pan/mdg: Remove midgard_opt_copy_prop_reg

This is known broken code, and the fix is highly nontrivial. It isn't
doing terribly much for perf, so just rip off the band-aid. Prevents a
dEQP regression, and likely fixes bugs in real apps too.

total instructions in shared programs: 96640 -> 97089 (0.46%)
instructions in affected programs: 27831 -> 28280 (1.61%)
helped: 33
HURT: 301
helped stats (abs) min: 1 max: 6 x̄: 3.64 x̃: 5
helped stats (rel) min: 1.96% max: 10.00% x̄: 6.48% x̃: 7.94%
HURT stats (abs)   min: 1 max: 18 x̄: 1.89 x̃: 1
HURT stats (rel)   min: 0.46% max: 15.00% x̄: 3.17% x̃: 2.38%
95% mean confidence interval for instructions value: 1.09 1.59
95% mean confidence interval for instructions %-change: 1.80% 2.63%
Instructions are HURT.

total bundles in shared programs: 45615 -> 45612 (<.01%)
bundles in affected programs: 11257 -> 11254 (-0.03%)
helped: 121
HURT: 146
helped stats (abs) min: 1 max: 7 x̄: 2.34 x̃: 1
helped stats (rel) min: 1.22% max: 23.33% x̄: 7.85% x̃: 5.26%
HURT stats (abs)   min: 1 max: 17 x̄: 1.92 x̃: 2
HURT stats (rel)   min: 0.42% max: 25.00% x̄: 5.17% x̃: 3.85%
95% mean confidence interval for bundles value: -0.34 0.31
95% mean confidence interval for bundles %-change: -1.69% 0.23%
Inconclusive result (value mean confidence interval includes 0).

total quadwords in shared programs: 76662 -> 77275 (0.80%)
quadwords in affected programs: 20148 -> 20761 (3.04%)
helped: 28
HURT: 275
helped stats (abs) min: 1 max: 4 x̄: 1.54 x̃: 1
helped stats (rel) min: 0.43% max: 25.00% x̄: 4.89% x̃: 2.50%
HURT stats (abs)   min: 1 max: 12 x̄: 2.39 x̃: 2
HURT stats (rel)   min: 0.51% max: 28.57% x̄: 5.18% x̃: 4.26%
95% mean confidence interval for quadwords value: 1.80 2.25
95% mean confidence interval for quadwords %-change: 3.64% 4.86%
Quadwords are HURT.

total registers in shared programs: 7078 -> 7081 (0.04%)
registers in affected programs: 1028 -> 1031 (0.29%)
helped: 62
HURT: 70
helped stats (abs) min: 1 max: 2 x̄: 1.11 x̃: 1
helped stats (rel) min: 8.33% max: 50.00% x̄: 15.03% x̃: 12.50%
HURT stats (abs)   min: 1 max: 2 x̄: 1.03 x̃: 1
HURT stats (rel)   min: 8.33% max: 66.67% x̄: 20.13% x̃: 11.25%
95% mean confidence interval for registers value: -0.17 0.21
95% mean confidence interval for registers %-change: -0.14% 7.38%
Inconclusive result (value mean confidence interval includes 0).

total threads in shared programs: 5032 -> 5036 (0.08%)
threads in affected programs: 31 -> 35 (12.90%)
helped: 12
HURT: 6
helped stats (abs) min: 1 max: 2 x̄: 1.08 x̃: 1
helped stats (rel) min: 100.00% max: 100.00% x̄: 100.00% x̃: 100.00%
HURT stats (abs)   min: 1 max: 2 x̄: 1.50 x̃: 1
HURT stats (rel)   min: 50.00% max: 50.00% x̄: 50.00% x̃: 50.00%
95% mean confidence interval for threads value: -0.43 0.87
95% mean confidence interval for threads %-change: 13.82% 86.18%
Inconclusive result (value mean confidence interval includes 0).

total loops in shared programs: 19 -> 19 (0.00%)
loops in affected programs: 0 -> 0
helped: 0
HURT: 0

total spills in shared programs: 0 -> 2
spills in affected programs: 0 -> 2
helped: 0
HURT: 1

total fills in shared programs: 0 -> 15
fills in affected programs: 0 -> 15
helped: 0
HURT: 1

Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10652>

pan/mdg: Use _output_ type for outmod printing

Fixes incorrect outmods printed for conversions.

Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10652>

docs: mark off VK_KHR_vulkan_memory_model for turnip

Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10669>

lavapipe: don't access pipeline blend state when it should be ignored

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10625>

lavapipe: don't access pipeline dsa state when it should be ignored

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10625>

lavapipe: don't access pipeline viewport state when it should be ignored

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10625>

lavapipe: ignore tess pipeline info if no tess shaders in pipeline

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10625>

lavapipe: don't unnecessarily flag dsa states for updating

these force a new dsa state to be created and bound, which isn't necessary
if the same value is being reset

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10625>

lavapipe: zero out the blend state info and flag for updating on null blend state

this still needs to be updated if there's no pipeline info available

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10625>

lavapipe: zero out the dsa state info and flag for updating on null dsa state

this still needs to be updated if there's no pipeline info available

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10625>

lavapipe: update more states on null multisample pipeline info

these all need to be unset to ensure expected functionality

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10625>

lavapipe: flag renderpasses as having color/zs attachments

it's useful to track this info for reuse

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10625>

lavapipe: set events to the unsignalled state on creation

this is otherwise uninitialized and not compliant with spec

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10625>

lavapipe: do not read sampler descriptor info during update if layout has immutables

this is illegal

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10625>

lavapipe: handle buffer sizes better in CmdBindTransformFeedbackBuffersEXT

according to spec, the pSizes array member is only used if the array is non-null
and the value is not VK_WHOLE_SIZE, otherwise this value is calculated based
on the buffer size - the offset

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10625>

pan/bi: Add two tuples to a clause when needed with NOSCHED

Fixes SuperTuxKart with BIFROST_MESA_DEBUG=nosched.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10661>

panfrost: Always write reloaded tiles when making CRC data valid

If CRC data is currently invalid and the current batch will make it
valid, write even clean tiles to make sure CRC data is updated.

Fixes: 8ba2f9f6985 ("panfrost: Create a blitter library to replace the existing preload helpers")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10566>

panfrost: Make pan_select_crc_rt a non-static function

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10566>

nir: add nir_isub_imm

Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10654>

gallium/inlines: remove atomic set from pipe_reference_init()

when an object is initialized with this, it should not be visible to any
other threads or contexts, so there should be no need to use an atomic set here

at the time of this commit, there are only two callers in the tree which pass
values != 1:
* zink uses a calculated number for framebuffer refcount on init (this is fine)
* aux/pb passes 0 on init (this is fine)

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10650>

Revert "gallium/util: Fix depth/stencil blit shaders"

This reverts commit 7ca72f172678116d29d254b786a9422b864aef3d.
Unlike what's stated in this commit, the depth or stencil components
have to be replicated on all channels, as specified in the
"Texture Sampling and Texture Formats" section of the TGSI doc
(docs/gallium/tgsi.rst).

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10649>

zink: clamp zs samplers to XXXX swizzle for all non-zero/one swizzles

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10649>

zink: do not ask glsl-compiler to unroll

We don't really need loops unrolled, so let's just disable this. This is
generally recommended for NIR drivers, but we can do even better; not
even unroll in NIR. And since we don't set
nir_shader_compiler_options::max_unroll_iterations, we're already there.

Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10487>

lavapipe: emit correct textures_used for texture-arrays

When we lower a texture-lookup with a dynamic index, we need to mark the
entire array as used, because we don't know better.

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10487>

v3d: enable NIR loop unrolling

The GL driver was getting loop unrolling from the GLSL compiler frontend,
but NIR unrolling is more sophisticated, so prefer that.

The only caveat is that loop unrolling is implemented in the Mesa state
tracker, so our backend won't have a chance to undo the optimization if
it causes us to lower thread count or spill, so we choose to be a bit more
conservative with the configuration than what we were doing with GLSL.

