • intel/tools: deduplicate zlib_inflate function

  • intel/tools: move ascii85_decode to common code

  • intel/decoder: free memory in error case

Ella Stanforth (4):

  • v3d/compiler: Implement load_output

  • v3d: enable framebuffer fetch

  • v3d: Fix fbfetch with discards.

  • v3d/compiler: Fixup output types for all 8 outputs

Emma Anholt (6):

  • mesa/ffvs: Skip doing redundant stores of .xyz when doing lighting calculation.

  • ci/anv: Enable testing with Vulkan video encode/decode.

  • docs: Drop some weird unhelpful text about DRI2.

  • egl: Apply autopep8.

  • egl: Retire NOK_swap_region support.

  • egl: Retire NV_post_sub_buffer support.

Emmanuel Gil Peyrot (1):

  • panvk: Initialize out array with the correct length

Eric Engestrom (159):

  • VERSION: bump to 25.1

  • docs: reset new_features.txt

  • docs: update calendar for 25.0.0-rc1

  • docs/android: drop libglapi.so now that it’s gone

  • ci/cuttlefish: drop `rm libglapi.so` now that it’s no longer loaded

  • gfxstream: drop unnecessary semi-colons

  • gfxstream: fix signedness of shifts

  • gfxstream: drop dead variables

  • gfxstream: use `range` variable for its intended purpose

  • gfxstream: mark unused variables as such

  • docs: update calendar for 25.0.0-rc2

  • llvmpipe/tests: include math.h for INFINITY

  • ci: don’t run on tag pipelines

  • ci: drop redundant condition

  • ci: only trigger the CI for release managers when pushing to staging branch

  • ci: run containers builds on staging branches

  • ci/yaml-toml-shell-py-test: don’t run on post-merge pipelines

  • ci/yaml-toml-shell-py-test: run on direct push pipelines

  • ci: finish sorting vars

  • ci: rename generate-env.sh to export-gitlab-job-env-for-dut.sh

  • docs: update calendar for 25.0.0-rc3

  • ci: debian-testing-ubsan is used by tests

  • llvmpipe/ci: drop fraction for asan tests that takes 1.5 min without fraction

  • docs: add release notes for 25.0.0

  • docs: add sha sum for 25.0.0

  • docs: update calendar for 25.0.0

  • docs: add release calendar for 25.0.x cycle

  • docs/releasing: fix “release schedule” subsections nesting

  • docs/release-calendar: add 25.1 branchpoint & rc dates

  • ci/alpine: pin the release to avoid random unexpected changes

  • ci/alpine: control wayland & wayland-protocols versions

  • ci: move shader-db test job from build jobs yaml to test jobs yaml

  • ci/build: remove a couple of unnecessary “override needs: to the same value”

  • ci/build: move .use-debian/x86_64_build out of the generic .meson-build and into the debian/x86_64 jobs

  • ci/build: split meson-build into build-for-tests and build-only

  • ci/build: add explicit build-for-tests or build-only to all jobs

  • ci/build: lower the delay to start build-only jobs

  • docs: update gitlab docs urls

  • vtn_bindgen2: keep the printf blob local to avoid LTO issues

  • lavapipe/ci: add vkd3d job

  • ci/build: build-test the dri2 code

  • ci/b2c: fix comment location

  • ci/b2c: explain better why we don’t clone mesa

  • ci/b2c: use B2C_JOB_TEMPLATE directly

  • ci/b2c: set default value for B2C_MACHINE_REGISTRATION_IMAGE in the job

  • ci/b2c: set default value for B2C_TELEGRAF_IMAGE in the job

  • ci/b2c: set default value for B2C_KERNEL_CMDLINE_EXTRAS in the job

  • ci/b2c: split B2C_JOB_VOLUME_EXCLUSIONS in the jinja template

  • ci/b2c: rename IMAGE_UNDER_TEST to B2C_IMAGE_UNDER_TEST

  • ci/b2c: pass through all the B2C_* variable without renaming them

  • ci/b2c: pass through all the CI_* variables as well

  • ci/b2c: set the registry proxy from the job

  • ci/b2c: use more readable “long” argument names

  • v3d/ci: mark traces humus/AmbientAperture and humus/DynamicBranching3 as flaky

  • docs: update calendar for 25.0.1

  • docs: add release notes for 25.0.1

  • docs: add sha sum for 25.0.1

  • wsi/x11: drop misleading reference to anv in var names

  • meson: simplify video-codecs option parsing

  • ci/deqp: backport fix for dEQP-VK.binding_model.buffer_device_address.*

  • meson: announce that clover is deprecated (slated for removal)

  • lvp/ci: document fixed tests

  • lvp/ci: skip tests that are timing out (>1 min)

  • lvp/ci: fix sorting of flakes

  • lvp/ci: remove duplicate flakes (noticed after sorting)

  • lvp/ci: document flakes seen over the last week

  • freedreno/ci: document fixed test

  • freedreno/ci: document flakes seen over the last week

  • nvk/ci: fix sorting of flakes

  • nvk/ci: document flakes seen over the last week

  • zink+nvk/ci: document new failures

  • zink+nvk/ci: fix sorting of flakes

  • zink+nvk/ci: document flakes seen over the last week

  • radv/ci: document flakes seen over the last week

  • zink+radv/ci: fix sorting of flakes

  • zink+radv/ci: document flakes seen over the last week

  • ci: document http proxy bug & disable farms relying on it

  • ci/container: fix image tags comment in trampoline script

  • ci/init-stage2: document that only lava jobs upload results to s3

  • ci/build: no need to list all the files that will go into the uploaded artifact

  • ci/piglit: drop usage of s3cp for a simple download

  • ci: always abort if the curl download fails

  • ci/baremetal: make sure we can follow redirects on s3 downloads

  • ci: do a regular GET request for /done files, instead of HEAD

  • ci: replace broken s3cp command with a simple curl call

  • ci: bump image tags

  • ci: drop placeholder-job tags to allow jobs to run

  • ci: document new llvmpipe & softpipe failures since the migration

  • ci: run shader-db & zink-lvp on kvm runners

  • ci: disable llvmpipe & virgl traces jobs

  • meson: do not compile libisaspec unless used

  • meson: do not compile libblake3 unless used

  • meson: do not compile libmesa_util_sse41 unless used

  • pick-ui: clean up formatting

  • pick-ui: fix enum value in test expectation

  • pick-ui: add missing field in test expectation

  • pick-ui: rename s/out/commit_message/ variable to make its contents clearer

  • pick-ui: fix parsing of multiple `backport-to:` lines

  • docs: update calendar for 25.0.2

  • docs: add release notes for 25.0.2

  • docs: add sha sum for 25.0.2

  • rpi/ci: disable traces jobs

  • ci: re-enable igalia (rpi) farm

  • ci: document who are the farm admins

  • docs: fix last references to gallium-drivers=swrast

  • ci: remove last uses of deprecated `swrast` alias for softpipe+llvmpipe

  • meson: drop deprecated `swrast` alias for softpipe+llvmpipe

  • meson: move special value `all` out of the middle of the list

  • rpi5/ci: sort flakes

  • rpi5/ci: drop duplicate flakes

  • ci: bump apitrace version

  • ci: drop packet.net tag on git archive job

  • ci: move aarch64 tag to .use-debian/arm64_build

  • ci: move android’s kvm requirement to .use-debian/x86_64_test-android

  • ci: add FDO_RUNNER_JOB_PRIORITY_TAG_* to control priority of generic freedesktop runners

  • docs: remove the last 24.3 releases

  • virgl: fix typo inverting a condition

  • docs: update calendar for 25.0.3

  • docs: add release notes for 25.0.3

  • docs: add sha sum for 25.0.3

  • ci/build: drop LTO from fedora build

  • ci: rename ci-tron priority tag to avoid conflict with the generic fdo runners

  • ci/fluster: use http proxy when checking for the vector files

  • ci/fluster: don’t overwrite FLUSTER_VECTORS_HOST_PATH to a different meaning

  • radeonsi/ci: update expectations

  • radv/ci: update expectations

  • freedreno/ci: update expectations

  • nvk/ci: update expectations

  • lvp/ci: update expectations

  • zink+nvk/ci: update expectations

  • zink+radv/ci: update expectations

  • zink+tu/ci: update expectations

  • zink+anv/ci: document a bunch of flaky glx tests that have been preventing merges all day

  • docs/ci: replace deprecated `pages` job with `pages: true`

  • docs/ci: add link to the website preview for convenience

  • docs/ci: add comment explaining what that long rule actually does

  • docs/ci: split pre-merge & merge pipeline rules

  • docs/ci: follow convention of only running jobs by default for Marge

  • ci: fix image tags indentation

  • ci/deqp: fix vulkan video build

  • VERSION: bump for 25.1.0-rc1

  • .pick_status.json: Update to d5ad7981401c2393cac38fc3215e8bbb97de06b9

  • .pick_status.json: Update to 1bf8542490679fa2a244e27fb2b04bbc3d122f74

  • meson: remove duplicate `deprecated` for `gallium-xa` option

  • meson: remove duplicate `deprecated` for `power8` option

  • .pick_status.json: Update to 5f3a3740dcc6d243f2ef14138fb1c09bcbb9b5fd

  • pick-ui: make `Backport-to: 25.0` backport to 25.0 *and more recent release branches*

  • .pick_status.json: Update to dd3e1190a2bdbc6b996152510407adb9a8cb5618

  • pick-ui: add missing dependency

  • [25.1 only] ci: don’t treat misleading-indentation warnings are error on alpine

  • VERSION: bump for 25.1.0-rc2

  • .pick_status.json: Update to 3493500abb78a4dc22aba14840bba5c777fde745

  • .pick_status.json: Update to eeffb4e674d10db9aefebeca91c2d87c1676b81e

  • VERSION: bump for 25.1.0-rc3

  • .pick_status.json: Update to 615d0c9669595adf114a705f5b8ee88277aa99f2

  • .pick_status.json: Update to 7f0de1a51212881c9a7614327bf3e1fbc9784ddb

  • ci: drop tracking of removed folder

  • .pick_status.json: Update to c434050a0088ec3f07d63fd1019aea541632faed

  • .pick_status.json: Update to 84b9c281fe82dd66f2552687cecb61a8e22809d0

Eric R. Smith (7):

  • panfrost: avoid potential divide by 0 calculating timer_resolution

  • panfrost: fix YUV center information for 422

  • panfrost: fix backward propagation of values in loops

  • panfrost: use an accessor function to read from bi_opcode_props

  • panfrost: consider xfb shader when calculating thread local storage size

  • panfrost,lima: use index size in panfrost minmax_cache

  • panfrost: fix transaction elimination crc valid calculation

Erico Nunes (2):

  • ci: re-enable lima farm

  • panvk: disable VK_EXT_image_drm_format_modifier for arch < 10

Erik Faye-Lund (57):

  • panvk: fix line-rasterization of bifrost

  • panvk: report strictLines as true

  • panvk/ci: add back incorrectly removed crash

  • pan/ci: add flaky tests to the flake-list

  • pan/ci: add fail from llvm 19 upgrade

  • pan/ci: add a couple of common flakes

  • panvk: correct number of read bytes for dynamic buffers

  • meson: rename meson_options.txt

  • panvk: report passing the VK CTS

  • panvk: rename helper

  • mesa/main: wire up glapi bits for EXT_multi_draw_indirect

  • pan/bi: use unreachable instead of DBG + assert

  • pan/bi: remove unused debug output

  • pan/genxml: rename field

  • panfrost: respect pipe_rasterizer_state::line_rectangular

  • panvk: disable shaderFloat16 on bifrost

  • docs/features: add missing panvk feature

  • docs/features: add VK_EXT_hdr_metadata

  • panvk: fix extension requirement

  • panvk: rework how we deal with extension-reqs

  • panvk: expose VK_KHR_display

  • panvk: expose EXT_display_control

  • panvk: correct VkPhysicalDeviceProperties::deviceName

  • panvk: enable KHR_line_rasterization support

  • panvk: add basic driconf infrastructure

  • panfrost: avoid accidental aliasing

  • panfrost: fix overflow-debugging

  • panfrost: use real array for panfrost_emit_plane

  • panvk: check for texture-compression support

  • panvk: expose textureCompressionBC when supported

  • mesa/main: fix regression in extension-checking

  • panvk/ci: disable some more slow tests

  • docs/features: update panvk support

  • panfrost: fixup typo in 16x sample-pattern

  • panfrost: correct tile-buffer size for some v7 GPUs

  • panvk/ci: move timeouts to crash

  • panfrost: properly align value

  • panfrost: allocate tile-buffer for dummy render-targets

  • panfrost: disable tile-pipelining when needed

  • panfrost: add color-attachment and msaa helpers

  • panvk: enable 8x and 16x msaa when supported

  • panvk: enable sampledImageIntegerSampleCounts for all MSAA formats

  • nir/lower_tex: use texture_mask instead of shifting on use

  • nir/lower_tex: avoid undefined-behavior

  • panvk: set shared_addr_format

  • panvk: enable KHR_spirv_1_4 on v10+

  • panvk: claim official conformance on v10

  • docs/panfrost: use anonymous hyperlinks

  • panvk: enable dualSrcBlend

  • docs/panvk: fixup extension support

  • docs/panvk: remove disabled extension

  • docs/panvk: fixup docs around float controls

  • docs/panvk: add VK_-prefix for extension name

  • docs/panvk: document ycbr in terms of extensions

  • docs/panvk: document EXT extension aswell

  • docs/panvk: add missing new features

  • panvk: support vulkan 1.2 on v10+

Ernst Persson (1):

  • intel/vulkan: Add bvh build dependency

Faith Ekstrand (165):

  • nvk,nak: Only use u64 texture handles with codegen

  • nvk: Only pass sampler handles when needed

  • nak: Add support for bound and cbuf textures

  • nak: Optimize bindless to cbuf textures on Volta+

  • nak: Fix cbuf textures

  • nak: Stop setting .EF on tex ops

  • nak: Rename MemEvictionPriority::Unchanged to LastUse

  • nak: Add more MemEvictionPriorities

  • nak: Print .dc for OpTld4::z_cmpr

  • nak: Add MemEvictionPriorities to tex ops

  • nvk: Fix scissor bounds

  • nvk: Fix a typo in a comment

  • nak/repair_ssa: Use a worklist for get_ssa_or_phi()

  • nvk: Rename nvk_descriptor_set::mapped_ptr

  • nvk: Respect VK_DESCRIPTOR_POOL_CREATE_HOST_ONLY_BIT_EXT

  • nvk: Implement descriptorBufferPushDescriptors

  • nvk: Pull shaders from the state command buffer in nvk_cmd_process_cmds()

  • nvk: Handle shader==NULL in nvk_cmd_upload_qmd()

  • nvk: Allow sparse loads on EDB buffers

  • nak: Handle sparse texops with unused color destinations

  • nvk: Use suld for EDB uniform texel buffers

  • nvk: Align UBO/SSBO addresses down rather than up

  • nak: Use suld.constant when ACCESS_CAN_REORDER is set

  • nvk: Use suld.constant for EDB uniform texel buffers

  • nvk: Constify instance and pdev pointers

  • ci: Remove some NVK vkd3d fails

  • nak: Only use suld.constant on Ampere+

  • nak: Use MemScope::GPU instead of MemScop::System

  • zink: Use the correct array size for signal_values[]

  • zink: Use persistent semaphores for PIPE_FD_TYPE_SYNCOBJ

  • nvk/nvkmd: Fix logging of VA bind addresses

  • nvk: Don’t bind a fragment shading rate image pre-Turing

  • nvk: Do not set INVALIDATE_SKED_CACHES pre-MaxwellB

  • nak: Handle tex ops with only one source

  • nak/nir: Don’t provide dummy backend2 tex srcs

  • nvk: Fix indentation in begin_end_query()

  • nouveau/class_parser: Make strided element functions const

  • nak/qmd: Drop some unnecessary .try_into().unwrap()

  • nak/qmd: Add a nak_get_qmd_cbuf_desc_layout() helper

  • nvk: Handle pre-Turing dispatch indirect commands

  • nvk: Only support deviceGeneratedCommandsMultiDrawIndirectCount on Turing+

  • nvk: Only support compute shader derivatives on Turing+

  • nak/nir: Re-materialize load_const instructions in use blocks

  • nvk/image: Prefer vk_image values over pCreateInfo

  • nvk/image: Drop some unneeded initializers

  • nvk: Fix capitalization of statistics

  • nak: Fix NAK_DEBUG=spill for large FS outputs

  • nak: Handle any->Mem parallel copies

  • nak: Add a new ConstTracker struct

  • nak: Don’t spill/fill const values

  • compiler/rust: Add u_printf_info to the rust bindings

  • nir: Add a get_io_index_src() helper

  • nir: Add a nir_opt_tex_skip_helpers optimization

  • nak: Set .NODEP on tex ops based on nir_opt_tex_skip_helpers()

  • zink: Don’t present to Wayland surfaces asynchronously

  • zink: Revert “zink: enable single-plane modifiers for generic 2D exports”

  • egl/dri2: Rework get_wl_surface_proxy()

  • egl/wayland: Pass the original wl_surface to kopper

  • util/box: Add a intersect_2d helper

  • iris: Use pipe_box helpers for damage calculations

  • zink: Use pipe_box helpers for damage calculations

  • vulkan: Add device address helpers to vk_buffer

  • nvk: Use the new buffer device address infrastructure

  • panvk: Use the new buffer device address infrastructure

  • hk: Use the new buffer device address infrastructure

  • vulkan/meta: Use vk_buffer.device_address directly

  • zink: Set needs_barrier after transitioning to QUEUE_FAMILY_FOREIGN

  • zink: Check queue families when binding image resources

  • spirv: OpAsmTargetINTEL is untyped

  • spirv: Update the JSON and headers

  • vulkan: Update XML and headers to 1.4.309

  • nouveau/winsys: Stop asserting that imported BOs are aligned

  • nvk: Allow rendering to linear images with unaligned strides

  • nil: Relax alignment requirements for linear images

  • nil: Split linear and tiled image creation

  • nvk,nil: Stop panicing in image creation

  • vtn: Support cooperative matrices in OpConstantNull

  • loader/nouveau: Fix the comment in nouveau_zink_predicate()

  • egl/x11: Re-order an if statement

  • egl/kopper: Update the EGLSurface size after kopperSwapBuffers()

