lavapipe: VK_EXT_dynamic_rendering_unused_attachments

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23297>

radv: advertise VK_EXT_dynamic_rendering_unused_attachments

It's a no-op and it's passing CTS.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23292>

vulkan: Update XML and headers to 1.3.251

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23292>

va/surface : Add Nv12 support for PRIME_2 imports

This patch will fix the resource allocation failed for NV12 buffer
format.

Signed-off-by: SureshGuttula <suresh.guttula@amd.com>
Reviewed-by: Leo Liu <leo.liu@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23175>

docs: document clang-format and how to use it

Signed-off-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23269>

ci: use meson to run clang-format

This will allow us to easily exclude vendored code when we enforce the
formatting of more components later on.

Signed-off-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23269>

meson: enable the clang-format target

.clang-format needs to exist in the root of the project for the target
to be generated, but since we don't have a global config it's a dummy,
empty file.

.clang-format-include lists the files (files! not folders) that are to
be formatted.

.clang-format-ignore lists the files to exclude, even if they are in the
include list above. Useful for vendored code for instance.

See https://mesonbuild.com/Code-formatting.html

Signed-off-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23269>

ci: color the diff for clang-format

Suggested-by: David Heidelberg <david.heidelberg@collabora.com>
Signed-off-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23269>

ci/zink+radv: document recent regressions

Signed-off-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23274>

ci/zink+radv: mark all spec@arb_copy_image@arb_copy_image-targets* as flaky after getting a bunch more of them

Signed-off-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23274>

ci/zink+radv: fix flakes definition

Signed-off-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23274>

v3d/v3dv/ci: adjust job fractions

So far we have 12 jobs for v3d-gl (OpenGL/ES and piglit), 1 job for
v3d-traces, and 10 jobs for v3dv-vulkan, but we only have 21 rpi4
devices for testing.

So let's reduce from 12 to 10 jobs in v3d-gl, so all jobs can run
simultaneously.

Also, as the ideal goal is that each job doesn't take more than 15
minutes, let's increase a little bit the fraction for v3dv, and include
a fraction for v3d-gl as well, so all jobs are ideally under the time
limit.

Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>
Reviewed-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23285>

radv/ci: skip tests that timeout since Vulkan loader 1.3

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23261>

radv/ci: update list of expected failures since Vulkan loader 1.3

Updating Debian to 12 also updated Vulkan loader 1.3 and these are
fixed now.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23261>

vc4: add specific stencil blit path

This implementation reinterprets the stencil data as a RGBA8888 texture.

Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23136>

vc4: allow tile-based blit for Z/S

Besides blitting color-based buffer, we can use the tile buffers to blit
also depth and stencil buffers.

This also fixes several piglit tests.

Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23136>

vc4: call blit paths in chain

Let each of the different blit paths to decide if they need to do the
blit operation based on the blit mask, and update the mask once the
operation is done.

This allows to call all the different versions without needing to check
if they success or not.

Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23136>

vc4: set blit mask correctly

Drivers expect either color or stencil/depth blit to be done, but not
both at the same time.

Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23136>

v3d: enable NIR compact arrays

Now that it is exposing GLSL 1.30, and we can read clipdistance arrays
in the fragment shader, let's enable this capability.

It fixes
`spec@glsl-1.30@execution@clipping@fs-clip-distance-interpolated,Crash`.

Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23232>

radv: rework configuring VGT_SHADER_STAGES_EN

For shader objects.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22991>

radv: Make shader related destruction happen before hw_ctx.

radv_destroy_shader_upload_queue waits for a semaphore, which will in turn
call query_reset_status on hw_ctx that will fail due to being already
destroyed.

Fix radv/amdgpu: amdgpu_cs_query_reset_state2 failed. (-9) spam in the logs
with RADV_PERFTEST=dmashaders.

Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23276>

amd: fix 64-bit integer color image clears

Fixes recent CTS dEQP-VK.api.image_clearing.*r64* since the Vulkan
specification has been clarified.

Only VK_FORMAT_R64_{UINT,SINT} are supported.

Cc: mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23250>

vulkan: complete the usage flags for video layouts

Otherwise Vulkan decoding with FFmpeg crashes on anv.

Reviewed-by: Dave Airlie <airlied@redhat.com>
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23184>

anv: Fix calculation of guardband clipping region.

