docs/relnotes/21.1.0.rst

8 Mesa 21.1.0 implements the OpenGL 4.6 API, but the version reported by
10 glGetIntegerv(GL_MINOR_VERSION) depends on the particular driver being used.
11 Some drivers don't support all the features required in OpenGL 4.6. OpenGL
15 Mesa 21.1.0 implements the Vulkan 1.2 API, but the version reported by
16 the apiVersion property of the VkPhysicalDeviceProperties struct
17 depends on the particular driver being used.
70 - [RADV] corruption in avatar after dying in Heroes of the Storm
84 - AMD hevc_vaapi ffmpeg encoding = wrong image width (48px black bar on the right)
168 - piglit-replay: JUnit file contains wrong links to the tracie dashboard
213 - ci: new traces runner needs dashboard links in the job log and junit
255 - osmesa: Pacify MSVC in the test code
256 - glx: Fix GLX_SGI_video_sync for the no-current-drawable case
279 - glx: Pull use_x_font out of the context vtable
280 - glx: Pull get_proc_address out of the context vtable
283 - dri: Explicitly handle all the config attributes
284 - dri: Fold attribMap into the code
286 - mesa: Remove the texture-from-pixmap state from struct gl_config
289 - mesa: Remove the pretense of aux buffer support
292 - ci: Bump the llvmpipe test timeout to 240 seconds
301 - dri: Don't tie the accum buffer's alpha-ness to the color buffer's
302 - glx: Stop pretending to validate the pbuffer fbconfig attributes
303 - glx: Don't downgrade the visual caveat from the server
308 - glx: Default sRGBCapable in the same place as the other config attribs
314 - glx: Lift sending the MakeCurrent request to top-level code
315 - Revert "glx: Lift sending the MakeCurrent request to top-level code"
349 - v3dv/pipeline: try to get the shader variant directly from the cache
353 - v3dv: Add support for the on-disk shader cache
405 - pan/bi: Implement spilling at the clause-level
412 - pan/bi: Factor nir_function_impl out of the context
451 - pan/decode: Remove dependency of decoder on the encoder
481 - pan/bi: Include ATEST datum in the instruction
523 - panfrost: Push uniforms required by the program
573 - pan/mdg: Rename bitcount8 to popcnt, fixing the unit
632 - pan/bi: Materialize \*DTSEL_IMM in the scheduler
728 - ci: correct the trace image URLs in the piglit summary
730 - ci: clean paths used in the piglit runner
734 - ci: tracie dashboard URLs only in the failure after the testcase
739 - ci: add libdrm to the x86_test-vk container
818 - nir: store the results of divergence analysis on loops
825 - v3dv: initialize render_fd at the top of physical_device_init
849 - st/nine: Add new function to know if we are the worker
851 - st/nine: Use the texture memory helper
852 - st/nine: Control the memfd virtual limit
869 - radeonsi: Limit the size of the in-memory shader cache
870 - radeonsi: fix leak when the in-memory cache is full
904 - st/nine: Allow to override the vram size
923 - radv: Use the surface offset from ac_surface instead of a plane offset.
929 - radv: Enable modifiers with the WSI.
948 - radv: Remove the format table.
960 - radv: Add DCC info to the metadata.
964 - ac/rgp: Only report double the prims per clock on GFX10.
982 - zink: Only set the needed number of scissors.
