pan/bi: Expose unit tested scheduler predicates

I want to move the tests to their own executable to integrate better
with meson.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12109>

pan/bi: Remove dated ASSERTED properties

This was used when we were only unit testing. Now that we have an actual
scheduler, it's pure noise.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12109>

pan/bi: Inline away bi_must_last

Totally trivial.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12109>

freedreno/ir3: Add build id to the disassembler test

This is required (at least for me on x86) to get the tool to pass it's
own test, otherwise it fails the build_id assertion.

Fixes: 1462b00391711ff86350e802c1a3e5075a3a1632
       ("freedreno/ir3: Add a unit test for our disassembler.")

Acked-by: Rob Clark <robclark@freedesktop.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12084>

zink: never use staging buffer for unsynchronized buffer maps

this is ultra broken, do not attempt

Fixes: 6bfbce0447b ("zink: rework buffer mapping")

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12111>

mesa/main: Check for fbo attachments when importing EGL images to textures

Fixes an assert when binding an fbo with a texture bound to one of its attachments,
if the texture was updated with an EGL image after it was bound.

Cc: mesa-stable
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11998>

iris: Split iris_upload_shader in two

Now the part that uploads the shader and the part that finishes the
creation of the shader are separated.  Each now has a more reasonable
number of parameters.

Suggested-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11229>

iris: Enable threaded shader compilation

There are a couple minor things that can be improved:

1. Eliminate (or reduce) the dynamic allocation of the
threaded_compile_job.

2. For apps like shader-db, improve the case where nr_threads=0.  Right
now this adds thread switching and mutex overhead.

3. Other performance improvements?  iris_uncompiled_shader::variants has
some special properties that make it ripe for replacement with a
lockless list.  Without gathering some data, it's hard to guess what
impact that could have.

v2: Fix whitespace and formatting issues.  Noticed by Ken.
s/threaded_compile_job/iris_threaded_compile_job/g.  Suggested by Ken.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11229>

iris: add sync_compile option

This can be useful to simplify debugging compiler issues.

Similar to 9445a4ab439 ("radeonsi: add radeonsi_sync_compile option").

v2: Actually query the driconf to set screen->driconf.sync_compile.
Noticed by Ken.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11229>

iris: Don't pass the shader key to iris_compile_[shader stage]

Use the copy of the key embedded in the compiled shader.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11229>

iris: Add the variant to the list as early as possible

I tried to find a way to break this into some smaller commits, but
everything is very intertwined. :(

When searching the variants list in the iris_uncompiled_shader, add the
new variant if it is not found.  This will be necessary for threaded
shader compilation.  This conceptually simple change had a bunch of
fallout.

Much of this was at least conceptually borrowed from radeonsi.

- Other threads might find a variant in the list before the variant has
  been compiled.  To accomdate this, add a fence.  Each thread will wait
  on the fence in the variant when searching the list.

- A variant in the list may fail compilation.  To accomodate this, add a
  flag.  All paths will examine iris_compiled_shader::compilation_failed
  before trying to use the variant.

- The race condition between multiple threads trying to create the same
  variant at the same time is handled *before* both thread spend the
  effort to compile the shader.  The means that iris_upload_shader
  cannot change shaders on the caller, so it does not need to return
  anything.

v2: Change "found" parameter of find_or_add_variant to "added."  This
inverts the values returned, and it probably makes uses of the returned
value more easily understood.  Always set the value in the called
function.  Suggested by Ken.

v3: Move shader->compilation_failed check to avoid shader != NULL test.
Rearrange some logic and add a comment in iris_update_compiled_tcs.
Suggested by Ken.  Don't call find_or_add_variant in
iris_create_shader_state.  See
https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11229#note_1000843
for more details.  Noticed by Ken.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11229>

iris: Allocate shader variant in caller of iris_upload_shader

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11229>

iris: Extract allocation bits from iris_upload_shader to iris_create_shader_variant

The added assertion in iris_create_shader_variant helped catch a bug in
the next commit.

v2: Drop (unnecessary) initialization of shader->assembly.res when
moving to iris_create_shader_variant.  Suggested by Ken.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11229>

iris: Ref count the uncompiled shaders

I tried /just/ ref counting the uncompiled shaders, but that is not
sufficient.  At the very least, it's a problem for blorp shaders that
only have variants (and no uncompiled shader).

