rusticl: advertize cl_khr_spirv_no_integer_wrap_decoration

It's supported with all drivers already.

Signed-off-by: Karol Herbst <git@karolherbst.de>
Reviewed-by: Nora Allen <blackcatgames@protonmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23708>

docs/cl: fix whitespace issues and add missing entries

Signed-off-by: Karol Herbst <git@karolherbst.de>
Reviewed-by: Nora Allen <blackcatgames@protonmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23708>

rusticl: advertize cl_khr_extended_versioning

We already implemented the OpenCL 3.0 core bits, but the extension also
has a `CL_DEVICE_OPENCL_C_NUMERIC_VERSION_KHR` query.

Signed-off-by: Karol Herbst <git@karolherbst.de>
Reviewed-by: Nora Allen <blackcatgames@protonmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23707>

rusticl/version: use cl_version instead of cl_uint and provide a From impl

Signed-off-by: Karol Herbst <git@karolherbst.de>
Reviewed-by: Nora Allen <blackcatgames@protonmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23707>

rusticl/device: sort cl_device_info queries

Signed-off-by: Karol Herbst <git@karolherbst.de>
Reviewed-by: Nora Allen <blackcatgames@protonmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23707>

rusticl/device: add intel usm queries DPCPP cares about

We don't implement them and we don't advertise the extension, but DPCPP
queries them regardless. We ultimately plan to implement the intel USM
extension. However until we do, just return 0 for those queries.

Signed-off-by: Karol Herbst <git@karolherbst.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23705>

gallium/hud: append results to files instead of overwriting them

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23706>

anv: do not explode on 32 bit builds

Fixes: 930e862af7a ("anv: add shaders for copying query results")
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/9213

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23686>

ci: implement farms handling trough files inside .ci-farms

 - Disable farm with `git mv .ci-farm{,-disabled}/$farm_name`.
 - Re-enable farm with `git mv .ci-farm{-disabled,}/$farm_name`.

NEVER MIX FARM MAINTENANCE WITH ANY OTHER CHANGE IN THE SAME MERGE REQUEST!

Reviewed-by: Martin Roukala (né Peres) <martin.roukala@mupuf.org>
Acked-by: Daniel Stone <daniels@collabora.com>
Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23629>

ci/crocus: depend on state of the Anholt farm

Reviewed-by: Martin Roukala (né Peres) <martin.roukala@mupuf.org>
Acked-by: Daniel Stone <daniels@collabora.com>
Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23629>

ci/amd: hide vaapi job dependent on Collabora farm when it's down

Fixes: ae9c67d773a7 ("ci/amd: add radeonsi-raven-va-full job to cover all VA-API tests")

Reviewed-by: Martin Roukala (né Peres) <martin.roukala@mupuf.org>
Acked-by: Daniel Stone <daniels@collabora.com>
Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23629>

ci/etnaviv: if farm is down, we expect no manual jobs can be triggered

Reviewed-by: Martin Roukala (né Peres) <martin.roukala@mupuf.org>
Acked-by: Daniel Stone <daniels@collabora.com>
Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23629>

ci/windows: move microsoft farm rules

No functional changes intended, align with other farms.

Reviewed-by: Martin Roukala (né Peres) <martin.roukala@mupuf.org>
Acked-by: Daniel Stone <daniels@collabora.com>
Acked-by: Jesse Natalie <jenatali@microsoft.com>
Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23629>

svga: lower images before ntt

ntt requires lowered images, so call gl_nir_lower_images first before
passing the shader to ntt.

Fixes piglit failures spec@glsl-4.30@execution@built-in-functions@cs*

Fixes: 0ac95418048 ("gallium: Drop PIPE_SHADER_CAP_PREFERRED_IR")

Reviewed-by: Neha Bhende <bhenden@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23697>

tgsi: remove unused functions and structs

These are no longer used and can be removed

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Christian Gmeiner <cgmeiner@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23637>

ci: enforce formatting for RADV & ACO

Signed-off-by: Eric Engestrom <eric@igalia.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23253>

aco: reformat according to its .clang-format

Signed-off-by: Eric Engestrom <eric@igalia.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23253>

radv: reformat according to its .clang-format

Signed-off-by: Eric Engestrom <eric@igalia.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23253>

radv,aco: tweaks to get clang-format to print nicer code

Signed-off-by: Eric Engestrom <eric@igalia.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23253>

clang-format: add explanation for anyone reading .clang-format-include

Signed-off-by: Eric Engestrom <eric@igalia.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23253>

egl: return correct error for EGL_KHR_image_pixmap

EGL_KHR_image_pixmap says:
> If <target> is EGL_NATIVE_PIXMAP_KHR, and <ctx> is not EGL_NO_CONTEXT,
> the error EGL_BAD_PARAMETER is generated.

