turnip: trim down host size of tu_descriptor_pool

There is no tu_bo array in tu_descriptor_set since bea6290ca09 ("turnip:
device global bo list").

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18152>

turnip: pCreateInfo->pPoolSizes[i] is wordy

pool_size is shorter.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18152>

ir3/spill: Fix extracting from a vector at the end of a block

If this happens then "after" is NULL, so we can't use it to get the
block, and the instruction is never moved at the end so we have to
create the split instructions before creating the collect to make sure
they are in the right order.

This happens when reloading a complex vector value that has been
coalesced at the end of a basic block, which apparently hasn't happened
until a gfxbench5 shader on zink hit this case. This fixes it.

Closes: #7054
Fixes: 613eaac7b53 ("ir3: Initial support for spilling non-shared registers")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18082>

util: Use literal instead call to list_inithead in u_queue.c

This is a more convenient way to struct list_head variable.
And removed the need to call list_inithead in global_init

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18092>

radeonsi/ci: make the running script easy to use

Changes:
- No path parameters are needed to run all tests.
- All test results are stored next to cloned repos instead of /tmp.
  This is a personal preference, though not necessary.

The new setup guide is going to be:
- mesa, piglit, deqp, and glcts directories must be next to each other.
- Add `PATH=$HOME/?/mesa/src/gallium/drivers/radeonsi/ci:$PATH` into
  `.bashrc`. Replace `?` with the proper path.
- Install Rust, which will include its package manager Cargo:
  https://www.rust-lang.org/tools/install
  - The installer will add the Cargo environment into `.bashrc`, which will
    add cargo into `PATH`.
- Restart bash to get the new `PATH`.
- Run: `cargo install deqp-runner`

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
The parameters setting the various paths won't be explained to keep
the guide short.

The path parameters and env vars can be removed if everybody stops using
them after this.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18155>

intel/compiler: Print more details when fs_visitor::validate() fails

Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18150>

wsi/x11: fix memleak in wsi_x11_connection_create()

Fixes: bbdf7e45b15f6345c5e5 ("wsi/x11: Hook up KHR_incremental_present")
Signed-off-by: Eric Engestrom <eric@igalia.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18177>

gallium: Remove util_make_fragment_tex_shader_xrbias

Unused.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17918>

gallium: Inline away util_make_fragment_tex_shader interp_mode

Everything sets it to linear.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17918>

gallium: Inline away util_make_fragment_tex_shader_writemask

Nothing actually uses writemasks, get rid of them.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17918>

tu: Enable bufferDeviceAddressCaptureReplay

Now that we are able to set iova from userspace we could support
the replay feature of BDA.

In order to prevent address clashing between ordinary allocations
and replayable ones we allocate:
- ordinary allocations - from the lowest address;
- replayable allocations - from the highest address.

Passes: dEQP-VK.binding_model.buffer_device_address.*replay*

Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15676>

tu: Switch to userspace iova allocations if kernel supports it

With MSM_INFO_SET_IOVA it's now possible for userspace to manually
set iova instead of relying on kernel.

In preparation to support bufferDeviceAddressCaptureReplay.

Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15676>

Revert "ci: disable unit tests"

As we disabled the glsl tests by meson options always, so
we reverts commit 4dd8c143a9b76b340da868f7b5186e268073fc27.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6897

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17928>

ci: disable glcpp tests for now

these are too flaky to continue running for now

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17928>

meson: add enable-glcpp-tests option

these are too intermittent to be left enabled on CI for now

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17928>

meson: Use different STRACEDIR folder for intel_devinfo_override_test

Fixes
--- stderr ---
strace: Can't fopen 'meson-logs/strace/intel_device_info_override_test_120_/log.10613': No such file or directory

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17928>

anv: Fixes struct anv_device::info is not initialized with struct anv_physical_device

Refactoring the function anv_device_set_physical out, so that it's can be called in unittests

Fixes: 356a60bd6c5 ("anv: Do not duplicate intel_device_info memory in each logical device")
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7092

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: José Roberto de Souza jose.souza@intel.com
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17928>

radv: move lowering the view index to radv_pipeline_link_shaders()

This is a link step because it needs to know if the previous stage is
a mesh shader.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18138>

radv: move lowering patch vertices to radv_pipeline_link_tcs()

This also moves merging the tess info.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18138>

radv: move assigning driver locations to the new linking helpers

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18138>

radv: rework linking shaders in NIR

Replace these ugly loops with something much cleaner and easier to
maintain.

