v3d/compiler: fix QPU scheduler TMU sequence shuffling

The QPU scheduler allows to move certain TMU instructions around and
since we enabled pipelining, we need to protect against the case where
doing this might break a TMU sequence. For example, this test:

dEQP-VK.rasterization.line_continuity.line-strip

Was generating this VIR:

mov tmud, t187
mov.pushz null, t176
mov.ifa tmua, t9
nop null; wrtmuc (img[0].p0 | 0x0)
mov tmut, t185
mov tmud, t180
mov.ifa tmusf, t183
nop null; thrsw

where we have a general TMU access (tmud,tmua) followed by an image
access (wrtmuc, tmut, tmud, tmusf), which the QPU scheduler was turning
into:

nop            ; nop               ; ldunifrf.rf22 (0xffffff00 / -nan)
nop            ; nop               ; wrtmuc (img[0].p0 | 0x0)
nop            ; nop               ; ldtmu.r2
add  r0, r2, 1 ; nop               ; ldtmu.r3
nop            ; nop               ; ldtmu.r4
nop            ; mov  tmud, r0
nop            ; mov.ifa  tmua, rf15
nop            ; mov  tmut, r4     ; thrsw
nop            ; mov  tmud, rf22
nop            ; mov.ifa  tmusf, r3

where it allowed the wrtmuc to move up and before the general TMU access,
leading to an incorrect TMU sequence.

Fix this by flagging TMUA writes (which are the sequence terminators for
general TMU accessess) as writing new TMU configuration, like we do for all
other TMU sequence terminators for textures and images.

Fixes: 197090a3fc ('broadcom/compiler: implement pipelining for general TMU operations')

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8954>

zink: add a texture barrier hook

this just flushes all the pending batches, which is enough since we
do our batch work synchronously and everything is awful

will be reworked when this is no longer the case

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8948>

zink: handle nir_texop_texture_samples

this is just a SpvOpImageQuerySamples

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8947>

zink: rework viewport handling

viewport can be transformed at draw time based on clip_halfz, so we actually
want to do the conversion from gallium->vulkan at that time

as a nice bonus, we get to delete a struct member

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8946>

turnip: enable inheritedQueries

Passes relevant CTS tests.

Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8824>

radv: fix memory leaks if a submission fails

If the device isn't lost, the application might want to re-submit.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8931>

zink: handle vote intrinsics in ntv

these are super simple and straightforward

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8950>

zink: add spirv_builder wrapper for vote intrinsics

these are all going to use the same mechanics, so we can simplify a bit

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8950>

ci: Add trailing slash to path for documentation preview

Fixes the documentation preview link on the MR Overview tab. Found by
comparing with the same thing in weston.

(It still appears only after all jobs in the pipeline have at least
started once though)

Fixes: ae8f9584f4c9 "CI: always expose docs artifacts"
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8927>

v3dv: support for depthBiasClamp

Gets tests like the following working:
dEQP-VK.dynamic_state.rs_state.depth_bias_clamp

Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8928>

aco: fix adjust_vertex_fetch_alpha

These offsets were wrong and didn't match the old behaviour.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Fixes: e8220e106bf ("aco: optimize AC_FETCH_FORMAT_SNORM alpha adjust")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8935>

radv: stop allocating useless ESGS scratch BO on GFX10+

On GFX10+, the ESGS ring is allocated in LDS. This saves one BO
when geometry shaders are used, and possibly two BOs if nothing
else (tess, scratch, etc) is used.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8926>

radv: fix double free when creating a fence failed

It's already freed in radv_destroy_fence().

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8938>

gallium: Document behavior of more lowering pipe caps.

Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8601>

v3d: Stop advertising support for flat shading.

The GL frontend can lower this weird GL feature away for us.  This should
fix redeclaration of the gl_Color/SecondaryColor as centroid, since that
case had been missed in the !flat special case here.

Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8601>

v3d: Stop advertising support for PIPE_CAP_*_COLOR_CLAMPED.

The GL frontend can lower away this deprecated GL feature for us.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8601>

v3d: Stop advertising support for PIPE_CAP_TWO_SIDED_COLOR.

The GL frontend can lower away this deprecated GL feature for us.

Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8601>

v3d: Clean up vestiges of alpha test lowering.