Shader-db results follow. Increase in instruction counts is expected due
to additional unrolling. We lose threads in very few shaders, but we
make up for this with the additional unrolling and reduced spilling. We
also managed to get 3 more shaders to compile successfully.

total instructions in shared programs: 13416427 -> 13461431 (0.34%)
instructions in affected programs: 96936 -> 141940 (46.43%)
helped: 58
HURT: 216
Instructions are HURT.

total threads in shared programs: 410626 -> 410598 (<.01%)
threads in affected programs: 56 -> 28 (-50.00%)
helped: 0
HURT: 14
Threads are HURT.

total loops in shared programs: 2121 -> 1708 (-19.47%)
loops in affected programs: 468 -> 55 (-88.25%)
helped: 446
HURT: 47
Loops are helped.

total uniforms in shared programs: 3676567 -> 3691185 (0.40%)
uniforms in affected programs: 25304 -> 39922 (57.77%)
helped: 23
HURT: 199
Uniforms are HURT.

total spills in shared programs: 5902 -> 5727 (-2.97%)
spills in affected programs: 285 -> 110 (-61.40%)
helped: 19
HURT: 0

total fills in shared programs: 13308 -> 13121 (-1.41%)
fills in affected programs: 301 -> 114 (-62.13%)
helped: 19
HURT: 0

total sfu-stalls in shared programs: 31860 -> 32856 (3.13%)
sfu-stalls in affected programs: 1692 -> 2688 (58.87%)
helped: 25
HURT: 196
Sfu-stalls are HURT.

total inst-and-stalls in shared programs: 13448287 -> 13494287 (0.34%)
inst-and-stalls in affected programs: 98404 -> 144404 (46.75%)
helped: 57
HURT: 217
Inst-and-stalls are HURT.

total nops in shared programs: 329276 -> 329551 (0.08%)
nops in affected programs: 2189 -> 2464 (12.56%)
helped: 58
HURT: 181
Nops are HURT.

LOST:   0
GAINED: 3

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10647>

broadcom/compiler: specify maximum thread count in compile strategies

Once we have exhausted compile strategies at 4 threads and we start
enabling lower thread counts, there is no point in starting compiles
with 4 threads for them, we know these will fail, so let's start at
2 in these cases.

This also has another nice implication: if the driver compiles at 4
threads and fails to register allocate, we were allowing it to try
with 2 threads, but this would only retry the register allocation
process and would not really recompile the shader with 2 threads. This
is not optimal, because at 2 threads we have more TMU fifo space for
each thread and we can do more TMU pipelining, so we were missing that
opportunity.

This improves performance in Sponza by ~1.5% and also seems to help
UE4 slightly.

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10647>

broadcom/compiler: refactor compile strategies

Until now, if we can't compile at 4 threads we would lower thread count
with optimizations disabled, however, lowering thread count doubles the
amount of registers available per thread, so that alone is already a big
relief for register pressure so it makes sense to enable optimizations
when we do that, and progressively disable them until we enable spilling
as a last resort.

This can slightly improve performance for some applications. Sponza,
for example, gets a ~1.5% boost. I see several UE4 shaders that also get
compiled to better code at 2 threads with this, but it is more difficult
to assess how much this improves performance in practice due to the large
variance in frame times that we observe with UE4 demos.

Also, if a compiler strategy disables an optimization that did not make
any progress in the previous compile attempt, we would end up re-compiling
the exact same shader code and failing again. This, patch keeps track of
which strategies won't make progress and skips them in that case to save
some CPU time during shader compiles.

Care should be taken to ensure that we try to compile with the default
NIR scheduler at minimum thread count at least once though, so a specific
strategy for this is added, to prevent the scenario where no optimizations
are used and we skip directly to the fallback scheduler if the default
strategy fails at 4 threads.

Similarly, we now also explicitly specify which strategies are allowed to do
TMU spills and make sure we take this into account when deciding to skip
strategies. This prevents the case where no optimizations are used in a
shader and we skip directly to the fallback scheduler after failing
compilation at 2 threads with the default NIR scheduler but without trying
to spill first.

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10647>

broadcom/compiler: add a compiler strategy to disable loop unrolling

Loop unrolling can increase register pressure significantly, leading to
lower thread counts and spilling.

