Mesa 25.0.7 Release Notes / 2025-05-28

Mesa 25.0.7 is a bug fix release which fixes bugs found since the 25.0.6 release.

Mesa 25.0.7 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 25.0.7 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: 592272df3cf01e85e7db300c449df5061092574d099da275d19e97ef0510f8a6  mesa-25.0.7.tar.xz
SHA512: 825bbd8bc5507de147488519786c0200afacf97dae621c80ead24b2c5dd55c5a442757ac8452698ae611e9344025465080795cf8f2dc4eb7ce07b5cc521b2b5c  mesa-25.0.7.tar.xz

New features

  • None

Bug fixes

  • RX9070 hard crash with Mafia Definitive Edition

  • RADV: Potential bug with vulkan fragment shader interpolation (on outputs from mesh shaders?)

  • In the game “Foundation” a buildings areas of effect is missing

  • ANV: Dota 2 May 22 2025 update crashing in vkCmdBindDescriptorSets with no validation error

  • [RADV][GFX9] Recent Mesa-git broken on AMD Vega 64 with ring sdma0 timeouts when launching DXVK games

  • Vulkan Video engages during playback of format which is not supported by my Fiji GPU

  • ACO: IR Validation error “SDWA operand selection size” triggered by compute shader on VEGA20

  • RADV: Gibberish output with llama.cpp (Vulkan compute) on Radeon VII (Vega20) with Mesa 25.1.0, works on 25.0.5

  • Blending broken in game SkinDeep

  • Radeon R5 (Mullins) H264 VA-API encoding acceleration doesn’t work

  • nvk: lib_rs_gen.py requires `rustfmt`

  • radv: vkd3d-proton test failure with predication + EXT_dgc

  • mesa-25.0.4 aborts Xserver due to ACO “Unsupported opcode” v_lshlrev_b16

Changes

Adam Jackson (2):

  • vtn: (Silently) handle FunctionParameterAttributeNo{Capture,Write}

  • vtn/opencl: Handle OpenCLstd_F{Min,Max}_common

Calder Young (2):

  • iris: Fix accidental writes to global dirty bit instead of local

  • iris: set dependency between SF_CL and CC states

Christian Gmeiner (1):

  • zink: Fix NIR validation error in cubemap-to-array lowering

Dave Airlie (1):

  • nvk: Fix compute class comparison in dispatch indirect

David Rosca (4):

  • radeonsi/vce: Fix bitstream buffer size

  • radeonsi/vce: Only send one task per IB

  • radeonsi/vce: Fix output quality and performance in speed preset

  • radv/video: Limit 10bit H265 decode support to stoney and newer

Ella Stanforth (1):

  • v3d/compiler: Fix ub when using memcmp for texture comparisons.

Eric Engestrom (3):

  • docs: add sha sum for 25.0.6

  • .pick_status.json: Mark 29d7b90cfcb67ecc2ff3e422dd7b38898abb1bbe as denominated

  • .pick_status.json: Update to 8965e60118fa17407c5bfcdca1fe2854ad2fb150

Erik Faye-Lund (1):

  • mesa/main: remove non-existing function prototype

Faith Ekstrand (2):

  • nvk: Allocate the correct VAB size on Kepler

  • nouveau/mme: Don’t install the HW tests

Georg Lehmann (2):

  • radeonsi: always lower alu bit sizes

  • aco: assume sram ecc is enabled on Vega20

Gurchetan Singh (1):

  • gfxstream: get rid of logspam in virtualized case

Hans-Kristian Arntzen (1):

  • radv: Consider that DGC might need shader reads of predicated data.

José Roberto de Souza (2):

  • anv: Implement missing part of Wa_1604061319

  • anv: Enable preemption due 3DPRIMITIVE in GFX 12

Karol Herbst (2):

  • nir: fix use-after-free on function parameter names

  • vtn: fix use-after-free on function parameter names

Lars-Ivar Hesselberg Simonsen (2):

  • panvk/v9+: Set up limited texture descs for storage use

  • panvk/v9+: Set up limited texture descs for storage use

LingMan (1):

  • entaviv/isa: Silence warnings about non snake case names

Lionel Landwerlin (4):

  • anv: enable preemption setting on command/batch correctly

  • anv/brw: stop turning load_push_constants into load_uniform

  • hasvk/elk: stop turning load_push_constants into load_uniform

  • anv: don’t use pipeline layout at descriptor bind

Marek Olšák (2):

  • winsys/amdgpu: fix running out of 32bit address space with high FPS

  • glsl: fix sampler and image type checking in lower_precision

Matt Turner (1):

  • gallivm: Use `llvm.roundeven` in lp_build_round()

Mel Henning (2):

  • nouveau/headers: Run rustfmt after file is closed

  • nouveau/headers: Ignore PermissionError in rustfmt

Mike Blumenkrantz (2):

  • llvmpipe: disable conditional rendering mem for blits

  • lavapipe: handle counterOffset in vkCmdDrawIndirectByteCountEXT

Natalie Vock (1):

  • driconf: Fix DOOM: The Dark Ages workaround name in 25.0.x

Olivia Lee (1):

  • util/u_printf: fix memory leak in u_printf_singleton_add_serialized

Patrick Lerda (1):

  • r600: fix pop-free clipping

Paulo Zanoni (1):

  • anv/trtt: don’t avoid the TR-TT submission when there is stuff to signal

Qiang Yu (1):

  • nir/opt_varyings: fix mesh shader miss promote varying to flat

Rhys Perry (1):

  • aco/gfx115: consider point sample acceleration

Rob Clark (1):

  • ci: Disable fd-farm

Samuel Pitoiset (5):

  • radv: fix fetching conditional rendering state for DGC preprocess

  • radv: fix conditional rendering with DGC and non native 32-bit predicate

  • radv: fix missing texel scale for unaligned linear SDMA copies

  • radv: fix capture/replay with sparse images and descriptor buffer

  • radv: add radv_disable_hiz_his_gfx12 and enable for Mafia Definitive Edition

Timothy Arceri (7):

  • st/mesa: fix _IntegerBuffers bitfield use

  • mesa/st: fix _BlendForceAlphaToOneDraw bitfield use

  • mesa/st: fix _IsRGBDraw bitfield use

  • mesa: fix _FP32Buffers bitfield use

  • mesa: update validation when draw buffer changes

  • mesa: extend linear_as_nearest work around

  • util: add workaround for the game Foundation