1013 - CI: windows: augment the timeout of building the windows container
1014 - CI: windows: split the layers to meet new registry requirements
1023 - panfrost: Don't skip the test with a 4k shader
1026 - panfrost: Don't memset the last attribute buffer entry twice
1027 - panfrost: Only allocate the extra attribute buffer entry on Bifrost
1033 - panfrost: Fix a polygon list corruption in the multi-context case
1034 - panfrost: Don't add the tiler BO when it's not accessed
1043 - panfrost: Keep the compiler inputs in the context
1044 - panfrost: Move the shader compilation logic out of the gallium driver
1045 - panfrost: Provide a helper to prepare the shader related parts of an RSD
1046 - panfrost: Use the pan_shader_prepare_rsd() helper
1048 - panfrost: Move the blend lowering code out of the gallium driver
1049 - panfrost: Move the blend logic out of the gallium driver
1052 - panfrost: Split the direct and indirect draw logic
1057 - pan/bi: Extend the bi_builder to support type variants correctly
1058 - panfrost: Add a knob to disable the UBO -> push constants optimization
1059 - panfrost: Allow passing an explicit UBO index for the sysval UBO
1060 - panfrost: Print the correct UBO size when dumping UBO information
1061 - panfrost: Don't count the special vertex/instance ID attributes on Bifrost
1062 - panfrost: Split the sampler and texture count
1067 - panfrost: Define the Surface and Surface-with-stride descriptors
1069 - panfrost: Use the descriptor allocators where appropriate
1071 - panfrost: Move the blend shader cache at the device level
1072 - panfrost: Use the blend shader cache attached to the device
1073 - panfrost: Don't reserve space in the color buffer for disabled RTs
1083 - panfrost: Add extra info to the pan_image_layout struct
1089 - panfrost: Patch the gallium driver to use pan_image_layout_init()
1091 - panfrost: Provide a helper to calculate the polygon list size
1094 - pan/midg: Use the sampler index passed to the texture instruction
1097 - panfrost: Add align info to the draw and draw padding definitions
1098 - panfrost: Add the early ZS pre frame mode
1104 - panfrost: Create a blitter library to replace the existing preload helpers
1106 - panfrost: Use the generic preload and FB helpers in the gallium driver
1107 - panfrost: Kill the old tile-buffer preload logic
1110 - panfrost: Don't advertise AFBC mods when the format is not supported
1123 - nir: Add a data pointer to the callback in nir_remove_dead_variables
1135 - spirv: Store SPIR-V version of the module
1159 - intel/compiler: Refactor the shader INTEL_DEBUG checks
1271 - etnaviv: tell the truth if alpha-test is supported
1373 - aco: don't decrease the vgpr_limit when encountering bpermute
1381 - aco: remove dead code for the handling of exec temporaries
1383 - nir: lower load_helper to is_helper if the shader uses demote()
1429 - turnip,freedreno/a6xx: tell hw the size of shared mem used by CS
1462 - turnip: enable infinities for f16 math and document the register
1527 - lavapipe: use the common icd generator
1557 - gallium: add a view mask to the draw command
1560 - draw: add interface to notify renderer of the current view index
1561 - draw: refactor out the instances drawing code
1563 - draw: pass the view index to the render driver
1564 - draw/vs: pass the view index to the vertex shader
1566 - llvmpipe: add the view index callback from draw
1577 - lavapipe: move to the common casting interfaces
1612 - drisw: move zink down the list below the sw drivers.
1642 - gallium/indices: Use "__restrict" to help the compiler
1655 - bin/post_version: convert the csv.reader into a concrete list
1690 - anv: Added the VK_EXT_sample_locations extension to the anv_extensions list
1695 - anv: Enabled the VK_EXT_sample_locations extension
1713 - gallium/tgsi: Rewrite the docs on source modifiers.
1715 - mesa/st: Make a single helper for the NIR-to-TGSI transfer.
1722 - ci/freedreno: Fix up the xfail/flake handling of a3xx texture functions.
1730 - mesa: Replace the float[4] unpack code with util/format's.
1731 - mesa: Make _mesa_unpack_rgba_block() use the u_format pack/unpack.
1739 - mesa: Move the rest of format_unpack.py out of code generation.
1740 - util/format: Simplify the generated unpack code.
1741 - swrast: Use util_format_write_4/4ub for the scattered pixel writes.
1743 - mesa/main: Replace the uint format packing code with util/format's.
1752 - util/bitset: Avoid dereferencing the bitset for size == 0.
1754 - freedreno/a6xx: Skip the body of emit_state if we're clean.
1758 - ci/freedreno: Detect cheza HFI errors and restart the run.
1762 - freedreno: Use a real type instead of void * for the fd_batch->key.
1763 - freedreno: Early-out from the resource write path when we're the writer.
1766 - mesa/st: Allocate the gl_context with 16-byte alignment.
1768 - ci/freedreno: Use the new nginx cache for trace downloads.
1769 - ci/freedreno: Use the http cache for artifacts downloads, too.
1771 - docs/ci: Document setting up the http cache for traces.