The existing guardband region calculation was mixing up x/y_min with
x/y_max in cmd_buffer_emit_viewport(), causing the calculated viewport
area to always be an empty region.  Luckily intel_calculate_guardband_size()
returns a non-empty but bogus guardband region in that case, so this
doesn't seem to have led to conformance regressions, but the
off-center guardbands could potentially impact performance in
geometry-heavy rendering.

Fixes: 893fa30afed10394f ("anv: Include scissors in viewport calculations")
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23174>

gitlab: add template for merge requests

Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22078>

bin/ci: mention requirements.txt

Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22078>

intel: Sync xe_drm.h

My second attempt at a minimally invasive reshuffle of the uAPI,
this time only forcibly aligning structures to 32-bits or 64-bits
depending on the members which follow, so that 64-bit userspace
is identical to the current unmerged kernel module, and the 32-bit
compat userspace aligns with that, and functions rather than
crashing.

Should work just fine with the current drm-xe-next Git tree, which
is rebased on 6.3.0-1, with a few extra changes, as of this commit.

Based on commit 2cd469458fcc24c5f345ad39721a1aedaf70ec0f ("drm/xe: Add explicit padding to uAPI definition")

Signed-off-by: Christopher Snowhill <kode54@gmail.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22652>

radeonsi: re-enable fp16_rtz for compute blits to fix PBO tests on gfx11

Fixes: bcdde99675d27 - radeonsi: don't convert to fp16 in the compute blit if not testing

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23216>

radeonsi: update test results and flakes

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23216>

radeonsi: reduce the supported compute grid size

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23216>

radeonsi/gfx11: use DISABLE_FOR_AUTO_INDEX to disable non-indexed prim restart

DISABLE_FOR_AUTO_INDEX disables primitive restart for non-indexed draws.
We set it in the preamble first, so that non-indexed draws can completely
ignore the primitive restart state.

There is a little bit of duplication that's needed to enclose the primitive
restart code in "if (index_size)" for indexed draws.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23216>

radeonsi: move emitting draws states out of si_emit_all_states

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23216>

radeonsi: don't enable WGP_MODE because of high cost of workgroup mem coherency

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23216>

radeonsi: remove the gl_SampleMask FS output if MSAA is disabled

It's better to remove the output than what the previous code did,
which only unset MASK_EXPORT_ENABLE.

Reviewed-by: Qiang Yu <yuq825@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23216>

radeonsi: cosmetic changes in si_shader.h

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23216>

radeonsi: reorder code in si_texture_create_object as preparation for the future

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23216>

radeonsi/gfx11: fix alpha-to-coverage with blending

Bug: https://gitlab.freedesktop.org/mesa/mesa/-/issues/8222

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23216>

ac/llvm: use LLVM 0/1 constants from ac_llvm_context instead of LLVMConstInt

Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23216>

ac/llvm: clean up visit_load_local_invocation_index and visit_load_subgroup_id

Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23216>

ac/llvm: clean up translation of nir_intrinsic_load_invocation_id

Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23216>

ac/surface: move determing ADDR_FMT_* into a helper function

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23216>

ac/surface: move CB format translation helpers here

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23216>

ac/gpu_info: give has_msaa_sample_loc_bug a more accurate name

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23216>

ci/windows: Update WARP to 1.0.7

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23267>

dxil: Don't run opt_remove_phis before lower_returns

nir_lower_returns will run nir_opt_remove_phis as part of its work

Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22913>

radv: Don't run opt_remove_phis before lower_returns

nir_lower_returns will run nir_opt_remove_phis as part of its work

Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Friedrich Vock <friedrich.vock@gmx.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22913>

nir: Add undef phi srcs when adding successors

Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22913>

nir_lower_returns: Optimize phis before beginning the pass

This adds a little extra work, since now dominance is computed and
blocks that don't just have then-return or else-return are looked at.
However it means that nir_lower_returns can now keep phis up to date
by inserting undefs without causing some phis to become non-trivial.

This ends up obviating a couple of tests for lower_returns.

Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22913>

panfrost: rename *.cc files to *.cpp

This extension caused them to be missed by clang-format.

Suggested-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Signed-off-by: Eric Engestrom <eric@igalia.com>
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23264>

panfrost: fix formatting of a couple of files that were missed

Signed-off-by: Eric Engestrom <eric@igalia.com>
Acked-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23264>

ci: show diff when clang-format check fails

Signed-off-by: Eric Engestrom <eric@igalia.com>
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23258>

ci: split clang-format list of folders for easier maintenance

Signed-off-by: Eric Engestrom <eric@igalia.com>
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23258>

mesa: Implement GL_CLEAR_TEXTURE flag

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7455
Signed-off-by: Mykhailo Skorokhodov <mykhailo.skorokhodov@globallogic.com>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21682>

wsi/display: drop unused parameters from local functions

These are all local functions, so it's trivial to add back any parameter
should they become useful.