  • nak: Insert the annotation in the right spot in assign_regs

  • nak: Don’t insert empty OpParCopy in assign_regs

  • nak: Always copy sources when handling vec/pack/mov ops

  • nak: Fix a SM check for OpPCnt

  • nak: Check num_regs(UGPR) instead of SM version

  • nak: Turing starts at SM73

  • nouveau/headers: Refactor class_parser

  • nouveau/headers: Drop unused Rust constants

  • nouveau/headers: Drop double-underscore from Rust names

  • nouveau/headers: Re-use Rust method types when possible

  • nvk: Reduce the size of nvk_image_view_capture

  • nvk: Free owned_gart_mem correctly

  • nvk: Fix a Volta check

  • nvk: Disable VK_EXT_post_depth_coverage on Maxwell A and earlier

  • nvk: Allocate QMDs from a heap on Maxwell A and earlier

  • nvk: Disable VK_EXT_device_generated_commands on Maxwell A and earlier

  • nvk: Don’t set filterMinmax properties prior to Maxwell B

  • nvk: Disable sparse buffer binding prior to Maxwell B

  • nouveau/mme/fermi: Don’t allow STATE and EMIT on the same op

  • nvk: Use the right sample mask for 8x/4pass on Maxwell A

  • nvk/nvkmd: Add a concept of incomplete pushes

  • nvk: Mark the push before an indirect push as incomplete

  • nak: hsetp2 and dsetp are slower on Volta

  • nvk: Bump the conformance version to 1.4.1.3

  • vulkan/wsi: Signal buffer memory object when blitting

  • venus: Assume wsi_mem->base_bo != NULL

  • venus: Don’t report global priorities if globalPriorityQuery is unsupported

  • venus: Only claim modifiers in WSI if the host driver supports it

  • venus: Set wsi_device::supports_scanout = false

  • compiler/rust: Add a nir_alu_type wrapper

  • compiler/rust: Add more NIR intrinsic getters

  • nak: Implement nir_intrinsic_convert_alu_types

  • nak/nir: Use correct rounding for fp64 -> fp16 conversions

  • nak,nir: Generalize nak_nir_split_64bit_conversions and move it to NIR

  • nak: Move some calc_instr_deps items to a new file

  • nak: Box our RegTrackers

  • nak: Improve WS abstractions in hw_runner

  • nak: Add a QMD heap to hw_runner

  • nak: Disable lea64 and f2fp.pack_ab tests pre-Volta

  • nvk: Disable vulkanMemoryModel on Kepler and earlier

  • nvk: Use max_image_dimension for maxFramebufferWidth/Height

  • nvk: Disable 32k images on Pascal A

  • nak: Move has_fixed_latency to Op

  • nak: Add and use a ShaderModel::needs_scoreboard() helper

  • nak: Add latency helpers to ShaderModel and use them

  • nak: Move SM70 encoding and legalize to a separate file

  • nak: Move exec_latency into the per-SM files

  • nak: Move latency information into the per-SM files

  • nak: Move sched_common.rs to reg_tracker.rs

  • nak: Add GPU generation helpers

  • nak: Use is_volta() instead of sm == 70

  • nak: Put the cycle count assert behind a debug flag

  • nak: Handle delays > 15

  • nak: Add an Op::no_scoreboard() helper

  • intel/compiler: Use nir_split_conversions()

  • nak: Add a ChannelMask type

  • nak: Add support for suld/st.b

  • nak,nir: Add an image_load_raw_nv intrinsic

  • nak: Use suld.b on Kepler if we have a format

  • nak: Allow predicates in nir_intrinsic_as_uniform

  • nak: Add a NAK_DEBUG=panic option

  • nvk: Call vk_device_finish() last in nvk_DestroyDevice()

  • nvk/nvkmd: Check the correct flag for the Kepler GART workaround

  • nil: Multiply by array_stride_B instead of adding

  • nak/hw_tests: Feed predicate/carry sources with 0/1 data

  • nak: Add a plop2 test

  • nak: Add False and True to IntCmpOp

  • nak: Lower texture inputs for Kepler B

  • nak/legalize: Add a helper for lowering ineg

  • nak: Add stubs for Kepler B

  • nak: Add stubs for Fermi and Kepler A

  • nak: Move some legalization helpers from sm50 to common code

  • nak/sm50: Add zero_reg() and true_reg() helpers

  • nak/sm70: Add zero_reg() and true_reg() helpers

  • nak: Get rid of RegRef::zero

  • nvk: Disable VK_EXT_descriptor_buffer pre-Maxwell

  • nak/qmd: Rework cbuf size suffix handling

  • nak/qmd: Add support for shifted cbuf addresses

  • nak/qmd: Add QMD version 4.0 for Hopper

  • nvk: Handle shifted QMD cbuf addrs in indirect command processing

  • nak/legalize: Take a RegFile in copy_alu_src_and_lower_fmod

  • nak/legalize: Take a RegFile in copy_alu_src_and_lower_ineg()

  • nak/sm70: Fix the bit74_75_ar_mod assert

  • nvk: Maxwell+ is now conformant

  • nak: Set lower_pack_64_4x16

Felix DeGrood (9):

  • vk/overlay-layer: fix regression in non-control pathway

  • intel/brw: support for dumping shader line numbers

  • anv: add INTEL_DEBUG=shaders-lineno

  • iris: add INTEL_DEBUG=shaders-lineno

  • drm-uapi: add eu_stall uapi

  • intel/perf: remove unnused argument from xe_perf_stream_read_error

  • intel/perf: add eu stall sampling support

  • util: add hash functions for u64 data type

  • intel/tools: create intel_monitor for sampling eu stalls

GKraats (2):

  • x11: give error messages if Xorg only supports DRI2 and mesa only DRI3

  • EGL: legacy-x11=dri2 should support hardware driver

Ganesh Belgur Ramachandra (1):

  • amd: use 128B compression for scanout images when drm.minor <63

Georg Lehmann (83):

  • nir/lower_poly_line_smooth: don’t emit control flow

  • nir/lower_poly_line_smooth: only smooth first color target

  • nir/lower_poly_line_smooth: support partial store_output

  • radv: remove radv_should_lower_poly_line_smooth

  • radv: inline radv_nir_lower_poly_line_smooth

  • nir/lower_poly_line_smooth: don’t reject fp16

  • nir/lower_poly_line_smooth: use intrinsics_pass

  • nir/opt_move: don’t move into critical sections

  • ac/nir/lower_ps: move exports after packing alu

  • nir/print: print large floats as mantissa + exponent

  • nir: range analysis for ffract

  • nir: fix range analysis for frcp

  • nir: fix frsq range analysis

  • nir: improve fsqrt range analysis

  • nir/opt_algebraic: optimize ffract(ffract(a))

  • nir/peephole_select: support demote for non CF HW

  • nir/peephole_select: handle demote and terminate in nir_opt_collapse_if

  • nir/peephole_select: don’t special case nir_opt_collapse_if + limit = ~0

  • nir/peephole_select: don’t include nir_search_helpers.h

  • nir/peephole_select: add options struct

  • nir/peephole_select: add option to allow discard without ~0 limit

  • nir/peephole_select: don’t completely ignore ifs with dont_flatten

  • nir: replace nir_opt_conditional_discard with nir_opt_peephole_select

  • radv: remove separate discard peephole select

  • nir/opt_algebraic: optimize b2f(a != 0) * a

  • nir/search_helpers: look through vecs in is_only_used_as_float

  • nir/search_helpers: check tex source type in is_only_used_as_float

  • nir/builder: add nir_shader_phi_pass

  • nir/opt_phi_precision: use nir_shader_phi_pass

  • nir/opt_remove_phis: use nir_shader_phi_pass

  • aco/insert_exec: fix continue_or_break on gfx6-7

  • nir: add a pass to optimize phis to 1bit

  • nir/opt_algebraic: optimize ineg(a) == ineg(b)

  • nir/opt_algebraic: optimize ineg(a) == #b

  • nir/opt_algebraic: 0 >= a -> 0 == a

  • nir/opt_algebraic: optimize DXBC boolean bcsel

  • nir/opt_algebraic: optimize more boolean bcsel with constants

  • nir/opt_algebraic: optimize dxbc boolean not

  • nir/opt_algebraic: optimize constant shift of DXBC booleans

  • nir/opt_algebraic: optimize b2i(a) != -b2i(b)

  • radv: use nir_opt_phi_to_bool

  • nir/opt_varyings: clean up nir_progress usage

  • radv/nir_apply_pipeline_layout: clean up progress handling

  • radv/nir_lower_ray_queries: use nir_foreach_function_impl

  • nir/opt_algebraic: optimize bit_count(a) != 0

  • nir/opt_algebraic: optimize bcsel of b2f and constants

  • nir/opt_algebraic: optimize b2i/b2f comparision with non 0/1 constants

  • nir/opt_algebraic: optimize ~a == ~b and ~a == #b

  • nir/opt_algebraic: push comparisons with constants into bcsel with constant

  • nir/opt_algebraic: optimize more ine/ieq(umin(b2i, ), 0)

  • nir/opt_algebraic: optimize d3d a ? b : 0

  • aco/optimizer: delete combine_and_subbrev

  • radv: remove outdated vectorize TODO

  • ac/nir/mem_access_bit_sizes: split unaligned vec3 lds access to allow more read2/write2

  • aco/opt_postRA: split try_optimize_scc_nocompare in two functions

  • aco/opt_postRA: allow try_optimize_scc_nocompare for all instructions

  • aco/opt_postRA: remove scc != 0 with multiple uses

  • aco/opt_postRA: remove scc == 0 for more opcodes

  • aco/isel: use s_mul_i32 instead of s_cselect_b32 for a ? b : 0

  • radv: enable invariant geom for DOOM(2016)

  • radv: add dcc_decompress_gfx11 in radv_graphics_state_key

  • ac/nir/lower_ps_late: consider dcc decompression for null exports

  • radv/gfx10+: remove null exports if discard isn’t used

  • aco: don’t assume that v_interp_mov_f32 flushes denorms

  • aco/gfx11.5: remove vinterp ddx/ddy path

  • aco/validate: fix scalar source validation for DPP and gfx11+ VINTERP

  • nir/opt_algebraic: create ubfe from (a & mask) >> c

  • aco/ra: disallow vcc definitions for pseudo scalar trans instrs

  • nir: add option to keep mul24_relaxed

  • aco: implement mul24_relaxed

  • ac/llvm: support mul24_relaxed

  • ac/nir: set has_mul24_relaxed

  • aco/insert_exec: reset exec temporary after combined p_demote + p_end_wqm

  • spirv: clamp/sign-extend non 32bit ldexp exponents

  • aco/gfx9+: use d16 global/scratch/buffer loads

  • spirv: fix cooperative matrix by value function params

  • aco/gfx10: simpler solution to avoid store instructions in clauses

  • aco: form mixed MTBUF/MUBUF clauses

  • nir/opt_algebraic: optimize open coded ffract

  • nir/opt_algebraic: disable fsat(a + 1.0) opt if a can be NaN

  • aco: set opsel_hi to 1 for WMMA

  • aco/insert_exec: only restore wqm mask after control flow if necessary

  • aco/insert_exec: reset temporary when recreating wqm mask from exact mask

Gert Wollny (1):

  • r600/sfn: gather info and set lowering 64 bit after nir_lower_io

Giovanni Mascellani (2):

  • llvmpipe: Remove an outdated comment about subclassing pipe_screen.

  • lvp: Remove some dead code.

Guilherme Gallo (30):

  • ci: Properly clean up rustup

  • ci: Remove cargo symlink workaround

  • ci/android: add missing pre/post build scripts

  • ci: setup-test-env: Prefer functions over aliases

  • ci: add _error_msg for internal messaging

  • ci: add support for structural tagging

  • ci: copy structural tag files to rootfs

  • ci/angle: add structured tag check to ANGLE build time

  • ci/angle: test-time structured tag checks

  • ci/angle: condense angle variables in one job

  • ci/angle: remove USE_ANGLE variable

  • ci/docs: add structured tagging documentation

  • ci/lava: Drop the repeating quotes on lava-test-case

  • ci/lava: Propagate errors in SSH tests

  • ci/lava: xtrace the lava_job_submitter call

  • ci/lava: Add U-Boot action timeout for rockchip DUTs

  • ci/lava: Properly detect VMWARE farm

  • ci: Specify the FARM variable for DUT jobs

  • ci: Simplify LAVA farm detection

  • ci/lava: Remove depthcharge-start timeout

  • ci/lava: Split boot action into deploy and boot

  • ci/lava: Tweak timeouts

  • ci/lava: Don’t print empty lines when changing sections

  • Revert “ci: setup-test-env: Prefer functions over aliases”

  • ci/bin: update_tag: improve tag load

  • ci/update_tag: fix linter errors

  • ci/lava: Fix LAVA lima jobs

  • ci/lava: Fix LAVA lima jobs

  • ci/lava: Fetch kernel modules from overlay

  • ci: Add some unit tests for the duration field

Gurchetan Singh (5):

  • gallium: drop const qualifier on return type

  • lavapipe: use quotes instead of angle bracket

  • gfxstream: check device exists before using it

  • gfxstream: refactor device initialization

  • gfxstream: follow the semantics desired by distro VK loader

Hans-Kristian Arntzen (3):

  • radv: Always allow sparse on normal GFX/COMPUTE/DMA queues.

  • radv: Repurpose radv_legacy_sparse_binding drirc

  • radv: Always set 0 dispatch offset for indirect CS.

Hyunjun Ko (12):

  • anv: Fix to set CDEF flter flag correctly for AV1 decoding

  • anv/video: clean-up duplicated code.

  • dri: fix a build error

  • kopper: implement to get sync values.

  • anv: fix maxDpbSlots and maxActiveReferencePictures for AV1 decoding.

  • anv: Add one more flag of VideoCapability for encoding.

  • anv: Do not support the tiling of DRM modifier if DECODE_DST

  • anv/ci: remove some expected failures of dEQP-VK.video.formats.*

  • vulkan/video: Do byte-alignment when building a h264 slice header

  • anv: Add stdSyntaxFlag values for h264/5 encoders

  • anv: Move rateControlMode to the video session.

  • anv: Use vk_video_derive_h265_scaling_list

Iago Toral Quiroga (9):

  • v3dv: implement sync debug option

  • v3dv: serialize jobs after any barrier when debug sync is set

  • v3dv: fix missing access bit flag when checking for texel buffer reads

  • mesa: fix RGBA_SIGNED_COMPONENTS for lowered signed luminance

  • v3dv: fix crash on 32-bit builds

  • v3dv: rename v3dv_cmd_buffer_merge_barrier_state

  • v3dv: make cmd_buffer_serialize_job_if_needed take a barrier state

  • v3dv: improve handling of trailing barriers

  • pan/va: fix FAU validation

Ian Romanick (38):

  • iris: Add missing nir_metadata_preserve in iris_lower_storage_image_derefs

  • crocus: Add missing nir_metadata_preserve in crocus_lower_storage_image_derefs

  • iris: Use nir_shader_intrinsics_pass in iris_lower_storage_image_derefs

  • crocus: Use nir_shader_intrinsics_pass in crocus_lower_storage_image_derefs

  • brw/copy: Fix handling of offset in extract_imm

  • brw/copy: Use extract_imm in try_constant_propagate_value

  • brw/copy: Allow constant propagation of some 64-bit integers

  • nir/algebraic: More (a == 0 || a == 1 || …) patterns

  • nir/algebraic: Optimize zero comparisons of umax or umin

  • nir/algebraic: Simplify equality comparisons of b2T with 1 or 0

  • nir/algebraic: Undistribute b2i from logic-ops

  • brw/print: Don’t let SHADER_OPCODE_FLOW affect indentation

  • brw: Fix typo in comment

  • brw/nir: Lower fsign again after last call to brw_nir_optimize

  • brw/opt: Move non-SSA register accounting after first brw_opt_split_virtual_grfs

  • brw: Add basic infrastructure for load_reg pseudo op

  • brw/copy: Prepare copy_propagation for load_reg

  • brw/coalesce: Prepare brw_opt_register_coalesce for load_reg

  • brw/algebraic: Constant folding for BROADCAST and SHUFFLE

  • brw: Add passes to generate and lower load_reg

  • brw/sat: Convert tests to use load_reg

  • brw/sat: Eliminate non-defs saturate propagation

  • brw/opt: Don’t call brw_opt_copy_propagation before brw_lower_load_reg

  • brw/nir: Fix source handling of nir_intrinsic_load_barycentric_at_offset

  • brw/nir: Eliminate default parameter to get_nir_src

  • brw/algebraic: Optimize derivative of convergent value

  • brw/copy: Refactor source modifier type checking

  • brw/copy: Copy prop -X into Y&1

  • brw/nir: Optimize b2f(not(X)) using logical operations instead of arithmetic

  • brw/nir: Allow b2f(not(X)) optimization on Gfx12.5+

  • brw/nir: Use offset() for all uses of offs in emit_pixel_interpolater_alu_at_offset

  • nir/algebraic: Allow fmin(a,a) optimization when flush denorm to zero is not set

  • brw/algebraic: Clear condition modifier on optimized SEL instruction

  • brw/algebraic: Don’t optimize float SEL.CMOD to MOV

  • elk/algebraic: Clear condition modifier on optimized SEL instruction

  • elk/algebraic: Don’t optimize float SEL.CMOD to MOV

  • brw/cmod: Fix some errors when propagating from CMP to ADD.SAT

  • brw/cmod: Don’t propagate from CMP to possible Inf + (-Inf)

Ivan A. Melnikov (1):

  • gallium/radeon: Make sure radeonsi PCI IDs are also included

Ivan Avdeev (2):

  • radv: add experimental support for AMD BC-250 board

  • radv,radeonsi: disable compute queue for BC250

Iván Briano (2):

  • anv: handle REMAINING_LAYERS in host image copy cases

  • brw: make HALT instruction act as barrier in new CSE pass

James Hogan (8):

  • glsl: Expose gl_ViewID_OVR back to GLSL 1.30

  • mesa: Fix multiview attachment completeness check

  • mesa: Fix FramebufferTextureMultiviewOVR num_views check

  • mesa: Consider NumViews to reuse FBO attachments

  • mesa: Handle GL_FRAMEBUFFER_INCOMPLETE_VIEW_TARGETS_OVR

  • mesa: Check views don’t exceed GL_MAX_ARRAY_TEXTURE_LAYERS

  • mesa: OVR_multiview framebuffer attachment parameters

  • mesa: Handle getting GL_MAX_VIEWS_OVR

Jan Alexander Steffens (heftig) (1):

  • gfxstream: Use proper log format for 32-bit Vulkan

Janne Grunau (8):

  • hk: Replace alloca with malloc in queue_submit

  • hk: Use rowPitch from VkImageDrmFormatModifierExplicitCreateInfoEXT

  • venus: Do not use instance pointer before NULL check

  • venus: virtgpu: Require stable wire format

  • asahi: build asahi_clc for -Dtools=asahi

  • asahi: Drop unnecessary idep_mesaclc dependency

  • panfrost: build panfrost_compile for -Dtools=panfrost

  • ci: Switch cross-builds to ‘-D tools=panfrost’

Jason Macnak (6):

  • gfxstream: Move snapshot decoder replay into VkDecoderGlobalState

  • gfxstream: Remove unused handling mappers

  • gfxstream: Move the handle replay buffer into BoxedHandleManager

  • gfxstream: Remove duplicated boxed handle func declarations

  • gfxstream: Update variable names to avoid -Wshadow error

  • gfxstream: Remove extra dispatch variable

Jeongik Cha (1):

  • gfxstream: Add AHARDWAREBUFFER_FORMAT_B8G8R8A8_UNORM in android_format_is_yuv

Jesse Natalie (2):

  • meson: Enable /Zc:preprocessor for MSVC

  • CI/Windows: Update container deps

Job Noorman (46):

  • freedreno/drm-shim: enable raytracing

  • ir3: fix emitting descriptor prefetches at end of preamble

  • ir3: add braces around complex if/else block

  • ir3/ra: handle phis with preferred regs first

  • ir3/parser: add helper to generate syntax errors based on gen

  • ir3/isa: fix (dis)asm of ldg.a/stg.a on a6xx

  • ir3: don’t create SRC2 for isam without .v

  • ir3/legalize: use (sy) for ray_intersection WAR hazards

  • ir3/lower_tess: make all NIR passes report progress

  • ir3: don’t use deprecated NIR_PASS_V anymore

  • ir3: reformat after previous commit

  • ir3/opt_prefetch_descriptors: fix crash after nir_progress rewrite

  • ir3: add reformatting commits to .git-blame-ignore-revs

  • nir/lower_phis_to_scalar: remove unused mem_ctx

  • nir/lower_phis_to_scalar: use nir_builder API where possible

  • nir/lower_phis_to_scalar: don’t create moves for undef sources

  • nir/lower_subgroups: use build_cluster_mask for quad mask

  • ir3/ra: prevent reusing parent interval of reloaded sources

  • ir3: clear instruction uses when cloned

  • ir3/sched: unblock a0.x/a1.x after last use

  • ir3: add ir3_cursor_current_block helper

  • ir3/cse: add support for mov a0.x/a1.x

  • ir3: remove hash table for a1.x

  • ir3: add helper to create STC

  • ir3: fix false dependencies of rpt instructions

  • ir3/sched: handle dependencies between stc and const reads

  • ir3: split immediate state from rest of const state

  • ir3: make const_imm_index_to_reg helper public

  • ir3: fix max const size calculation for the binning pass

  • ir3: lower immediates to const regs in preamble on a7xx

  • ir3: keep inputs at start block when creating empty preamble

  • ir3/legalize: fix off-by-one error in kill_sched

  • ir3/legalize: take wrmask into account for delay updates

  • ir3: don’t sync every TCS/GEOM block

  • ir3: run opt_if after opt_vectorize

  • ir3: make shpe a terminator

  • ir3/ra: assign interval offsets to new defs after shared RA

  • ir3: add ir3_aggressive_coalesce helper

  • ir3/ra: create merge sets for splits/collects inserted for shared RA

  • ir3/opt_preamble: take alias.rt into account for rewrite cost

  • ir3: remove spaces in shader stats

  • ir3/cp: add option to disable immediate to const lowering

  • ir3/cp: ignore alias sources for sam.s2en

  • ir3: run cp after ir3_imm_const_to_preamble

  • ir3/ra: add helper for getting a dst interval

  • ir3/ra: ignore phis handled by shared RA

John Anthony (2):

  • panvk: Avoid division by zero for vkCmdCopyQueryPoolResults

  • panvk: Enable VK_EXT_direct_mode_display

Jon Hunter (1):

  • freedreno/registers: Fix gen_header.py for older python3 versions

Jordan Justen (5):

  • intel/dev: Add BMG PCI IDs (0xe210, 0xe215, 0xe216)

  • intel/dev: Stop checking hwconfig values at driver runtime

  • tools/intel_dev_info: Print hwconfig discrepancies

  • intel/dev: Ignore hwconfig difference due to WA 18040209780

  • intel/dev: Add BMG 0xe211 PCI ID

Jose Fonseca (1):

  • glapi: Make _GLAPI_EXPORT a no-op on Windows.

