review.tizen.org Git - platform/upstream/mesa.git/log

asahi: Do not use memctx for pools / meta cache

ralloc is not thread-safe, so we can't use dev->memctx for allocating
context-specific things without locking. On top of that, we always
need to explicitly clean up pools anyway since we need to unref the BOs,
so there is no point to using a memctx.

And since pools need to be explicitly cleaned up, the meta cache code
needs explicit cleanup, so add that and drop memctx from there too.

Signed-off-by: Asahi Lina <lina@asahilina.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21348>

asahi: Fix shader key cloning overreads

We call agx_get_shader_variant through with casted inner shader key
types, so it has to make sure to only copy as much of the union as is
actually valid.

Signed-off-by: Asahi Lina <lina@asahilina.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21348>

turnip: fix a null descriptor set dereference

Fixed
dEQP-VK.pipeline.pipeline_library.graphics_library.misc.other.null_descriptor_set_in_monolithic_pipeline.

Fixes: cb3872f2cdc ("tu: Implement VK_EXT_descriptor_buffer")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21328>

turnip: fix a major leak with GPL LTO

This allows dEQP-VK.pipeline.pipeline_library.* to run to finish.

Fixes: e9f5de11d40 ("tu: Initial implementation of VK_EXT_graphics_pipeline_library")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21328>

tu: Don't expose KHR_present_id,KHR_present_wait without KHR_swapchain

VK_KHR_present_id and VK_KHR_present_wait depend on VK_KHR_swapchain
being present, which is not present at least on Android/KGSL.

Fixes:
src/vulkan/util/vk_extensions.h:450: void assert_device_extensions_requirements(
const struct vk_device_extension_table *, const struct vk_instance_extension_table *):
assertion "!device_ext->KHR_present_id || device_ext->KHR_swapchain" failed

Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21345>

anv,hasvk: flush what UNIFORM_READ flushes on SHADER_READ

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/8287
Cc: mesa-stable
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21310>

egl/kopper: Pass ancillary invalidate flush flags down to gallium.

We can just add the flags to the kopper interface, since it's private to
Mesa. This gets us depth/stencil invalidation on swapbuffers, which is
critical for tiler performance.

glmark2-es2 -b texture (windowed) goes from 1650 to 1930 fps on
zink+turnip with ZINK_DEBUG=rp.

Part of #7321 (we're still a little behind freedreno's 2180 fps)

Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21317>

egl/kopper: Use the kopper private interface for swapBuffers.

I'm going to be introducing an extra flag to indicate EGL vs GLX behavior
in the private interface.

Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21317>

egl: Add a note explaining the swapBuffers badness in dri2_x11_copy_buffers().

I tried to drop the swapBuffers path, but it turns out it's being taken by
softpipe/llvmpipe, and the tests are passing. The piglit egl-copy-buffers
test even passes on zink, but you end up with a bad display because of an
un-preserved back buffer.

Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21317>

egl/kopper: Add assert for no kopper in dri2_copy_region.

This shouldn't be doing a swapBuffers, that's not what this function is
supposed to do. But also, we shouldn't be doing this from zink, which the
swap was introduced for, because we don't implement the extension. Cleans
up some strangeness from 3c4be122cc49 ("egl: implement more hooks for
swrast")

Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21317>

Revert "meson: Fix Asahi build on macOS"

This reverts commit 4ca4a05627e40e8212a8ea957f8ef2f697e4e322.

Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Asahi Lina <lina@asahilina.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21185>

meson: don't pass vk wsi args where they don't belong

Only code that cares about Vulkan WSI should get the corresponding
arguments passed. Otherwise, the Vulkan headers might end up including
other headers that we don't have the correct dependencies passed for.
So let's give those a dedicated variable, and only pass that where it's
actually needed.

Fixes: b39958a3a18 ("anv,nir: Move the ANV YCbCr lowering pass to common code")
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/8193
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21185>

vulkan: prefer vulkan_core.h over vulkan.h

If we include vulkan.h, we risk including the WSI bits as well, which we
don't need here. Only trouble can follow from including these where
they're not needed.

So let's include vulkan_core.h in these places instead.