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10647>

v3d: move NIR compiler options to GL driver

The Vulkan driver was already creating and using its own set of options, so
the ones defined in the compiler are only used with GL, which is confusing.
Move them to the GL driver.

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10647>

v3dv: setup loop unrolling

We set the maximum at 16 iterations (the GL compiler chooses 32
iterations for the GLSL front-end loop unrolling pass) because we
have observed a bunch of shaders from Sascha Willems that spill
significantly with 32, leading to massive performance degradation,
while 16 avoids spilling and doesn't seem to cause visible
performance degradation compared to cases that unroll 32 without
spilling.

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10647>

broadcom/compiler: add a loop unrolling pass

Right now this is useful for Vulkan onnly, because GL gets loop
unrolling from the GLSL compiler and/or mesa state tracker
NIR front-ends.

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10647>

Revert "CI: Disable Panfrost and radeonsi"

Lab is up and running again.

This reverts commit de1a20f4bec789c690e5eeba728073ca9c9cfc08.

Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10663>

radv: fix missing ITERATE_256 for D/S MSAA images that are TC-compat HTILE

To make them readable by shaders, only needed on GFX10+.

This also fixes corruption with Control and MSAA.

Cc: 21.1 mesa-stable
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/10610>

i965: support only color formats with memory objects

There are issues with depth support and lack of support for
combined depth+stencil, let's support only color formats for now.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/4712
Fixes: f73aeca0ce3 ("i965: Initial implementation for EXT_memory_object_*")
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Rohan Garg <rohan.garg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10646>

turnip: Move the extension tables to tu_device.c

Following intel's lead in 27d49670.  In the dEQP-VK.info.* tests, this
bumps apiVersion from 1.1.128 to 1.1.177.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10635>

turnip: Switch to the shared vulkan ICD generator.

One less python script to maintain.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10635>

gallium/tgsi_exec: Simplify bounds checks on the const file.

We were doing two < 0 checks, when we can just treat the value as unsigned
and check against our unsigned size limit.  Cuts 2k of text from the
various inlined copies of this function.

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10605>

tgsi_exec: Mark the store file default case as unreachable.

It isn't reached, and shouldn't be, and it's not like returning NULL
(instead of the &null case for TGSI_FILE_NULL) will do anything good.

Shaves another tiny bit of dead code off of release builds.

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10605>

tgsi_exec: Drop unused destination dimension support.

As you can see, nothing used the index2d value we computed.  Turns out
this does remove some unused code from release builds.

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10605>

gallium/tgsi_exec: Drop the unused dst_datatypes from dest stores.

It was just for the disabled check_inf_or_nan debugging, and if you want
to do that then you should do it with an assert at the end of
exec_instruction using tgsi_opcode_infer_type().

Despite the fact that this operand was dead, it actually reduces generated
code in release builds.

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10605>

zink: remove incorrect border-swizzle assumption

According to the Khronos issue[1], it seems like RADV and NVIDIA is in
the wrong, so let's turn the logic around here.

This makes us do the right thing for Lavapipe, and we should probablt
remove RADV from this list once !9731 lands.

[1]: https://github.com/KhronosGroup/Vulkan-Docs/issues/1421

Fixes: cdb9a4775bf ("zink: set PIPE_CAP_TEXTURE_BORDER_COLOR_QUIRK")
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10321>

vtn: Use relaxed 24bit opcodes for CL 24bit math

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10549>

nir: Add relaxed 24bit opcodes

These are equivalent to the 32bit opcodes if there are no more efficient
24bit opcodes available, but inputs are guaranteed to already be 24bit,
so the 24bit opcodes can be used instead if they exist and are efficient.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10549>

llvmpipe: Fix optimization loop to actually loop

Reviewed-by: Dave Airlie <airlied@redhat.com>
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10549>

freedreno/a5xx: Fix up border color pointers.