There shouldn't be any functional changes.

No fossils-db changes (tested on NAVI21, VEGA10 and POLARIS10).

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18138>

radv: add radv_remove_point_size() linking helper

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18138>

radv: remove useless parameter from radv_link_shaders()

It can be accessed from the pipeline key instead.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18138>

radv: use NIR_PASS with radv_export_implicit_primitive_id

For consistency.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18138>

radv: remove unused parameter in radv_lower_io_to_mem()

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18138>

radv: move dumping NIR shaders to the previous similar loop

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18138>

radv: move lowering FS intrinsics to radv_postprocess_nir()

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18138>

radv: add a radv_postprocess_nir() helper

This looks cleaner.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18138>

radv: copy the whole radv_shader_info struct for merged stages

We do something similar for shader args and this allows to remove this
ugly hack when postprocessing NIR shaders.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18138>

radv: copy linked I/O info for merged stages in radv_fill_shader_info

To copy info to merged stage at the same place.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18138>

radv: remove useless check when creating the noop FS

If the pipeline type is graphics, VK_PIPELINE_CREATE_LIBRARY_KHR
shouldn't never be set, so this was redundant.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18138>

r600/sfn: Don't use broken idiv lowering

NIR has two implementations of lower_idiv, keyed on the imprecise_32bit_lowering
flag. This flag is misleading: the results when setting this flag "imprecise",
they're completely wrong for some values. Non-constant highp integer division is
rare enough this shouldn't noticeably harm performance of real applications.

Address r600 portion of #6555.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18154>

r300: Set more shadow sampler lowering in precompiles.

This extends the commit 17cea74b8cd3b1a56d923edeb40772b3e8b18ab2
also for rectangular textures. They are otherwise lowered to MOV
dst temp[0].0000 and deadcode elimitation goes crazy.

Fixes shader-db with tesseract shaders from https://gitlab.freedesktop.org/mesa/mesa/-/issues/6771

Signed-off-by: Pavel Ondračka <pavel.ondracka@gmail.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18168>

Update docs/drivers/llvmpipe.rst

to include a link to the LLVM Snapshots for Fedora.

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14537>

vc4: store tex sampler in proper register

When unpacking the texture sample result ensure it is moved to the
proper expected dest register.

This fixes incorrect texturing in Chromium using PixiJS framework.

CC: mesa-stable
Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>
Reviewed-by: Eric Engestrom <eric@igalia.com>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18122>

ci: enable lima farm

Signed-off-by: Erico Nunes <nunes.erico@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18178>

lavapipe: Set ss_dirty in emit_state

Set ss_dirty to false when updating the samplers.

Signed-off-by: Konstantin Seurer <konstantin.seurer@gmail.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18161>

meson: drop duplicate `lib` in lib name

The default lib prefix is `lib`, so adding `lib` to the lib name just
results in `liblibbroadcom_v3d.a` for instance.

Those are internal, build-temporary static libraries that end up being
linked into the real user-visible libraries, so it doesn't really
matter, but we might as well just clean up their names anyway.

Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Juan A. Suarez <jasuarez@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18171>

vulkan/wsi/x11: detect Xwayland via the XWAYLAND extension

This is more robust than checking for XRandR output names.

Signed-off-by: Simon Ser <contact@emersion.fr>
References: https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/954
Reviewed-by: Michel Dänzer <mdaenzer@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17790>

bin/perf-annotate-jit.py: Update to Python 3.

Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12926>

st_pbo/compute: pre-clamp loaded geometry based on coord components

if only some of these components will be used, clamp the rest to enable
better shader optimizing

Acked-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18118>

st_pbo/compute: pre-clamp shader geometry to 16bit

no point doing this in the shader

Acked-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18118>

st_pbo/compute: fix 1D coord dimension by pre-trimming vectors

cc: mesa-stable

Acked-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18118>

st_pbo/compute: fix 1D_ARRAY offsets

these come in as yoffset, but the constant data must use 0 since the sampler
view is created with the offset

cc: mesa-stable

Acked-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18118>

st_pbo/compute: handle download failures with fallback

Acked-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18118>

panfrost: Assert that blend shaders are nontrivial

Even if the driver doesn't *use* trivial blend shaders, building and compiling
blend shaders is expensive. We shouldn't be building blend shaders that should
never be used.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17841>

panfrost: Include mask in replace blend shader name

Helpful to disambiguate blend shaders with different colour masks used for the
same format/replace operation.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17841>

panfrost: Simplify blitter blend shader creation

We don't need blending in the blitter. That means blend shaders are only needed
on Midgard. Simplify accordingly.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17841>

panfrost: Invert no_colour to enabled

This way, blend info that's left zeroed out defaults to disabled. Fixes an
assertion failure in KHR-GLES31.core.draw_buffers_indexed.color_masks due to
the blend CSO max_rt being set unreliably. This is also probably clearer.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17841>

panfrost: Don't compile empty blend shaders

This is silly. They'll be masked out later, but it's still pointless, and adds
noise when grepping BIFROST_MESA_DEBUG=shaders,internal for "pan_blend" when
debugging why an app is using blend shaders.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17841>

panfrost: Avoid blend shader when not blending

On Midgard, we need a "blend" shader even if blending is disabled, if the format
isn't blendable. This is inefficient. Bifrost solves this by decoupling the
format conversion from the blending, allowing opaque (unblended) output to any
format without a blend shader or fragment key.

Unfortunately, our blend code is from the Midgard era -- I wrote an early
version of nir_lower_blend when I was still in high school! -- so we've been
using blend shaders for opaque output even on Bifrost. Whoops!

In SuperTuxKart, reduces blend shader calls by 30%, translating to a 15%
reduction in i-cache misses.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17841>

panfrost: Promote blend shader outputs 8->16-bit

..on Bifrost and later, where the conversion hardware makes this reasonable.
This saves us from inserting a pile of conversions in the compiler to lower away
the 8-bit input/output. This also generates substantially better code.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17841>

panfrost: Don't saturate in Bifrost blend shaders

It's unnecessary since the hardware already does the conversion for us.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17841>

panfrost: Set blit output variable types correctly

The type of the output variable will propagate through the store_output
intrinsic's src_type field to the BLEND instruction's register format field. On
Valhall, the register format for a BLEND comes from the instruction -- the
register format specified in the conversion descriptor (used on Bifrost) is
ignored. That means it has to match.

Previously, we always used a blend shader for integer rendering. Since blend
shaders ignore the register format of the BLEND instruction, that masked this
issue. That also means we don't need to backport this.

Will prevent a regression from the following commit.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17841>

panfrost: Handle untyped_color_outputs on Bifrost

For untyped_color_outputs, we need to ignore the type of the colour output in
the shader and instead use the type from the format. We have all the information
to do this at blend descriptor pack time, but not at shader compile time. This
means we need a (somewhat expensive) fixup in this edge case to ingest
NIR-to-TGSI. This will prevent a regression from the rest of the series.