Jose Maria Casanova Crespo (3):

  • v3dv/ci: add new flakes

  • glapi: import noop_array and public stubs earlier.

  • v3dv: avoid TFU reading unmapped pages beyond the end of the buffers

José Roberto de Souza (20):

  • intel: Initialize upper 32bits of drm_xe_sync.handle

  • intel/dev: Improve max_cs_threads documentation

  • intel/dev: Call intel_device_info_update_after_hwconfig() from common code

  • intel/common: Retry GEM_CONTEXT_CREATE when PXP have not finished initialization

  • anv: Remove protected memory support from compute queue

  • intel: Sync xe_drm.h

  • anv: Move code adding protected memory type to common code

  • anv: Add support to create protected bo and protected exec_queue in Xe KMD

  • iris: Add support to create protected bo and protected exec_queue in Xe KMD

  • intel: Add function to check if PXP is supported in Xe KMD

  • iris: Replace BO_ALLOC_* macros by a enum

  • intel/hwconfig: Sync hwconfig with IGT

  • intel/hwconfig: Remove ignored intel_hwconfigs from apply_hwconfig_item()

  • intel/dev/xe3: Set max_slices and max_subslices_per_slice using hwconfig

  • intel/perf: Update intel_perf to match xe_drm.h

  • drm-uapi: Sync xe_drm.h

  • intel: Program XY_FAST_COLOR_BLT::Destination Mocs for gfx12

  • intel: Fix the MOCS values in XY_FAST_COLOR_BLT for Xe2+

  • intel: Fix the MOCS values in XY_BLOCK_COPY_BLT for Xe2+

  • intel/tools: Fix batch buffer decoder

Juan A. Suarez Romero (31):

  • broadcom/compiler: move stores to the end of shader

  • Revert “st/mesa: move VS & TES output stores to the end before unlowering IO”

  • broadcom/ci: add new failures/flakes

  • v3dv: take into account GS when enabling line smooth

  • v3dv/ci: disable rpi5 job

  • vulkan: don’t leak debug utils label name

  • v3dv: duplicate key for texel_buffer cache

  • vc4/ci: update expected results

  • broadcom/simulator: use string copy instead of memcpy

  • vc4/ci: update expected results

  • v3dv/ci: Skip tests causing OOM

  • Revert “v3dv/ci: disable rpi5 job”

  • v3d/v3dv/vc4: review all expected timeouts

  • v3dv: remove src_format from blit render pass creation

  • v3dv: don’t batch regions with different depth offsets

  • v3dv: include depth offset on image view creation

  • vc4: check instruction before setting flags

  • v3dv: asserts struct is always non null

  • v3dv: check dynamic offset output

  • v3dv: asserts push constants data is valid

  • vc4: initialize register

  • vc4: add assertion on constant_fold

  • vc4: assert there are sources when emitting texture

  • broadcom/cle: assert attribute has a value

  • vc4: use safe iterator to remove instructions

  • broadcom/compiler: use safe iterator to remove instructions

  • broadcom/compiler: don’t use VLA on emit alu

  • broadcom/compiler: initialize register

  • v3dv: don’t check if DRM device is master

  • v3d(v)/ci: update expected results

  • ci: include duration in the CustomLogger

Julia Zhang (5):

  • vulkan: handle device memory report requests

  • radv: add import and export handle_type in radv_alloc_memory

  • radv: add obj_id to radeon_winsys_bo

  • radv: emit device memory report for device memory events

  • radv: advertise VK_EXT_device_memory_report

Jung-uk Kim (1):

  • FreeBSD: Disable support for “-mtls-dialect” for FreeBSD

Juston Li (3):

  • anv: xe: fully initialize drm_xe_sync addr/handle union

  • iris: xe: fully initialize drm_xe_sync addr/handle union

  • wsi/common: android: disable KHR_present_[wait/id]

K900 (1):

  • meson: support building with system libgbm

Karmjit Mahil (3):

  • loader/wayland: Fix missing timespec.h include

  • tu: Fix Perfetto build error with vk_buffer

  • tu: Fix segfault in fail_submit KGSL path

Karol Herbst (49):

  • ci/windows: Bump Vulkan SDK for SPIRV-Tools

  • clc: use SetUseHighestVersion when linking spirvs

  • mesa_clc: drop spirv version workaround

  • rusticl/mem: set bind flags for gl imports

  • rusticl/mesa: add PipeContext::device_reset_status

  • rusticl/queue: check device error status

  • clc: bump SPIR-V target to 1.6

  • rusticl/kernel: call nir_lower_variable_initializers earlier

  • rusticl: support SPIR-V 1.5 and 1.6

  • rusticl/mem: do not apply offset with in copy_image_to_buffer

  • rusticl/mesa: add buffer and texture variant for resource_copy_region

  • rusticl/mem: Buffer::copy_to_image layering

  • rusticl/mem: Image::copy_to_buffer layering

  • rusticl/mem: Image::copy_to_image layering

  • rusticl/mem: Image::write layering

  • rusticl/mem: accelerate Buffer::copy_rect

  • rusticl/mem: accelerate Buffer::write_rect

  • rusticl/mem: set num_samples and num_mip_levels to 0 when importing from GL

  • rusticl/platform: advertise all extensions supported by all devices

  • rusticl/util: add missing comment and assert to char_arr_to_cstr

  • intel/brw, lp: enable lower_pack_64_4x16

  • nir: Do not eliminate dead writes to shared memory in called functions.

  • rusticl/program: implement CL_INVALID_PROGRAM_EXECUTABLE check in clGetProgramInfo

  • rusticl/program: pass options by reference

  • rusticl/program: loop over all devices inside Program::build

  • rusticl/program: rework build_nirs so it only touches devices we care about

  • rusticl/program: fix building kernels

  • rusticl/program: simplify active_kernels check

  • rusticl/kernel: rename CSOWrapper to SharedCSOWrapper

  • rusticl/queue: make it unncessary to keep QueueContext Send

  • rusticl/queue: cache bound CSO

  • rusticl/mesa: remove Sync from PipeContext

  • nir/serialize: fix decoding of is_return and is_uniform

  • vtn: Support the UniformDecoration capability.

  • zink: don’t apply the map_offset when mapping a staging resource in zink_buffer_map

  • iris: remove all clover support code

  • freedreno: remove all clover support code

  • llvmpipe: remove all clover support code

  • gallium: stop filling ir_target in various drivers

  • gallium: stop using PIPE_BIND_COMPUTE_RESOURCE in drivers

  • gallium: stop implementing set_compute_resources in various drivers

  • nouveau: ignore req_input_mem

  • rusticl/device: fix panic when disabling 3D image write support

  • nir_lower_mem_access_bit_sizes: fix negative chunk offsets

  • nak: fix handling of delays > 15

  • r600: fix r600_buffer_from_user_memory for rusticl

  • iris: parse global bindings for every gen

  • iris/xe: fix compute shader start address

  • iris/xe: take the grids variable_shared_mem into account

Kenneth Graunke (37):

  • brw: Drop unused defines

  • brw: Eliminate fs_inst::shadow_compare

  • brw: Replace fs_inst::pi_noperspective with a logical control source

  • brw: Drop FB_WRITE_LOGICAL_SRC_DST_DEPTH source

  • brw: Replace fs_inst::last_rt with a logical control source

  • brw: Replace fs_inst::target field with logical FB read/write sources

  • brw: Use correct builder size for MEMORY_FENCE/INTERLOCK virtual opcodes

  • brw: Change destination of memory fences to UD type

  • brw: Eliminate the BTI source from MEMORY_FENCE/INTERLOCK opcodes

  • brw: Add latencies for HDC/RC memory fences

  • brw: Lower MEMORY_FENCE and INTERLOCK in lower_logical_sends

  • brw: Drop INTERPOLATE_AT mlen handling from size_read()

  • brw: Drop unnecessary mlen/header_size on virtual GET_BUFFER_SIZE op

  • nir: Eliminate dead writes to shared memory at the end of the program

  • brw: Rename shared function enums for clarity

  • isl: Delete redundant “use separate stencil?” check

  • isl: Drop compile time “use separate stencil” checks.

  • intel: Delete devinfo->must_use_separate_stencil

  • intel: Delete devinfo->has_surface_tile_offset

  • intel: Move devinfo->has_negative_rhw_bug into the elk compiler

  • intel: Move devinfo->has_compr4 into the elk compiler

  • intel/dev: Set minimum HS URB entries to 0.

  • intel/dev: Set max_wm_threads to 0 in the Gfx9+ devinfo structs

  • intel/dev: Rework device info macros for Gfx8+

  • intel/dev: Set a higher minimum number of URB entries for GS

  • intel: Use devinfo->urb.min_entries[GS and TCS] for setting URB configs

  • intel: Move unlit centroid workaround into the elk compiler

  • intel/decoder: Decode compute shaders in EXECUTE_INDIRECT_DISPATCH

  • brw: Make a helper to emit UNDEF for temporaries containing small types

  • brw: Emit UNDEF as needed in SSA-style builder helpers

  • brw: Skip unnecessary UNDEFs for comparisons

  • brw: Use a smaller type for masked sub-32-bit shift values

  • brw: Avoid regioning restrictions for u2u16/i2i16 narrowing conversions

  • brw: Track the largest VGRF size in liveness analysis

  • brw: Use live->max_vgrf_size in register coalescing

  • brw: Use live->max_vgrf_size in pre-RA scheduling

  • brw: Don’t assert about MAX_VGRF_SIZE in brw_opt_split_virtual_grfs()

Kenny Levinsen (1):

  • device-select: Support linux-dmabuf feedback

Kevin Chuang (2):

  • anv/bvh: Fix encoder handling sparse buffer

  • anv/bvh: Fix copy shader handling sparse buffer

Konstantin (1):

  • nir/tests: Do not rely on __LINE__

Konstantin Seurer (68):

  • nir: Stop using instructions for debug info

  • spirv: Handle NonSemantic.Shader.DebugInfo.100

  • nir: Add variable debug info to instructions

  • nir/lower_vars_to_ssa: Annotate defs with variable names

  • vulkan: Stop using strings for BVH build pipeline keys

  • vulkan/meta: Remove object types from vk_meta_object_key_type

  • vulkan/meta: Stop using strings for meta keys

  • hk: Stop using strings or common key types for meta keys

  • radv/meta: Stop using strings for meta keys

  • lavapipe: Fix maintainance7 descriptor set limits

  • vulkan/cmd_queue: Simplify freeing cmd_queue entries

  • vulkan/cmd_queue: Add VK_CMD_TYPE_COUNT

  • vulkan/radix_sort: Stop force-unrolling loops

  • gallivm: Remove loop limiting

  • lavapipe: Implement some functions required by the common BVH framework

  • lavapipe: Use the common BVH framework

  • radv: Optimize fs builtins using static gfx state

  • gallivm: Split nir prepasses into aos/soa

  • gallivm/nir/aos: Remove the dependency on lp_bld_nir.c

  • gallivm/nir/soa: Remove the dependency on lp_bld_nir.c

  • gallivm/nir/soa: Lower bools to i1

  • gallivm/nir/soa: Implement robusst uniform loads without controlflow

  • gallivm/nir/soa: Select more IO to gather/scatter intrinsics

  • lavapipe: Move nir passes to a new directory

  • lavapipe: Lower descriptor sets in NIR

  • lavapipe: Initialize the compiler options of the noop fs

  • llvmpipe: Do not use coroutines when they are unnecessary

  • nir: Rename in-bounds-agx to in-bounds

  • nir: Do not emit amul if it is unsupported

  • lavapipe: Optimize buffer robustness

  • gallivm/nir/soa: Do not lower vectors to llvm arrays

  • nir/divergence_analysis: Handle load_const_buf_base_addr_lvp

  • gallivm/nir/soa: Use divergence analysis

  • lavapipe: Lower push constants in NIR

  • gallivm: Only guard tex/image ops if the exec mask can be zero

  • gallivm/nir/soa: Skip bounds checking for in-bounds access

  • gallivm/nir/soa: Properly skip empty else branches

  • lavapipe: Remove uniform inlining

  • vulkan: Add utilities for triggering renderdoc captures

  • radv: Lower ray query vars to structs

  • radv: Implement multidimensional ray query arrays

  • llvmpipe: Skip draw_mesh if the ms did not write gl_Position

  • nir: Test nir_minimize_call_live_states

  • nir/sweep: Fix handling instructions with debug info

  • nir/print: Do not print debug information when gathering it

  • gallivm: Create a debug builder and add GALLIVM_DEBUG=symbols

  • llvmpipe: Annotate functions with debug information

  • gallivm: Handle nir_instr_debug_info

  • gallivm: Emit debug info for definitions

  • gallivm: Add a debug variable for the exec mask

  • gallivm: Run nir_lower_load_const_to_scalar

  • lavapipe: Enable debug information if GALLIVM_DEBUG=symbols is set

  • radv/meta: Change the return type of get_r32g32b32_format to VkFormat

  • ci: Do not build hk on alpine

  • asahi: Only require IOKit for tooling

  • hk: Fix building without the gallium driver

  • spirv: Emit code for NonSemantic.DebugPrintf if supported

  • nir: Turn the format string index into a const index

  • radv: Handle nir_intrinsic_printf

  • clc: Print errors when initializing clang fails

  • clc: Allow bitfields

  • clc,libcl: Clean up CL includes

  • radv: Fix rayTracingPositionFetch with multiple geometies

  • lavapipe: Prefetch 56 bytes of node data during ray traversal

  • lavapipe: Run nir optimizations on ray tracing pipelines

  • lavapipe: pre-load tmax

  • lavapipe: Do not emit aabb handling if no isec shader is used

  • radv: Return VK_ERROR_INCOMPATIBLE_DRIVER for unsupported devices

Lakshman Chandu Kondreddy (1):

  • freedreno: Add support for Adreno623 GPU

Lars-Ivar Hesselberg Simonsen (9):

  • panfrost: Do not evaluate_per_sample for non-MSAA

  • Revert “panfrost: remove is_blit flag”

  • Revert “panfrost: fix hang by using MALI_PIXEL_KILL_WEAK_EARLY in color preload”

  • panvk: Set missing shader_modifies_coverage flag

  • panfrost: Use RUN_COMPUTE over RUN_COMPUTE_INDIRECT

  • panvk: Use RUN_COMPUTE over RUN_COMPUTE_INDIRECT

  • vulkan/wsi/wayland: Avoid duplicate colorspace entry

  • panvk: Add barrier for interleaved ZS copy cmds

  • vk/sync: Fix execution only barriers

Leder, Brendan Steve (1):

  • amd/vpelib: Reformat index variables and update enum

Leonard Göhrs (2):

  • etnaviv/ci: add pengutronix LAVA lab with one i.MX8MP device

  • ci: re-enable pengutronix farm

LingMan (2):

  • meson: Update pest subproject family

  • meson: Sync subproject version numbers in `packagefiles` with their `.wrap` equivalents

Lionel Landwerlin (137):

