Mesa 24.2.3 Release Notes / 2024-09-18

Mesa 24.2.3 is a bug fix release which fixes bugs found since the 24.2.2 release.

Mesa 24.2.3 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 24.2.3 implements the Vulkan 1.3 API, but the version reported by the apiVersion property of the VkPhysicalDeviceProperties struct depends on the particular driver being used.

SHA checksums

SHA256: 4ea18b1155a4544a09f7361848974768f6f73c19d88f63de2ec650be313b2d0c  mesa-24.2.3.tar.xz
SHA512: 7a1ace23568d1907b778a2859f97c8988a414ba74e02e1fb5af6f95f768e1b1a2dfdaf412b0d655678ed915d28273953fd1236ebcd87553a1880f1a7f3ea4d44  mesa-24.2.3.tar.xz

New features

  • None

Bug fixes

  • ANV: scalar FS validation failure is always hit when playing the steam game “Black Myth: Wukong Benchmark Tool”

  • [REG 24.1.7->24.2.1, bisected] RADV: uVkCompute’s mad_throughput benchmark hangs

  • Color attribute not restored properly when using glPushAttrib/glPopAttrib calls

  • 69c772e4ea8f (egl: link with libgallium directly) breaks VAAPI - OpenGL interop on radeonsi

  • [v3d] performance regression in GTK’s NGL renderer from 24.1 to 24.2

  • Failed to cross-compile Mesa-24.2 & Rusticl with bindgen 0.65.1 and rustc 1.80.0

  • v3d gbm_bo_unmap() crashes when used by TigerVNC

Changes

Boris Brezillon (1):

  • pan/kmod: Don’t cap VM bind operations to one

Daniel Stone (1):

  • ci/alpine: Fix shellcheck errors

Dave Airlie (5):

  • vl/bitstream: use an int32_t for se encoding.

  • radv/video: handling encoding both sps and pps in same buffer

  • radv: Fix radeon_enc_code_ue with values over 2^16

  • radv/video: fix encode reference slot counting

  • radv/video/enc: report pps overrides in feedback for h265

David Heidelberg (3):

  • nir_lower_mem_access_bit_sizes: Assert when 0 components or bits are requested

  • freedreno/ir3: Use nir_lower_mem_access_bit_sizes instead custom lowering

  • ci/freedreno: move disabled a530 entries back to main gitlab-ci.yml

David Rosca (4):

  • meson/megadriver: Add megadriver_libdir argument

  • targets/va: Build va driver into libgallium when building with dri

  • targets/vdpau: Build vdpau driver into libgallium when building with dri

  • radeonsi: Disable EFC on VCN 2.2

Dylan Baker (3):

  • docs: update sha sums for 24.2.2

  • iris: Run checks that do not require resources before creating them

  • anv: if queue is NULL in vm_bind return early

Eric Engestrom (4):

  • .pick_status.json: Update to 8b272c8d8c419ecb7aee0257563c9489b675f4ef

  • .pick_status.json: Mark 5632a6e24f9053385e01a6464599ef4ba00e0c98 as denominated

  • .pick_status.json: Update to 45377dc5c46c4f449307c7efc28a1b66a57cf6aa

  • .pick_status.json: Update to ad3e6bb06a8e598be2381dfe2f5947f872b76bcd

Georg Lehmann (3):

  • nir/opt_sink: do not sink load_ubo_vec4 out of loops

  • nir/opt_sink: do not sink inverse_ballot out of loops

  • nir/instr_set: fix fp_fast_math

Iván Briano (1):

  • anv: be consistent about aux usage with modifiers

Jesse Natalie (1):

  • d3d12: Fix shader selector hash to hash array instead of pointer-to-array

Jordan Justen (1):

  • intel/dev: Fix warning for max_threads_per_psd when devinfo->verx10 == 120

Jose Maria Casanova Crespo (1):

  • v3d: v3d_resource Use LINEAR layout for importing with INVALID modifier

Juan A. Suarez Romero (2):

  • Revert “v3d: never replace a mapped bo”

  • v3d: do not rebind a sampler view already rebound

Karol Herbst (1):

  • rusticl: do not use CL vector types in bindings and code

Kenneth Graunke (1):

  • intel/brw: Use NUM_BRW_OPCODES in can_omit_write() check

Konstantin Seurer (4):

  • gallium,st/mesa: Add and set pipe_image_view::is_2d_view_of_3d

  • lavapipe: Implement VK_EXT_image_2d_view_of_3d with sparse textures

  • lavapipe: Do not adjust imageGranularity for different block sizes

  • radv: Work around broken terrain in Warhammer III

Lionel Landwerlin (3):

  • anv: selectively disable binding table usage on Gfx20

  • brw: use a builder of the size of the physical register for uniforms

  • brw: fix vecN rebuilds

Lucas Stach (1):

  • etnaviv: emit all PA shader attributes

Mary Guillemard (1):

  • panvk: Ensure to clear dirty dynamic state in panvk_cmd_draw

Mike Blumenkrantz (2):

  • zink: delete erroneous kopper assert

  • zink: fix sparse bo deallocation

Mohamed Ahmed (1):

  • nvk: Use stride in the explicit modifier case for linear images

Rhys Perry (3):

  • nir/opt_if: fix fighting between split_alu_of_phi and peel_initial_break

  • nir/opt_loop: skip peeling if the break is non-trivial

  • nir/opt_loop: skip peeling if the loop ends with any kind of jump

Rob Clark (1):

  • freedreno/drm: Fix ring_heap flags

Rohan Garg (1):

  • intel/compiler: use the correct cache enum for loads and stores

Samuel Pitoiset (6):

  • radv: fix allocating sparse descriptor buffers in the 32-bit addr space

  • radv: disable shaders linking with ESO when nextStage of VS/TES isn’t present

  • radv/rt: skip shaders cache for pipelines created with the capture/replay flag

  • radv: fix lowering the view index to an input varying for FS

  • radv,aco: fix legacy vertex attributes when offset >= stride on GFX6-7

  • radv: fix lowering VS inputs when offset >= stride on GFX6-7

Sviatoslav Peleshko (2):

  • brw: Fix mov cmod propagation when there’s int signedness mismatch

  • mesa: Reset vbo attributes after flushing them to Current in glPopAttrib

Tapani Pälli (1):

  • iris: fix issues with memory object updates via glBufferSubData

llyyr (1):

  • vulkan/wsi/wayland: fix suboptimal flag being ignored with explicit sync