Fixes: b39958a3a18 ("anv,nir: Move the ANV YCbCr lowering pass to common code")
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21185>

gallium: plumb resolve attachments through from frontends -> pipe_framebuffer_state

some drivers may find this useful

Reviewed-by: Adam Jackson <ajax@redhat.com>
Acked-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18695>

radv/ci: add missing expected failures with RADV_PERFTEST=gpl on GFX1100

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21334>

hasvk: use Vulkan runtime's robust buffer access

Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21338>

anv: use Vulkan runtime's robust buffer access

Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21338>

frontends/va: Map VAEncCodedBufferType buffer as PIPE_MAP_READ

VAEncCodedBufferType is used for reading back encoded data.
Mapping it for read instead of write speeds up reading
the data on CPU.

On radeonsi this will result in VRAM copy to staging buffer
in cached GTT, making the CPU read much faster.

Signed-off-by: David Rosca <nowrep@gmail.com>
Reviewed-by: Ruijing Dong <ruijing.dong@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20376>

fix: clover/llvm: replace llvm::None with std::nullopt for LLVM 17+

llvm::None was deprecated and builds started failing with

error: ‘None’ is not a member of ‘llvm’

Instead of using the temporarily available include in ADT which would
add a deprecation warning to the build, directly replace llvm::None with
the recommended std::nullopt

This change takes only effect with LLVM 17 or newer.

Reference: https://github.com/llvm/llvm-project/blob/d4f38ef288c3a4cf2318182c8585a5c7e760877a/llvm/include/llvm/ADT/None.h
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21125>

ir3: Consider dst type in ubo_vec4 to ldc lowering

The dst type could be either 16b or 32b.

Fixes validation failure in dEQP-VK.subgroups.* tests which deal with
16b types.

validation fail: (type_size(instr->cat6.type) <= 16) == !!((instr->dsts[0])->flags & IR3_REG_HALF)
-> for instruction: MESA: info: 0023:0000:000: ldc.offset0.base0 hssa_23 (wrmask=0x3), ssa_1, ssa_22

Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21316>

pvr: Add support for VK_ATTACHMENT_LOAD_OP_LOAD.

Signed-off-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21280>

pvr: Upload spm load programs to device.

The programs are currently unused but will be needed for the spm
background object load op.

Signed-off-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21279>

pvr: Add SPM load usc empty programs

Signed-off-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21279>

iris: Drop iris_cache_flush_for_render

Before dropping this function, handle the two callers of this function:

* The call in iris_blorp.c is redundant. The required cache flushes are
  already handled by the callers of blorp functions. Delete this.

* The call in iris_resolve.c is still providing a benefit because it
  calls iris_emit_buffer_barrier_for internally. Inline the needed
  barrier.

Cc: 23.0 <mesa-stable>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21303>

iris: Flush caches for aux-mode changes more often

Memory accesses can get corrupted when there's a disagreement between:
* the aux-mode of existing cache lines for a surface and
* the aux-usage in that surface's RENDER_SURFACE_STATE object

We have already prevented hardware from seeing this conflict for
rendering operations, but due to how the L3 is shared among multiple
clients in gfx12 (e.g., sampler engine, render engine, etc.), we need to
expand the scope of the existing solution. Now, before any access of a
compressible resource, we make sure to flush the prior aux-mode from the
caches.

The majority of changes here refactor things for use in a new function,
flush_previous_aux_mode. The remaining change calls that function from
within iris_resource_prepare_access.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6558
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7625
Cc: 23.0 <mesa-stable>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21303>

iris: Update comment in iris_cache_flush_for_render

Update the comment to reflect the fact that iris no longer switches
between CCS_E and CCS_D.

Cc: 23.0 <mesa-stable>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21303>

meson: Combine duplicated c_args and cpp_args

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Dylan Baker <dylan@pnwbakers.com>
Reviewed-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19185>

meson: Split c_cpp_args from pre_args

pre_args should not include compiler options

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19185>

meson: drop `TODO: opengl`, it's done

Suggested-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Eric Engestrom <eric@igalia.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19600>

meson: only build glsl when needed

Signed-off-by: Eric Engestrom <eric@igalia.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19600>

meson: only build libglsl_util when needed

Signed-off-by: Eric Engestrom <eric@igalia.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19600>

meson: only build the loader when needed

Signed-off-by: Eric Engestrom <eric@igalia.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19600>

meson: only build mapi when needed

Signed-off-by: Eric Engestrom <eric@igalia.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19600>

meson: move float64_glsl_file one meson.build up

anv uses it.