  • anv: fixup missing compiler dependency on tests

  • intel: move internal shader compile to vtn_bindgen2

  • compiler: drop vtn_bindgen

  • intel_clc: remove NIR output support

  • spirv: remove spirv_library_to_nir_builder

  • brw: fixup scoreboarding for find_live_channels

  • anv: reuse helper for compute push constants

  • anv: increase general state pool

  • anv: track the first 2MB of unused VA

  • anv: use heap size to program generate state heap

  • anv: add a helper for getting gfx push constant addresses

  • brw/anv: rework push constants for mesh/task shaders

  • brw: enable A64 pulling of push constants

  • anv: use A64 messages for push constants loads on Gfx12.5+

  • nir: add a high precision conversion unorm->float

  • anv: add source hashes for BVH building shaders

  • vulkan/wsi: propagate protected swapchain to images

  • vulkan: allow support for protected surfaces

  • anv: support protected surfaces with display platform

  • anv,driconf: Add sampler coordinate precision workaround for Dynasty Warriors

  • genxml: make component packing an array

  • genxml: add convenience dwords for packing components

  • brw: fix indentation

  • brw: remove nr_attribute_slots from vs_prog_data

  • brw: port vs input to lower_64bit_to_32_new

  • brw: update vulkan max attribute limit

  • brw: add a max HW vertices attribute limit

  • brw: enable vertex fetching component packing

  • blorp: emit 3DSTATE_VF

  • anv: disable VF statistics for memcpy

  • anv: enable vertex fetching component packing

  • anv: ensure Wa_16012775297 interacts correctly with Wa_18020335297

  • brw: use meaningful io locations for system values

  • brw: add support for no VF input slot compaction

  • brw: add documentation about slot compaction & component packing

  • brw: fix component packing starting index

  • anv: move RT stage bits to main header

  • anv: move reg_mask push constant field to gfx

  • anv: hold a prepacked COMPUTE_WALKER instruction on CS pipelines

  • anv: make gfx state flushing available externally

  • anv: make compute state flush helper visible

  • runtime: sort push constant layouts

  • anv: avoid memory type changes with INTEL_DEBUG=noccs

  • anv/ci/adl: update fail expectation for video

  • anv: fixup compute walker storage length

  • nir: add options to lower only some image atomics to global

  • brw: factor out base prog_data setting

  • brw: store source_hash in prog_data

  • anv: switch to use brw’s prog_data source_hash

  • isl: report tiling address swizzles

  • isl: add usage for software detiling

  • isl: centralize supported tilings in a single function

  • isl: select a tiling for shader detiling

  • isl: add support for R64 storage image lowering

  • isl: add a helper to report what dimensions a tiling supports

  • nir: track lowered image intrinsics to globals

  • brw: include UGM fence when TGM + lowered image->global

  • brw: add support for 64bit storage images load/store

  • brw: add support for texel address lowering

  • anv: rename compressed format emulation helpers

  • anv: add mapping for VBO formats in format mapping

  • anv: add a is_sparse for image format support checks

  • u_trace: pass tracepoint flags to the read_timestamp callback

  • intel/ds: rework RT tracepoints

  • anv: fix missing 3DSTATE_PS:Kernel0MaximumPolysperThread programming

  • brw: optimize load payload with immediate headers

  • brw: avoid setting up the sampler header bits when unused

  • spirv: fix racy build

  • vulkan/runtime: ensure robustness state is fully initialized

  • vulkan/runtime: pass robustness state to preprocess vfunc

  • vulkan/runtime: add a multialloc vk_shader allocator

  • vulkan/runtime: store flags on descriptor set layouts

  • anv/iris: centralize TBIMR drirc

  • iris: remove duplicate TessellationDistributionMode programming

  • anv/iris: add drirc keys to disable VF/TE distribution

  • anv/apply_layout: split binding table building into its own function

  • intel/genxml: add a genX RT include header

  • genxml: simplify genX_rt_pack.h

  • brw: make intel_shader_enums.h opencl importable

  • anv/brw: move INTEL_MSAA_* flag computation to the compiler

  • anv: break down Wa_16014912113 in need/apply parts

  • anv: fixup indentation around Wa_16014912113

  • brw: avoid calling lower_indirect_derefs multiple times

  • anv: fill runtime buffer device_address field

  • anv: move index buffer entry point out of genX code

  • anv: move vertex buffer storage to 64bit address + mocs

  • anv: move xfb buffer storage to 64bit address + mocs

  • anv: track protection on anv_address

  • anv: use addresses for buffer<->image copies

  • anv: simplify internal blorp helper

  • anv: fix non page aligned descriptor bindings on <Gfx12.0

  • lavapipe: fill buffer address

  • vulkan/runtime: rely on vk_buffer::device_address

  • brw: fix spilling for Xe2+

  • brw: ensure VUE header writes in HS/DS/GS stages

  • anv: Set limit_trig_input_range option for Company of Heroes 3

  • anv: avoid early lower of the fp64 code

  • blorp: assert that shaders don’t spill

  • blorp: relax depth/stencil<->color copy restriction

  • vulkan: add helper for color/depth-stencil capable formats

  • anv: relax depth/stencil<->color copy restrictions

  • anv: fix end of pipe timestamp query writes

  • anv: disable replication when we don’t have both VS/FS stages

  • brw: always write the VUE header

  • anv: limit implict write with drirc

  • nir: add support for lowering non uniform texture offsets

  • elk: stop using intel_nir_lower_texture

  • brw: don’t lower tg4 offsets without LOD

  • intel: move lower_texture to brw

  • brw: move texture offset packing to NIR

  • anv: enable non uniform texture offset lowering

  • anv: wire VkAccessFlagBits3KHR flags in internal helpers

  • anv: expose VK_KHR_maintenance8 support

  • intel/genxml: remove ISA fields

  • intel/genxml: add more engine tagging on instructions

  • intel/genxml: fixup engine filtering

  • intel/genxml: define post-sync operations for MI_FLUSH_DW

  • intel/genxml: add MI_FLUSH_DW to blitter engine

  • anv/genxml: use special genX video pack files

  • anv/hasvk: consider timeline semaphore support stable

  • docs: remove unused env variable

  • anv/hasvk: sort out debug options

  • anv: consolidate environment variables

  • anv: add shader-hash debug option

  • brw: fix shuffle with scalar/uniform index

  • anv: relax restriction on variable count descriptors

  • anv: fix self dependency computation

  • brw: fix Wa_22013689345 emission

  • iris: update Wa_1607156449 to use WA infrastructure

  • anv: update Wa_1607156449 to use WA infrastructure

  • intel/dev: remove ADLN references

  • anv: remove ALWAYS_INLINE from globally visible functions

  • anv/iris: implement Wa_18040903259

  • anv: use companion batch for operations with HIZ/STC_CCS destination

  • anv: force fragment shader execution when occlusion queries are active

  • intel: fix null render target setup logic

  • brw: add pre ray trace intrinsic moves

Lorenzo Rossi (8):

  • nvk: fix preprocess buffer alignment

  • nvk: Fix MSAA sparse residency lowering crash

  • nir: support shared atomics in nir_lower_atomics

  • nvk, nak: Implement shaderSharedInt64Atomics

  • nak: Fold bool-int-bool conversions

  • nak: Flatten AttrAccess into instructions

  • nak: Fix SM50 rounding-mode encoding edge-case

  • nak: Simplify shl64 lowering on Maxwell

Loïc Minier (1):

  • freedreno: check if GPU supported in fd_pipe_new2

Loïc Molinari (5):

  • perfetto: Let MESA_TRACE_FUNC() take printf-like format arguments

  • panfrost: Add CPU traces

  • pan/kmod: Add drmIoctl() wrapper pan_kmod_ioctl() with CPU trace

  • docs: Add Panfrost to the list of drivers with CPU traces

  • mesa: Add CPU traces

Lucas Fryzek (4):

  • anv: Release correct bo for RT scratch

  • vulkan/runtime: Add object type to DMR API

  • anv: Implement VK_EXT_device_memory_report

  • anv: Expose VK_EXT_device_memory_report

Lucas Stach (19):

  • etnaviv: drm: fix instruction limit for cores with instruction cache

  • etnaviv: drop double masking in etna_emit_load_state

  • etnaviv: split large multi-state updates into multiple batches

  • etnaviv: emit PS start and end PC states only on shader changes

  • etnaviv: correct and rename shader range register check

  • etnaviv: replace open-coded shifts in shader range registers with macros

  • etnaviv: place FS right behind VS in unified instruction memory

  • etnaviv: allow larger shaders with unified instruction memory

  • kmsro: look for graphics capable screen as renderonly device

  • etnaviv: rs: fix slow/fast clear transitions

  • etnaviv/ci: remove dEQP-GLES2.functional.polygon_offset.default_enable fail

  • etnaviv: fix ETNA_MESA_DEBUG=no_early_z

  • etnaviv: Update headers from rnndb

  • etnaviv: enable forwarding ZW fragcoord components from RA to SH

  • etnaviv/ci: remove fragcoord related fails on GC7000

  • etnaviv: use pipe_resource in etna_resource_status

  • etnaviv: split PIPE_BUFFER resources from other types of resources

  • include: update GL headers from the registry

  • etnaviv: add context flush sw query

Ludvig Lindau (1):

  • panvk: Get flush_id once per submit

Maaz Mombasawala (3):

  • svga: Add all tgsi double instructions for shader codegen checks

  • nir: Add option to preserve double immediates in tgsi shader.

  • svga: Check float type when emitting atomic instructions

MaciejDziuban (3):

  • vulkan: handle use_default_scaling_matrix_mask in h264 decoder

  • vulkan: Add default scaling lists for H265

  • radv: Use vk_video_derive_h265_scaling_list

Marek Olšák (128):

  • amd: drop support for LLVM 15, 16, 17

  • ci/debian-ppc64el: don’t build AMD drivers due to having only LLVM 15

  • gallium,st/mesa: allow reporting compile failures from create_vs/fs/.._state

  • nine: change the vtxbuf dirty mask to bool

  • nine: always update vertex buffers after updating vertex elements

  • nine: remove unused last_vtxbuf_count

  • nine: bind exactly the number of vertex buffers as vertex elements need

  • cso_context: add cso_get_vertex_elements_for_bind, letting the caller bind it

  • gallium/u_threaded,st/mesa: add a merged set_vertex_elements_and_buffers call

  • radeonsi: add assertion requiring binding vertex elements before vertex_buffers

  • radeonsi: require that count in set_vertex_buffers matches vertex elements state

  • radeonsi: don’t set num_vertex_buffers and don’t unbind in set_vertex_buffers

  • radeonsi: simplify bind_vertex_elements due to being before set_vertex_buffers

  • radeonsi: print why draws are rejected

  • util: remove glthread enablement from app profiles

  • mesa: don’t build st_format_test on Windows

  • glapi: remove dead code

  • glapi: inline functions or use equivalent alternatives

  • glapi: don’t export always-private functions

  • mesa: remove a glapi workaround for Mesa 10.5 and older

  • glapi: fix build dependencies by putting all xml/py files into a single list

  • glapi: remove support for dynamically-registered functions

  • glapi: use an assertion in SET_by_offset instead of doing nothing

  • glapi: clarify the meaning of static_data.functions

  • glapi: remove check_table tests

  • glapi: generate static offsets from the list of GL functions automatically

  • glapi: remove unused functions from dispatch tables

  • glapi: move legacy imaging functions to the end of dispatch tables

  • glapi: verify that aliased functions don’t have entries in dispatch tables

  • glapi: just use _gloffset_COUNT_ everywhere, which is always the function count

  • glthread: handle glFlush with HasExternallySharedImages more efficiently

  • glthread: rename scripts to match the names of generated files

  • glapi: disable python escape sequences in strings that use invalid ones

  • glapi: use static_data.libgl_public_functions directly

  • glapi: remove is_static_entry_point wrapper

  • glapi: inline entry_current_get()

  • glx: don’t call GL functions directly, use the current dispatch instead

  • glx: fix hardcoded use of dispatch table index in glAreTexturesResidentEXT

  • glx: stop exporting GL functions from libGLX_mesa.so

  • glx: make it more obvious what libglapi_bridge is

  • glapi: remove unused python code

  • mesa: allocate GLmatrix aligned to 16 bytes

  • Remove osmesa

  • gallium/u_blitter: remove UTIL_BLITTER_ATTRIB_COLOR, use a constant buffer

  • gallium/u_blitter: change blitter_attrib from union to struct

  • meson: never use static glapi because shared glapi is also static and better

  • glx/xlib: switch glapi from static to shared (which is also static)

  • meson: remove with_shared_glapi variable (it’s always true)

  • glapi: remove static glapi (it’s dead code now)

  • glapi: remove duplicated generated header glapitable.h

  • glx: fix build with -Dlegacy-x11=dri2

  • egl: use the current dispatch to execute glFlush instead of get_proc_address

  • glapi: remove extension definitions that will likely never be implemented

  • ac/gpu_info: use max_good_cu_per_sa for computation of max_scratch_waves

  • ac,radeonsi: don’t set num_slots/src/dest_type/write_mask when they’re set automatically

  • Revert “ac/nir: clamp vertex color outputs in the right place”

  • ac/nir/cull: extract a helper calling accept_func

  • ac/nir/ngg: add an option to skip viewport-based culling

  • ac/cmdbuf: split meta_*_policy to dcc and htile variables

  • ac/cmdbuf: rework CB/DB cache controls for better perf

  • winsys/amdgpu: don’t use 32-bit address space for IBs

  • radeonsi: use si_is_buffer_idle everywhere

  • radeonsi: move buffer reallocation to a separate function

  • radeonsi: enable NGG culling when the shader writes the viewport index

  • radeonsi: enable Z/S caching in GL2 by default except FurMark

  • radeonsi: use the restrict keyword for draw parameters

  • radeonsi: lower IO only if io_lowered == false

  • radeonsi: reflect blitter VS in si_context::num_vertex_elements

  • radeonsi: don’t cull front/back faces in the hw if the shader culls them

  • ac,radeonsi: define all SDMA DCC fields & use them, enable compressed writes

  • radeonsi: remove clover support

  • glx: don’t generate indirect GLX dispatch for ARB_imaging functions

  • glapi: make a separate copy of entry.c for each lib

  • glapi: replace mapi_func type with identical _glapi_proc

  • glapi: simplify codegen macros

  • glapi: merge all shared-glapi source files into one .c file

  • glapi: remove unused _mesa_glapi_Dispatch

  • glx: add a test to verify exported symbols

  • glapi: get the list of public GL functions from libgl-symbols.txt

  • glx: remove the hack that forced exporting GL functions from libGL

  • glapi: remove the GLAPI_PREFIX macro, just use gl as the function prefix

  • glapi: simplify headers of generated files

  • glapi: simplify mapi_abi.py

  • glapi: rename dispatch stubs to use function names instead of numbers

  • glapi: remove noop_generic callbacks

  • glapi: remove the option to set the nop handler

  • mesa: move dispatch.h and marshal_generated.h generation to glapi/gen

  • mesa: inline main/meson.build

  • glx: stop exporting EXT_texture_object functions from libGLX_mesa.so

  • glapi: fix x86 32-bit asm dispatch regression

  • amd: update addrlib

  • amd/addrlib: remove the DCC page fault workaround

  • ac/nir/cull: cull small prims using a point-triangle intersection test

  • glsl: return failure from gl_nir_validate_first_and_last_interface_explicit_locations

  • glsl: return failure from varying_matches_assign_locations

  • glsl: return failure from remove_unused_io_vars

  • glsl: stop calling nir_opt_combine_stores (redundant with nir_opt_vectorize_io)

  • glsl: remove a deprecated comment about nir_compact_varyings

  • glsl: always return true at the end of link_varyings

  • glsl: move code after link_varyings into link_varyings

  • st/mesa: don’t do nir_remove_dead_variables on in/out twice

  • nir/opt_algebraic: lower 16-bit imul_high & umul_high

  • ac: define physical VGPRs for fake hw overrides

  • ac/nir/cull: rename skip_viewport_culling -> skip_viewport_state_culling

  • ac/nir/cull: always do frustum culling, skip only small prim culling

  • ac/surface: remove 64K_2D modifier with 64B max compressed blocks for gfx12

  • radeonsi: work around a primitive restart bug on gfx10-10.3

  • radeonsi: always use ACO callbacks to scalarize/vectorize 16-bit ALU

  • radeonsi: expose 16-bit NIR types for ALU, MEM, and LDS (no inputs/outputs)

  • radeonsi/ci: update gfx11 failures

  • radeonsi/ci: don’t run GTF tests (they have been removed from glcts)

  • radeonsi: lower load/store bit sizes before load/store vectorization

  • radeonsi: add a VOP3P swizzle requirement for 16-bit packed math

  • radeonsi/ci: add gfx12 failures and flakes

  • radeonsi: make si_shader_selector::main_shader_part_* an iterable union

  • radeonsi: add ACO-specific main shader parts

  • winsys/amdgpu: don’t add VM_ALWAYS_VALID buffers into the BO list

  • radeonsi: determine VM_ALWAYS_VALID accurately

  • ac/gpu_info: increase the attribute ring size for gfx12

  • ac/gpu_info: remove has_tmz_support function

  • ac/gpu_info: add payload_entry_size into ac_task_info

  • ac/gpu_info: add 256 to payload_entry_size to increase future task shader perf

  • ac/nir: set X=0 for task->mesh shader dispatch when Y or Z is 0

  • ac/surface: make gfx12_estimate_size reusable by gfx6

  • ac/surface: select 3D tile mode without overallocating too much for gfx6-8

  • radeonsi: always scalarize shared memory instructions

  • radeonsi: use nir_opt_shrink_vectors

  • radv: fix incorrect patch_outputs_read for TCS with dynamic state

Mark Collins (4):

  • tu/kgsl: Remove zero CB queue submission fast path

  • tu/kgsl: Revert “Remove zero CB queue submission fast path”

  • tu/kgsl: Fix KGSL syncobj lifetime in no CB submit

  • ir3/a7xx: Add post-RA pass to track liveness and insert (last)

Martin Krastev (7):

  • svga/ci: disable vmware farm

  • svga/ci: enable vmware farm

  • svga/ci: enable vmware farm

  • svga/ci: enable vmware farm

  • svga/ci: update EXTERNAL_KERNEL_TAG to new kernel

  • svga/ci: disable vmware farm

  • docs/svga: Add steps how to get VMware Workstation Pro on Linux

Martin Roukala (né Peres) (42):

  • ci: fix the artifact name

  • ci: be explicit about the fact HW jobs do not need linter artifacts

  • ci/test/b2c: do not download the debian/*_test-(gl|vk) artifacts

  • ci-templates: update to the latest upstream version

  • ci/build-deqp: delegate the deqp main commit check to github

  • ci/test: add arm32 b2c jobs

  • ci/debian/test-vk: remove software-properties-common

  • ci/debian/test-vk: trim down the container on a per-arch basis

  • ci/image-tags: bump the debian base tag

  • ci/b2c: add support for diskless DUTs

  • ci/b2c: de-duplicate the download of install.tar

  • etnaviv/ci: convert from baremetal to CI-Tron

  • ci: re-enable austriancoder’s farm

  • zink/ci: use the debian-built-testing for nvk

  • zink/ci: update the radv expectations

  • radv/ci: bump tahiti’s cpu cores

  • radv/ci: update expectations

  • freedreno/ci: update expectations

  • ci/b2c: fix the S3 artifact for amd64 manual vk/gl

  • radv/ci: document more Tahiti VKCTS flakes

  • turnip/ci: re-introduce the `multiviewport` flakes

  • zink/ci: mark query-rgba-signed-components as fixed on more platforms

  • zink/ci: document more RADV flakes

  • radv/ci: add testing on RAPHAEL

  • radv/ci: reduce Renoir concurrency in vkcts

  • radv/ci: set a tight timeout on vkcts-navi31

  • radv/ci: document more flakes

  • zink/ci: document more radv flakes

  • zink/ci: document more GA106 flakes

  • zink/ci: increase the a750 job’s timeout to 18 minutes

  • radv/ci: mark a whole subset of tests as flaky on tahiti

  • zink/ci: document more RADV flakes

  • zink/ci: document more NVK GA106 flakes

  • ci/b2c: allow jobs to opt-in into the new mars setup command

  • radv/ci: opt-in the new mars setup command

  • etnaviv/ci: opt-in the new mars setup command

  • nvk/ci: opt-in the new mars setup command

  • freedreno/ci: opt-in the new mars setup command

  • i915g/ci: opt-in the new mars setup command

  • r300g/ci: opt-in the new mars setup command

  • radv/ci: reduce the timeout of vkcts-tahiti to a more sensible time

  • radv/ci: add hawaii to CI

Mary Guillemard (95):