1772 - ci/lava+baremetal: Add an xserver to the root fs.
1783 - ci/freedreno: bump VK coverage to 1/4 of the CTS.
1789 - mesa/st: Always precompile the first shader variant.
1790 - mesa/st: Assume that the default variant is always first in the list.
1800 - gallium: Flip the default value of PIPE_CAP_SHAREABLE_SHADERS.
1801 - virgl: Drop a context dependency from part of the shader compile path.
1802 - v3d/qpu: Avoid leaking memory in the QPU disasm test.
1804 - ci: Allow better customization of the name of the artifacts for minio.
1806 - ci/freedreno: Drop the "arm64" in front of job names.
1811 - ci/a5xx: Update the piglit expectations.
1813 - ci: Move the dEQP and traces expectations to the per-driver CI dirs.
1814 - ci: Move the piglit expectations lists to the per-driver CI dirs.
1818 - softpipe: Fix the const buffer overflow check.
1819 - mesa: Get the FXT1 compressor/decompressor off of GL types.
1820 - mesa: Move the FXT1 compressor/decompressor to util/
1823 - ci/piglit: Stop including the test counts at the end of expectations.
1824 - ci/iris: Move the traces yml file to the driver-specific dir.
1829 - freedreno: Use the mesa/st frontend lowering of GL_CLAMP.
1831 - freedreno/a5xx+: Drop the unused no_decode_srgb flag.
1836 - ci/lava: Move the per-driver gitlab-ci.yml to each driver.
1837 - ci/lava: Move the driver expectation files to the per-driver CI dir.
1838 - tgsi_exec: Roll the loops for condmask handling.
1841 - ci/freedreno: Switch the fastboot boards to using nfsroot.
1844 - ci/a5xx: Increase the gles3/31 coverage.
1848 - lima: stop encoding the texture format in the shader key
1850 - lima: upload the shader to a BO at shader creation
1855 - freedreno/a5xx: Flush depth at the end of sysmem, like a6xx does.
1857 - u_format: Mark the generated pack/unpack src/dst args as restrict.
1860 - mesa/st: Fix precompile misses on compat GL VSes writing to color outputs.
1862 - ci/freedreno: Add three more a5xx flakes from the last day.
1863 - freedreno/a5xx: Fix the texel buffer alignment requirement.
1864 - freedreno/a5xx: Fix the max texture buffer size.
1865 - ci/panfrost: Disable t860/radeonsi testing while the runners are struggling.
1867 - ci/freedreno: Switch the piglit testing to the new piglit runner.
1870 - ci/freedreno: Mark the rest of the glx_arb_sync_control@timing as flakes.
1875 - ci/bare-metal: Move the db820c lockup detect to the right boot script.
1877 - freedreno/a6xx: Rename the RB_BLIT_INFO.INTEGER field to SAMPLE_0.
1879 - freedreno: Move the ir3 linked shader cache to the context.
1886 - ci/android: Build the v3dv driver.
1890 - freedreno: Move max-tf-vtx calculation to just the HW that needs it.
1891 - freedreno: Move the primitives generated/written updates after the draw.
1896 - freedreno/a6xx: Use the frontend userclip lowering.
1900 - ci: Drop the custom db820c kernel/dtb from the kernel+rootfs.
1902 - ci/freedreno: Rename a306-test and a530-test to drop "arm64" from the name.
1906 - ci/freedreno: Switch to the trimmed glxgears trace.
1907 - ci/freedreno: Fix up the a5xx border color flake annotation.
1912 - nir_to_tgsi: Use ARL instead of UARL in the !native_integers case.
1971 - docs/zink: document the independentBlend requirement for GL3
2083 - compiler/glsl: correct the number of string-arguments
2129 - vulkan/device_select: Only call vkGetPhysicalDeviceProperties2 if the device supports it.