This is in prepartion for using the live shader cache.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11229>

iris: Merge iris_create_[shader stage]_state funcs into iris_create_shader_state

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11229>

iris: Unify iris_create_[shader stage]_state functions

v2: Merge old iris_create_shader_state into the new function.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11229>

iris: Unify iris_delete_[shader stage]_state functions

This is in prepartion for using the live shader cache.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11229>

iris: Fix return type of iris_compile_*

Trivial.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11229>

iris: Add a comment for iris_uncompiled_shader::nir

This would have saved me a few minutes searching around.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11229>

util/queue: Don't crash in util_queue_destroy when init failed

This simplifies the error exit paths for drivers that use these queues.

v2: Move allocation of queue->jobs after initializing the mutxes and
condition variables.  Noticed by Ken.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11229>

st/texture: refactor get_src_format() to be more useful

take more params in order to enable better reuse

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11982>

mesa/st: add format-finding capabilities to pbo get_dst_format()

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11982>

mesa/st: enable calling st_choose_format() purely for translation

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11982>

mesa/st: break up st_choose_matching_format()

having a function that just returns the equivalent format is handy

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11982>

mesa/st: break up st_GetTexSubImage

this makes the function more readable and enables a lot of the code
to be reused for other purposes

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11982>

ir3: Preserve gl_ViewportIndex in the binning shader

Fixes dEQP-VK.draw.shader_viewport_index.* with TU_DEBUG=forcebin.

Fixes: efff7342204 ("turnip: multiViewport and VK_EXT_shader_viewport_index_layer")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12104>

gallium/tests: Fix warning calculating absdiff

With clang building tests:

../src/gallium/tests/trivial/compute.c:1215:29: warning: taking the absolute value of unsigned type 'unsigned int' has no effect [-Wabsolute-value]
                        if (abs(((uint32_t *)x)[j] -
                            ^
../src/gallium/tests/trivial/compute.c:1215:29: note: remove the call to 'abs' since unsigned values cannot be negative
                        if (abs(((uint32_t *)x)[j] -

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/12087>

loader/dri3: create linear buffer with scanout support

If we have a different GPU dealing with display, we fallback to
exchanging linear buffers with the compositor. We should specify in
creating the linear buffer that this could be used for display.

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/4706
Cc: mesa-stable
Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11942>

tu: delay decision of forcing sysmem due to subpass self-dependencies

DXVK always inserts vertex stage subpass self-dependency for every
subpass regardless of whether there actually would be a barrier.
This effectively disabled gmem rendering with DXVK.

Thus we delay the decision to disable gmem rendering until we
see a barrier with vertex stages.

Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12038>

docs/ci: Update http cache config to let Authorization headers pass through

When downloading traces from a restricted repository, the Authorization
header will be set. Previous to this change, the http proxy would drop
it.

Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6441>

ci/freedreno: Test with non-redistributable traces

Add a job to test with traces that we cannot redistribute, listed in a
separate file. Since those traces might not be accessible by everyone,
this job is created only when the pipeline is triggered by `marge-bot`.

This job is optional because otherwise it could be blocking a merge
request of someone who cannot really debug the issue due to lack of
access to these traces.

The documentation available under `docs/ci` goes into more details
explaining the rationale behind optional traces.

Signed-off-by: Antonio Caggiano <antonio.caggiano@collabora.com>
Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6441>

gallium/vl: don't leak fd in vl_dri3_screen_create

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12092>

radeonsi: preserve derivatives after discards for ARB shaders

Contrary to the GLSL spec, the ARB fp spec doesn't explicitely
states that derivatives are undefined after discards.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/5018
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11972>

mesa: move gl_program::is_arb_asm to shader_info

This will be used from radeonsi in the next commit.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11972>

lavapipe: do not assert on more than 32 samplers

We can have more than 32 samplers, but the code below will assert in that
case. The return value is not used for samplers, so let's just return
zero early and be done with it.

Fixes: c18ff60087e ("lavapipe: emit correct textures_used for texture-arrays")
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11845>

vulkan/wsi/x11: do not inherit last_present_mode

Under XWayland, the first present after a window resize is sometimes
completed with COPY (seems to happen when the previous present with the
old size is pending; not really sure).  The following presents are
completed with FLIP.