  • panfrost: Fix group priorities in drm-shim

  • panfrost: Fix PROGRESS_LOAD destination register

  • pan/bi: Properly encode LEA_BUF_IMM

  • pan/bi: Remove shift lanes invalid encodings

  • pan/bi: Fix invalid CLPER encoding

  • pan/bi: Remove b3210 from valid swizzle

  • pan/bi: Use 2D dimension with TEX_FETCH with CUBE on Valhall

  • pan/decode: Fix indirect branch calculation for 64-bit

  • panfrost: Properly name CSF instruction UMIN32 source 0

  • panvk: Disallow unknown GPU models early in physical device init

  • pan/genxml: Stop using “constant” for struct fields in xml defs

  • pan/genxml: Rework gen_pack.py to support OpenCL

  • panfrost: Add base of OpenCL C infrastructure

  • bi: Optimize scratch access

  • panvk: Integrate libpan

  • panfrost: Integrate libpan in gallium driver

  • panvk: Add create_shader_from_binary

  • panvk: Move TLS preparation logic to cmd_dispatch_prepare_tls

  • panvk: Expose calculate_task_axis_and_increment on CSF

  • panvk: Implement precomp dispatch

  • panfrost: Implement precomp dispatch on Gallium

  • panfrost,panvk: Wire printf and abort support

  • ci: Transition to precomp-compiler

  • ci: Enable mesa-clc and precomp-compiler on debian-arm32

  • pan/bi: Fix scratch access optimization

  • panvk: Switch JM copy queries to CLC

  • panvk: Switch JM clear queries to CLC

  • pan/bi: Document removed instructions on v11

  • pan/bi: Do not run bi_fuse_small_int_to_f32 on v11+

  • pan/bi: Disallow dst swizzle optimization in case of conversion

  • pan/bi: Add round mode modifier to FADD

  • pan/bi: Stop using V2F32_TO_V2F16 on Valhall

  • pan/bi: Lower SWZ.v4i8 to multiple MKVEC.v2i8 on v11+

  • pan/bi: Stop using S16_TO_F32 in nir_texop_lod computation on v11+

  • pan/bi: Stop using V2F16_TO_V2S16 for barycentric_at_offset on v11+

  • pan/bi: Lower removed instructions in algebraic on v11+

  • pan/bi: Lower FROUND.v2s16 in nir_lower_bit_size on v11+

  • pan/bi: Lower hadd on v11+

  • pan/bi: Handle LD_VAR_BUF_IMM encoding changes on v11+

  • pan/bi: Properly extract primitive facing on v11+

  • pan/genxml: Define RUN_IDVS staging registers in an enum

  • panfrost: Remove write to TSD_3 in Gallium driver

  • pan/genxml: Define RUN_COMPUTE staging registers in an enum

  • pan/genxml: Define RUN_FRAGMENT staging registers in an enum

  • panfrost: Rework cs_sr_regXX to be a macro

  • panfrost: Switch Gallium driver to use cs_sr_regXX

  • panfrost: Use CSIF info for CSF registers count

  • panfrost: Avoid hard crash when major arch is unknown

  • panfrost: Rename CSF MOVE into MOVE48

  • panfrost: Fix FLUSH_CACHE2 other definition

  • panfrost: Rename CS ADD_IMMEDIATEXX to ADD_IMMXX

  • pan/genxml: Use DCD Flags in Draw struct on v9+

  • pan/bi: Disallow FAU special page 3 and WARP_ID on message instructions

  • pan/bi: Add unit tests for FAU special page 3 and WARP_ID

  • pan/bi: Run nir_lower_bit_size after algebraic

  • pan/bi: Run bifrost_nir_lower_algebraic_late while there is progress

  • pan/bi: Lower FREXPE.v2f16 and FREXPM.v2f16 on v11+

  • pan/bi: Fix out of range access in bi_instr_replicates

  • pan/bi: Ensure we select b0 with halfswizzle in va_lower_constants

  • nir: Add Panfrost specific shader_output intrinsic

  • pan/bi: Use shader_output intrinsic for IDVS

  • pan/bi: Add support for IDVS2 on Avalon

  • pan/bi: Lower IADD.v4s8 in algebraic on v11+

  • pan/bi: Allow no_psiz variant with IDVS2

  • pan/genxml: Rename UMIN32 opcode to COMPARE_SELECT32

  • pan/genxml: Add v12 definition

  • pan/genxml: Add v13 definition

  • pan/genxml: Build libpanfrost_decode for v12

  • pan/genxml: Build libpanfrost_decode for v13

  • pan/lib: Build for v12

  • pan/lib: Build for v13

  • pan/clc: Build for v12

  • pan/clc: Build for v13

  • panvk: Support 64x64 meta tile size for v12 in cmd_preload_render_area_border

  • pan/lib: Adapt CRC calculation to align to 64x64 on v12+

  • panfrost: Disable hierarchy levels based on multiple of tile_size

  • panvk: Use spd variant instead of pos_points when checking for vs shader presence

  • pan/earlyzs: Default to FORCE_EARLY for ZS update on v11+

  • panvk: Implement Z/S dependency tracking on v11+

  • panvk: Add v12 support

  • panvk: Add v13 support

  • panfrost: Add v12 support to the Gallium driver

  • panfrost: Add v13 support to the Gallium driver

  • panfrost/ci: Add Mali-G720 current expectations

  • panfrost/ci: Add Mali-G725 current expectations

  • panfrost: Advertize Mali-G720 support

  • panfrost: Advertize Mali-G725 support

  • panfrost: Update the release note to mention G720/G725 addition

  • panvk: Fix inverted ZS read flags in DCD2 on v11+

  • panfrost: Take tiler memory budget into account in pan_select_tiler_hierarchy_mask

  • panvk: reset dyn_bufs map count to 0 in create_copy_table

  • panvk: Take rasterization sample into account in indirect draw on v10+

  • panvk: Take resource index in valhall_lower_get_ssbo_size

  • pan/bi: Properly lower add/sub with saturation on v11+

  • pan/genxml: Fix inverted logic on nr_regs

Matt Turner (22):

  • gallium: Fix typos

  • glsl: Fix typos

  • glx: Fix typos

  • intel: Fix typos

  • mapi: Fix typos

  • vulkan: Fix typos

  • gallium: Return NULL, not false, from functions returning pointers

  • nir: Return NULL, not false, from functions returning pointers

  • iris: Initialize pointer with NULL, not false

  • intel/compiler: Use FALLTHROUGH

  • glsl: Use FALLTHROUGH

  • glsl: Add missing break

  • intel/compiler: Use unreachable instead of assert(!”…”)

  • intel/isl: Use unreachable instead of assert(!”…”)

  • anv: Use unreachable instead of assert(!”…”)

  • hasvk: Use unreachable instead of assert(!”…”)

  • intel/decoder: Remove assert(!”…”) in recoverable condition

  • intel/compiler: Add missing breaks

  • intel/compiler: Use correct enum type

  • iris: Remove ignored qualifier

  • intel/isl: Remove ignored qualifier

  • anv: Remove ignored qualifier

Maíra Canal (4):

  • v3dv: VK_EXT_acquire_drm_display doesn’t require a DRM master fd

  • v3dv: don’t overwrite the primary fd if it’s already set

  • broadcom/simulator: Expose V3D revision number in the simulator interface

  • broadcom/simulator: Fix Indirect CSD jobs for V3D 7.1.6+

Mel Henning (31):

  • nvk: Use hw support for instancing on PASCAL_B+

  • nak/opt_copy_prop: Fix IAdd3 overflow check

  • nak/opt_copy_prop: Add force_alu_src_type

  • nak/opt_copy_prop: Force alu src for IAdd2X/IAdd3X

  • driconf: force_vk_vendor on Deep Rock Galactic+NVK

  • nak: Add OpLea

  • nak,nir: Add 32-bit nir_op_lea_nv and use it

  • nak: Add OpLeaX

  • nak,nir: Add 64-bit lea_nv

  • nouveau/mme: Print ip in mme_tu104_dump

  • nak: Add static cycle count statistic

  • nak: Add an occupancy statistic

  • nak/spill_values: Make Spill take &mut self

  • nak: Add spill/fill statistics

  • nvk: Don’t zero imported memory

  • nvk: Add NVK_DEBUG=trash_memory

  • vulkan: Relax bda assert for zero-size buffers

  • nak: Add a simple postpass instruction scheduler

  • nak: Calc static cycle count in instr_sched

  • nak: Assert instr_sched matches calc_instr_deps

  • nak/instr_sched: Barriers activate after 2 cycles

  • nvk: SET_PS_{REGISTER,WARP}_WATERMARKS

  • nvk: Support blackwell in max_warps_per_mp_for_sm

  • nak/spill_values: Spill constants across edges if needed

  • nak: Handle idp4 ureg latencies

  • nvk: SET_STATISTICS_COUNTER at start of meta_begin

  • nvk: Override render enable for blits and resolves

  • wsi/headless: Override finish_create

  • nak: Remove hfma2 src 1 modifiers

  • nak: Add Src::is_unmodified() helper

  • nak: Check that swizzles are none

Mi, Yanfeng (2):

  • anv: Support putting image base address and image params in surface state

  • anv: add emulated 64bit integer storage support

Michael Cheng (5):

  • anv: Fix missing Perfetto trace for as build

  • Revert “anv: Fix missing Perfetto trace for as build”

  • vulkan : Pass in number of tlas/blas being built

  • anv: Change as_build to show num tlas/blas

  • vulkan: add leaves and ir_leaves sizes to encode for utrace.

Michel Dänzer (1):

  • egl/glx/sw: Check xcb_query_extension_reply return value for MIT-SHM

Mike Blumenkrantz (97):

  • zink: emit SpvCapabilityDemoteToHelperInvocation for IsHelperInvocation

  • zink: implement ops for KHR_shader_subgroup

  • aux/trace: trace tex2d_from_buf for samplers/images

  • gallium: add a pipe_tex2d_from_buf struct

  • rusticl: stop clearing shader images after every dispatch

  • zink: also refcount needs_present from frontbuffer flush

  • zink: delete some old DGC remnants

  • zink: guard rebar check against fallback heap detection

  • ac/surface: always allow LINEAR modifier for color formats

  • radv: fix error reporting for VkExternalMemoryTypeFlagBitsKHR

  • radv: stop blocking non-2D import/export ops

  • radv: print stringname for VkExternalMemoryHandleTypeFlagBits error

  • driconf: move a glthread viewperf option to global

  • driconf: add GL_EXT_shader_image_load_store for viewperf to fix crashing

  • zink: only enable unsynchronized_texture_subdata with HIC

  • zink: never try to oom flush during unsync texture upload

  • zink: apply layer/depth to clear handling

  • zink: wait on tc fence before checking for fd semaphore

  • zink: handle buffer import/export

  • zink: verify that adding a dmabuf bind actually chooses a modifier

  • zink: support cl_gl_sharing if dmabuf is supported

  • egl/x11: delete some dri2 remnants

  • zink: force cached mem for streaming uploads

  • zink: always fully unwrap contexts

  • zink: use a separate ralloc ctx for zink_program objects

  • zink: put zink_program::reference on separate cacheline

  • dri: delete INVALIDATE extension

  • zink: enable single-plane modifiers for generic 2D exports

  • zink: clamp UBO sizes instead of asserting

  • anv: support all dimensions of image for LINEAR dmabufs

  • aux/trace: truncate descriptor unbinds

  • zink: always check submit_count to disambiguate when checking/waiting

  • zink: don’t reset all batch states when stalling

  • zink: alloc bo ids for non-slab bos too

  • zink: explicitly check usage in buffer barriers

  • svga/ci: disable vmware farm

  • llvmpipe: pass layer count to rast clear

  • gallium: fix pipe_framebuffer_state::view_mask

  • mesa: add error handling for OVR_multiview

  • mesa: avoid creating incomplete surfaces when multiview goes out of range

  • zink: improve precision on changes to depth bias between draws

  • zink: delete zink_batch_state::ref_lock

  • zink: split set_vertex_buffers to avoid conditionals

  • zink: start using ctx->vertex_state_changed again

  • zink: use VKCTX for vertex buffer binds and delete unused screen local

  • zink: only add fb surf refs on unbind

  • radv: rewrite radv_get_line_mode() conditional

  • radv: store num_attributes to shader info

  • radv: store vertex prolog simple input check to cmdbuf on vs bind

  • radv: roll line topology dynamic state changes into existing rast samples flag

  • radv: eliminate a memset in radv_get_vbo_info()

  • radv: inline some vertex descriptor functions

  • radv: split out dynamic vertex input descriptor writing

  • radv: move non_trivial_format calc to dynamic VI bind

  • radv: get vbo info directly into dgc upload

  • radv: don’t unnecessarily flag prolog recalc when binding VBOs

  • zink: disable reordering on compute contexts

  • zink: extract some shader image resource binding code

  • zink: extract some samplerview bind/unbind code

  • zink: support crazy CL buffer-to-texture extension

  • st/drawpixels: move sv unref out to callers

  • gallium: eliminate frontend refcounting from samplerviews

  • gallium: make pipe_sampler_view::reference non-atomic

  • mesa: remove st_sampler_view::private_refcount

  • aux/trace: set sampler_view_release pointer

  • gallium: delete tests

  • util/tests: move u_half_test into half_float_test

  • gallium: delete u_cache

  • meson: add i915 to ‘all’ build for gallium-drivers

  • egl/x11: always expose EXT_swap_buffers_with_damage

  • egl: delete invalidate_available flag

  • egl: move _EGLDisplay DriverData association into dri2_display_create

  • egl: hoist dri2 display creation up a level

  • egl/x11: split out dri2 init entirely

  • egl: move kopper detection to display creation

  • egl/x11: hoist up dri2_get_xcb_connection call

  • egl/x11: simplify a kopper check on init

  • egl/x11: hoist up swrast/zink driver_name setting

  • egl/x11: hoist and document dri3_x11_connect() during init

  • egl: delete dri2_egl_driver_fail

  • egl/x11: consolidate LIBGL_DRI3_DISABLE use on init

  • egl/x11: unify swrast/kopper/dri3 paths a bit

  • gallium: delete pipe_surface::width and pipe_surface::height

  • zink: fix refcounting of zink_surface objects

  • zink: block streaming cached uploads on unsynchronized/persistent maps

  • zink: implement unsynchronized staging uploads for buffers

  • zink: block inferred loading for swrast

  • gallium/util: check nr_samples in pipe_surface_equal()

  • tu: check for valid descriptor set when binding descriptors

  • meson: deprecate gallium-nine

  • meson: deprecate gallium-xa

  • zink: don’t set shared block stride without KHR_workgroup_memory_explicit_layout

  • zink: stop setting ArrayStride on image arrays

  • zink: use implicit stride in ntv for temp vars

  • ci: update VVL to current week

  • zink: verify that surface exists when adding implicit feedback loop

  • egl: fix sw fallback rejection in non-sw EGL_PLATFORM=device

Mohamed Ahmed (7):

  • nvk: Add NVK_MAX_IMAGE_PLANES for images

  • nvk: Add NVK_MAX_SAMPLER_PLANES for samplers

  • nil: Use multiplanar init_info during image creation

  • nil: Force smallest block size for images meant for Vulkan Video

  • nvk: Set NIL_IMAGE_USAGE_VIDEO_BIT for images meant for video usage

  • nil: Add an API to clamp max image alignment

  • nvk: Implement VK_MESA_image_alignment_control

Mykhailo Skorokhodov (1):

  • drirc/anv: force_vk_vendor=-1 for Drive Beyond Horizons

Natalie Vock (30):

  • vulkan/bvh: Move leaf builder code to header

  • vulkan/bvh: Add option to override leaf builder SPIR-Vs

  • vulkan/bvh: Add optional fine-grained instance node bounds calculation

  • radv/bvh, vulkan/bvh: Move AccelerationStructureInstance to vk_build_helpers

  • radv/bvh: Remove unused build_instance helper

  • radv/bvh: Prefix RADV-specific node functions with radv_

  • radv/bvh: Add custom leaf node builder

  • mailmap: Update my name

  • radv/rt: Don’t allocate the traversal shader in a capture/replay range

  • aco/ra: Use iterators for linear VGPR copy extraction

  • aco/ra: Use struct for parallelcopies

  • aco/ra: Add option to skip renaming for parallelcopies

  • aco/ra: Handle temps fixed to different regs in different operands

  • aco/tests: Add tests for precolored operands in different regs

  • aco/ra: Assert operands only clear their own id

  • radv/rt: Limit monolithic pipelines to 50 stages

  • radv/rt: Guard leaf encoding by leaf node count

  • radv/rt: Flush L2 after writing internal node offset on GFX12

  • ac/llvm: Don’t use getTriple() on LLVM21+

  • aco: Add get_temp_reg_changes helper

  • aco/spill: Invert reloads map

  • aco: Add RegisterDemand(Temp) constructor

  • aco/spill: Allow spilling live-through operands

  • radv/rt: Flush CP writes from the common BVH framework with INV_L2 on GFX12

  • vulkan/bvh: Move first PLOC task_count fetch inside PHASE

  • aco: Fix RT VGPR limit on Navi31/32, GFX11.5, GFX12

  • aco: Make private_segment_buffer/scratch_offset per-resume

  • radv: Add radv_emulate_rt drirc and enable for Indiana Jones TGC

  • radv: Add radv_enable_float16_gfx8 drirc and enable for Indiana Jones TGC

  • radv/ci: Test FP16 for GFX8

Nikita Popov (2):

  • clover: Don’t include libclc headers

  • clover: Fix MSVC build

Nikolas Zimmermann (1):

  • etnaviv: drm: Add DRM_RDWR permissions to drmPrimeHandleToFD() call.

Olivia Lee (1):

  • panfrost: allow promoting sysval UBO to push constants

Patrick Lerda (27):

  • r600: fix r600_init_shader_caps() has_atomics issue

  • r600: fix evergreen_emit_vertex_buffers() related cl regression

  • r600: fix cayman sfn_nir_legalize_image_load_store ssa dominance

  • r600: prepare the lds constant buffer to be shared

  • r600: implement a conformant gl_VertexID with base offset

  • r600: fix the indirect draw 8-bits path

  • r600: fix emit_image_size() range base compatibility

  • r600: update evergreen_convert_border_color()

  • r600: update cayman_convert_border_color()

  • r600: fix cayman main non-deterministic behavior problem

  • r600: update the software fp64 support

  • radeonsi: fix clear_depth_stencil refcnt imbalance

  • r600: move stores to the end of shader when required

  • r600: fix clear_depth_stencil refcnt imbalance

  • r600: fix textures with swizzles limited to zero and one

  • r600: implement EXT_window_rectangles

  • r600: remove deprecated NIR_PASS_V

  • r600: fallback to util_blitter_draw_rectangle when required

  • r600: fix pa_su_vtx_cntl rounding mode

  • r600: fix points clipping

  • r600: enable ARB_compute_variable_group_size

  • r600: clean up not used fields detected by clang

  • i915: fix i915_set_vertex_buffers() related refcnt imbalance and remove redundancies

  • i915: fix slab_create() related memory leaks

  • i915: fix nir_to_tgsi() related memory leak

  • i915: fix draw_create_fragment_shader() related memory leak

  • mesa_interface: fix legacy dri2 compatibility

Patrick Nicolas (1):

  • radv/video: Add low latency encoding

Paul Gofman (1):

  • radv/amdgpu: Fix hash key in radv_amdgpu_winsys_destroy().