We were forgetting to increment in the loop, but also it looks from blob
dumps on Pixel 2 like all the pointers it emitted were shifted up by 3
compared to our xml, and that's the same shift that a6xx uses for its
pointers.  None of the tests seem to use more than one
border-color-requiring texture, so it's hard to tell.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9904>

CI: Disable Panfrost and radeonsi

The Cambridge office is having connection difficulty; disable the jobs
until it comes back.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10658>

freedreno/tools: Fix async flush vs fdperf/computerator

They need to wait on the ready fence to ensure the submit has been
flushed to the kernel.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10626>

mesa: NOTE! Default branch is now main

To update your local repository to use the new default branch, these
commands may help:

$ git fetch origin
$ git checkout master
$ git branch -m main
$ git branch --set-upstream-to=origin/main
$ git remote set-head origin --auto

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Eric Engestrom <eric@engestrom.ch>

docs: Rename master branch to main

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Eric Engestrom <eric@engestrom.ch>

docs/releasing.rst: Rename master branch to main

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Eric Engestrom <eric@engestrom.ch>

issue_templates/Bug Report: Rename master branch to main

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Eric Engestrom <eric@engestrom.ch>

.gitlab-ci.yml: Use main branch for gitlab ci

Reworks:
 * Fix mesa/mesamaster typo to mesa/mesa main (anholt)
 * Use $CI_DEFAULT_BRANCH (eric_engestrom)

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Eric Engestrom <eric@engestrom.ch>

bin/pick: Rename master branch to main

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Dylan Baker <dylan@pnwbakers.com>
Reviewed-by: Eric Engestrom <eric@engestrom.ch>

commit_in_branch_test.py: Rename branch master to main

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Dylan Baker <dylan@pnwbakers.com>
Reviewed-by: Eric Engestrom <eric@engestrom.ch>

docs: remove doxygen support

It seems building the doxygen docs has bit-rotted over time, and now
generates a set of empty modules, apart from some basic descriptions.

Since Mesa is mostly implementing externally documented APIs, I don't
think it makes a whole lot of sense trying to fix this, and I think the
presence of these files might confuse users who try them out.

So let's just get rid of this.

Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10611>

ci/lavapipe: Add fractional NIR stress test job.

I think between the disk cache unit tests and testing that we can really
serialize/deserialize NIR, this covers what I cared about for testing disk
caching.

Closes: #3597
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10629>

ci/llvmpipe: Add testing of gles3/31/gl.

llvmpipe is a pretty important driver, we should be fully testing it.
Also, enable some options to stress test some NIR internals.

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10629>

turnip: Demote API version to 1.1.

We don't support major 1.2 required extensions like timeline semaphores.
Fixes many complaints in the dEQP-VK.info.vulkan1p2.* group.

We were originally bumped to 1.2 in 75755e0eba17 ("turnip: Pretend to
support Vulkan 1.2") but hopefully that build issue has been fixed in the
entrypoint reworks since then.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10471>

venus: populate VK_ERROR_OUT_OF_HOST_MEMORY if applied

Fix dEQP-VK.wsi.android.swapchain.simulate_oom.* failures.

Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10639>

freedreno/a5xx: Remove ppgtt hack

This should no longer be needed after !7773, which fixes the issue that
lead to the crash. Sorry for not fixing the issue earlier ;)

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10645>

aux/cso_cache: add handling for save/restore of compute states

just shader and samplers for now

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10616>

panfrost: Fix format definitions to match gallium expectations

Gallium wants the depth or stencil component replicated on all .XYZW.
That's easily done on pre-v7 since we can forge all the swizzles we
want, but Bifrost v7 only supports a few combinations, so we have to
combine the user swizzle with our own 'replicate' swizzle to make it
work. Note that v7 has a trick to make border color work when the GRBA
order is chosen: they apply the red border color to the green component.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10612>

android: pan/bi: add bi_opt_constant_fold.c to Makefile.sources

Fixes the following building error:

FAILED: out/target/product/x86_64/obj/SHARED_LIBRARIES/gallium_dri_intermediates/LINKED/gallium_dri.so
...
ld.lld: error: undefined symbol: bi_opt_constant_fold
>>> referenced by bifrost_compile.c:3105 (external/mesa/src/panfrost/bifrost/bifrost_compile.c:3105)

Fixes: 1cb11969bec0 ("pan/bi: Add simple constant folding pass")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10603>

anv: Allow storage on all formats that support typed writes

In particular, this gives us B8G8R8A8_UNORM storage support which is
useful for writing WSI images from compute shaders.  These formats can
only be accessed in a spec-compliant way by decorating the variable
NonReadable in the SPIR-V (writeonly in GLSL).  If the client doesn't so
decorate the variable, it'll get the null surface state where reads
return 0 and writes are ignored.