When a swapchain is created with an old swapchain, and
old_chain->last_present_mode is FLIP, chain->last_present_mode is set to
FLIP as well.  This causes the new swapchain to be marked
VK_SUBOPTIMAL_KHR, which is sticky, if the first present is completed
with COPY.

Instead of inheriting, treat each swapchain as independent.  We will
miss the case where an old swapchain is flipping but a new swapchain is
copying.  But swapchain reallocation normally happens in response to
present engine state change.  If the newly allocated swapchain is
copying, another reallocation is unlikely to fix that.

Fixes: 61309c2a727 ("vulkan/wsi/x11: Return VK_SUBOPTIMAL_KHR for X11")
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12030>

zink: rework buffer mapping

this collects a bunch of changes which need to be made but which can't
be split up incrementally without breaking things:
* move map offset to zink_transfer::offset for consistent handling
* fix stream uploader map rules to only apply for discard cases so
  the stream uploader doesn't explode
* fix the staging buffer map rules to apply to any non-staging readback
* stop making huge staging buffers for readback
* break out the DONTBLOCK case since this is only coming from qbo readback
* add explicit read-only sync for maps to stall on writes

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12090>

zink: use MAP_ONCE for qbo readback

this implies that the buffer can be unmapped

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12090>

pan/va: Allow floating-point swizzles on ATEST

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Align staging registers on Valhall

This handles the following from the Valhall specification (that I wrote):

    If multiple subsequent staging registers are accessed, the base must
    be aligned to 2. However, even if 4 registers are accessed, it is
    not necessary to align to 4, only to 2. This restriction allows the
    hardware to use a 64-bit data path without handling unaligned
    access, which is more efficient. This restriction does not apply if
    only a single register is accessed.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Clarify the logic of bi_reconverge_branches

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Clean up and export bi_reconverge_branches

Decides when we need "branch reconvergence" (canonical term), the
logical opposite of "back-to-back execution" (non-canonical term, this
is old code in Bifrost terms). So invert the return value, rename, and
export so we can use it when packing Valhall instructions.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Add branch_offset immediate

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Garbage collect stuff in bi_layout.c

Predates clause scheduling.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Fix UBO push with nir_opt_shrink_vectors

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Add constant folding unit test

I just played with the implementation, let's ensure I didn't break it.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Refactor constant folding for testability

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Use bi_apply_swizzle in constant folding

Much more legible now.

Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Add helper to swizzle a constant

Instead of open-coding special cases in multiple places, just keep a
canonical version that handles every case.

Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Add strip_index helper

Needed to correctly lower sources to moves, used in the FAU lowering.
Technically, the issue is already present on Bifrost, but it's hidden
because Bifrost packing doesn't validate the absense of unsupported
modifiers. Valhall packing adds more safety, which caught this issue.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Rename CLPER_V7 back to CLPER

v6 is really the oddball here. CLPER on v9 supports a superset of v7.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Rename NOP.i32 to NOP

Simpler and matches the syntax of Valhall I've picked out.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Fix typo in FAU enum

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Model Valhall special values as FAU

Not sure if this is "really" FAU but it's convenient to treat it as
such.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Model RSCALE for Valhall

Faster than FMA_RSCALE. We'll want to optimize this (e.g. for exp2).

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Model *ADD_IMM instructions in IR

Needed for Valhall.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Remove unused BIR_FAU_HI

Redundant with offset.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Add discard flag to bi_index

Needed to model Valhall instructions. Should also be useful to RA if we
ever get around to doing something SSA based.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: DCE after bifrost_nir_lower_algebraic_late

Needed for sat_signed to fuse, since we run modifier prop before backend
DCE.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Constify BIR manipulation

For use in Valhall packing.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Remove redundant check in clamp fusing

Already checked above.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Add instruction unit test macro

Checks for instruction equality.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Add instruction equality helper

Useful for unit testing.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Add helpers for unit testing

At some point I should stop reinventing GTest but, look, writing tests
at all is big for me, one thing at a time, ok? 😋

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Output binaries from standalone compiler

Useful for shader replacement.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Only call clause code on Bifrost

Valhall will have its own simpler code path.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Add quirks for Mali G78

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Do more mesa/st stuff in standalone compiler

From freedreno.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Zero initialize shader_info

Fixes the following assert with the standalone compiler

bifrost_compiler: ../src/panfrost/bifrost/bi_opt_push_ubo.c:134: bi_opt_push_ubo: Assertion `ctx->info->push.count == 0' failed.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Parse file names in standalone compiler

Would like to compile compute shaders too.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Remove unused option

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

pan/bi: Remove unused pointer from bi_instr

Yikes.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12026>

zink: use dynamic line stipple

save those pipeline bits!