Paulo Zanoni (10):

  • brw: increase brw_reg::subnr size to 6 bits

  • brw: don’t always set cond_modifier on parsed assembly instructions

  • brw: don’t mark instructions read from text assembly as compacted

  • brw: add instructions missing from is_control_flow()

  • brw: extend the NOP+WHILE workaround

  • drirc/anv: DiggingGame.exe needs force_vk_vendor=-1

  • intel/i915: restrict the RAM size restrictions to Anv

  • anv/xe: detect the newer xe.ko memory reporting model and act accordingly

  • anv: restore the old behavior of up to 75% of RAM for the system heap

  • anv: add ANV_SYS_MEM_LIMIT for debugging system memory restrictions

Pavel Ondračka (22):

  • i915/ci: use debian-build-testing instead of debian-testing

  • r300: remove some dead code in redeon_program_alu

  • r300: do not limit maximum TEX group for R300/R400

  • r300: copy propagate constant swizzles

  • ci: fix debian-build-testing BUILDTYPE

  • ci: disable LTO for nightly debian-build-testing

  • i915/ci: update expectations

  • i915: rework shader compile failures reporting

  • mesa: properly signal report_compile_error to drivers

  • nine: set pipe_shader_state.report_compile_errors

  • r300: stop reporting compile failures in finalize_nir

  • r300: forward all compile failures to state tracker

  • r300: do not include newline in the error messages

  • r300,i915: update CI expectations

  • r300: fix temps counting for shader-db stats

  • r300/ci: add dEQP job with RADEON_DEBUG=notcl

  • r300: remove usage of NIR_PASS_V

  • r300: fix INV and BIAS presubtract on R300/R400

  • r300: remove finalize_nir

  • r300: remove support for tgsi_texcoord

  • r300: fix crash when creating surfaces

  • r300/ci: update expectations

Peyton Lee (6):

  • radeonsi/vpe: check reduction ratio

  • radeonsi/vpe: support geometric scaling

  • amd/gmlib: add gmlib for radeonsi

  • radeonsi/vpe: vpe support tonemapping

  • radeonsi/vpe: vpe support hdr input

  • radeonsi/vpe: fix background issue

Philipp Zabel (3):

  • etnaviv/ml: Drop duplicated function reorder_for_hw_depthwise()

  • etnaviv/ml: Fix padding input/output tensor zero points

  • teflon: Log (un)supported operations

Pierre-Eric Pelloux-Prayer (16):

  • radeonsi: update si_need_gfx_cs_space upper bound

  • winsys/amdgpu: treat cs overflow as context lost

  • radeonsi: disable dcc when external shader stores are used

  • tc: add missing TC_SENTINEL for TC_END_BATCH

  • mesa/st: call _mesa_glthread_finish before _mesa_make_current

  • tc: flag closed batches

  • st/mesa: fix nir_load_per_vertex_input parameter

  • radeonsi: cache u_trace_perfetto_active value

  • radeonsi: tag perfetto conditions as unlikely

  • radeonsi: add pid/tid to the vk_queue_submit event

  • radeonsi: guard perfetto support with ifdef in si_draw

  • ac/nir: fix nir_metadata value of ac_nir_lower_image_opcodes

  • radeonsi/video: require has_image_opcodes for full modifier support

  • radeonsi: use composed swizzle in cdna_emu_make_image_descriptor

  • winsys/amdgpu: disable VM_ALWAYS_VALID

  • radeonsi: fix potential use after free in si_set_debug_callback

Pohsiang (John) Hsu (7):

  • d3d12: fix incorrect IsRefUsedByCurrentPic marking for P Frame

  • d3d12: log all the field of dpb buffer for better diagnostic, cosmetic space removal

  • d3d12: initialize ReconstructedPicTexture

  • d3d12: fix start code prevention in write_sei_nalu()

  • d3d12: fix d3d12_video_nalu_writer_h264::write_slice_svc_prefix

  • d3d12: use log2_max_pic_order_cnt_lsb_minus from upper layer

  • d3d12: use log2_max_pic_order_cnt_lsb_minus4 from upper layer for h264

Qiang Yu (78):

  • lavapipe: fix min_vertex_pipeline_param

  • gallium: fix ddebug and noop screen caps init

  • gallium,clover: add compute caps used by clover only

  • r600: init compute caps without ir_type param

  • radeonsi: init compute caps without ir_type param

  • gallium,mesa: remove ir_type param when get_compute_param

  • gallium: add pipe_shader_caps and pipe_compute_caps

  • gallium: copy shader and compute caps for ddebug/noop/trace

  • tgsi: add tgsi_exec_init_shader_caps

  • gallivm: add gallivm_init_shader_caps

  • draw: add draw_init_shader_caps

  • agx: init shader and compute caps

  • crocus: init shader and compute caps

  • d3d12: init shader and compute caps

  • etnaviv: init shader caps

  • freedreno: init shader and compute caps

  • i915: init shader caps

  • iris: init shader and compute caps

  • lima: init shader caps

  • llvmpipe: init shader and compute caps

  • nouveau/nv30: init shader caps

  • nouveau/nv50: init shader and compute caps

  • nouveau/nvc0: init shader and compute caps

  • panfrost: init shader and compute caps

  • r300: init shader caps

  • r600: init shader and compute caps

  • radeonsi: init shader and compute caps

  • softpipe: init shader and compute caps

  • svga: init shader and compute caps

  • tegra: init shader and compute caps

  • v3d: init shader and compute caps

  • vc4: init shader caps

  • virgl: init shader and compute caps

  • zink: init shader and compute caps

  • nine: GET_SHADER_CAP use pipe_shader_caps

  • lavapipe: change min shader param to use pipe_shader_caps

  • gallium,mesa: replace get_shader_param with pipe_shader_caps access

  • clover: replace get_compute_param with pipe_compute_caps

  • gallium,mesa: replace get_compute_param with pipe_compute_caps

  • rusticl: replace get_shader_param with pipe_shader_caps

  • rusticl: replace get_compute_param with pipe_compute_caps

  • asahi: remove shader and compute get param

  • crocus: remove shader and compute get param

  • d3d12: remove shader and compute get param

  • etnaviv: remove shader get param

  • freedreno: remove shader and compute get param

  • i915: remove shader get param

  • iris: remove shader and compute get param

  • lima: remove shader get param

  • llvmpipe: remove shader and compute get param

  • nouveau/nv30: remove shader get param

  • nouveau/nv50: remove shader and compute get param

  • nouveau/nvc0: remove shader and compute get param

  • panfrost: remove shader and compute get param

  • r300: remove shader get param

  • r600: remove shader and compute get param

  • radeonsi: remove shader and compute get param

  • softpipe: remove shader and compute get param

  • svga: remove shader and compute get param

  • tegra: remove shader and compute get param

  • v3d: remove shader and compute get param

  • vc4: remove shader get param

  • virgl: remove shader and compute get param

  • zink: remove shader and compute get param

  • gallium/aux: remove aux shader param get function

  • gallium: remove screen shader and compute get param callback

  • doc,src: replace doc and comments for shader and compute cap

  • ac/surface: fix radv import dmabuf from radeonsi

  • gallium/ddebug: add missing modifier callback

  • radeonsi: fix has_non_uniform_tex_access info

  • radeonsi,util: add more usage for AMD_FORCE_SHADER_USE_ACO

  • radeonsi: fix GravityMark corruption when use aco

  • egl: decouple dmabuf import/export cap from xserver support

  • dri: dmabuf cap does not rely on winsys multibuffer support

  • glx/egl/x11: fix x11_dri3_check_multibuffer

  • kopper: remove unused function definition

  • dri,egl,glx: remove redundant usage of HAVE_X11_DRM

  • llvmpipe/ci: change for oes_egl_image_external_essl3

Rebecca Mckeever (27):

  • pan/format: Use HW version to determine siting for YUV 422 formats

  • pan/texture: Only use plane_chroma_2p for chroma planes

  • vk/meta: Extend copy/fill/update helpers to support YCbCr

  • util/hash_table: Add _mesa_hash_table_u64_replace()

  • panvk: Allow a 32-bit binding value in desc id key and use 64-bit keys

  • panvk: Move single-plane views of multiplane formats to pview.planes[0]

  • panvk: Change immutable_samplers to panvk_sampler **

  • panvk: Create helper function for sampler descriptor emission

  • panvk: Update panvk_get_desc_stride prototype

  • panvk: Move mali_texture_packed structs in panvk_image_view to a union

  • panvk: Use multiple sampler planes and one texture descriptor per plane

  • panvk: Fix assertion in is_disjoint()

  • panvk: Add YCbCr sampler NIR lowering pass

  • panvk: Split get_format_properties into format features helper functions

  • panvk: Report formats not supported by HW as unsupported

  • panvk: Enable YCbCr support for v10+

  • vk/image: Add vk_image_can_be_aliased_to_yuv_plane() helper

  • panvk: Use vk_image_can_be_aliased_to_yuv_plane() helper

  • panvk: Allow 3-byte formats

  • panfrost: Add BI_OPCODE_WMASK to bi_instr_uses_helpers

  • panvk: Enable VK_KHR_shader_subgroup_uniform_control_flow for v10+

  • panvk: Enable VK_KHR_shader_maximal_reconvergence for v10+

  • panvk: Enable VK_KHR_shader_quad_control for v10+

  • panvk: Add STORAGE_IMAGE_BIT feature for formats supporting sampled images

  • panvk: Enable shaderStorageImageExtendedFormats

  • panvk: Remove lower_tg4_broadcom_swizzle from panvk_preprocess_nir()

  • panvk: Support shaderImageGatherExtended

Renato Pereyra (1):

  • perfetto/android: align datasource names with tooling expectations

Rhys Perry (41):

  • nir,aco,radv: add align_mul/offset to buffer_amd intrinsics

  • nir/opt_offsets: don’t check NUW for unswizzled buffer_amd

  • ac/nir/ngg: update bit_sizes_int

  • nir/load_store_vectorize: handle load_buffer_amd/store_buffer_amd

  • ac/nir: set memory_modes for lowered TES input loads

  • radv: don’t use bit_sizes_int to skip nir_lower_bit_size

  • radv: move nir_opt_shrink_vectors later

  • radv: DCE before nir_opt_shrink_vectors

  • nir/load_store_vectorize: fix sorting of vectors in add_to_entry_key

  • nir/algebraic: optimize ishl(iadd(iadd(a, #b), c), #d)

  • radv: vectorize descriptor loads

  • radv: vectorize lowered shader IO

  • nir/use_dominance: invalidate nir_metadata_instr_index

  • nir/opt_move: invalidate nir_metadata_instr_index

  • nir/lower_io_arrays_to_elements: invalidate metadata

  • nir/find_array_copies: invalidate nir_metadata_instr_index

  • nir/linking_helpers: invalidate metadata in nir_link_opt_varyings

  • radv/rt: correctly preserve metadata in move_rt_instructions

  • nir: add NIR_DEBUG=extended_validation

  • nir: add NIR_DEBUG=invalidate_metadata

  • nir/validate: assert that if condition dominates use

  • aco: validate operands fixed to definitions

  • aco/ra: reverse renaming of operands outside update_renames

  • aco/ra: unconditionally call undo_renames

  • radeonsi: fix invalidation of metadata in si_nir_lower_abi

  • ac/nir: don’t cross swizzle elements when vectorizing buffer_amd intrinsic

  • ac/nir: set higher alignment for some swizzled store_buffer_amd

  • ac/nir: fix tess factor optimization when workgroup barriers are reduced

  • aco: insert dependency waits in certain situations

  • radv: improve radv_get_max_waves for multi-wave workgroups with LDS

  • radv: don’t assume WGP mode in radv_get_max_waves

  • amd/drm-shim: add gfx1201

  • aco/ra: fix free register counting when moving variables

  • radv/winsys: increase gfx12 vgprs for null winsys

  • radv/winsys: set gart_page_size for null winsys

  • radv/winsys: set has_distributed_tess for null winsys

  • aco/gfx12: don’t use second VALU for VOPD’s OPX if there is a WaR

  • aco: combine VALU lanemask hazard into VALUMaskWriteHazard

  • aco/gfx11: create waitcnt for workgroup vmem barriers

  • aco: fix get_temp_reg_changes with clobbered operands

  • aco: swap the correct v_mov_b32 if there are two of them

Rob Clark (17):

  • freedreno+tu: Add new virtgpu caps

  • tu: Suballoc VkEvent BOs

  • tu: Don’t emit SP_PS_2D_WINDOW_OFFSET on a6xx

  • tc: Add missing tc_set_driver_thread()

  • freedreno: Wait for imported syncobj fences to be available

  • ci: Re enable fd-farm

  • freedreno/ci: Disable traces job until piglit is fixed

  • ci: Re enable fd-farm

  • ir3: Comment re-indent

  • tu/vdrm: Fix userspace fence cmds

  • tu: Avoid extraneous set_iova

  • tu: Add some func traces

  • util: Add drmSyncobj shim

  • panvk: Remove dependency on vk_device::drm_fd

  • vulkan: Use syncobj shim

  • virtio/vdrm: Add vtest backend

  • tu: vdrm vtest support

Robert Mader (4):

  • llvmpipe: Take offset into account when importing dmabufs

  • llvmpipe: Free dummy_dmabuf on shutdown

  • gallivm: Re-add check for passmgr before disposing it

  • meson: Bump minimum version to 1.3.0

Rohan Garg (7):

  • anv: refactor add_aux_surface_if_supported to use a common variable

  • anv,blorp,isl: handle compressed CPS surfaces through the depth stencil hw

  • anv: CPB surfaces that are used as color attachments or for stores cannot be compressed

  • anv: no need to handle transitions for CPB surfaces

  • anv: separate fast clear handling for compressed CPS

  • isl: enable CPB compression

  • anv: re enable compression for CPS surfaces on platforms other than Xe

Roland Scheidegger (3):

  • llvmpipe: don’t assert on exceeding if_stack size

  • llvmpipe: Fix overflow issues calculating loop iterations for aniso

  • llvmpipe: Fix alpha-to-coverage without dithering

Ruijing Dong (1):

  • radeonsi/vcn: vcn5 roi fix

Ryan Mckeever (5):

  • pan/format: Update format flags to follow HW spec

  • pan/format: Add PAN_BIND_STORAGE_IMAGE flag

  • panvk: Enable KHR_format_feature_flags2 and use them

  • panvk: shaderStorageImageReadWithoutFormat support

  • panvk: shaderStorageImageWriteWithoutFormat support

Sagar Ghuge (12):

  • intel/genxml: Update BLEND_STATE_ENTRY structure

  • anv: Enable simpleFloatBlendEnable on Xe3+

  • iris: Enable simpleFloatBlendEnable on Xe3+

  • blorp: Enable SimpleFloatBlendEnable on Xe3+

  • intel/compiler: Drop primitive leaf desc load code

  • anv: Exclude non-standard block shapes on Xe2+

  • intel/compiler: Zero out the header for texel fetch

  • anv: CPS LOD Compensation Enable is deprecated on Xe2+

  • intel/compiler: Add support for MSAA typed load/store messages

  • intel/compiler: Lower sample index into coord for MSRT messages

  • intel/compiler: Don’t lower 64bit data memory access on LSC

  • intel/compiler: Fix stackIDs on Xe2+

Samuel Pitoiset (227):

  • ac,radv,radeonsi: add new GFX12_DCC_WRITE_COMPRESS_DISABLE tiling flag

  • radv/meta: add missing pipeline lookups

  • radv/nir: update radv_nir_opt_tid for derivative group quads

  • radv: implement derivative group quads on GFX12

  • radv: advertise computeDerivativeGroupQuads on GFX12

  • radv/meta: stop using string keys also for DGC and query objects

  • util/disk_cache: add a new helper to create a disk cache

  • vulkan/runtime: allow to use a different disk cache

  • radv: fix caching on-demand meta shaders

  • radv: fix adding the BO to cmdbuf list when starting conditional rendering

  • radv: fix fetching draw vertex data from counter buffers with transform feedback

  • radv: remove redundant drawCount == 0 for indirect mesh/task draws

  • radv: use radv_indirect_dispatch() more

  • radv: rework passing dispatch info via radv_dispatch_info

  • radv: rework passing draw info via radv_draw_info

  • radv: do not keep track of the streamout binding buffer

  • nir: adjust number of components for cmat_muladd_amd

  • radv/nir: add a struct for parameters to cooperative matrix lowering

  • radv/nir: add cooperative matrix lowering for GFX12

  • radv: advertise VK_KHR_cooperative_matrix on GFX12

  • radv/meta: use BDA for query resolves

  • radv/meta: compute the destination addr earlier for query resolves

  • radv/meta: simplify creating buffers for R32G32B32 operations

  • radv: remove unused device memory init/finish helpers

  • radv/video: pass addr to send_cmd()

  • radv/rmv: pass addr to log_resource_bind_locked()

  • radv: pass addr to radv_copy_buffer()

  • radv: rename radv_buffer::bo_va to addr

  • radv: compute radv_buffer::addr at bind time

  • radv: use radv_buffer::addr more

  • radv/video: fix missing image offset when computing VA

  • radv/meta: use BDA for clear HTILE mask

  • radv/meta: use BDA for copying VRS rates to HTILE

  • radv: use BDA for the uploaded parameters with DGC

  • radv: simplify determining VBO size

  • radv/meta: remove the heuristic that prefers CP DMA for GTT BOs

  • radv/meta: disable conditional rendering for fill/update buffer operations

  • radv/meta: inline copy_buffer()

  • radv: rename fill/copy memory helpers

  • radv/meta: add radv_{copy,fill,update}_memory() helpers

  • radv: compute VBO addr at bind time

  • radv: stop relying on VkBuffer for VBO

  • ac/gpu_info: add gfx12_supports_dcc_write_compress_disable

  • radv: add initial DCC support on GFX12

  • vulkan: add descriptor buffer support to compute astc decoder

  • radv/meta: switch to descriptor buffers

  • radv: remove radv_buffer_{init,finish}() helpers

  • radv: remove radv_buffer_view_{init,finish}() helpers

  • radv: fix adding the VRS image BO to the cmdbuf list on GFX11

  • radv/meta: add BOs to cmdbuf list earlier for image copy operations

  • radv/meta: pass the buffer addr to SDMA image buffer copy operations

  • radv/meta: use radv_copy_memory() instead of radv_copy_buffer()

  • radv/meta: remove the buffer dependency for image copy operations

  • radv: stop using image binding offset when exporting BO metadata

  • radv: remove redundant zero initialization when creating images

  • radv: rename radv_image::bindings::bo_va to addr

  • radv: compute radv_image::bindings::addr at bind time

  • radv: use radv_image::bindings::addr more

  • radv: fix adding the BO for unaligned SDMA copies to the cmdbuf list

  • vulkan: constify vk_acceleration_structure_get_va()

  • radv: use radv_CmdDispatchIndirect() in the accel struct path

  • radv/meta: use radv_copy_memory() for the FMASK copy

  • radv: use radv_copy_memory() for accel structure updates

  • radv/meta: inline radv_copy_buffer()

  • radv/meta: rename image<->buffer copies helpers

  • radv/video: fix adding the query pool BO to the cmdbuf list

  • radv: stop computing the UUID using the physical device cache key

  • radv: fix missing SQTT barriers for fbfetch color/depth decompressions

  • radv: reserve bits explicitly for cache key structs

  • vulkan: filter duplicate pNext struct at device creation

  • radv: replace radv_image::shareable by vk_image::external_handle_types

  • radv: remove meaningless TODOs in radv_GetDeviceImageMemoryRequirements()

  • radv: handle OOM error properly when selecting image modifier

  • radv/meta: inline radv_meta_get_view_layer()