Reported-by: Rohan Garg <rohan@garg.io>
Closes: https://gitlab.freedesktop.org/mesa/mesa/issues/166
Fixes: a1c4a8a3c855d52fbfef "egl: Add support for EGL_KHR_image."
Signed-off-by: Eric Engestrom <eric@engestrom.ch>
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/454>

iris: Support MTL modifier MC_CCS

Signed-off-by: Jianxun Zhang <jianxun.zhang@intel.com>
Reviewed-by: Nanley Chery <nanley.g.chery@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20327>

intel/isl: Add MTL MC CCS modifier into modifier info

Signed-off-by: Jianxun Zhang <jianxun.zhang@intel.com>
Reviewed-by: Nanley Chery <nanley.g.chery@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20327>

iris: Support I915_FORMAT_MOD_4_TILED_MTL_RC_CCS_CC modifier

Signed-off-by: Jianxun Zhang <jianxun.zhang@intel.com>
Reviewed-by: Nanley Chery <nanley.g.chery@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20327>

intel/isl: Add MTL RC CCS CC modifier into modifier info

Signed-off-by: Jianxun Zhang <jianxun.zhang@intel.com>
Reviewed-by: Nanley Chery <nanley.g.chery@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20327>

iris: Support I915_FORMAT_MOD_4_TILED_MTL_RC_CCS modifier

Signed-off-by: Jianxun Zhang <jianxun.zhang@intel.com>
Reviewed-by: Nanley Chery <nanley.g.chery@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20327>

intel/isl: Add MTL RC CCS modifier into modifier info

Signed-off-by: Jianxun Zhang <jianxun.zhang@intel.com>
Reviewed-by: Nanley Chery <nanley.g.chery@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20327>

include/uapi: Update drm_fourcc.h from drm kernel

From:
ba57b9b11f78530146f02b776854b2b6b6d344a4
Merge: 959294e47953 24335848e543
Author: Dave Airlie <airlied@redhat.com>
Date:   Fri Jun 9 16:43:35 2023 +1000

    Merge tag 'drm-intel-gt-next-2023-06-08' of
    git://anongit.freedesktop.org/drm/drm-intel into drm-next

    UAPI Changes:

    - I915_GEM_CREATE_EXT_SET_PAT for Mesa on Meteorlake.

Signed-off-by: Jianxun Zhang <jianxun.zhang@intel.com>
Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Nanley Chery <nanley.g.chery@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20327>

iris: Fix memory alignment when importing dmabuf (GFX12.5)

Fix the failure in modifier tests on MTL:

../src/intel/common/intel_aux_map.c:609: intel_aux_map_add_mapping:
Assertion `(address & get_page_mask(main_page_size)) == 0' failed.

Signed-off-by: Jianxun Zhang <jianxun.zhang@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Nanley Chery <nanley.g.chery@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20327>

dri: Replace usage of boolean/TRUE/FALSE with bool/true/false

First do the find/replace in src/gallium/frontends/dri/ folder,
then fixes the compiling errors manually:
Because of this, the prototype of functions in include/GL/internal/dri_interface.h
are changed

Acked-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23582>

radv/rt: Enable RT pipelines on GFX10_3+ excluding vangogh

Reviewed-by: Friedrich Vock <friedrich.vock@gmx.de>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23677>

Revert "radv: Enable ray tracing pipelines by default"

This partially reverts commit 374bd4e1bef25ba2aeef6016ac35b41df07f3a99.