Tested-by: Simon Ser <contact@emersion.fr>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10624>

isl: document format fields

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/10624>

anv: Check offset instead of alloc_size for freeing surface states

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10624>

gallium/u_vbuf: avoid dereferencing NULL pointer

When I last time fixed this, I missed that continuing here would make us
leak pointers in the translate state, which is what made this avoid a
crash in the first place.

That's not great, we need to set *some* pointer in this case. The
obvious option would be NULL, but that means that the translate-code
also needs to support NULL-pointers here.

Instead, let's point to a small, static buffer that contains enough
zero-data for the largest possible vertex attribute. This avoids having
to add more NULL-checks.

Fixes: a8e8204b186 ("gallium/u_vbuf: support NULL-resources")
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7773>

radeon/vcn: Use the correct pitch for chroma surface.

The pitch of the chroma plane isn't necessarily half that of the
luma plane, as tiling (and presumably even linear) swizzle modes
apply some alignment.

Fixes: 35e25ea1d07 ("ac/surface: allow non-DCC modifiers for YUV on GFX9+")
Reviewed-by: James Zhu <James.Zhu@amd.com>
Reviewed-by: Leo Liu <leo.liu@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10638>

turnip: implement VK_KHR_vulkan_memory_model

No handling of Acquire/Release because at the moment scheduler
works as if any barrier is Acq+Rel.

Instead of removing scoped_barrier with scope/mode that for TCS
corresponds to a control_barrier or a memory_barrier_tcs_patch in
ir3_nir_lower_tess_ctrl - remove them in emit_intrinsic_barrier.
And do the same for memory_barrier_tcs_patch and control_barrier.
While in any case hw fence/barrier shouldn't be emitted for them,
they still affect ordering of stores, and in feature ir3 backend
may want to have that information.

Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9054>

ir3: update bar/fence bits in accordance to blob

On a6xx blob uses .l rather differently from a5xx.

Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9054>

ir3: memory_barrier also controls shared memory access order

nir_intrinsic_memory_barrier has the same semantic as memoryBarrier()
in GLSL, which is:

GLSL 4.60, 4.10. "Memory Qualifiers":
 "The built-in function memoryBarrier() can be used if needed to
 guarantee the completion and relative ordering of memory accesses
 performed by a single shader invocation."

GLSL 4.60, 8.17. "Shader Memory Control Functions":
 "The built-in functions memoryBarrier() and groupMemoryBarrier() wait
 for the completion of accesses to all of the above variable types."

Fixes tests:
 dEQP-VK.memory_model.message_passing.core11.u32.coherent.fence_fence.atomicwrite.device.payload_local.image.guard_nonlocal.workgroup.comp
 dEQP-VK.memory_model.message_passing.core11.u32.coherent.fence_fence.atomicwrite.device.payload_nonlocal.workgroup.guard_local.image.comp

Fixes: 819a613a ("freedreno/ir3: moar better scheduler")

Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Reviewed-by: Samuel Iglesias Gonsálvez <siglesias@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9054>

docs: do not generate redirects on error

The build-finished event is also triggered when there's an error. I
somehow got the second argument wrong, and ended up ignoring the case.
This can lead to new exceptions being thrown due to missing files, that
ends up hiding the real problem.

Fixes: 64a4ba9e1ce ("docs: add an extension to generate redirects")
Reviewed-by: Eric Engestrom <eric@engestrom.ch>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10407>

zink: only emit ImageBuffer cap if needed

Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10371>

zink: emit cap early

Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10371>

zink: emit sample-shading cap early

Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10371>

zink: remove needless shader-info from context

There's no longer any code reading this, so let's get rid of it. It
always rubbed me a bit the wrong way, because this repeated some
information already present in the context.

Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10371>

zink: emit cap early

We have enough information to emit this cap early, so let's do that.

Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10371>

zink: do not check for varying output for fragment shaders

This will make us emit these caps needlessly, possibly on implementations
not supporting the enabling features.

Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10371>

zink: always enable fixed shader-caps

This is required for a bunch of stuff that can occur in any Vulkan
shader stage, not just these few. So let's always emit this cap.

Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10371>

intel/decoder: add gen4/5 geometry state decode

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10642>

intel/decoder: fixup batch decoder for binding tables on gen4/5

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10642>

radv: expose 2/3rd of total memory as VRAM and 1/3rd as GTT on APUs

A bunch of games complain when the VRAM size is too small. The most
compatible solution seems to return memory heaps like a dGPU.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/3423
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/9774>

radv: adjust the computation of the total usage of memory used

internal_usage is the memory allocated by the current process (intent)
while system_usage is the memory allocated globally (actual).

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/9774>

gallium/u_threaded: don't set resource pointers to NULL after driver calls

The pointers won't be used at that point. Just decrement the refcounts.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.eric.pelloux-prayer@amd.com>
Acked-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10606>

gallium/u_threaded: rewrite slot layout to reduce wasted space

A lot of space was wasted due to 16-byte alignment for slots. This new
layout tries to match glthread. Highlights:

- the slot size changed to 8 bytes (was 16), so less padding
- the header size changed to 4 bytes (was 8), so some calls can use
  the remaining 4 bytes in the slot for parameters
- draw merging merges up to 307 draws (was 256) due to space savings
- parameters in structures are sorted based on implicit type alignment
  (uint8_t first, pointers last) to make it easier to utilize the 4 bytes
  after the header and to remove holes
- some structures use smaller types for parameters than pipe_context where
  it's safe (e.g. clear uses float instead of double for depth)

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.eric.pelloux-prayer@amd.com>
Acked-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10606>

gallium/u_threaded: handle sampler views == NULL better

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.eric.pelloux-prayer@amd.com>
Acked-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10606>

gallium/u_threaded: move base_valid_buffer_range to transfer where it belongs

This saves 8 bytes per resource.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.eric.pelloux-prayer@amd.com>
Acked-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10606>

Revert "gallium/u_threaded: align batches and call slots to 16 bytes"

This reverts commit 3b1ce49bc1e5aff87805b0bab255885c84bf5052.

It will be completely rewritten, but let's revert this first.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.eric.pelloux-prayer@amd.com>
Acked-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10606>

gallium: Reset attachments to ST_ATTACHMENT_INVALID when revalidating

It is *super* *confusing* to leave this initialized to zero, i.e.
ST_ATTACHMENT_FRONT_LEFT.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10491>

mesa: Ignore the depth buffer when computing framebuffer floatness

Not that Z32F is especially common, but we shouldn't consider it to
imply that the color buffers are also float, which is what floatMode
is meant to mean.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10491>

mesa: Remove unused _mesa_{create,destroy}_visual

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10491>

util/math: change ROUND_DOWN_TO to return a uint64_t

For 32-bit builds.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10622>

intel/genxml: fix raster op fields on gen4/5

These should be unsigned integers

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10636>

intel/gemxml: move blitter command to render on gen4/5

The blitter commands don't show up in INTEL_DEBUG=bat, but on
gen4/5 they are emitted on the render engine ring so just change
the XML to reflect that.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10636>

intel/genxml: rewrite the prefilterop xml to be more consistent.

This uses a prefix at Ken's suggestion and aligns it across gens

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10636>

intel/genxml: align gen4/5 xml for store data immediate

Just align with the gen6 and later xml

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10636>

intel: fix MI builder for pre-gen7

MI_LOAD_REGISTER_MEM is only available on gen7+, so avoid build errors
on earlier generations.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10636>

zink: export PIPE_CAP_SPARSE_BUFFER_PAGE_SIZE

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10509>

zink: implement sparse buffer creation/mapping

sparse buffers are not cpu-readable, so any mapping requires that they
use a staging buffer, either the stream uploader for writes or a manual
copy for readback

future work here should attempt to resolve two perf issues:
* sparse allocations should be allocated dynamically using a suballocator
  or some other, more useful strategy
* readback shouldn't allocate a huge staging buffer

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10509>

zink: add a pipe_context::resource_commit hook

so many structs

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10509>