Open source graphics drivers Mesa 26.0.1 released with various bug fixes and a security fix

Aitor Camacho (7):
– wsi/metallic: Expose extra colour areas if occasion extension enabled
– kk: Fill pipelineUUID
– kk: Repair shader uint32_t worth serialization
– kk: Accurately launch pipeline handles at shader destroy
– kk: Repair compute pipeline cache
– kk: Transfer gfx pipeline information to the data struct inside kk_shader
– kk: Repair graphics pipeline serialization

Alyssa Rosenzweig (1):
– brw: drop buggy SLM optimization

Anna Maniscalco (1):
– freedreno/frequent: set has_astc_hdr true for a7xx targets

Benjamin Otte (1):
– lavapipe: Repair options for nonsubsampled ycbcr codecs

Daniel Schürmann (1):
– nir/clone: Repair cloning oblique name directions

Danylo Piliaiev (1):
– ir3: Align TCS per-patch output to 64 bytes to forestall stale reads

Emma Anholt (1):
– ir3/ra: Repair DOUBLE_ONLY restrict stress computation.

Eric Engestrom (7):
– docs: add sha sum for 26.0.0
– .pick_status.json: Replace to 03d2cc2b2ae5341409ee1fab74e98134a6df0511
– bin/gen_release_notes: repair help for python 3.14
– pick-ui: add `Backport-to: *` as a synonym to `Cc: mesa-stable`
– .pick_status.json: Mark 7dd7731ac710b0c7213f6bb466b33f62eca80604 as denominated
– docs: add launch notes for 26.0.1
– VERSION: bump for 26.0.1

Religion Ekstrand (6):
– pan/clear: Cease packing undefined bits in colours
– nir/gather_info: Add help for panfrost tile load/retailer intrinsics
– panvk: Create each Z/S descriptors, even for separate Z/S
– panvk/preload: Cease assuming 32 registers
– panvk/jm: Refactor BeginRendering()
– panvk: Additionally load output attachments with LOAD_OP_NONE+STORE_OP_NONE

Frank Binns (2):
– pvr/ci: transfer some timing out exams from fails to skips
– pvr: Repair alloc callbacks utilization when liberating body buffers

Ian Romanick (8):
– spirv: Use STACK_ARRAY as an alternative of NIR_VLA
– nir: Use STACK_ARRAY as an alternative of NIR_VLA
– brw: Name nir_opt_algebraic_late in brw_nir_create_raygen_trampoline
– brw: Name nir_opt_algebraic_late later in brw_postprocess_nir_opts
– elk: Name nir_opt_algebraic_late in elk_postprocess_nir
– brw/cmod: Do not propagate from CMP to ADD if there’s a write between
– elk/cmod: Do not propagate from CMP to doable Inf + (-Inf)
– elk/cmod: Do not propagate from CMP to ADD if there’s a write between

Janne Grunau (3):
– asahi: Use GPU for buffer copies in useful resource_copy_area()
– asahi: Implement clear_buffer utilizing libagx_fill*
– hk: Use aligned vector fill in hk_CmdFillBuffer if doable

Jarred Davies (2):
– pvr: Repair allocating the required scratch buffer house for tile buffers
– pvr: Add lacking help for tile buffers to SPM EOT packages

Jesse Natalie (1):
– meson: Embrace DirectX-Headers dependency for all VK Home windows builds

Jianxun Zhang (1):
– anv: Restrict modifier disabling workaround to particular GTK variations

José Roberto de Souza (1):
– intel/perf: Add HSW verx10 to intel_perf_query_result_write_mdapi()

Juston Li (1):
– anv: set lacking protected bit for protected depth/stencil surfaces

Konstantin Seurer (2):
– radv: Repair setting the viewport for depth stencil FS resolves
– vulkan/cmd_queue: Fixup stride for multi attracts

Lars-Ivar Hesselberg Simonsen (2):
– panvk: Repair dcd_flags1 soiled bit
– pan/genxml/v13: Repair HSR Prepass typo

Leon Perianu (1):
– pvr: repair format desk properties duplicate

Lionel Landwerlin (8):
– anv: flush render caches on first pipeline choose
– anv: repair nested command buffer relocations
– anv: add lacking fixed cache invalidation for descriptor buffers
– isl: repair 32bit math with 4GB buffer dimension
– anv: apply the identical ccs disabling for Xe3 than Xe2
– anv: disable ccs modifier reporting when ccs modifiers are disabled
– anv: soiled descriptors after blorp operations
– anv: take away snprintf for aux op transition

Mary Guillemard (1):
– hk: Repair crash in hk_handle_passthrough_gs

Matt Turner (4):
– brw/cse: repair `operands_match` corrupting non-IMM register information
– brw/cse: use copies in `operands_match` as an alternative of in-place modification
– elk/cse: repair `operands_match` corrupting non-IMM register information
– elk/cse: use copies in `operands_match` as an alternative of in-place modification

Mike Blumenkrantz (2):
– zink: repair damaged compiler assert
– zink: solely do pre-sync switch barrier after a renderpass

Natalie Vock (3):
– radv/rt: Solely use ds_bvh_stack_rtn if the stack base is feasible to encode
– radv: Initialize nir_lower_io_to_scalar progress variable
– radv/nir: Accurately deal with workgroup sizes not aligned to 32

Nick Hamilton (5):
– pvr: Repair incorrect subpass merging optimisation
– pvr: Rename pvr_render_input_attachment
– pvr: Add lacking help for protect attachments
– pvr: Replace CI fails checklist after render go fixes
– pvr: Add help for fragment go by shader

Olivia Lee (1):
– hk: repair passthrough GS key invalidation

Pavel Ondračka (2):
– r300: align macro-tiled stride-addressed textures in X
– mesa: implement FRAMEBUFFER_RENDERABLE internalformat question

Rhys Perry (3):
– aco: repair gfx6-8 store_scratch() with operate calls
– aco: reset all vgpr_used_by_vmem_ in resolve_all_gfx11
– aco: resolve hazards earlier than calls

Robert Mader (1):
– lavapipe: allow dmabuf import for planar drm codecs

Ryan Zhang (1):
– panvk: guard towards NULL tips to keep away from crash

Samuel Pitoiset (5):
– ac,radv,radeonsi: use appropriate swizzle/pitch for depth-only pictures with SDMA
– radv: repair potential corruption after FMASK decompression on GFX6-8
– radv/meta: repair depth/stencil resolves with totally different areas
– ac/nir: repair writemask for twin supply mixing on GFX11+
– radv: repair potential GPU hangs with secondaries on switch queue

Tapani Pälli (1):
– util: deliver again repair to keep away from strict aliasing bugs in xxhash

Timothy Arceri (2):
– mesa: add _mesa_lookup_state_param_idx() helper
– st/glsl_to_nir: ensure that the variant has the right areas set

Wei Hao (1):
– radeonsi: repair threaded shader compilation ending after context is destroyed

Yiwei Zhang (2):
– venus: workaround a gcc-15 lifeless retailer elimination (DSE) bug
– venus: sync protocol for strict aliasing compliance