2139 - r600/sfn: Keep array registers alive for the whole shader
2146 - r600/sfn: Add the position input as varying
2147 - r600/sfn: Fix FS inputs when reading from the same position
2149 - r600/sfn: Use the constant buffer ID when given
2167 - r600/sfn: Allow any channel for the helper invocation evaluation
2174 - r600/sfn: handle querying the number of layers in cube arrays
2179 - nir: Add r600 specific intrinsic for loading the tesselation coords
2203 - r600/sfn: drop the local register mep
2217 - r600/sfn: Only fetch the constants that are needed in fdd*
2235 - gitlab-ci: build the iris gallium driver as well
2285 - v3dv: fix disabling Early Z for the whole frame
2296 - v3dv: handle D/S buffer to image copies with the texel buffer path
2297 - v3dv: batch copies in the copy_buffer_to_image_blit path
2311 - broadcom/compiler: disallow reading two uniforms in the same instruction
2331 - broadcom/compiler: drop the destination for unused ldunifa
2336 - broadcom/compiler: add an additional sanity check assert to the ldvary fixup
2340 - v3dv: call util_cpu_detect() when initializing the instance
2364 - broadcom/compiler: add a definition for the unifa skip distance
2379 - intel/compiler: Enable the ability to emit CMPN instructions
2380 - intel/compiler: Make the CMPN builder work like the CMP builder
2394 - nir/algebraic: Apply addition property of equality to the other ordering too
2412 - pan/bi: Improve interoperability of the command-line disassembler
2417 - panfrost: Use the correct NIR options for OpenCL on Bifrost
2425 - pan/bi: Add w0 to the 'h01' swizzle bucket
2437 - panfrost: Fix the tile size assertion
2441 - panfrost: Add the tiler heap to fragment jobs
2442 - pan/bi: Return the size of the last clause from bi_pack
2445 - pan/bi: Don't check liveness unless the index is valid
2446 - pan/bi: Use the correct size for UBO loads
2454 - pipe-loader,gallium/drm: Fix the kmsro pipe_loader target
2459 - panfrost: Only add resource checksum BOs to the batch once
2465 - pan/decode: Allow frame shader DCDs to be in another BO than the FBD
2479 - nv50,nvc0: clear internal vbo masks based on the trailing slots
2521 - st/mesa: adapt for the case where buffers are not supported in frag
2549 - anv: use helper function to get the buffer size
2556 - nouveau: no modifier != the invalid modifier
2600 - intel/genxml,anv,iris: Drop the legacy compute path from gen125.xml
2603 - anv: Take the set of stages to flush in flush_descriptor_sets
2604 - anv: Only flush descriptors used by the pipeline
2630 - anv: Clean up the vk_device on the CreateDevice error path
2638 - anv: Use the common extension table struct
2641 - vulkan-overlay-layer: Use the new dispatch tables
2649 - anv: Use the common dispatch framework
2656 - vulkan: Add a truly common VK_EXT_debug_report implementation
2657 - anv: Switch to the common VK_EXT_debug_report
2658 - turnip: Use the common dispatch framework
2661 - turnip: Switch to the common VK_EXT_debug_report
2665 - v3dv: Switch to the common VK_EXT_debug_report
2667 - radv: Switch to the common VK_EXT_debug_report
2671 - vulkan: Make the debug_report implementation internal
2674 - vulkan: Drop the type_prefix parameter from gen_extensions
2676 - nir: Drop the lower_mem_constant_vars declaration
2679 - anv: Pull the patch version from the XML
2690 - spirv: Store the nir_function in vtn_function
2691 - spirv: Delete the impl for prototype-only functions
2693 - nir: Fix parameter order in the bcsel-of-shuffle optimization
2698 - intel/mi_builder: Create a context in the tests
2706 - intel/mi_builder: Drop the gen\_ prefix
2722 - turnip: Move the CreateRenderPass wrapper to common code
2747 - anv/apply_pipeline_layout: Rework the early pass index/offset helpers
2748 - anv/apply_pipeline_layout: Lower UBO loads in the early pass
2749 - anv/apply_pipeline_layout: Run DCE between the early and late passes
2755 - anv: Rework the 64bit_bounded_global resource index format
2756 - anv: Zero out the last dword of UBO/SSBO descriptors in the shader
2759 - anv/apply_pipeline_layout: Rework the desc_addr_format helper
2760 - anv/apply_pipeline_layout: Use the new helpers for early lowering