Although the register_format field is also present on Valhall blend descriptors,
it is ignored so we don't need the fixup there.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17841>

pan/bi: Handle info.fs.untyped_color_outputs on Valhall

Colour outputs in TGSI are untyped so we have to ignore the register format on
the store_output. Luckily, Valhall gives us the auto32 escape hatch to deal with
it, and Bifrost doesn't care anyway. This will avoid regressions from native int
output without corrupting the compiler for GLSL and SPIR-V shaders that are
well-typed.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17841>

pan/bi: Don't allow ATEST to take a temporary

Clause scheduler edition of db2bdc1dc37 ("pan/bi: Require ATEST coverage mask
input in R60"). ATEST wants to read r60, which can't work if its input isn't
even in a register.

When per-sample shading isn't in use, prevents regressions in:

   KHR-GLES31.core.sample_variables.mask.*

These tests previously passed because per-sample shading was forced. It's
not clear whether the bug addressed in this patch is possible to hit "in the
wild", i.e. without the optimizations in this series that allow us to use
per-pixel shading in more cases.

No shader-db changes.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17841>

pan/bi: Don't reorder image loads across stores

Fixes flaking in
dEQP-GLES31.functional.image_load_store.cube.qualifiers.volatile_r32i due to
image reads being moved past a BARRIER.

To make this more robust/optimal, we probably need scheduling information
(coherent/volatile/etc) added to instructions like ACO does. That's left for a
future extension, for now I just want the test to stop flaking.

Fixes: 569e5dc7450 ("pan/bi: Schedule for pressure pre-RA")
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17841>

asahi: Fix warning building for macOS 12.0

Unnecessary rename that breaks forward compatibility... but Apple says
this is just NULL. Do the simpler thing. Note that the argument is a
mach_port_t, which is a natural_t == uint32_t in userspace... even
though it's a pointer in the kernel. Although Apple's docs claim that
kIOMasterPortDefault is NULL, it's really just 0.

../src/asahi/lib/agx_device.c:290:35: warning: 'kIOMasterPortDefault' is deprecated: first deprecated in macOS 12.0 [-Wdeprecated-declarations]
      IOServiceGetMatchingService(kIOMasterPortDefault, matching);
                                  ^~~~~~~~~~~~~~~~~~~~
                                  kIOMainPortDefault

Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18121>

tu: Assert that if draw state is enabled it has valid iova

GPU faults at draw state loading completely could completely bring
down the board. (Happened on HDK888)

Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18149>

r300: merge MOVs with MULs or ADDs in merge channels

Shader-db stats with RV530:
total instructions in shared programs: 166499 -> 164362 (-1.28%)
instructions in affected programs: 80056 -> 77919 (-2.67%)
total temps in shared programs: 21658 -> 21565 (-0.43%)
temps in affected programs: 1780 -> 1687 (-5.22%)

Signed-off-by: Pavel Ondračka <pavel.ondracka@gmail.com>
Reviewed-by: Filip Gawin <filip@gawin.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17560>

r300: merge MOVs into ADD using the 0 swizzle

Shader-db stats with RV530:
total instructions in shared programs: 169509 -> 166013 (-2.06%)
instructions in affected programs: 99126 -> 95630 (-3.53%)
total presub in shared programs: 10975 -> 10758 (-1.98%)
presub in affected programs: 744 -> 527 (-29.17%)
total temps in shared programs: 21722 -> 21649 (-0.34%)
temps in affected programs: 1350 -> 1277 (-5.41%)

Signed-off-by: Pavel Ondračka <pavel.ondracka@gmail.com>
Reviewed-by: Filip Gawin <filip@gawin.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17560>

r300: don't merge w channel in fragment shaders

Skip the merge if one of the instructions writes just w channel
and we are compiling a fragment shader. We can pair-schedule it together
later anyway and it will also give the scheduler a bit more flexibility.