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12011>

ci: add another zink job with timelines disabled

this codepath needs to be maintained for *vague handwaving*

Acked-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12089>

zink: add env var to disable timelines

for ci use only

Acked-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12089>

zink: export PIPE_CAP_TGSI_BALLOT

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11990>

zink: remove VK_EXT_shader_subgroup_ballot from device info

restrict this to 1.2 only

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11990>

zink: implement compiler handling for subgroup ballot builtins/intrinsics

these are all lowered and unremarkable

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11990>

zink: lower subgroup ballot instructions

this handles rewrites of ballot-related instructions/builtins to spirv
vec4 types from the native GLSL U64 type

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11990>

zink: change descriptor flushing to assert

there should never be flushing due to pool depletion; instead, trigger an
oom flush and stall to replenish the pool after the draw/compute

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12012>

etnaviv: fix gbm_bo_get_handle_for_plane for multiplanar images

Implement resource_get_param for PIPE_RESOURCE_PARAM_NPLANES and fix
resource_get_handle to walk to the correct linked resource for
multiplanar images, allowing gbm_bo_get_handle_for_plane to be called
with plane > 0.

This fixes an assert that is triggered when a wayland client tries
to send weston an NV12 dmabuf, for example:

  weston: .../mesa/src/gbm/backends/dri/gbm_dri.c:752: gbm_dri_bo_get_handle_for_plane: Assertion `plane == 0' failed.

Fixes: 788f6dc85781 ('Revert "gallium/dri: fix dri2_from_planar for multiplanar images"')
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12037>

virgl: Enable caching for sampler views and render targets

Signed-off-by: Rohan Garg <rohan.garg@collabora.com>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11241>

Revert "Revert "virgl: Cache depth and stencil buffers""

This reverts commit a8e75bb73c6f8ce94c4f598a2f4763b12188991c.

Signed-off-by: Rohan Garg <rohan.garg@collabora.com>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11241>

virgl: Add more meta data to cached resources

By expanding the meta data about resources in the cache we can match more
resources such as 3D textures, samplers and render targets.

Signed-off-by: Rohan Garg <rohan.garg@collabora.com>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11241>

pan/va: Integrate the tests into meson test

This way we will get testing in CI. Invoke as

   meson test --suite=panfrost

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12025>

pan/va: Add disassembler test harness

Uses the same set of cases. This is a standalone C program because the
easy way of hooking into the disassembler from Python with subprocesses
was slow. This seems cleaner anyway.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12025>

pan/va: Add assembler test harness

Integration regression testing. Nothing fancy.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12025>

pan/va: Add negative test cases for the assembler

These are lines of assembly that look valid but are not, and should
raise a parser error but not otherwise crash the assembler or produce
invalid code.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12025>

pan/va: Add dis/assembler test cases

These are valid pairs of hexdumped assembled instructions and the
corresponding disassembly, to be used to regression test both the
assembler and the disassembler.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12025>

pan/va: Add disassembler generator

When we bring up the Valhall compiler in Mesa, we will like to have a
disassembler in native code, so we shouldn't write our disassembler in
Python. Instead, we write a disassembler generator in Python with mako
templates, which will produce a va_disasm_instr entrypoint from the
architecture defined in ISA.xml and valhall.py.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12025>

pan/va: Check for FAU conflicts in the assembler

Logic described in the "Uniform/constant restrictions" section of the
Valhall specification. (You know, my Valhall specification. Is there
another one?)

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12025>

pan/va: Add Valhall assembler

This Python script acts as a standalone assembler. It takes Valhall
instructions with Mesa-flavour syntax, parses them, errors out if there
are syntax errors, and writes out an assembled binary if there are not.
It also is available as a programmatic interface for automated testing.