2761 - anv/apply_pipeline_layout: Use the new helpers for images
2770 - intel/tools: Handle milti-LRI in the batch decoder
2771 - intel/tools: Handle GT_MODE in the batch decoder
2780 - anv: Clean up anv_descriptor_pool::base on the error path
2781 - anv: Clean up anv_semaphore::base on the error path
2787 - anv: Use the same re-order mode for streamout as for GS
2788 - intel/isl: Fix isl_color_value_unpack to match the prototype
2793 - darwin: Use the system libexpat
2795 - darwin: Use the system libunwind
2810 - CI: Use a sha for the Windows SPIRV-LLVM-Translator dependency
2817 - microsoft/clc: Move inline samplers to the end of the variable list
2836 - d3d12: Use CreateDXGIFactory2 and use the debug flag when appropriate
2867 - nir_opt_deref: ptr_as_array(deref_cast<T*>(x))[0] isn't the same as x[0] if the cast has alignment
2977 - ci/v3dv: add flaky test in the skip list
3035 - iris: Reconfigure the URB only if it's necessary or possibly useful
3036 - iris: Move the URB size checks into iris_update_compiled_xs
3042 - iris: Store a list of shader variants in the shader itself
3052 - iris: Reference the shader variant for last_vue_map as well
3053 - iris: Pin surface state buffers after possibly updating the clear color
3063 - i965: Rename the rest of intel_* functions to brw_*
3069 - tnl: Call _mesa_matrix_analyse to make sure the inverse MVP is updated
3075 - iris: Use different shader uploaders for precompile vs. draw time
3083 - Half-revert "gallium/dri2: Pass the resource that corresponds to the plane"
3093 - radeon/vcn: clean the message buffers and their indexes logic
3104 - radeonsi: replace the hard coded video decode and encode caps
3110 - virgl: move new added field to the end.
3120 - intel/perf: link queries back to the gen_perf_config object
3147 - intel/perf: drop the special READ_REG operator
3171 - anv: track the end of the command buffers
3190 - intel/mi_builder: use device info to use the right CS prefetch size
3191 - anv: use the device size of CS prefetch to pad secondary buffer calls
3214 - renderonly: close the gpu fd when destroying renderonly
3215 - etnaviv: don't try to copy PIPE_BUFFER with the 3D engine
3217 - Revert remaining half of "gallium/dri2: Pass the resource that corresponds to the plane"
3283 - radeonsi: trim the size of si_vgt_param_key and si_vgt_stages_key
3290 - radeonsi: inline the last use of si_get_vs_state
3292 - radeonsi: enable the GS tri strip adj workaround with primitive_restart
3293 - radeonsi: clear dirty_atoms and dirty_states only if we entered the emit loop
3314 - mesa: eliminate the switch statement for STATE_TEXGEN
3339 - radeonsi: move if (sctx->vertex_buffers_dirty) into the upload function
3354 - gallium/u_upload_mgr: eliminate all atomics for the upload buffer
3357 - gallium/util: optimize pipe_vertex_buffer_reference binding the same buffer
3367 - radeonsi: mark shader_pointers dirty once outside the upload descriptors loop
3391 - mesa: don't count buffer references for the context that created them
3393 - radeonsi: simplify the NGG culling condition in si_draw_vbo
3395 - radeon: decrease the size of radeon_cmdbuf by switching prev fields to uint16
3401 - winsys/amdgpu: expand the slab allocation range to [256 B, 1 MB]
3405 - winsys/amdgpu: move amdgpu_winsys_bo::cpu_ptr into the u.real union
3406 - winsys/amdgpu: move amdgpu_winsys_bo::is_shared to the u.real union
3407 - winsys/amdgpu: move amdgpu_winsys_bo::is_user_ptr to the u.real union
3408 - winsys/amdgpu: move amdgpu_winsys_bo::use_reusable_pool to the u.real union
3442 - mesa: optimize the dual source blend error checking using a bitmask
3452 - mesa: optimize set_varying_vp_inputs by precomputing the conditions
3462 - radeonsi: for tess, determine the minimum num_patches before optimizing tg size
3466 - radeonsi: fix the value of uses_bindless_samplers
3474 - glthread: ignore the return value of glUnmapBuffer, don't sync, and return true
3475 - i915: use align_calloc for the context to fix m32 crashes
3476 - radeon,r200: use align_calloc for the context to fix m32 crashes
3477 - nouveau_vieux: use align_calloc for the context to fix m32 crashes
3482 - mesa: don't compute the inverted projection matrix if not used