Shader-db stats with RV530:
total instructions in shared programs: 169522 -> 169509 (<.01%)
instructions in affected programs: 14170 -> 14157 (-0.09%)
total temps in shared programs: 21712 -> 21722 (0.05%)
temps in affected programs: 324 -> 334 (3.09%)

Signed-off-by: Pavel Ondračka <pavel.ondracka@gmail.com>
Reviewed-by: Filip Gawin <filip@gawin.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17560>

r300: generalize the merge_movs pass

To allow a simple extension with more merging combinations in the
later commits.

Signed-off-by: Pavel Ondračka <pavel.ondracka@gmail.com>
Reviewed-by: Filip Gawin <filip@gawin.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17560>

r300: run dataflow optimizations in separate loops

Constant folding first, than copy propagate simple movs, after that
we run the merge movs pass and finally peephole. The important part
is to do the copy propagate for the whole program before running merge
movs.

We no longer check the return from merge_movs so convert it to void.

Shader-db changes with RV530:
total instructions in shared programs: 155361 -> 154787 (-0.37%)
instructions in affected programs: 67920 -> 67346 (-0.85%)
total temps in shared programs: 20836 -> 20773 (-0.30%)
temps in affected programs: 711 -> 648 (-8.86%)
total presub in shared programs: 8226 -> 8202 (-0.29%)
presub in affected programs: 223 -> 199 (-10.76%)
total temps in shared programs: 20836 -> 20773 (-0.30%)
temps in affected programs: 711 -> 648 (-8.86%)

Signed-off-by: Pavel Ondračka <pavel.ondracka@gmail.com>
Reviewed-by: Filip Gawin <filip@gawin.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17560>

r300: check for identical saturate mode when merging MOVs

Signed-off-by: Pavel Ondračka <pavel.ondracka@gmail.com>
Reviewed-by: Filip Gawin <filip@gawin.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17560>

r300: fix negate mask computation when merging movs

The main problem here is we can have a negate bit set for an unused
channel, so we can't just OR together the negates when channel merging.
Right now the bug is hidden because how we run the pass order, but
that will change in a later commit. Add some helpers for merging of the
negates, they will be also used more in a later commits. As a bonus
construct the new source separatelly and only rewrite the original
instructions after checking that the final swizzle is valid.

Signed-off-by: Pavel Ondračka <pavel.ondracka@gmail.com>
Reviewed-by: Filip Gawin <filip@gawin.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17560>

r300: allow constant swizzles with inline constant

This will prevent a regression in the number of inlined constants in
a later commit. Constructs like 4.000000 (0x48).w110 works just fine.

There is a small behavioral change. We would previously allow positive
and negative same-value contants to be produced, e.g.,
4.000000 (0x48).w-w__ and this would be later split into some extra
movs in the dataflow swizzle pass. We now explicitly check that the
final swizzle is valid while inlining. So there is a minor decrease
in inlined constants and in the total instructions.

total lits in shared programs: 4328 -> 4194 (-3.10%)
lits in affected programs: 554 -> 420 (-24.19%)
total instructions in shared programs: 155488 -> 155361 (-0.08%)
instructions in affected programs: 5707 -> 5580 (-2.23%)

Additonally, a fix for pair translation is needed since the constant
inlining can now produce swizzles like this: 4.000000 (0x48).w-0-0-_
so we have to teach pair translation to also ignore the sign for zero
swizzle.

Signed-off-by: Pavel Ondračka <pavel.ondracka@gmail.com>
Reviewed-by: Filip Gawin <filip@gawin.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17560>

microsoft/clc: Handle a null clc logger

Fixes: 55b3980e ("microsoft/compiler: Add dxil_logger type and parameter to nir_to_dxil")
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7076
Reviewed-by: Michael Tang <tangm@microsoft.com>
Tested-by: Prodea Alexandru-Liviu <liviuprodea@yahoo.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18127>

vulkan/device_select_wayland: fix a memory leak with DRM device handling

Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18076>

radv/ci: Remove host build related fails

!17028 removed acceleration structure host build support.

Signed-off-by: Konstantin Seurer <konstantin.seurer@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18142>

aco: Improve SCC nocompare optimization when SCC is clobbered.