Signed-off-by: Eric Engestrom <eric@igalia.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19600>

lavapipe: enable linear filtering for depth formats

this seems to work according to cts

Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20859>

intel: Implement Wa_16011448509

"Use 3DSTATE_CONST command for individual shaders instead of
3DSTATE_CONST_ALL COMMAND"

On gen 12.0 platforms, 3DSTATE_CONSTANT_ALL command is not processed
correctly in certain cases.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21301>

zink: block LINEAR filtered blits for zs formats

this is illegal, and the u_blitter path has to be taken to guarantee
enough accuracy that the strictest piglit tests pass

cc: mesa-stable

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21314>

Revert "zink: always use NEAREST for zs blits"

This reverts commit 067545eb9af79414054e4bef3ab48af003758663.

this is good enough for CI but not enough for anholt's piglit tests

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21314>

vk/runtime: turn vk.xml extension requirements into asserts

More specifically, turn
<extension name="VK_KHR_foo" requires="VK_KHR_bar">
into
assert(!ext->KHR_foo || ext->KHR_bar);

Signed-off-by: Eric Engestrom <eric@igalia.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21261>

vk/runtime: keep track of supported instance extensions

Signed-off-by: Eric Engestrom <eric@igalia.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21261>

vk/util: keep track of extension requirements

Signed-off-by: Eric Engestrom <eric@igalia.com>
Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21261>

dzn: Enable KHR_storage_buffer_storage_class

It's required for VK1.1 and Mesa's SPIR-V parser handles it for us.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21261>

intel/fs: bound subgroup invocation read to dispatch size

This is to avoid out of bound register accesses (potentially leading
to hangs) when the dispatch size is smaller than when is reported in
the NIR subgroup_size.

v2: Implement bounding with a mask (since workgroup sizes are powers of 2) (Faith)

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Fixes: 530de844ef4d ("intel,anv,iris,crocus: Drop subgroup size from the shader key")
Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21282>

zink: enable renderpass optimizations by default for selected drivers

CI has been running this for months on turnip, so let's give it some
broader testing

also add a ZINK_DEBUG=norp for testing

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21296>

aco: Don't use vcmpx with DPP.

V_CMPX+DPP returns 0 with reads from disabled lanes, unlike V_CMP+DPP (RDNA3 ISA doc, 7.7)

Fixes: baab6f18c91 ("aco: Optimize branching sequence during SSA elimination.")
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20537>

nir/from_ssa: Use more helpers in resolve_parallel_copies

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21299>

nir: Fix typos in the from-SSA pass comments

Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21299>

nir: Fix merge_set_dump() to compile again

This #if 0'd debug code has been broken since -Werror=vla was added.

Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21299>

nir: Print divergence information for registers as well as SSA defs

This patch causes us to print "con" and "div" for registers as well as
SSA defs. We print it on both register declarations, and destinations.
The latter isn't strictly necessary, but it is handy to be able to see
e.g. a convergent value being assigned to a divergent register without
having to constantly refer back to definitions that might be much
earlier in the program. I originally printed it for sources as well,
but that got to be a bit wordy, so I dropped that.

Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21299>

meson: remove unused USE_FOO_ASM defines

The usage of these defines was removed, so let's remove the definitions
as well.

Reviewed-by: Eric Engestrom <eric@igalia.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21212>

anv/blorp: use existing function to convert the op to a string

Signed-off-by: Rohan Garg <rohan.garg@intel.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21283>

isl: fix some documentation

Signed-off-by: Rohan Garg <rohan.garg@intel.com>
Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21283>

anv: reuse the VK_IMAGE_ASPECT_PLANES_BITS_ANV macro

Signed-off-by: Rohan Garg <rohan.garg@intel.com>
Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21283>

anv: drop unused headers

Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21283>

aco: new 16bit VOP3 opcodes can use opsel

No Foz-DB changes on gfx11.

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20705>

ci/venus: Skip tests risking out of memory issues

These two dEQP tests:
* dEQP-VK.api.object_management.max_concurrent.graphics_pipeline
* dEQP-VK.api.object_management.max_concurrent.compute_pipeline

Require a lot of memory and might kill crosvm or the renderer server because of EOM
issues. Skip them for now.

Acked-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Signed-off-by: Corentin Noël <corentin.noel@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21306>

radv: ignore registering pipeline libaries with SQTT

They aren't executable pipelines and they might not contain all
shader stages.

This fixes a crash when generating RGP captures with GPL.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21235>

zink: remove incorrect trailing comma

A trailing comma in a list is not valid JSON. Let's drop it.