  • radv/meta: remove useless assertion in when copy VRS to HTILE

  • radv: check HTILE compression for depth/stencil images per level

  • radv: enable DCC fast clears for 8bpp/16bpp on GFX11

  • radv: fix re-emitting fragment output state when resetting gfx pipeline state

  • docs: add missing RADV_PERFTEST=video_encode description

  • radv: fix trap handler exception options

  • radv: use radv_emulate_rt() more

  • radv: remove redundant radv_instance::drirc::rt_wave64

  • docs: adjust NGG culling options description

  • spirv: move workarounds to an inner struct in spirv_to_nir_options

  • spirv/tests: initialize compiler options in constructor

  • spirv/tests: add a test for NonSemantic.DebugBreak

  • spirv/tests: add a test for force_ssbo_non_uniform

  • spirv/tests: add a test for force_tex_non_uniform

  • spirv/tests: add a test for lower_terminate_to_discard

  • ci: update VKCTS main to ba86fb95004331f2cf571dd9adefe2458290ee11

  • radv: switch to device address from vk_buffer

  • ci/b2c: fix passing B2C_* variables

  • radv/ci: re-enable ET2C emulation testing on non-native GPUs

  • radv: add RADV_DEBUG=pso_history

  • ac,radv: add a workaround for a hw bug with primitive restart on GFX10-GFX10.3

  • radv/ci: enable RADV_PERFTEST=video_{decode,encode} on few GFX9+ GPUs

  • aco/tests: use GFX1201 instead of GFX1200

  • radv/winsys: use real info for GFX12 in the null winsys

  • radv: fix a GPU hang with inherited rendering and HiZ/HiS on GFX1201

  • radv/amdgpu: fix device deduplication

  • meson: add build-radv-tests option

  • ci: enable build-radv-tests for debian-clang and debian-vulkan

  • radv/winsys: enable has_timeline_syncobj for the null winsys

  • radv: add a small framework for RADV specific tests

  • radv/tests: add a test to verify that pipelineCacheUUID is invariant

  • radv/tests: add a test to verify that pipeline hash matches RGP<->Fossilize

  • radv/tests: add few tests that verify drirc options

  • radv: update conformance version

  • aco: do not apply OMOD/CLAMP for pseudo scalar trans instrs

  • radv/ci: enable RADV_PERFTEST=video_decode,video_encode on TAHITI,HAWAII and POLARIS10

  • ac/rgp: few fixes for GFX11.5

  • ac/rgp: add GFX12 definitions

  • ac/rgp: bump version to 1.6

  • ac/rgp: bump instrumentation API version to 1.5

  • ac/sqtt: fix registers programming for GFX12

  • radv: enable RGP on GFX12

  • radv: emit a dummy PS state for noop FS on GFX12

  • radv: track redundant register writes for PA_SC_HISZ_CONTROL on GFX12

  • radv: fix creating pipeline binary from the traversal shader

  • radv: use radv_sdma_emit_nop() more

  • radv: add more SDMA emit helpers

  • radv: add a helper to know whether compute queue is enabled

  • radv: remove meaningless comment when resetting SQTT trace

  • radv: add a helpers to know whether video decode/encode queues are enabled

  • docs: update documentation for RADV_PERFTEST=video_decode,video_encode

  • radv: fix bpe for the stencil aspect of depth/stencil copies on transfer queue

  • radv: fix compresed depth/stencil copies on transfer queue

  • radv: remove useless parameter to radv_sdma_get_buf_surf()

  • radv: cleanup passing the aspect mask for SDMA operations

  • radv: remove radv_force_pstate_peak_gfx11_dgpu=true for Helldivers 2

  • radv: use PM4 for setting specific graphics registers in the preamble

  • radv/meta: fix color<->depth/stencil image copies

  • radv/video: use a pointer to write the total task size

  • radv/video: rework command buffer emission

  • ci: uprev vkd3d-proton to 078f07f588c849c52fa21c8cfdd1c201465b1932

  • radv/ci: remove vkcts-navi21-llvm-valve completely

  • radv/ci: remove radv-stoney-flakes.txt

  • radv/ci: drop aco suffix for CI files

  • radv/ci: rename radv-gfx1200 to radv-gfx1201

  • radv/ci: delete empty radv-hawaii-skips.txt

  • radv: replace radeon_set_reg_seq by a macro

  • radv: switch all emit helpers to macros

  • radv: move the optimized context reg macros with other similar ones

  • radv: rework radeon_set_uconfig_perfctr_reg_seq to use amd_ip_type

  • radv: add new helper to emit PKT3_EVENT_WRITE for sampling queries

  • radv: add radeon_event_write() macros

  • radv/ci: fix renaming the VKCTS job for tahiti

  • radv: rework the shader pointer emit as macros

  • radv: apply some cosmetic changes for future begin/end CS sequences

  • radv: add a helper to emit indirect buffer for draws/dispatches

  • radv: add a helper to emit SPM muxsel

  • radv: slightly change the COND_EXEC for sampling performance counters

  • radv: use radv_cs_write_data_imm() more

  • radv: add more helpers to start/stop perfcounters

  • radv/video: slightly change radv_vcn_sq_header()

  • radv: add a helper to emit a PKT3_COPY_DATA with an immediate

  • radv: add a helper to emit PM4 commands to a CS

  • radv: skip FCE for comp-to-single fast clears with DCC MSAA

  • radv: rework radv_fast_clear_flush_image_inplace()

  • radv: add radv_fmask_color_expand()

  • radv: rework radv_handle_color_image_transition()

  • radv: inline radv_fast_clear_flush_image_inplace()

  • radv: make sure to always decompress FMASK before expanding it

  • radv: disable TC-compatible CMASK with {FMASK,DCC}_DECOMPRESS

  • ac,radv,radeonsi: use PM4 for shadowed registers

  • radv: do not trigger FCE or FMASK decompress on compute queue

  • radv: add queue family assertions when doing decompression passes

  • radv: remove useless use of radv_image_use_comp_to_single()

  • radv: only enable HTILE for depth/stencil attachment images

  • radv: rework command buffer emission with begin/end sequences

  • radv: switch back radeon_cmdbuf to use 32-bit counters

  • radv/ci: stop skipping one memory test due to timeouts

  • radv/ci: remove all skips for STONEY

  • ac/surface: fix selecting preferred alignments for HiZ/HiS on GFX12

  • Revert “radv: program SAMPLE_MASK_TRACKER_WATERMARK optimally for GFX11 APUs”

  • Revert “radeonsi/gfx11: program SAMPLE_MASK_TRACKER_WATERMARK optimally for APUs”

  • radv: fix ignoring conditional rendering with vkCmdResolveImage()

  • radv: add new helper to suspend/resume user conditional rendering

  • radv: rework suspend/resume user conditional rendering

  • ac,radv: remove has_scheduled_fence_dependency

  • ac/gpu_info: bump required DRM minor version to 3.42.0 (kernel 5.15+)

  • radv: remove useless FDCC_ENABLE bitfield clear on GFX12

  • radv: tidy up radv_emit_raster_state()

  • radv: restore DB_DFSM_CONTROL properly when POPS isn’t used

  • radv: tidy up radv_emit_db_shader_control()

  • radv: split framebuffer color state emission for GFX12

  • radv: split framebuffer depth/stencil state emission for GFX12

  • radv: split null framebuffer state emission for GFX12

  • radv: stop emitting CB_FDCC_CONTROL to zero on GFX11-GFX11.5

  • radv: do not emit the VRS surface VA when it’s not enabled

  • radv: add a fuction to emit the VRS surface on GFX11

  • radv: determine if HiZ/HiS is enabled earlier on GFX12

  • radv: add a workaround for buggy HiZ/HiS on GFX12

  • radv: apply the workaround for buggy HiZ/HiS on GFX12 for DGC

  • radv: add before/after draw functions for DGC

  • radv: move emitting raster and depth/stencil state slightly earlier

  • radv: emit conservative raster mode as part of the MSAA state

  • radv: configure COVERAGE_TO_SHADER_SELECT only if conservative rast is enabled

  • radv: track more MSAA related register writes

  • radv: regroup emitting all MSAA states in one function

  • radv: add clip rects state bit for emitting discard rectangles

  • radv: remove an old workaround for D3D9 with DXVK 2.3.0 and older

  • radv: move the disable_trunc_coord drirc at instance/pdev level

  • radv: move emitting more fb registers when rendering begins

  • radv: use consecutive registers for PA_SC_WINDOW_SCISSOR_{TL,BR}

  • radv: track redundant DB_RENDER_OVERRRIDE register writes on GFX12

  • radv: reduce the number of emitted DWORDS for MSAA 8x user sample locs

  • radv: tidy up radv_emit_raster_state()

  • radv: tidy up radv_emit_hw_ngg()

  • radv: add macros for paired context registers on GFX12

  • radv: use paired context regs when optimal on GFX12

  • radv: only enable DCC for invisible VRAM on GFX12

  • ci: update VKCTS main to 76c1572eaba42d7ddd9bb8eb5788e52dd932068e

  • radv: allocate the SPM BO in GTT for faster readback

  • radv: print more error messages during SPM initialization

  • ac/perfcounter: add support for GFX12

  • radv/sdma: simplify configuring the number of uncompressed DCC blocks

  • radv/sdma: use SDMA5_DCC_xxx bitfields

  • radv/sdma: remove redundant check for compression when getting metadata

  • radv/sdma: use the correct helper to get the number type field

  • radv/sdma: add a new flag to know if the surface is compressed

  • radv/sdma: add support for compression on GFX12

  • radv: set radv_disable_dcc=true for WWE 2k23

  • radv: fix re-emitting VRS state when rendering begins

  • radv: do not clear unwritten color attachments with dual-source blending

  • radv: disable SINGLE clear codes to workaround a hw bug with DCC on GFX11

  • radv: fix GPU hangs with image copies for ASTC/ETC2 formats on transfer queue

Saroj Kumar (2):

  • mesa: Add GL_EXT_protected_textures support

  • radeonsi: Move buffer descriptor slot to the beginning

Sasha Finkelstein (1):

  • vtn_bindgen2: Fix memory corruption

Serdar Kocdemir (6):

  • gfxstream: Track more fence functions on host

  • gfxstream: Add VK_KHR_multiview support

  • gfxstream: track pipeline layouts on decoder

  • gfxstream: wrap semaphore functions on the host

  • gfxstream: Add dispatcher validity checks

  • gfxstream: Wrap vkEnumerateInstanceExtensionProperties for host

Sergi Blanch Torne (8):

  • ci: disable Collabora’s farm due to maintenance

  • Revert “ci: disable Collabora’s farm due to maintenance”

  • ci: typo in debian-android in .build-for-tests-jobs

  • ci: disable Collabora’s farm due to maintenance

  • Revert “ci: disable Collabora’s farm due to maintenance”

  • ci: disable Collabora’s farm due to maintenance

  • Revert “ci: disable Collabora’s farm due to maintenance”

  • Uprev Piglit to ebdf60e0d4b0dc23e79373cb923fdee023eb3b2b

Sergii Ushakov (1):

  • gfxstream: Emulate DMABUF with OPAQUE_FD

Seán de Búrca (8):

  • rusticl/mem: don’t write more supported image formats than requested

  • rusticl/mem: don’t create svm_pointers slice from null raw pointer

  • rusticl/mem: use cl_slice::from_raw_parts in place of std::slice

  • rusticl: rename CheckedPtr::copy_checked to match primitive method

  • rusticl: mark CheckedPtr::write_checked as unsafe

  • rusticl: add debug assertions to avoid truncating casts

  • rusticl: correct calculation of maximum allocation size

  • rusticl: cap max alloc size to i32 to avoid overflowing gallium

Sil Vilerino (13):

  • d3d12: Fix HEVC range extension pic params validation

  • pipe: Add profiles for HEVC 422 8/10b and 444 10b

  • d3d12: Add support for Y210, Y410, YUY2 and HEVC 422 8/10b, HEVC 444 10b profiles

  • d3d12: Add NULL initialization for d3d12_video_enc::m_pVideoTexArrayDPBPool

  • d3d12: Add some missing members initialization for d3d12_video_buffer

  • d3d12: Increase DPB video texture array pool size for async queue depth

  • d3d12: Fix array of texture DPB cap detection

  • d3d12: Fix warning 4305 truncation from type1 to type2

  • nir.h: Fix warning C4800 forcing value to bool ‘true’ or ‘false’

  • u_thread.h: Fix warning C4800 forcing value to bool ‘true’ or ‘false’

  • d3d12: Fix warning C4800 forcing value to bool ‘true’ or ‘false’

  • d3d12: Enable warnings C4056, C4305, C4351, C4756, C4800, C4291, C4020, C4624, C4309, C5105, C4024, C4189

  • d3d12: Cache the texture array cap requirement in encoder creation for calls to d3d12_video_create_dpb_buffer

Simon Ser (9):

  • gbm: fix get_back_bo() failure with gbm_surface and implicit modifiers

  • pvr: replace dup() with os_dupfd_cloexec()

  • freedreno: replace dup() with os_dupfd_cloexec()

  • iris: replace dup() with os_dupfd_cloexec()

  • lavapipe: replace dup() with os_dupfd_cloexec()

  • venus: replace dup() with os_dupfd_cloexec()

  • panvk: replace dup() with os_dupfd_cloexec()

  • libsync: replace dup() with os_dupfd_cloexec()

  • vulkan/wsi/x11: replace dup() with os_dupfd_cloexec()

Stéphane Cerveau (2):

  • anv: fix error code in GetPhysicalDeviceVideoFormatProperties

  • radv: video: rework maxActiveReferenceSlot/MaxDpbSlots

Sushma Venkatesh Reddy (1):

  • intel/tools: Improve memory allocation failure handling in aubinator_error_decode_xe

Sviatoslav Peleshko (3):

  • anv: Add full subgroups workaround for the shaders that use shared memory

  • drirc: Apply assume_full_subgroups_with_shared_memory to Resident Evil 2

  • vulkan/wsi/headless: Remove unnecessary wsi_configure_image()

Tapani Pälli (25):

  • intel/genxml/anv: fix the layout of call stack handler struct

  • intel/dev: reduce warning noise from urb settings

  • intel/common: fix mi_builder_test issue

  • anv: handle non-wsi images in anv_layout_to_aux_state

  • anv: tighten condition for changing barrier layouts

  • anv: apply cache flushes on pipeline select with gfx20

  • iris: wait for imported fences to be available in iris_fence_await

  • intel/compiler: add a spec note about L1WT types being uncached

  • iris: remove dead code that cannot get hit anymore

  • intel/dev: update mesa_defs.json from internal database

  • anv: restrict TessellationDistributionLevel for Wa_16025857284

  • iris: restrict TessellationDistributionLevel for Wa_16025857284

  • intel/dev: reduce warning noise from urb settings II

  • isl: add usage field to isl_buffer_fill_state_info

  • iris: pass down buffer usage for isl_buffer_fill_state

  • anv: pass down buffer usage for isl_buffer_fill_state

  • isl/iris/anv: setup L1CacheControl based on surface and buffer usage

  • compiler/glsl: check that bias is not used outside fragment stage

  • intel/dev: update mesa_defs.json from internal database

  • mesa: clamp texbuf query size to MAX_TEXTURE_BUFFER_SIZE

  • mesa: various fixes for ClearTexImage/ClearTexSubImage

  • iris: force reallocate on eglCreateImage with GFX >= 20

  • iris: make sure to not mix compressed vs non-compressed

  • anv: put parenthesis to the set_sampler_size equation

  • intel/dev: update mesa_defs.json from internal database

Taras Pisetskyi (1):

  • anv,driconf: Add sampler coordinate precision workaround for EVE Online

Thomas H.P. Andersen (1):

  • nvk: use a valid allocation scope

Tim Keller (1):

  • dril: Check for null config in dril_target.c

Timothy Arceri (14):

  • util/disk_cache: dont create multidisk cache dir if unused

  • util/u_idalloc: fix util_idalloc_sparse_alloc_range()

  • ci: move llvmpipe fails to flakes

  • mesa: fix reuse of deleted buffer object

  • mesa: fix reuse of deleted texture object

  • mesa: fix potential race condition in with TexObjects

  • mesa: fix reuse of deleted sampler object

  • mesa: fix potential race conditions in with FrameBuffers

  • mesa: fix potential race condition in with RenderBuffers

  • mesa: fix potential race condition in with ATIShaders

  • mesa: fix potential race condition in with Programs

  • nir: fix uniform cloning helper

  • glsl: fix regression in ubo cloning

  • util/driconf: add force_gl_depth_component_type_int workaround

Timur Kristóf (94):

  • radv: Move buffer related NIR meta shaders to radv_meta_nir.c

  • radv: Move blit NIR shaders to radv_meta_nir.c

  • radv: Move blit2d NIR shaders to radv_meta_nir.c

  • radv: Move buffer-image copy and clear NIR shaders to radv_meta_nir.c

  • radv: Move clear NIR shaders to radv_meta_nir.c

  • radv: Move VRS HTILE copy NIR shader to radv_meta_nir.c

  • radv: Move DCC retile NIR shader to radv_meta_nir.c

  • radv: Move expand depth stencil NIR shader to radv_meta_nir.c

  • radv: Move DCC decompress NIR shader to radv_meta_nir.c

  • radv: Move FMASK copy NIR shader to radv_meta_nir.c

  • radv: Move FMASK expand NIR shader to radv_nir_meta.c

  • radv: Move resolve NIR compute shaders to radv_meta_nir.c

  • radv: Move resolve NIR fragment shaders to radv_meta_nir.c

  • radv: Move resolve NIR fs to radv_meta_nir.c

  • radv: Move NIR helpers from radv_meta.c to radv_meta_nir.c

  • radv: Move NIR specific function declarations to radv_meta_nir.h

  • radv: Rename get_global_ids to radv_meta_nir_get_global_ids.

  • radv: Add missing copyright info to radv_meta_buffer.c

  • util/enum_operators: Don’t define anything for OpenCL

  • nir: Add struct names where they were missing.

  • nir: Add missing extern “C” to nir_defines.h

  • nir: Add forward declarations of relevant structs to nir_defines.h

  • nir: Move some enums and structs to nir_defines.h

  • nir: Move nir_tcs_info to separate file.

  • nir: Move nir_shader_compiler_options to separate file.

  • nir: Don’t include full nir.h in nir_xfb_info.h

  • nir: Don’t include the full nir.h when not necessary.

  • spirv: Don’t include full nir.h in nir_spirv.h

  • vk: Don’t include full nir.h in headers.

  • glsl: Don’t include full nir.h where not necessary.

  • ac/nir: Don’t include nir.h in headers anymore.

  • ac/nir: Move surface related NIR functions to separate file.

  • aco: Don’t include nir.h in aco_interface.h anymore.

  • ac: Don’t include full nir.h anymore.

  • radv: Add missing includes and remove unnecessary NIR includes.

  • hk: Don’t include full nir.h in hk_shader.h

  • zink: Don’t include full nir.h where not necessary.

  • ttn: Don’t include full nir.h where not necessary.

  • nak: Don’t include full nir.h in nak.h

  • nvk: Don’t include full nir.h in nvk_shader.h

  • nir: Don’t include xxhash.h in nir.h, only where it is used.

  • nir: Don’t include bitscan.h in nir.h, it’s not actually used.

  • nir: Don’t include u_printf.h in nir.h, only where necessary.

  • nir: Don’t include u_format.h in nir.h, it’s not actually used.

  • nir: Remove struct keyword from nir.h where possible.

  • nir: Add comment to indicate that NIR_PASS_V is deprecated.

  • nir: Add bool return value to nir_fixup_deref_modes.

  • nir: Add bool return value to nir_fixup_deref_types.

  • nir: Don’t use deprecated NIR_PASS_V macro anymore.

  • radv: Stop using deprecated NIR_PASS_V with core NIR passes.

  • radv: Add bool return value to radv_nir_lower_abi.

  • radv: Add bool return value to radv_nir_apply_pipeline_layout.