Signed-off-by: Eric Engestrom <eric@engestrom.ch>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20566>

ci: uprev virglrenderer to include changes needed for GL 4.6 support

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23196>

features: Update virgl features

Depending on host features we now support OpenGL 4.6

Signed-of-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23196>

virgl: enable ARB_gl_spirv

This is handled in the guest and can just be enabled.

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23196>

virgl: signal support for group vote and draw parameters

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23196>

virgl: Submit drawid_offset if is not zero

This is needed in the case of multi-draw to evaluate the right gl_DrawID
in the shader.

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23196>

virgl: Fix IB upload when a start >0 is given

We don't need to add an offset in the buffer, because we submit
the offset where the data was written to to the host. The
correction of this offset is also not needed and results in draw
errors.

Fixes: 0cf5d1f22620d67659bbd632a2400c3a6956a011
  gallium: remove PIPE_CAP_INFO_START_WITH_USER_INDICES and fix all drivers

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23196>

anv: opportunistically align VMA to 2Mb

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Tested-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23198>

anv: further reduce pool alignments

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Tested-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23198>

intel: reduce minimum memory alignment on Gfx12.5

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Tested-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23198>

anv: update aux-tt alignment requirements for MTL

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Tested-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23198>

iris: add a comment about aux-tt alignment requirements

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Tested-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23198>

intel: add alignment helper for aux map

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Tested-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23198>

zink: don't propagate psiz in quads emulation gs

this is pointless

haha.

cc: mesa-stable

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23240>

zink: don't remove psiz from linked shaders if the consumer reads it

it's valid for psiz to be piped through successive stages, so don't
delete it unnecessarily

cc: mesa-stable

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23240>

meson: simplify another "any of" check

Signed-off-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22491>

ci/intel: reuse iris_file_list instead of copying its definition

Signed-off-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22456>

bin/ci: fix mistakenly hardcoded repo name in get_gitlab_project()

It didn't matter so far because all callers always pass "mesa".

Signed-off-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23223>

ci/amd: only define AMDGPU_GPU_ID for the duration of the call

This avoids the risk of accidentally forgetting to overwrite it again in
a later call.

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/23249>

ci/amd: move AMD-specific LD_PRELOAD to AMD config

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/23249>

ci: uprev vkd3d-proton to 2.9

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23162>

radv: fix copying 2D to 3D images

CTS is testing 2D to 3D image copies but the checks are incomplete and
we used to only copy the first slice.

This should fix
dEQP-GLES31.functional.copy_image.non_compressed.*.texture2d_array_to_texture3d
with ANGLE.

Cc: mesa-stable
Suggested-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23231>

zink/ci: skip arb_texture_buffer_object@texture-buffer-size-clamp* with RADV

These tests can run OOM and might hang.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23248>

ci: Allow zink-radv jobs to be manual when uprev piglit

It is required for the automatic uprev of piglit in mesa to have the
"^zink-radv-.*-valve" jobs in the pipeline that checks an uprev candidate. So,
when there is a change in the file where the revision of piglit is set, those
jobs will be listed as manual.

Signed-off-by: Sergi Blanch Torne <sergi.blanch.torne@collabora.com>
Reviewed-by: Martin Roukala <martin.roukala@mupuf.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21193>

frontends/va: add default intra idr period

Use default value in case h264->intra_idr_period is not set by app.
Also to avoid dividing by 0 for later calculations in case it's not set.

Signed-off-by: Boyuan Zhang <boyuan.zhang@amd.com>
Acked-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: Ruijing Dong <ruijing.dong@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22803>

ci_run_n_monitor: add ability to specify the pipeline to use, instead of auto-detecting it

The auto-detection code currently looks for a repo called "mesa" in the
current user's fork (ie. the user providing the api token), which is great for
the common use case, but sometimes needs to be able to be overridden, such as
when running a pipeline in another fork than one's own, when working with
someone else in their fork.