Reviewed-by: Friedrich Vock <friedrich.vock@gmx.de>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23677>

radv: Add the BOs of all shaders in a RT pipeline

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23681>

aco/opcodes: delete wrong comment copy pasted from NIR

Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21882>

aco: remove v_cvt_pkrtz_f16_f32_e64 when it's actually VOP2

More consistent with how we handle all the opcodes which were moved to VOP3.

Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21882>

aco/opcodes: move v_cndmask_b32 back to the VOP2 list

Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21882>

aco/spill: Use arena allocator for spills

Gives a marginal speedup.

Co-authored-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12000>

aco/spill: Use arena allocator for next use distances

Speedup of aco::spill(): ~50%

Co-authored-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12000>

nouveau: eliminate busy waiting on fences

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19543>

anv: convert most pc in genX_cmd_buffer to use pc helper

Some are left, batch_set_preemption does not have devinfo pointer
and IndirectStatePointersDisable does not have corresponding ANV bit.

Signed-off-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/23583>

anv: change most pipe controls in gfx8_cmd_buffer to use pc helper

One using a flag (PSDSyncEnable) that has no corresponding ANV bit.

Signed-off-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/23583>

anv: change pipe control in indirect draw gen to use pc helper

Signed-off-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/23583>

anv: change pipe controls in genX_gpu_memcpy to use pc helper

Signed-off-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/23583>

anv: change pipe control in genX_pipeline to use pc helper

Signed-off-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/23583>

anv: change pipe controls in genX_state to use pc helper

Signed-off-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/23583>

anv: convert genX_query pipe controls to use pc helper

Signed-off-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/23583>

anv: implement invalidate part of emit_apply_pipe_flushes with helper

Signed-off-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/23583>

anv: implement flush part of emit_apply_pipe_flushes with helper

Signed-off-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/23583>

anv: wrap pipe control emission to a set of helper functions

This makes it possible to have HW specific rules and WA's implemented
in a central place. Also all pipe controls will get anv_debug_dump_pc.

Signed-off-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/23583>

virgl: Fix stack overflow in virgl_bind_sampler_states

Triggered on hosts with more than 32 samplers.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/9207
Fixes: 63c4c559 ("virgl: obtain supported number of shader sampler views from host")
Signed-off-by: Tim Pambor <tp@osasysteme.de>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23678>

radv: add dgc_emit_index_buffer()

For emitting VK_INDIRECT_COMMANDS_TOKEN_TYPE_INDEX_BUFFER_NV.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23584>

radv: add dgc_emit_draw_indexed() helper

For emitting VK_INDIRECT_COMMANDS_TOKEN_TYPE_DRAW_INDEXED_NV.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23584>

radv: add dgc_emit_draw() helper

For emitting VK_INDIRECT_COMMANDS_TOKEN_TYPE_DRAW_NV.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23584>

radv: add dgc_emit_vertex_buffer() helper

For emitting VK_INDIRECT_COMMANDS_TOKEN_TYPE_VERTEX_BUFFER_NV.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23584>

radv: add dgc_emit_push_constant() helper

For emitting VK_INDIRECT_COMMANDS_TOKEN_TYPE_PUSH_CONSTANT_NV.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23584>

radv: add dgc_emit_state() helper

For emitting VK_INDIRECT_COMMANDS_TOKEN_TYPE_STATE_FLAGS_NV.
The scissor workaround for GFX9 is only needed if the state is emitted,
so move it there as well.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23584>

radv: remove unused radv_dgc_token struct

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23584>

ac/radeonsi: add av1 defaults header file from radeonsi

This just moves this header file so radv can use it.

Acked-by: Leo Liu <leo.liu@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23640>

lavapipe: correctly update shader object per-stage push constant sizes

Fixes: 8b3022c9180 ("lavapipe: implement EXT_shader_object")

Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23663>

lavapipe: fix shader binary binding with mesh shaders

Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23663>

lavapipe: more fixes for sample shading

this fixes the case where a draw without sample shading precedes
a draw with sample shading without changes to the sample mask

Fixes: cc9e9580531 ("lavapipe: fix DS3 min sample setting")

Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23663>

aux/trace: add methods for mesh shaders

Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23663>

clc: static assert that clc_optional_features has no padding

Signed-off-by: Karol Herbst <git@karolherbst.de>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23685>

meson: Remove reference to removed SWR driver

Fixes: e2de00876a7 ("gallium/swr: Remove common code and build options")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23679>

bin/pick-ui: use asyncio.new_event_loop

Instead of .get_event_loop, which is deprecated when there isn't a
running loop (like in our case).