  • radv: Add bool return value to ray tracing NIR lowerings.

  • ac/nir: Add bool return value to ac_nir_lower_ls_outputs_to_mem.

  • ac/nir: Add bool return value to ac_nir_lower_hs_inputs_to_mem.

  • ac/nir: Add bool return value to ac_nir_lower_hs_outputs_to_mem.

  • ac/nir: Add bool return value to ac_nir_lower_tes_inputs_to_mem.

  • ac/nir: Add bool return value to ac_nir_lower_es_outputs_to_mem.

  • ac/nir: Add bool return value to ac_nir_lower_gs_inputs_to_mem.

  • ac/nir: Add bool return value to ac_nir_lower_task_outputs_to_mem.

  • ac/nir: Add bool return value to ac_nir_lower_mesh_inputs_to_mem.

  • ac/nir: Add bool return value to ac_nir_lower_legacy_vs.

  • ac/nir: Add bool return value to ac_nir_lower_legacy_gs.

  • ac/nir/ngg: Add bool return value to ac_nir_lower_ngg_nogs.

  • ac/nir/ngg: Add bool return value to ac_nir_lower_ngg_gs.

  • ac/nir/ngg: Add bool return value to ac_nir_lower_ngg_mesh.

  • radv: Don’t use deprecated NIR_PASS_V macro for AMD common NIR passes.

  • compiler/clc: Stop using deprecated NIR_PASS_V macro.

  • vulkan/runtime: Don’t use deprecated NIR_PASS_V macro.

  • hk: Don’t use deprecated NIR_PASS_V macro anymore.

  • glsl: Don’t use deprecated NIR_PASS_V macro anymore.

  • nak: Don’t use deprecated NIR_PASS_V macro anymore.

  • nvk: Don’t use deprecated NIR_PASS_V macro anymore.

  • radv: Use flush postamble on GFX7 with different flags.

  • ac/nir/ngg: Run copy propagation.

  • ac/nir/ngg: Improve reuse of position value.

  • ac/nir/ngg: Remove inputs_needed_by_*

  • ac/nir/ngg: Prepare deferred shader part before adding culling code.

  • ac/nir/ngg: Gather info about what the deferred shader part uses.

  • ac/nir/ngg: Use deferred info for compacted arguments.

  • ac/nir/ngg: Remove cleanup_culling_shader_after_dce.

  • nir/xfb: Preserve some xfb information when gathering from intrinsics.

  • nir/opt_varyings: Fix assertion when deduplicating TCS outputs.

  • radv: Use buffers_written mask when gathering XFB info.

  • radv: Call nir_opt_undef too after nir_opt_varyings.

  • radv: Remove radv_streamout_info::num_outputs.

  • nir/print: Fix variable mode for arrayed output load intrinsics.

  • radv: Add radv_foreach_stage to ForEachMacros again.

  • radv: Inline radv_graphics_shaders_link_varyings_{first/second}.

  • radv: Refactor loops in radv_graphics_shaders_link_varyings.

  • radv: Move preparation and fixup to separate loops in varying optimization.

  • radv: Don’t call nir_opt_varyings a second time when unnecessary.

  • radv: Clear dirty flag for MSAA state after emitting it.

  • radv: Clear dirty flag for clip rects state after emitting it.

Tomeu Vizoso (4):

  • egl/surfaceless: Only choose drivers that expose the graphics capability

  • kopper: Explicitly choose zink

  • etnaviv/ml: Use etna_buffer_resource instead of etna_resource

  • etnaviv: Release screen->dummy_desc_reloc.bo

Trigger Huang (2):

  • radeonsi: Fix perfcounter start event in si_pc_emit_start

  • radeonsi: Change program seqnece for perf counters

Valentine Burley (84):

  • amd/ci: Revert to 6.6 kernel on Raven

  • zink/ci: Add a fraction for zink-venus-lvp

  • khronos-update: Update ANDROID guards in vk_android_native_buffer.h

  • zink/ci: Make zink-venus-lvp-full a nightly job

  • ci/intel-gpu-freq: Add Xe support

  • ci: Use new kernel with Intel Xe driver

  • ci/lava: Allow passing extra cmdline arguments

  • intel/ci: Drop redundant BOOT_METHOD variables

  • intel/ci: Add newer i915/ADL-P firmware to rootfs

  • intel/ci: Load Xe instead of i915 on ADL

  • intel/dev: Provide a toggle to avoid warnings about unsupported devices

  • intel/ci: Use INTEL_XE_IGNORE_EXPERIMENTAL_WARNING to reduce warnings

  • intel/ci: Update expectations for Xe

  • amd/ci: Fix fraction for radv-stoney-angle-full

  • ci/android: Don’t build desktop GL CTS

  • ci/lava: Don’t build VK-main for arm64

  • ci: Allow building ANGLE for multiple platforms

  • ci/android: Build and use ANGLE

  • ci/android: Check ANGLE version

  • ci: Move debian-android up to the build-for-tests stage

  • freedreno/ci: Unify naming for a306 jobs

  • zink/ci: Rename a618 suite and expectation files

  • ci/angle: Rework building ANGLE (again)

  • ci/lava: Build ANGLE for arm64

  • ci/android: Add build section for Cuttlefish

  • ci/angle: Use lld-19 for linking ANGLE

  • intel/ci: Fix manual rules for ANGLE jobs

  • ci/angle: Uprev ANGLE

  • ci: Update expectations from latest nightly

  • turnip/ci: Add nightly ANGLE jobs on a618 and a660

  • ci/container: Include ANGLE in the arm64 test-gl container

  • turnip/ci: Rename valve-freedreno-turnip-manual-rules

  • turnip/ci: Add a nightly ANGLE job on a750

  • radv/ci: Don’t start X11 for ANGLE

  • anv/ci: Don’t start X11 for ANGLE

  • ci: Simplify downloading kernel for crosvm

  • ci: Don’t download the kernel image in lava_build.sh

  • intel/ci: Update GuC firmware for ADL-S and ADL-N

  • ci: Use new kernel that supports more Mediatek devices

  • anv/ci: Update expectations from latest nightly

  • intel/ci: Honor device-specific FDO_CI_CONCURRENT variables

  • intel/ci: Add brask and nissa

  • anv/ci: Migrate anv-adl-angle job to brask

  • intel/ci: Migrate intel-adl-cl and intel-adl-skqp to nissa

  • anv/ci: Increase parallelism of zink-anv-adl

  • lavapipe: Update driverVersion

  • anv/ci: Remove fixed test from xfails

  • zink/ci: Run more traces on Tiger Lake

  • zink/ci: Add trace testing on Alder Lake

  • anv/ci: Append -vk suffix to VKCTS job names

  • ci: Update kernel to include i.MX8MP dtb

  • panforst/ci: Migrate the G57 GL job to MT8195

  • panfrost/ci: Add a Piglit job on G57

  • panfrost/ci: Shorten Piglit job names

  • panfrost/ci: Pin g610-gl job to 4GB DUTs

  • tu: Switch to device address from vk_buffer

  • ci: Enable Perfetto in debian-no-libdrm for Turnip build testing

  • zink/ci: Work around recent OOM issues in zink-anv-adl

  • ci/deqp: Delete more uncompressed caselist files

  • ci: Add missing kvm runner tags

  • ci/android: Don’t delete ninja after building LLVM

  • ci/android: Keep the LLVM install when rebuilding

  • ci/android: Update to LLVM 19 for Android

  • ci/android: Temporarily disable building llvmpipe

  • radv/ci: Delete obsolete vkcts-stoney-valve job

  • amd/ci: Rename AMD jobs to follow unified naming convention

  • ci: Disable the Google freedreno farm

  • panvk/ci: Migrate the G52 VK job to MT8186

  • ci: Make it possible to use ANGLE traces on other architectures

  • radv/ci: Update ANGLE version used for traces

  • ci/android: Remove platform-tools from test-android container

  • ci/android: Add section for downloading Android CTS

  • ci/container: Disable debian/arm32_test-vk container

  • ci/container: Remove double build sections in test-* containers

  • ci/container: Move calling strip-rootfs.sh to common scripts

  • ci/android: Add build section for uninstalling build software

  • ci/container: Drop unnecessary variables for image paths

  • intel/ci: Convert iris-kbl-piglit to deqp-runner suite

  • ci/piglit: Drop redundant PIGLIT_PROFILES variable

  • ci/piglit: Consolidate HWCI_TEST_SCRIPT for piglit traces

  • ci/piglit: Remove piglit-runner.sh script

  • ci/piglit: Consolidate identical skip lists for X11 and gbm

  • ci/lava: Consolidate piglit trace job definitions

  • zink/ci: Work around recent OOM issues in zink-anv-tgl

Vasily Khoruzhick (20):

  • lima: ppir: handle ffma in the backend

  • lima: ppir: improve readability of ppir represantation dump

  • lima: ppir: fixup src node when replacing src for select and load_reg

  • lima: ppir: print index of the node that breaks node_to_instr

  • lima: ppir: fix diassembling atan and combiner codegen definition

  • lima: ppir: assert on unexpected pipeline dest for fmul and vmul

  • lima: ppir: add codegen for mov and mul on combiner unit

  • lima: ppir: use combiner unit for mul

  • lima: ppir: duplicate fneg and fabs if its source is an intrinsic

  • lima: ppir: fix regalloc bugs

  • lima: ppir: assign an index for discard block

  • lima: ppir: introduce an optimizer

  • lima: ppir: try scheduling root nodes into the same instruction

  • lima: ci: update deqp CI expectations

  • lima: ppir: add compactification pass

  • lima: ppir: reuse uniform load in instruction if possible

  • lima: ppir: reuse load_temp/store_temp nodes if possible

  • lima: ppir: assign actual index to discard block

  • lima: ppir: optimize branches

  • lima: ppir: try inserting nodes into successor instr for uncond branch

Vignesh Raman (2):

  • s3_upload: improve url validation and error message

  • ci: Uprev kernel to 6.14

Visan, Tiberiu (2):

  • amd/vpelib: Fix studio range

  • amd/vpelib: Apply normalization for full range

Vlad Zahorodnii (3):

  • egl/wayland: Damage whole surface using wl_surface_damage_buffer()

  • vulkan/wsi/wayland: Damage whole surface using wl_surface_damage_buffer()

  • vulkan/wsi/wayland: Document why wl_surface_damage() code path ignores provided damage

Xaver Hugl (6):

  • increase required wayland-protocols version to 1.41

  • vulkan/wsi: implement support for VK_EXT_hdr_metadata on Wayland

  • vulkan/wsi: handle the compositor not supporting extended target volume better

  • vulkan/wsi: don’t use sRGB if the compositor doesn’t support it

  • vulkan/wsi: validate HDR metadata to not cause protocol errors

  • vulkan/wsi: warn once when HDR metadata is skipped because of protocol errors

Yinjie Yao (2):

  • gallium/pipe: Increase hevc max slice to 600

  • frontends/va: Handle properly when decoding more slices than limit

Yiwei Zhang (124):

  • venus: scrub disallowed ycbcr features for rgba10x6

  • venus: further sanitize image props for rgba10x6

  • venus: refactor more to image format props sanitization

  • venus: enable VK_EXT_external_memory_acquire_unmodified if needed

  • venus: use dedicated allocation for ANB image memory import

  • venus: emulate a second graphics queue on Android

  • venus: sync venus protocol for below extensions

  • venus: group private data together with other 1.3 exts

  • venus: fix to handle pipeline flags2 from maint5

  • venus: support VK_EXT_blend_operation_advanced

  • venus: refactor to share more codes between pipeline state fillings

  • venus: support VK_EXT_sample_locations

  • venus: fix sampler locations feats and props scrub

  • venus: sync protocol to v1.3.302

  • Support 5 more promoted extensions

  • venus: fix maintenance5 props init and create flags2

  • venus: drop unused codes

  • venus: sync protocol for v1.4.307 release and update promoted entries

  • venus: updated to use core types for promoted extensions

  • venus: support VK_KHR_shader_subgroup_rotate

  • venus: support VK_KHR_shader_float_controls2

  • venus: support VK_EXT_pipeline_protected_access

  • venus: support VK_EXT_pipeline_robustness

  • venus: support VK_KHR_map_memory2

  • venus: support VK_KHR_global_priority

  • venus: support VK_KHR_dynamic_rendering_local_read

  • venus: a few tiny naming fixes

  • venus: support VK_KHR_maintenance6

  • venus: update second queue emulation for 1.4 requirement

  • venus: deprecate a few useless micros

  • venus: advertise 1.4 support

  • venus: fix 2 entry points from maint6

  • venus: fix sample location info validity

  • venus: limit second queue emulation to android framework

  • venus: temporarily disable 1.4 support

  • venus: suppress a few -Wmaybe-uninitialized

  • venus: drop unnecessary struct

  • venus: use sharing_mode from common vk_image

  • venus: fix image format cache miss with AHB usage query

  • venus: sync latest protocol v3 support for host copy

  • venus: extend image format cache for host copy props

  • venus: implement host image copy commands

  • vulkan/util: clean up copy_property and fix for setter

  • venus: enable VK_EXT_host_image_copy support

  • venus: re-enable 1.4 support

  • venus: align on wsi frontends support

  • venus: support wsi maintenance1 extensions

  • venus: sync protocol for the passthrough extensions

  • venus: added passthrough extension support - Part I

  • venus: added passthrough extension support - Part II

  • venus: added passthrough extension support - Part III

  • venus: added passthrough extension support - Part IV

  • venus: added passthrough extension support - Part V

  • venus: relax the requirement for sync2

  • venus: fix an obsolete protocol sync earlier

  • venus: fix a memory corruption in query records recycle

  • venus: sync protocol for ray tracing support

  • venus: add a debug option for ray tracing support

  • venus: support VK_KHR_deferred_host_operations

  • venus: add stubs for accel struct host commands

  • venus: implement VK_KHR_acceleration_structure - Part I

  • venus: implement VK_KHR_acceleration_structure - Part II

  • venus: implement VK_KHR_acceleration_structure - Part III

  • venus: enable VK_KHR_acceleration_structure

  • venus: support VK_KHR_ray_query

  • venus: prepare push template for ray tracing pipeline

  • venus: implement VK_KHR_ray_tracing_pipeline commands

  • venus: enable VK_KHR_ray_tracing_pipeline

  • venus: support VK_KHR_ray_tracing_position_fetch

  • venus: support VK_KHR_ray_tracing_maintenance1

  • lavapipe: set availability bit for accel struct host queries

  • lavapipe: fix accel struct device query copy

  • venus: use common memory report implementation

  • venus: fix to ignore dstSet for push descriptor

  • venus: extend async descriptor set alloc coverage

  • venus: relax 2 assertions for prime blit path

  • venus: sync protocol support for maint7

  • venus: sync protocol for accel struct indirect build encoding fix

  • pan/kmod: set DRM_RDWR for exported dma-bufs

  • venus: drop vn_call usage on apis without any returns

  • venus: sync latest protocol for more extension support

  • venus: add a few more trivial extensions

  • venus: support VK_EXT_filter_cubic

  • venus: support VK_EXT_legacy_dithering

  • venus: support VK_EXT_depth_bias_control

  • venus: support VK_EXT_depth_clamp_control

  • venus: support VK_EXT_attachment_feedback_loop_dynamic_state

  • venus: support VK_EXT_nested_command_buffer

  • venus: default to passthrough ray tracing support

  • venus: add a new debug option to revive memory budget support

  • venus: properly enable display platform extensions

  • venus: back out display control

  • venus: support VK_EXT_debug_report

  • venus: sync latest protocol for 2 more extensions

  • venus: support VK_EXT_pipeline_library_group_handles

  • venus: support VK_EXT_image_sliced_view_of_3d

  • panvk: disable VK_KHR_shader_quad_control

  • panvk: fix dependency for EXT_display_control

  • vulkan: update ALLOWED_ANDROID_VERSION for api level 34

  • vulkan: update ALLOWED_ANDROID_VERSION for api level 35

  • venus: rename common vk object base member to vk

  • venus: explicitly get vn_device from vk_device

  • venus: use common vk_command_pool

  • venus: use vk_command_pool internals

  • venus: use common vk_command_buffer

  • venus: use vk_command_buffer internals

  • venus: use common cmd pool_link

  • docs: demote VK_KHR_shader_relaxed_extended_instruction

  • venus: fix unexpected ring alive status expire upon owner thread switch

  • venus: fix ahb usage caching

  • venus: fix maint4 multi-planar memory requirements

  • venus: improve image memory requirement cache for image aliasing

  • venus: enable VK_EXT_debug_utils

  • virgl/venus/vtest: align capset and protocol with virglrenderer

  • panvk/csf: rework cache flush reduction

  • panvk: fix memory requirement query for aliased disjoint image

  • venus: support VK_KHR_maintenance7

  • venus: sync latest protocol

  • venus: support VK_EXT_buffer_device_address

  • venus: support VK_KHR_shader_relaxed_extended_instruction

  • venus: refactor format properties cache to be extensible

  • venus: support VK_EXT_multisampled_render_to_single_sampled

  • docs: update venus driver page

  • venus: fix missing renderer destructions

Yogesh Mohan Marimuthu (7):

  • winsys/amdgpu: do not use rcs->csc

  • winsys/amdgpu: make csc context as array

  • winsys/amdgpu: amdgpu_cs_context is csc, amdgpu_cs is acs

  • winsys/radeon: struct radeon_cmdbuf is rcs instead of cs for consistency

  • winsys/amdgpu: same_queue variable should be set if there is only one queue

  • winsys/amdgpu: userqueue multi ctx jobs are guaranteed to be in sequence

  • winsys/amdgpu: userq non imported fence can be ignored for same ip_type

Yurii Kolesnykov (1):

  • Get rid of 5 remaining references to glapitable.h

Zan Dobersek (11):

  • tu/a750: invalidate vertex state before CP_DRAW_INDIRECT_MULTI

  • tu: make tu_debug_flags enum 64-bit

  • tu: use query index when retrieving performance query iovas

  • tu: performance query result writes must use dedicated union type

  • tu/a7xx: disable preemption during performance query measurement

  • freedreno: add common implementation of perfcntr-based derived counters

  • tu: support exposing derived counters through VK_KHR_performance_query

  • tu: fix zero-index perfcntr collection for derived counters

  • tu: disable logic operations for float and sRGB formats

  • freedreno/registers: add useful A6XX_SP_TP_MODE_CNTL bitfields

  • tu: allow D3D-compatible texture coordinate rounding

Zhao, Jiali (1):

  • amd/vpelib: Fix studio output CSC

forbiddenlake (2):

  • docs: Fix HTML build with Sphinx 8.2

  • ci/alpine: upgrade sphinx and hawkmoth to the latest versions

irql-notlessorequal (7):

  • hasvk: Fix non-functioning version override.

  • elk: ensure VUE header writes in HS/DS/GS stages

  • elk: always write the VUE header

  • hasvk: Pre-plumbing needed for VK_KHR_maintenance5

  • hasvk: Implement VkPipelineCreateFlags2KHR support

  • hasvk: Implement VK calls and formats.

  • hasvk: Advertise VK_KHR_maintenance5

lcagustini (1):

  • panvk: Advertise support for VK_EXT_border_color_swizzle

liuqiang (1):

  • intel/brw: Remove redundant condition in components_read()

llyyr (1):

  • vulkan/wsi/wayland: initialize surface colorspace with PASS_THROUGH_EXT

sarbes (1):

  • lima: add genxml for texture descriptor