While this attempts to handle syntax errors, it does not check
semantics. It will happily compile programs that fault, provided each
instruction locally 'looks' plausible.

The code itself is quite small, despite supporting most of the known
ISA, because the syntax is regular and the heavylifting is done by
ISA.xml and valhall.py.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12025>

pan/va: Assert no instructions are duplicated

Caught a bug in the published PDF. Whoops!

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12025>

pan/va: Add ISA.xml parser and support code

This Python module parses the ISA.xml file added in the previous
commits, extracts all the useful information, and combines it with
extra annotations. In total, it provides a programmatic way to interface
with the Valhall instruction set.

   from valhall import instructions, enums

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12025>

pan/va: Add initial ISA.xml for Valhall

This handwritten file is the product of over a hundred hours of
reverse-engineering and represents the sum of what I've learned about
the Valhall architecture. It will be used in the next commits as the
backbone of a Valhall toolchain.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12025>

llvmpipe: correct the debug information printed with GALLIVM_PERF=nopt

GALLIVM_PERF_NO_OPT is "belond" to GALLIVM_PERF environment variable,
not GALLIVM_DEBUG. when GALLIVM_PERF=nopt is passed to llvmpipe,
"-sroa -early-cse -simplifycfg -reassociate -constprop -instcombine -gvn"
should not be printed. Those llvm optimation techniques should only be
printed when the default(-O2) optimation is enabled.

$ GALLIVM_PERF=nopt GALLIVM_DEBUG=dumpbc glmark2

Before apply this patch:

  Invoke as "opt -sroa -early-cse -simplifycfg -reassociate -mem2reg -constprop -instcombine -gvn ir_setup_variant_0.bc | llc -O2 [-mcpu=<-mcpu option>] [-mattr=<-mattr option(s)>]"

After apply this patch:

  Invoke as "opt -mem2reg ir_fs304_variant0.bc | llc -O0 [-mcpu=<-mcpu option>] [-mattr=<-mattr option(s)>]"

Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Signed-off-by: suijingfeng <suijingfeng@loongson.cn>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11994>

gallium: Reset {d,r}Priv in dri_unbind_context

The code in dri_make_current just checks the value of the pointers
to decide to update texture_stamp or not. This is buggy since a new
allocated drawable could share the same address with the previous
released drawable. Fix the stale pointer issue by always resetting
these pointers to NULL in dri_unbind_context.

v2:
   Move the reset codes to the end of the function.

Signed-off-by: Lepton Wu <lepton@chromium.org>
Cc: mesa-stable
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12050>

anv/android: Rework our handling of AHardwareBuffer imports

The current code we have for this is a bit of a mess, likely due to
trying too hard to put it in anv_android.c.  The external_format bit in
anv_image, for instance, really means "quit creation early" which is
something we want to do for AHardwareBuffer imports regardless of
whether or not they use a native format.  It gets set both by declaring
an AHardwareBuffer external handle type and by VkExternalFormatANDROID.
However, VkExternalFormatANDROID is only allowed for AHardwareBuffer
imports.  If we ever did get an external format outside the context of
an AHardwareBuffer import, we would end up with a useless partially
created image.

When we detect an AHardwareBuffer import, we punt off to a function in
anv_android.c that does nothing interesting but call anv_create_image
with AUX disabled and external_format = true.  The aux disable here is
useless because the actual isl_surf layout is done by resolve_ahw_image
which also sets ISL_SURF_USAGE_DISABLE_AUX_BIT.  As far as external
formats go, anv_image_from_external() sets it regardless of whether or
not there is actually an external format.

This commit replaces anv_image::external_format with anv_image::from_ahb
which is the thing we actually want to track for this.  We delete
anv_image_from_external and a bunch of the external_format handling
because it's all useless.  The end result is massively simpler and,
while it appears to blur the boundary between Android code and the rest
of the driver, it makes the whole flow more obvious.

Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12040>

tu: Triage some CTS failures

These should be fixed by the linked VK-GL-CTS change. I couldn't
reproduce the other renderpass failure on a650 though.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12065>

panvk: Implement vkCmdClearAttachments

For now it only supports one RT and color attachments.

Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12019>