3483 - mesa: don't compute the ModelView * Projection matrix if not used
3484 - mesa: add _NEW_MATERIAL to reduce the weight of _NEW_LIGHT_CONSTANTS
3489 - mesa: remove the fixed-func vert prog dependency on all texture states
3502 - mesa: don't update the fixed-func frag prog on irrelevant _NEW_COLOR changes
3516 - ac: set the TCC line size for Aldebaran
3532 - mesa: clear reserved parameter storage because it's stored in the shader cache
3538 - radeonsi: select the optimal micro tile mode at clear regardless of fast clear
3540 - amd/addrlib: add back the incorrect original DCC checking
3546 - radeonsi: set the clear/copy cache policy based on L2 cache size
3553 - radeonsi: set compute/cpdma sync flags in the outermost caller
3559 - radeonsi: use the optimal packet order before draw packets for VS flushes too
3563 - radeonsi: don't use a constant buffer for the copy_image compute shader
3564 - radeonsi: decrease the maximum variable block size
3565 - radeonsi: pack the variable block size in one SGPR, 10 bits per component
3566 - amd: fix parsing the last dword of DMA_DATA packets
3568 - radeonsi: determine accurately whether the framebuffer state has DCC MSAA
3577 - radeonsi: turn the loops over color buffers into while loops in si_clear
3581 - gallium/pb: pass an optional winsys pointer to the buffer destroy function
3611 - radeonsi: indent the code for TC-compatibility HTILE transition
3619 - ac/surface: only apply the 3D swizzle mode tuning to gfx10+
3621 - ac/surface/tests: fix a random segfault in the modifier test
3625 - ac/surface: limit the number of swizzle modes that can have displayable DCC
3626 - ac,radeonsi: rewrite DCC retiling without the DCC retile map
3630 - radeonsi: move binding the internal compute shader into si_launch_grid_internal
3636 - glsl: pack 16-bit uniforms in the NIR linker
3653 - radeonsi: make the gfx9 DCC MSAA clear shader depend on the number of samples
3685 - ir3: Don't count (nopX) towards the wrong category
3705 - android: mesa: Move the FXT1 compressor/decompressor to util/
3715 - android: panfrost: Use the blend shader cache attached to the device
3730 - microsoft/spirv_to_dxil: Add extra lowering functions according to the docs on nir_inline_functio…
3739 - ci: Run 'time' in the background and propagate signals to test process
3758 - ci: Add strace to the x86_build docker image
3821 - util/set: add the found param to search_or_add
3849 - zink: break out some of the u_blitter setup into util function
3932 - radv: zero the bo descriptor array when allocating a new set
3958 - zink: flush gfx/compute batches when the other pipeline needs resource sync
3981 - zink: add PIPE_BIND_QUERY_BUFFER to the all-purpose resource creation path
3995 - zink: make the xfb_query_pool into an array
3997 - zink: put SO_OVERFLOW queries on the primgen list
4002 - zink: hook up cs invocation queries to the compute batch
4009 - zink: add push constant value to indicate whether the current draw is indexed
4015 - zink: add util function for submitting the compute batch
4030 - zink: create a VkPipelineCache object on the screen and use it
4034 - zink: move command pool to the batch
4068 - zink: just end the current renderpass in zink_batch_no_rp()
4073 - zink: simplify some of the qbo direct buffer write code
4083 - zink: rename 'has_draw' flag on batches and set it when the batch is used
4084 - zink: move gfx pipeline creation closer to the bind point
4095 - zink: ralloc the main context
4104 - zink: discard pending clears during blit/copy if we'll overwrite the data
4106 - zink: optimize the remaining read cases of applying pending clear calls
4107 - zink: move all the clear stuff to zink_clear.h
4168 - zink: apply only the pending zs clear bits during deferred clears
4205 - zink: move surface refs to the end of descriptor updating
4209 - zink: add a null sampler view descriptor hash to the screen
4214 - zink: move descriptor set allocation near the top of update_descriptors
4215 - zink: only batch-reference the program in use once per descriptor update
4218 - zink: don't double iterate all the per-batch sets on reset
4226 - zink: break out all the descriptor pool/layout stuff into a new struct