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23472>

bin/pick: Add support for adding notes on patches

This is pretty useful for keeping track of why a patch isn't landed, or
who I'm waiting on feedback from.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23472>

bin/pick: use lineboxes to make the UI clearer

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23472>

bin/pick: fix issue where None for nomination_type could fail

We have an assumption it's never None, so use a special value in the
Enum instead.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23472>

rusticl/spirv: Key optional clc features when caching.

Sadly I can't use serde yet, so I have to do this nonsense.

Signed-off-by: Karol Herbst <git@karolherbst.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23532>

clc: add commment to clc_optional_features to ensure no padding exists

Hopefully this is good enough.

Signed-off-by: Karol Herbst <git@karolherbst.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23532>

docs: upgrade bootstrap to 5.3.0

The release version of Bootstrap 5.3.0 is out, let's upgrade from the
alpha-version we were using.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23423>

nv50/ir: resolve -Woverloaded-virtual=1 warnings

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23656>

nv50/ir: use override

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23656>

zink: more anv ci flakes

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23683>

anv: avoid requiring ordered memory planes for explicit import

The spec does not have such requirement, but anv requires it for
validating the offset. However, for DRM_FORMAT_YVU420, chroma channels
can be swapped upon import to match B/R channel order of
VK_FORMAT_G8_B8_R8_3PLANE_420_UNORM.

This fixes some sw codec path in Instagram when interop with gpu.

v2: fix image memory requirement for re-ordered explicit import

Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Emma Anholt <emma@anholt.net> (v1)
Reviewed-by: Matt Tuner <mattst88@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23643>

asahi: Use bitfield_extract for texture lowering

This makes descriptor crawls a lot easier to read, which is good because more
are coming.

Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23351>

agx: Implement bitfieldExtract natively

We have a bfeil instruction which mostly maps to the GLSL thing, so use it with
the appropriate lowering.

Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23351>

nir/builder: Add ubitfield_extract_imm helper

We have a ubfe_imm helper that creates ubfe ops. Not all drivers support ubfe,
however, as it requires SM5 semantics. A few drivers support oly
ubitfield_extract. They should still get the convenience of an _imm helper, so
add a symmetric helper.

It might be nice to unify these helpers into a single helper that asserts its
inputs do not overflow (such that the two ops become equivalent) and emits
either ubfe or ubitfield_extract depending on the underlying driver. That is
left for future work as it's unclear exactly what naming/semantics we want.

Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23351>

pvr: add missing frag to geom dependency for jobs targetting same render target

Seen as a firmware assert when using a debug build of the firmware
and tested against:

dEQP-VK.pipeline.monolithic.render_to_image.core.1d_array.huge.width_layers.r8g8b8a8_unorm_d16_unorm

Signed-off-by: SoroushIMG <soroush.kashani@imgtec.com>
Acked-by: James Glanville <james.glanville@imgtec.com>
Reported-by: Luigi Santivetti <luigi.santivetti@imgtec.com>
Reviewed-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23651>

microsoft/clc: Add unreachable() to fix 'may be unitialized' warning

  In function ‘lower_load_kernel_input’,
      inlined from ‘clc_nir_lower_kernel_input_loads’ at ../src/microsoft/clc/clc_nir.c:205:28:
  ../src/microsoft/clc/clc_nir.c:169:7: warning: ‘base_type’ may be used uninitialized [-Wmaybe-uninitialized]
    169 |       glsl_vector_type(base_type, nir_dest_num_components(intr->dest));
        |       ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
  ../src/microsoft/clc/clc_nir.c: In function ‘clc_nir_lower_kernel_input_loads’:
  ../src/microsoft/clc/clc_nir.c:151:24: note: ‘base_type’ was declared here
    151 |    enum glsl_base_type base_type;
        |                        ^~~~~~~~~

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23666>

r300: remove unused opcodes from r300_tgsi_to_rc

Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Filip Gawin <filip.gawin@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23642>

r300: remove unused LIT lowering

Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Filip Gawin <filip.gawin@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23642>

r300: remove unused ROUND lowering

Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Filip Gawin <filip.gawin@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23642>

r300: remove unused DST lowering

Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Filip Gawin <filip.gawin@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23642>

r300: remove unused POW lowering

Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Filip Gawin <filip.gawin@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23642>

r300: remove unused FLR lowering

Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Filip Gawin <filip.gawin@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23642>

r300: move CEIL lowering to NIR

Also remove unused backend CEIL lowering.