We had an unnecessary case in our uniforms upload switch statement, since
we no longer advertise the cap.

Fixes: 8ad931808e2d ("v3d: do not report alpha-test as supported")
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8601>

vc4: Stop advertising support for VS color clamping.

This is special GL behavior that the frontend is now able to handle for
us.

Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8601>

vc4: Stop advertising support for PIPE_CAP_TWO_SIDED_COLOR.

The GL frontend is perfectly able to handle this for us.

Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8601>

vc4: Remove vestiges of alpha test lowering.

We stopped reporting the alpha test screen cap, and stopped using the
value in the key, so now shrink the key.  This gets another switch case
out of the hot uniforms upload path.

Fixes: 1404b8b1e5b6 ("vc4: do not report alpha-test as supported")
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8601>

mesa/st: Assume that the default variant is always first in the list.

Since we always precompile now, we can remove some checks for presence of
the first variant or for it being the wrong key.

Acked-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8601>

mesa/st: Always precompile the first shader variant.

Every driver wants the chance to see the default variant of the shader in
the hopes of avoiding draw-time shader compiles, even if they have some
NOS compiled into the shader by st/mesa.  Before, every driver had to
avoid st/mesa lowering options to get this behavior, even if they would
like to have the shared lowering code, and some drivers were just missing
the behavior because nobody had debugged that it was missing yet.

Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8601>

freedreno/a5xx: Don't forget to count our custom blits against queries.

Noticed when comparing the a5xx and a6xx paths.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8873>

freedreno: Remove FD_STAGE_* in favor of a "disable_all" flag.

The only paths are "prepare queries for rendering" and "finish all
queries", everything else is handled by set_active_query_state().

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8873>

freedreno: Force updating active queries on batch reordering.

The state of queries for the batch we've just switched to may not reflect
the current active flag, so make sure that we update at the start of the
next draw.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8873>

ci/virgl: Fix GLES31 testing on desktop GL.

The version overrides leaked into the virglrenderer server, breaking
compiling of all sorts of shaders.  We already get GLES3.1 when running
virgl on llvmpipe GL without doing any overrides.

Reviewed-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8854>

features: mark off GL 4.3 for zink

\o/

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8941>

zink: GLSL 430

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8941>

zink: enable compute

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8781>

zink: export compute-specific shader/compute caps

this doesn't enable the compute cap itself, it just handles all the other
caps

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8781>

zink: add launch_grid pipe_context hook for compute handling

we're reusing update_descriptors, so this is trivial

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8781>

zink: flush gfx/compute batches when the other pipeline needs resource sync

if compute or gfx use a resource which is already in use by the other pipeline,
we need to ensure tha we flush that pipeline now in order to have commands execute
in the expected order

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8781>

zink: handle descriptor set updates for compute operations

this is almost identical, the difference being that we need to explicitly
use the compute batch/program in a few places

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8781>

zink: handle memory barriers for compute batch

this is a bit different since the regular functions cycle the gfx batch
array, but the compute batch isn't in an array

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8781>

zink: setup compute batch and add handling

this is separate to the gfx batch array,
so we need separate struct members in zink_context for it

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8781>

zink: break out descriptor updating into separate function

this is a massive block of code that we need to reuse for compute shaders
because we absolutely don't want to be copying it around

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8781>

zink: add compute programs and pipelines

these are similar to the gfx variants, but mostly smaller

some things are implemented as placeholders for now to make the code
similar to gfx programs/pipelines and thus easier to read/review

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8781>

zink: rename pipeline_cache_entry -> gfx_pipeline_cache_entry

prep for compute pipelines

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8781>

zink: explicitly get shader stage from shader during binding setup in draw

the iterator matches the shader stage currently, but this won't always be the
case

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8781>

zink: make allocate_descriptor_set() take more params instead of a gfx_program

this makes the function reusable between different types of programs

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8781>

zink: refactor batch creation

by moving this into its own function, we can reuse it for creating a compute
batch, which will not be part of the gfx batch array

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8781>

zink: take a pipe_reference param in zink_batch_reference_program

this lets us reuse the function for gfx and compute programs without much
churn, since in C we can use a pointer to the first member of a struct and
a pointer to the struct interchangeably