When SCC is clobbered between s_cmp and its operand's writer,
the current optimization that eliminates s_cmp won't kick in.

However, when s_cmp is the only user of its operand temporary,
it is possible to "pull down" the instruction that wrote the operand.

Fossil DB stats on Navi 21:

Totals from 63302 (46.92% of 134906) affected shaders:
CodeSize: 176689272 -> 176418332 (-0.15%)
Instrs: 33552237 -> 33484502 (-0.20%)
Latency: 205847485 -> 205816205 (-0.02%); split: -0.02%, +0.00%
InvThroughput: 34321285 -> 34319908 (-0.00%); split: -0.00%, +0.00%

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16266>

aco: Support s_cselect_b64 in SCC no-compare optimization.

This was simply left out by accident when I wrote this.

Fossil DB stats on Navi 21:

Totals from 70165 (52.01% of 134906) affected shaders:
CodeSize: 246375656 -> 245814396 (-0.23%)
Instrs: 46519773 -> 46379458 (-0.30%)
Latency: 385159303 -> 385089261 (-0.02%); split: -0.02%, +0.00%
InvThroughput: 66490172 -> 66487867 (-0.00%); split: -0.00%, +0.00%

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16266>

aco/optimizer_postRA: Don't try to optimize dead instructions.

Also delete them when they are already dead in process_instruction().

No Fossil DB changes.

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16266>

ci: implement stress testing jobs in ci_run_n_monitor.py

Can be triggered by `--stress`.

Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18014>

ci: GraphQL was already implemented in ci_run_n_monitor.py

Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18014>

ci: traces: update freedreno and broadcom to brotli compressed traces

Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18156>

zink: add gfx pipeline lookup shortcut

if the final hash was the same as the last-used hash for this program,
it should be safe to reuse the last pipeline object in the presence of
dynamic vertex input since the number of permutations is low

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18135>

zink: track which stages in gfx programs are using inlined uniforms

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18135>

zink: split resource tracking into two sets for buffer/texture

this will make the hash lookups slightly faster to cut down overhead

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18135>

zink: make batch internal api static again

hooray

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18135>

zink: merge all the batch state sets onto the struct

less memory fragmentation

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18135>

zink: use PIPE_CAP_RGB_OVERRIDE_DST_ALPHA_BLEND

this eliminates the bespoke handling I wrote to do the same thing a while ago

cc: mesa-stable

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18135>

zink: add is_buffer param to zink_batch_resource_usage_set

this allows constants to be specified, reducing register pressure during
descriptor binds

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18135>

zink: move zink_batch_resource_usage_set to be static inline

this has to go in zink_resource.h to avoid infinite include dependency
conflicts

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18135>

zink: remove c++ guard from zink_resource.h

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18135>

zink: add more c++ guards to headers

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18135>

zink: make zink_bo.h compile in c++

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18135>

zink: constify shader module variable

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18135>

zink: pass shader stage as param for gfx module update

less register pressure

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18135>

zink: use dynarray instead of list for internal shader cache

this is less likely to obliterate cpu caches

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18135>

zink: change u_foreach_bit to regular for loop in zink_gfx_program_update

too hot to be using this

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18135>

zink: make program cache values more intelligible

instead of magic constants

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18135>

zink: cache dynamic render vs renderpass pipelines separately

removes comparisons from hash lookup

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18135>

zink: use stage comparison template to ignore vertices_per_patch

no need to compare if this won't be used

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18135>

zink: add templates for pipeline stage comparison

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18135>

zink: use a list for free batch states

this ensures "fairer" reuse and avoids having any states sitting idle
for too long with resources attached

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18135>

zink: split out first-time shader module creation

this simplifies the code a bit and makes it more flexible

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18135>

zink: split out gfx shader module creation from cache searching

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18135>

zink: split out nonseamless/inline shader module info calculation

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18135>