Fixes: f7b2dbb2bdc ("zink: relax bresenhamLines requirement for non-strictLine drivers")
Acked-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21211>

zink: correct companies in requirements

These contributors don't work for Zink itself, but various companies
who's working on Zink. Add the correct ones.

Acked-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21211>

ci: Enable the hasvk Vulkan driver in the fedora-release job

To match the Fedora packaging.

While at it, sort the Vulkan drivers alphabetically.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21236>

ci: Enable i915 Gallium driver in fedora-release job

To match the Fedora packaging.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21236>

ci: Re-enable intel-clc in fedora-release job

Fedora 36 has new enough llvm-spirv-translator.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21236>

ci: Update Fedora image to 36

Fedora 34 is EOL.

With Fedora 37, ccache hangs in a futex syscall in CI for some reason,
so let's go with 36 for now.

v2:
* No need to add -Wno-error=stringop-overflow anymore.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21236>

clover: Reserve vector memory in make_text_section

This isn't strictly required, but it works around
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100366 , and it might
avoid a memory reallocation.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21236>

anv/grl: Use union for reinterpreting integer as float

Fixes strict aliasing violations flagged by GCC 12:

../src/intel/vulkan/grl/include/GRLOCLCompatibility.h: In function ‘float as_float(uint32_t)’:
../src/intel/vulkan/grl/include/GRLOCLCompatibility.h:182:13: warning: dereferencing type-punned pointer will break strict-aliasing rules [-Wstrict-aliasing]
  182 |     return *reinterpret_cast<float*>(&i);
      |             ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
../src/intel/vulkan/grl/include/GRLOCLCompatibility.h: In function ‘float3 as_float3(int3)’:
../src/intel/vulkan/grl/include/GRLOCLCompatibility.h:187:13: warning: dereferencing type-punned pointer will break strict-aliasing rules [-Wstrict-aliasing]
  187 |     return *reinterpret_cast<float3*>(&i3);
      |             ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../src/intel/vulkan/grl/include/GRLOCLCompatibility.h:187:13: warning: dereferencing type-punned pointer will break strict-aliasing rules [-Wstrict-aliasing]
../src/intel/vulkan/grl/include/GRLOCLCompatibility.h: In function ‘float4 as_float4(int4)’:
../src/intel/vulkan/grl/include/GRLOCLCompatibility.h:192:13: warning: dereferencing type-punned pointer will break strict-aliasing rules [-Wstrict-aliasing]
  192 |     return *reinterpret_cast<float4*>(&i4);
      |             ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../src/intel/vulkan/grl/include/GRLOCLCompatibility.h:192:13: warning: dereferencing type-punned pointer will break strict-aliasing rules [-Wstrict-aliasing]

Fixes: 5f948503e40c ("anv: Import GRL")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21236>

zink: set ZINK_DESCRIPTORS=db for radv jobs

there's no testing for this in ci, and radv has a unique heap/memory
setup which catches allocation regressions

Reviewed-by: Martin Roukala <martin.roukala@mupuf.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20968>

anv: enable task redistribution

Disabling is no longer needed after "intel/compiler/mesh: use
slice id of task urb handles in mesh shaders".

This reverts commit 4eaecd79657f0cd99f6536fd55a7a14958d6624b.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7141

Reviewed-by: Caio Oliveira <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21007>

intel/compiler/mesh: use slice id of task urb handles in mesh shaders

When mesh shader is spawned on a different slice than the originating
task shader, then input task urb handle can come from a different
slice, so masking this information off will load data from the current
slice, instead of the one where real data are.

Reviewed-by: Caio Oliveira <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21007>

anv: implement VK_EXT_vertex_input_dynamic_state

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21026>

anv: move 3DSTATE_VERTEX_ELEMENT emission to dynamic path

Prep work for VK_EXT_vertex_input_dynamic_state

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21026>

anv: remove copied information from runtime graphics state

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21026>

anv: fix how unset gl_Viewport & gl_Layer are handled in mesh case

See also: c6f69eea6ac ("anv/pipeline: Properly handle unset gl_Layer and gl_ViewportIndex")

Reviewed-by: Caio Oliveira <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17620>

intel/compiler: replace gl_Layer & gl_ViewportIndex by 0 in fs if ms doesn't write it

Reviewed-by: Caio Oliveira <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17620>

radv: disable DCC for mipmaps on GFX11

It seems broken but can't really figure out why and DCC levels aren't
interleaved on GFX11. Skipping DCC initialization for levels seems to
also fix it but seems safer to disable completely, as a hotfix.