Signed-off-by: Eric Engestrom <eric@igalia.com>
Acked-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23230>

util: move pot functions to use existing macros

Signed-off-by: Rohan Garg <rohan.garg@intel.com>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20153>

util: revert back to ALIGN since it moved to util

Signed-off-by: Rohan Garg <rohan.garg@intel.com>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20153>

util: migrate alignment functions and macros to use ALIGN_POT

Signed-off-by: Rohan Garg <rohan.garg@intel.com>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20153>

util: fix ROUND_DOWN_TO alignment type

Alignments can only be unsigned. Cast alignment to uint64_t to keep MSVC
happy.

Signed-off-by: Rohan Garg <rohan.garg@intel.com>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20153>

freedreno: set alignment to next POT

Signed-off-by: Rohan Garg <rohan.garg@intel.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20153>

ac/surface: make sure alignment is a POT

Signed-off-by: Rohan Garg <rohan@garg.io>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20153>

iris: correctly set alignment to next power of two for struct size

We're currently aligning the offset to the size of the data structure
itself when the upload manager actually expects a POT. Ideally this
would be the next POT that's greater than the size of the structure.

Fixes: c24a574e6c78 ("iris: Don't allocate a BO per query object")

Signed-off-by: Rohan Garg <rohan.garg@intel.com>
Reviewed-by: Sagar Ghuge <sagar.ghuge@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20153>

docs/relnotes: add sha256sum for 23.1.1

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23241>

docs: add release notes for 23.1.1

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23241>

docs: update calendar for 23.1.1

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23241>

wsi/win32: Handle acquiring an image while one is already acquired

In the case where the app is attempting to acquire both (since the
min/max num images is 2) swapchain images, we need to wait on the
fence for the next image in the chain.

Reviewed-by: Giancarlo Devich <gdevich@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23222>

v3dv: fix align() computation for pixel formats with non-POT block sizes

Fixes hundreds of dEQP-VK.api.copy_and_blit.* tests when including the
assert that the alignment in align() is valid, as added in !20153.

Fixes: 3ba839bf734f37f50dfc ("v3dv: align compressed image regions to block size")
Signed-off-by: Eric Engestrom <eric@igalia.com>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23224>

radv: Stop using radv_get_int_debug_option

We can use debug_get_num_option instead. Bool options are changed to use
debug_get_bool_option.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23146>

nir: Add intrinsics for multisampling on AGX

sample_mask_agx maps to the AGX instruction used to write out a sample mask.

api_sample_mask_agx is a system value that returns the value of glSampleMask
(or its Vulkan equivalent), used to lower glSampleMask (etc).
This is distinct from sample_mask_in, which we map to the hardware thing and
AND with this as a lowering.

sample_positions_agx is a system value returning the sample positions in a
packed fixed-point format matching the hardware register, used to lower
gl_SamplePositions.

Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Asahi Lina <lina@asahilina.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23040>

ci/zink: document new zink-radv-navi10-valve failures

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23234>

aco/tests: add fix_derivs_in_divergent_cf tests

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22636>

aco/tests: improve performance of declaration parsing

Unlike \S, \w only matches characters which are valid in identifiers. This
seems to be much faster, especially for longer identifier names.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22636>

radv: use fix_derivs_in_divergent_cf

fossil-db (navi21):
Totals from 3256 (2.40% of 135636) affected shaders:
MaxWaves: 65430 -> 64366 (-1.63%)
Instrs: 3517336 -> 3517724 (+0.01%); split: -0.12%, +0.13%
CodeSize: 18963788 -> 18946904 (-0.09%); split: -0.14%, +0.05%
VGPRs: 172464 -> 175872 (+1.98%); split: -0.02%, +2.00%
Latency: 33643792 -> 33643179 (-0.00%); split: -0.12%, +0.12%
InvThroughput: 5912965 -> 5934404 (+0.36%); split: -0.09%, +0.46%
VClause: 60268 -> 60275 (+0.01%); split: -0.12%, +0.13%
SClause: 125227 -> 125241 (+0.01%); split: -0.09%, +0.10%
Copies: 253452 -> 254638 (+0.47%); split: -1.29%, +1.76%
Branches: 100951 -> 100953 (+0.00%); split: -0.00%, +0.00%
PreSGPRs: 186403 -> 185641 (-0.41%)
PreVGPRs: 153751 -> 156915 (+2.06%); split: -0.06%, +2.12%

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22636>

aco: implement texture samples with strict WQM coordinates

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22636>

aco: implement strict_wqm_coord_amd

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22636>

aco: add MIMG_instruction::strict_wqm

This lets us use linear VGPRs for part of the texture sample's address.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22636>