4234 - zink: move descriptor binding out of the update codepath
4260 - zink: determine whether the vulkan driver requires mesa flush wsi
4265 - zink: unset ctx->program pointers when an unref destroys the object
4308 - zink: apply Delete All The Code methodology to the ubo/ssbo variables
4311 - ci/panfrost: disable the rest of these jobs temporarily
4383 - zink: refactor clears a little to track a bitfield of enabled clears on the context
4406 - lavapipe: use the passed offset for CmdCopyQueryPoolResults
4446 - aux/draw: pass the full draw params through to draw_instances()
4447 - aux/draw: pass the full draw params through to draw_pt_arrays_restart()
4483 - radv: set gfx pipeline vtx_emit_num to the number of sgprs
4484 - radv: track whether drawid is used on the pipeline struct
4485 - radv: track whether baseinstance is used on the pipeline struct
4512 - zink: only emit SpvCapabilitySampleMaskPostDepthCoverage if the mode is set
4554 - aux/trace: dump all the blend state members
4573 - zink: always copy the nir shader before compiling
4591 - zink: only reset query on suspend if the query has previously been stopped
4605 - iris: Fix the depth aspect aux usage in iris_blit
4657 - ac/rgp: make the max gap between shader code a warning
4658 - ac/rtld: make ac_rtld_upload returns the code size
4660 - radeonsi/sqtt: keep a copy of the uploaded shader code
4669 - mesa/hash: make the mtx non-recursive
4685 - radeonsi: use SI_CONTEXT_FLAG_AUX when recreating the aux context
4686 - radeonsi: do not recreate the aux context from the aux context
4687 - radeonsi: only recreate the aux_context when soft recovery failed
4688 - radeonsi: re-create the aux context in si_create_context
4707 - aco: move VADDR to the end of the operand list
4717 - aco: don't consider a phi trivial if same's register doesn't match the def
4723 - aco: fix WQM for texture instructions with args before the coordinates
4726 - aco: disable a*1.0 optimization if the instruction is precise
4836 - aco: allow SDWA sels smaller than the operand size
4991 - ci: Ensure that jobs inherting the ci-deqp jobs artifact meson logs
4997 - virgl: Support the ETC1_RGB8 format as virglrenderer supports it
5005 - radeon/vcn: release si buffer for encoding at the end.
5013 - anv: Invalidate the correct AUX-TT entry
5035 - radv: do not invalidate the L2 metadata cache on compute queues
5038 - radv: add a comment explaining the micro tile mode resolve
5041 - radv: restore invalidating the vector cache for internal meta operations
5042 - radv: flush L2 for images affected by the pipe misaligned issue on GFX10+
5043 - ci: exclude one CTS test that timeout most of the time for RADV CI
5048 - radv: fix overflow when computing the SQTT buffer size
5055 - radv: use the range aspect mask in FMASK color expand
5057 - radv: only decompress the depth/stencil aspect that needs to be resolved
5059 - radv,aco: fix shifting input VGPRs for the LS VGPR init bug on GFX9
5066 - ac/surface: store HTILE mip info into the surface
5067 - radv: use the image view range when fast clearing depth
5068 - radv: check if HTILE is enabled per-level instead of the entire image
5071 - radv: add support for fast clearing levels of the HTILE buffer
5072 - radv: teach radv_htile_enabled() about the number of HTILE levels
5076 - radv/winsys: move the initial BO domain to radeon_winsys_bo
5078 - radv: fix waiting on the last enabled RB for occlusion queries
5079 - radv/winsys: use an array for the global BO list instead of a list
5080 - radv/winsys: remove the radv_amdgpu_winsys_bo::ws indirection
5084 - radv/winsys: simplify the user fence logic for submission
5090 - radv/winsys: move the debug_all_bos check outside of the add/del helpers
5092 - radv/winsys: add buffer_make_resident() to the API
5093 - radv/winsys: add the resident BOs to the list of BOs at submit time
5094 - radv/winsys: enable the global BO list unconditionally
5095 - radv: use the global BO list from the winsys
5096 - radv: fix printing the debug option names
5100 - radv: do not overallocate the SQTT buffer
5101 - radv: adjust an error message related to the SQTT buffer size
