Mesa 25.0.1 Release Notes / 2025-03-05

Mesa 25.0.1 is a bug fix release which fixes bugs found since the 25.0.0 release.

Mesa 25.0.1 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 25.0.1 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: 49eb55ba5acccae91deb566573a6a73144a0f39014be1982d78c21c5b6b0bb3f  mesa-25.0.1.tar.xz
SHA512: 1ecb1b90c5f78de4c61f177888543778285731faccc6f78d266d4b437f7b422a78b705a6e9fc6c9eab62c08f2573db5dd725eaa9cc9e5bedcaa7d8cfe6b47a1f  mesa-25.0.1.tar.xz

New features

  • None

Bug fixes

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

  • GLmatrix needs aligned malloc

  • Lavapipe crashes if no Position is output in mesh shader

  • [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”

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

Changes

Benjamin Lee (4):

  • panfrost: remove NIR_PASS_V usage for noperspective lowering

  • panfrost: fix large int32->float16 conversions

  • panfrost: fix condition in bi_nir_is_replicated

  • panfrost/va: remove swizzle mod from LDEXP

Caio Oliveira (1):

  • brw: Fix size in assembler when compacting

Daniel Schürmann (5):

  • aco/scheduler: always respect min_waves on GFX10+

  • 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/ssa_elimination: insert parallelcopies for p_phi immediately before branch

  • aco/assembler: Fix short jumps over chained branches

Dave Airlie (1):

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

David Rosca (2):

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

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

Dylan Baker (2):

  • iris: Correctly set NOS for geometry shader state changes

  • iris: fix handling of GL_*_VERTEX_CONVENTION

Emmanuel Gil Peyrot (1):

  • panvk: Initialize out array with the correct length

Eric Engestrom (8):

  • docs: add sha sum for 25.0.0

  • .pick_status.json: Update to b331713f20148852370a4fae5c2830d46801eb3b

  • .pick_status.json: Update to 55c476efed01121b3a64a58c304aae8ef9a79475

  • .pick_status.json: Mark b85c94fc891fe9d73b3a032aea8a6a71b8e6173b as denominated

  • .pick_status.json: Update to 4348253db5232b7be4db0a0ff47b31d51bc8f534

  • .pick_status.json: Update to fbc55afbdfc93a82c69f1cd6a1f4abbed96cfd19

  • .pick_status.json: Mark 5461ed5808421a8ffb79bdaa1449265f3e8f40a5 as denominated

  • .pick_status.json: Update to 45e771f4fbe4245b252c6360e55776080f0bf458

Erik Faye-Lund (1):

  • mesa/main: wire up glapi bits for EXT_multi_draw_indirect

Faith Ekstrand (12):

  • nak: Only use suld.constant on Ampere+

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

  • zink: Use persistent semaphores for PIPE_FD_TYPE_SYNCOBJ

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

  • nvk: Do not set INVALIDATE_SKED_CACHES pre-MaxwellB

  • 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+

  • zink: Don’t present to Wayland surfaces asynchronously

  • egl/dri2: Rework get_wl_surface_proxy()

  • egl/wayland: Pass the original wl_surface to kopper

Georg Lehmann (1):

  • aco/insert_exec: fix continue_or_break on gfx6-7

Gert Wollny (1):

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

Guilherme Gallo (3):

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

  • ci/lava: Propagate errors in SSH tests

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

Hans-Kristian Arntzen (1):

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

Hyunjun Ko (1):

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

Iago Toral Quiroga (1):

  • pan/va: fix FAU validation

James Hogan (5):

  • 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

Job Noorman (1):

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

Juan A. Suarez Romero (2):

  • v3dv: duplicate key for texel_buffer cache

  • broadcom/simulator: use string copy instead of memcpy

Karol Herbst (3):

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

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

  • intel/brw, lp: enable lower_pack_64_4x16

Kevin Chuang (2):

  • anv/bvh: Fix encoder handling sparse buffer

  • anv/bvh: Fix copy shader handling sparse buffer

Konstantin Seurer (1):

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

Lars-Ivar Hesselberg Simonsen (2):

  • panfrost: Use RUN_COMPUTE over RUN_COMPUTE_INDIRECT

  • panvk: Use RUN_COMPUTE over RUN_COMPUTE_INDIRECT

Lionel Landwerlin (2):

  • anv: fix missing 3DSTATE_PS:Kernel0MaximumPolysperThread programming

  • vulkan/runtime: ensure robustness state is fully initialized

Lorenzo Rossi (1):

  • nvk: Fix MSAA sparse residency lowering crash

Marek Olšák (1):

  • mesa: allocate GLmatrix aligned to 16 bytes

Mary Guillemard (1):

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

Mike Blumenkrantz (6):

  • zink: wait on tc fence before checking for fd semaphore

  • zink: always fully unwrap contexts

  • zink: clamp UBO sizes instead of asserting

  • llvmpipe: pass layer count to rast clear

  • gallium: fix pipe_framebuffer_state::view_mask

  • mesa: avoid creating incomplete surfaces when multiview goes out of range

Natalie Vock (1):

  • radv/rt: Don’t allocate the traversal shader in a capture/replay range

Patrick Lerda (3):

  • r600: fix evergreen_emit_vertex_buffers() related cl regression

  • r600: fix the indirect draw 8-bits path

  • r600: fix emit_image_size() range base compatibility

Paulo Zanoni (1):

  • brw: extend the NOP+WHILE workaround

Peyton Lee (1):

  • radeonsi/vpe: check reduction ratio

Pierre-Eric Pelloux-Prayer (2):

  • tc: add missing TC_SENTINEL for TC_END_BATCH

  • mesa/st: call _mesa_glthread_finish before _mesa_make_current

Rhys Perry (1):

  • ac/nir: fix tess factor optimization when workgroup barriers are reduced

Roland Scheidegger (1):

  • llvmpipe: Fix alpha-to-coverage without dithering

Samuel Pitoiset (3):

  • radv/video: fix adding the query pool BO to the cmdbuf list

  • radv: fix missing SQTT barriers for fbfetch color/depth decompressions

  • radv: fix re-emitting fragment output state when resetting gfx pipeline state

Tapani Pälli (2):

  • iris: wait for imported fences to be available in iris_fence_await

  • iris: remove dead code that cannot get hit anymore

Yiwei Zhang (2):

  • venus: fix image format cache miss with AHB usage query

  • venus: relax the requirement for sync2

Yogesh Mohan Marimuthu (1):

  • winsys/amdgpu: same_queue variable should be set if there is only one queue