Single regressed gnome-shell shader due to fceil followed by f2i32
where before nir_lower_int_to_float would recognize that we already
have integer and emit mov instead of trunc for the f2i32. We can
clean this up easily once we move ntt to the backend.

Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Filip Gawin <filip.gawin@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23642>

r300: remove unused SSG lowering

Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Filip Gawin <filip.gawin@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23642>

r300: remove unused SIN/COS lowering

Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Filip Gawin <filip.gawin@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23642>

radv/rt: Stop forcing wave32 by setting compute_subgroup_size

We end up reporting the wrong subgroup size this way.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23572>

iris: Drop GPGPU Tex Invalidate restriction for TGL+

According to the HW docs, TGL+ no longer requires that a CS stall be
added to a texture cache invalidate done in the compute pipeline.

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18725>

iris: Drop the RT flush for PIPE_BARRIER_TEXTURE

The render target flush would have been needed if it was possible to:

1) pollute the render cache and write to the data port in one draw
   call.

2) perform a subsequent operation that assumed the render cache was
   up-to-date.

However, this is not possible for the two glMemoryBarrier barrier bits
that get translated to this pipe barrier:

* GL_TEXTURE_FETCH_BARRIER_BIT is only used for sampling operations.
  It's possible to pollute the render cache and data cache with writes
  to a texture in one draw call (1). However, the GL spec states that
  apps cannot assume that any existing render caches are up-to-date for
  sampling the written locations immediately afterwards. Apps are
  required to use glTextureBarrier before the sampling operation, so
  requirement #2 is not satisfied.

* GL_PIXEL_BUFFER_BARRIER_BIT could be used for a PBO upload (2), but
  it's not possible to pollute the render cache and data cache with a
  PBO access in one draw call. PBOs cannot be bound to framebuffers
  for rendering, so requirement #1 is not satisfied.

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> (v1)
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18725>

Revert "iris: Add missed tile flush flag"

This reverts commit 0523607ebb108d8c90bbda9c6564b66a0a6250e6.

The issue that commit worked around seems to have been fixed as of
commit 1c8b4940ebf ("iris: Emit flushes for push constant source
buffers"). I could no longer reproduce it from that point onward with
this revert applied.

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18725>

intel/blorp: Avoid 32bpc fast clear sampling issue

For 32bpc formats, the ICL+ sampler fetches the raw clear color dwords
used for rendering instead of the converted pixel dwords typically used
for sampling. The CLEAR_COLOR struct page documents this for 128bpp
formats, but not for 32bpp and 64bpp formats.

In blorp_copy, map R11G11B10_FLOAT to R8G8B8A8_UINT instead of R32_UINT.
This will cause the sampler to fetch the clear color pixel, allowing
drivers to keep clear color support enabled during copies.

If iris is forced to convert blits to copies, this patch fixes the
following test on gfx12:

  dEQP-GLES3.functional.fbo.color.repeated_clear.blit.rbo.r11f_g11f_b10f

At the moment, both iris and anv won't hit this issue outside of
blorp_copy. This is due to the read/write access restrictions they
currently place on texture views that reinterpret the surface format.

Cc: mesa-stable
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/8964
Reviewed-by: Jianxun Zhang <jianxun.zhang@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23604>

nir: add missed nir_cmp_imm-helpers

Seems I missed these in my previous round, let's fix them up now!

Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23461>

nir: add and use nir_imod_imm

Just a short-hand, really. Makes the code a bit easier to read.

Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23461>

nir: add and use nir_fdiv_imm

Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23461>

nir: use nir_imm_{true,false}

Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23461>

nir: isub -> iadd_imm

Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23461>

nir: use more imm-helpers

Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23461>