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8781>

zink: ignore compute batch when starting/ending batches

compute batches don't have queries

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8781>

zink: handle nir_intrinsic_memory_barrier_shared in ntv

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8781>

zink: add handling for shared atomic ops in ntv

these are mostly the same as ssbo except they use the shared block variable

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8781>

zink: implement shared load/store nir ops in ntv

these access the data in the shared block variable at an offset

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8781>

zink: handle COMPUTE glsl variables

we can just use existing helpers for these

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8781>

zink: handle COMPUTE setup in ntv

addressing mode, shared block, and execution modes all need to be handled
here

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8781>

zink: handle COMPUTE bindings in compiler/ntv

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8781>

zink: add spirv_builder function for emitting a 3word literal exec mode

used by compute

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8781>

zink: wait on compute batch when necessary during transfer map

now that we have the capability to wait on specific batches, we can
check whether we need to sync with the compute queue here before trying
to map resources

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8781>

zink: make zink_batch_reference_resource_rw return usage info

now this provides some info regarding which batches need to be flushed,
allowing gfx batches and compute batches to share resources

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8781>

zink: make get_resource_usage() public

also cover compute batch id (still not yet used)

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8781>

zink: bump resource usage flags to allow 5 batches

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8781>

zink: add defines for compute batch and gfx batch count

this will be used for handling batch usage with compute

also bump some struct field sizes based on this

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8781>

intel/stub: plug some gaps in our ioctl faking

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8934>

drm-shim: report support for timeline semaphores

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8934>

anv: only signal wsi fence BO on last command buffer

I'm pretty sure this doesn't fix anything because the WSI code only
use a single VkSubmitInfo, but better be safe.

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Fixes: ccb7d606f1a293 ("anv: Use submit-time implicit sync instead of allocate-time")
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8934>

anv: break up internal queueing function

We want more control over batching things up when possible to minimize
the round trips to i915. This breaking up of one big internal function
into helpers will be useful later.

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8934>

anv: add a comment describing has_relocs field

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8934>

vulkan/lavapipe: add missing VKAPI_ATTR/CALL

Acked-by: Jose Fonseca <jfonsec@vmware.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8823>

aco: remove loop to flag loop blocks as WQM

This is no longer necessary.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8446>

aco: rewrite setting of Exact_Branch

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8446>

aco: do not flag all blocks WQM to ensure we enter all nested loops in WQM

This should no longer be necessary since the mark_block_wqm() we use to
flag break conditions as WQM now adds block to the worklist. With them
added to the worklist, get_block_needs() will add WQM to block_needs.

Adding WQM to block_needs here without adding the block to the worklist
(like we do here) can cause issues because it does not ensure that the
predecessors' branches are in WQM (needed for it to be possible to
transition to WQM in the block). This happened in an Overwatch shader.

No fossil-db changes.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Fixes: 661922f6ac9 ("aco: add block to worklist in mark_block_wqm()")
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/4066
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8446>

ci: Append build image tag to LAVA tag used for minio path

This automatically ensures the LAVA kernel / device-tree / rootfs files
are updated along with the build image, so any relevant changes in the
latter are reflected in the former.

v2:
* Use the same definition of DISTRIBUTION_TAG in .lava-test:amd64 as in
  kernel+rootfs_amd64
v3:
* Extend comment about $MESA_BASE_TAG in .lava-test:amd64 (Eric Anholt)

Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7988>

ci: Incorporate base image tag into dependent image tags

Using new .incorporate-base-tag+templates-commit & .set-image-base-tag
templates.

This automatically ensures dependent images get rebuilt along with the
base image, no more need to manually bump the dependent image tags.

Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7988>

ci: Add and use .set-image template to construct docker image name

Instead of open-coding everywhere.

Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7988>

ci: Move FDO_DISTRIBUTION_TAG assignment to template

This will be useful in more places later. Should have no functional
effect for now.

Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7988>

ci: Use MESA_IMAGE_TAG everywhere

Instead of just "TAG".

Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7988>

ci: Use MESA_ namespace for image variables in Windows jobs

Instead of a separate WINDOWS_ namespace.

Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7988>

ci: Set GALLIVM_PERF=nopt,no_filter_hacks for llvmpipe-gles2

Gives us 48 more passed tests, and possibly slightly shorter job times.

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Acked-by: Daniel Stone <daniels@collabora.com>
Acked-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4986>

ci: Set GALLIVM_PERF=no_filter_hacks for llvmpipe-piglit-quick_gl

Gives us 13 more passed tests.