Fixes DCC issues with Hi-Fi Rush, Sonic Frontiers, Hogwarts Legacy
and probably more.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/8230
Cc: mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21186>

ci/etnaviv: Update deqp xfails for gc2000.

More of a mixed bag here than gc7000, but at least the status is stable
in the last 3 nightly runs.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21294>

ci/etnaviv: Drop stale xfails from gc7000.

These have passed in 3/3 of the last nightly runs. Congrats, etnaviv!

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21294>

ci/radv: Add a skip for navi21-llvm for a test that consistently timeouts.

I had seen this in my last round of work, but forgot to git add the skips
file.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21294>

zink: verify compressed format layer count when creating surfaces

this is illegal for some cases, and mesa/st is equipped to handle
per-layer copying here

fixes #8283

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21292>

zink: don't handle mutable init on surface creation with tc enabled

using the cmdbuf during this call is illegal and causes desync, thus
the initialization has to be deferred until the surface is bound

fixes #7579

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21292>

zink: const-ify a surface param

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21292>

zink: break out pipe_surface init for new surface creation

no functional changes

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21292>

zink: account for null surface when trying to retain clears on fb bind

not currently possible but will be soon

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21292>

zink: only try for a fb rebind if fb binds exist in rebind_image()

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21292>

zink: fix indentation of rebind_image()

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21292>

crocus: switch gen4/5 tiling flags to follow suggestions.

Fixes: 6043f66dd332 ("crocus: disable Y tiling for render targets properly.")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21300>

broadcom/ci: mark test as flaky

It passed once here: https://gitlab.freedesktop.org/gdevi/mesa/-/jobs/36317041

Signed-off-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21295>

Revert "freedreno/a5xx: Fix clip_mask"

This reverts commit 2dfebf34874b5365156d254c2c1ba2ecc5262deb.

It causes GPU hangs in piglit tests like
spec@glsl-1.20@execution@clipping@vs-clip-vertex-enables, for reasons I'm
totally unclear on. The commit was not necessary, because the frontend
lowering already handles disabled clip planes by storing 0.0 to the
corresponding clipdist array element in that shader variant. Add a note
to that effect.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21298>

ci/freedreno: Update manual-run xfails for a530.

While I'm having a hard time stabilizing most of the test list on this HW
due to the clip-enable GPU hangs leaking into random other tests, these
have been consistent in the last 4 runs.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21298>

d3d12: Don't clear d3d12_shader_key

The shader key structure is quite large and memsetting it to zero to be
able to create or often simply find an existing shader is responsible
for a large portion of CPU usage during benchmarks.

This change is more surgical about what, when, and how things get
cleared.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21247>

nir: Check sampler_binding is valid when lowering tex shadow

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21247>

d3d12: Add unions to encompass shader key stage vars, use in hashing

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21247>

d3d12: Use varying comparison function for TESS stage key compare

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21247>

util/disk_cache: Handle OS'es without d_type in struct dirent

Needed to build on Solaris

Fixes: f58e6fee745 ("util/disk_cache: delete more cache items in one go when full")
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21110>

zink: always use NEAREST for zs blits

LINEAR is illegal, and swapping filtering at this point still
seems to be conformant given questionable GL spec requirements for
LINEAR zs filtering

cc: mesa-stable

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21203>

radv/rra: Hide deferred accel struct data destruction behind an env var

Keeping around copies of the BVHs in CPU memory can cause issues with
Applications creating a large amount of acceleration structures (Control).
This commit adds back the old path of copying acceleration structures
while still keeping the deferred, possibly more accurate path around.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20700>

radv/rra: Find copy memory index when initializing the trace state

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20700>

anv/hasvk: handle a SAMPLED_READ/STORAGE_READ access flags

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: mesa-stable
Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21277>

zink: free resource objects' views array during destruction

since the array is not ralloc managed, it has be explicitly freed.

cc: mesa-stable

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21285>

zink: stop creating pipeline library cache for non-optimal_key drivers

currently the nin-optimal codepath doesn't use this at all and this just
leaks memory due to not handling non-tcs generated shaders.

Fixes: 487ac6dbd60 ("zink: implement cross-program pipeline library sharing")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21285>