5102 - radv: add support for resizing the SQTT buffer automatically
5103 - ac/rgp: append the number of seconds to the generated RGP file
5105 - radv: only make the WSI images resident if the global BO list is used
5106 - radv/winsys: set use_global_list inside the critical section
5107 - radv: only apply the MRT output NaN fixup to non-meta shaders
5108 - radv: create the start/stop CS for SQTT dynamically
5113 - radv: do not allow to capture SQTT on the compute queue
5116 - radv: use the pipeline key as hash for pipeline bind markers
5118 - radv: make the border color BO a resident buffer
5119 - radv: make the trace BO a resident buffer
5120 - radv: make the TMA/TBA BOs resident buffers
5121 - radv: emit the trap handler registers earlier
5125 - radv: store a pointer to the code in radv_shader_variant
5128 - radv: do not scale the depth bias for D16_UNORM depth surfaces
5132 - ac/rgp: report the number of memory operations per clock
5133 - ac/rgp: report the number of primitives per clock
5135 - radv: always select the first active CU when profiling with SQTT
5138 - Revert "radv: do not overallocate the SQTT buffer"
5139 - radeonsi,radv: do not overallocate the SQTT buffer size
5140 - radv: remove useless decompression of the DS resolve attachment
5142 - ac/sqtt: fix determining if the trace is complete on GFX10+
5143 - radv: double the SQTT buffer size when it is resized
5144 - radv: trigger a new SQTT capture automatically after resizing the buffer
5145 - radv: bump the initial SQTT buffer size to 32MB per SE
5164 - radv: do not enable TC-compat CMASK if the image isn't readable by a shader
5167 - radv: only configure the CMASK tiling for TC-compat on GFX8
5168 - radv: initialize TC-compat CMASK images with the DCC clear code
5171 - radv: extend the dirty bits to 64-bit
5176 - radv: gather if the FS uses perspective or linear interpolations
5191 - radv: use a sampled image descriptor for reads for the MSAA color decompress
5202 - aco: implement a workaround for the image load DCC hw bug on GFX10.3
5212 - radv: do not clamp framebuffer dimensions to the minimum dimension
5222 - ac/surface: copy the HTILE equations to the surface
5224 - ac/surface: store the HTILE pitch to the surface
5225 - radv: expose R8_UINT as the only supported format for VRS attachments
5229 - radv: configure the VRS HTILE encoding size
5230 - radv: do not use the whole HTILE buffer for depth when VRS is used
5231 - radv: update the HTILE clear word when VRS is used
5234 - radv: handle the VRS attachment subpass
5235 - radv: bind our internal depth buffer when not provided by the app
5238 - radv: configure the VRS combiners when an attachment is used
5246 - radv/winsys: fix allocating the number of CS in the sysmem path
5247 - radv/winsys: fix resetting the number of padded IB words
5250 - radv: fix computation of the number of user SGPRS for NGG GS state
5302 - anv: toggle on sample shading if it is set in the shader
5327 - util/disk_cache: use a new cache dir for the single file cache feature
5333 - util/disk_cache: move cache tests to the util directory
5338 - util/disk_cache: make use of the new compression helpers
5374 - aco: Implement the new tessellation I/O related NIR intrinsics.
5377 - ac/llvm: Implement the new tessellation intrinsics.
5387 - radv: Calculate tess patches and LDS use outside the backend compilers.
5388 - radv: Determine tcs_in_out_eq in radv_pipeline instead of the compiler.
5390 - radv: Reorder some NIR optimizations in preparation for the I/O changes.
5400 - aco: Use s_setprio 3 at the beginning of every VS and TES.
5413 - Revert "ci/panfrost: disable the rest of these jobs temporarily"
5414 - Revert "ci/panfrost: Disable t860/radeonsi testing while the runners are struggling."
5418 - gallium/dri2: Pass the resource that corresponds to the plane
5445 - aco/ra: Fix print_regs using the wrong constant to check for blocked slots
5448 - aco/ra: Add helper to get a PhysRegInterval for the register demand
5449 - aco: Fix vector::reserve() being called with the wrong size
5466 - lima: add precompile debug flag
5519 - radv: memset the alignment hole in cache_entry to 0
5538 - nir/from_ssa: don't check for interference within the same set