(GALLIVM_PERF=nopt breaks some tests here)

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Acked-by: Daniel Stone <daniels@collabora.com>
Acked-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4986>

ci: Set GALLIVM_PERF=no_filter_hacks for llvmpipe-piglit-quick_shader

Gives us 1 more passed test.

(GALLIVM_PERF=nopt breaks some tests here)

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Acked-by: Daniel Stone <daniels@collabora.com>
Acked-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4986>

ci: Disable two radeonsi jobs

The machine to which these boards are connected to is having trouble
keeping up when the rootfs are expanded. This is causing jobs to time
out and fail.

So as a mitigation measure reduce the load by disabling two of these
jobs until the root problem is solved.

Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8930>

radeon/vcn: enable dynamic dpb Tier2 support

On VCN3 with codec AV1 and VP9

Signed-off-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: James Zhu <James.Zhu@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8916>

radeon/vcn: implement dynamic dpb Tier2 support

Fill up the t2 message buffers based on reference lists, so to
avoid unnecessary allocation of the buffers.

Signed-off-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: James Zhu <James.Zhu@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8916>

radeon/vcn: add dynamic dpb Tier2 message buffer interface

Tier2 will dynamically manange the dpb buffers, so that can
save even more VRAMs

Signed-off-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: James Zhu <James.Zhu@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8916>

radv: fix printing the debug option names

radv_dump_enabled_options() doesn't like holes. While we are at it,
use ull everywhere.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8925>

aco: optimize AC_FETCH_FORMAT_SNORM alpha adjust

This looks like it was copied from LLVM, which didn't have a fmax
intrinsic.

fossil-db (GFX8):
Totals from 43 (0.03% of 140385) affected shaders:
CodeSize: 49660 -> 49488 (-0.35%)
Instrs: 10434 -> 10348 (-0.82%)
Cycles: 41736 -> 41392 (-0.82%)
VMEM: 13793 -> 13719 (-0.54%)

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8918>

radv: Fix vram override with fully visible VRAM.

Fixes: cf2eebdf4f1 ("radv,gallium: Add driconf option to reduce advertised VRAM size.")
Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8915>

gallium/tgsi-to-nir: Take property NUM_CLIPDIST_ENABLED into account

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8919>

gallium/tgsi_to_nir: Handle SAMPLE_MASK output in FS

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8919>

radv: use the global BO list from the winsys

We had two different implements for the global BO list, one in RADV
and one in the winsys. This will also allow to make more BOs resident.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8868>

radv/winsys: enable the global BO list unconditionally

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8868>

radv/winsys: add the resident BOs to the list of BOs at submit time

The number of resident BOs is currently 0.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8868>

radv/winsys: add buffer_make_resident() to the API

To make a BO resident (or to evict one).

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8868>

radv/winsys: set use_global_list to avoid adding a BO twice

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8868>

radv/winsys: move the debug_all_bos check outside of the add/del helpers

The add/del helpers will be used to implement the global BO list
directly in the winsys.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8868>

pan/bi: Don't check liveness unless the index is valid

Otherwise we will read out-of-bounds when the index is ~0.

Fixes: b8f042c9bb0 ("pan/bi: Dead code eliminate per-channel")
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8923>

v3dv: initialize render_fd at the top of physical_device_init

This fixes an uninitialized variable warning.

Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8902>

gallium/u_tests: test no-op fragment shader instead of NULL fragment shader

radeonsi stopped supporting NULL fragment shaders. This makes the test pass.

Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8906>

tgsi_to_nir: translate FBFETCH

Tested by u_tests.

Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8906>

tgsi_to_nir: translate SAMPLEID

Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8906>

st/mesa: do scissored clears on depth/stencil as well when supported

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Tested-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8905>

zink: support nir_intrinsic_memory_barrier_buffer

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8884>

panfrost: Fix clear color packing for 12-byte formats

Make the case for 6 bytes the same as for 8 while we're at it.

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8886>

radv: Update JSON generator if Windows

Use vulkan_radeon.dll, and apply current working directory.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8911>

nv50,nvc0: add scissored clear support

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8901>

nv50: add PIPE_CAP_NIR_IMAGES_AS_DEREF to unsupported list

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8901>