review.tizen.org Git - platform/upstream/mesa.git/log

intel/mi_builder: optimize 64bit immediate register loads & memory stores

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Marcin Ślusarz <marcin.slusarz@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8525>

anv: don't disable KHR_performance_query in debug mode

This is useful to verify that queries are built properly and avoiding
to deal with i915-perf (typically under simulation).

v2: Remove likely() (Marcin)

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Marcin Ślusarz <marcin.slusarz@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8525>

v3d: fix dest offset in TFU setup

It is using the source level instead of the destiny level (base_level)
to compute the dest offset.

This fixes `framebuffer-blit-levels draw rgba -auto -fbo` piglit test.

Fixes: 976ea90bdca ("v3d: Add support for using the TFU to do some blits.")
Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8491>

freedreno/ir3/parser: add cat7 support

Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8420>

st/mesa: optimize binding and unbinding shader images

- use local variable num_images
- only unbind the number of shader images that we have bound

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8180>

st/mesa: unbind sampler views, images, and vertex buffers after meta ops

v2: use a null array to unbind sampler views

Reviewed-by: Eric Anholt <eric@anholt.net> (v1)
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8180>

cso_context: remove ability to restore VBs, const bufs, sampler views, images

This is meant to decrease CPU overhead and reduce the use of cso_context when
better solutions are possible.

These functions are removed. Gallium frontends should just use pipe_context:
- cso_set_sampler_views
- cso_set_shader_images
- cso_set_constant_buffer*

Additionally, vertex buffers are passed through cso_context but they are not
saved there. The only remaining role of cso_context with regard to vertex
buffers is to decide whether to go through u_vbuf or skip it. u_vbuf also
can't save and restore vertex buffers anymore.

This is how the states are restored for meta ops:
- in st/mesa, set st->dirty |= flags to re-bind whatever state has been
touched
- outside of st/mesa, there is new interface st_context_iface::invalidate_-
_state that external modules can use to essentially set st->dirty through
it (like HUD)

Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Zoltán Böszörményi <zboszor@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8180>

st/mesa: replace st->pipe with pipe in a few places

Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Zoltán Böszörményi <zboszor@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8180>

st/mesa: don't use cso_context to set const bufs, sampler views and images

These cso_context functions will be removed and they are no longer needed
by st/mesa. They also cause CPU overhead because they save states.

Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Zoltán Böszörményi <zboszor@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8180>

st/mesa: don't use cso_context to restore VBs, etc. for PBO glTexSubImage

These cso_context capabilities will be removed.

Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Zoltán Böszörményi <zboszor@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8180>

st/mesa: don't use cso_context to restore VBs, etc. for PBO glReadPixels

These cso_context capabilities will be removed.

Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Zoltán Böszörményi <zboszor@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8180>

st/mesa: don't use cso_context to restore VBs, sampler views for glDrawTex*OES

These cso_context capabilities will be removed.

Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Zoltán Böszörményi <zboszor@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8180>

st/mesa: don't use cso_context to restore VBs, sampler views for glDrawPixels

These cso_context capabilities will be removed.

Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Zoltán Böszörményi <zboszor@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8180>

st/mesa: don't use cso_context to restore VBs for glClear

These cso_context capabilities will be removed.

Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Zoltán Böszörményi <zboszor@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8180>

st/mesa: don't use cso_context to restore VBs, sampler views for glBitmap

These cso_context capabilities will be removed.

Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Zoltán Böszörményi <zboszor@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8180>

gallium/pp: don't use cso_context to restore VBs, constbuf 0 and sampler views

These cso_context capabilities will be removed.

Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Zoltán Böszörményi <zboszor@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8180>

gallium/hud: don't use cso_context to restore VBs, constbuf 0 and sampler views

These cso_context capabilities will be removed.

Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Zoltán Böszörményi <zboszor@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8180>

gallium/api: add state invalidate interface as alternative to cso_save/restore

Some cso_context save/restore capabilities will be removed to decrease
CPU overhead. This is the alternative solution that e.g. the HUD will use.

Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Zoltán Böszörményi <zboszor@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8180>

gallium/tests: stop using cso_set_sampler_views

It will be removed.

Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Zoltán Böszörményi <zboszor@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8180>

st/xa: stop using cso_set_sampler_views

It will be removed.

Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8180>

st/nine: stop using cso_set_sampler_views

It will be removed.

Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8180>

ci: mark some sparse CTS as expected failures on RAVEN

Same as Vega10.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8518>

android: r600/sfn: add sfn_nir_lower_64bit.cpp to Makefile.sources

Fixes the following building errors:

FAILED: out/target/product/x86_64/obj/SHARED_LIBRARIES/gallium_dri_intermediates/LINKED/gallium_dri.so
...
ld.lld: error: undefined symbol: r600::r600_nir_split_64bit_io(nir_shader*)
>>> referenced by sfn_nir.cpp:981 (external/mesa/src/gallium/drivers/r600/sfn/sfn_nir.cpp:981)
...
ld.lld: error: undefined symbol: r600::r600_nir_64_to_vec2(nir_shader*)
>>> referenced by sfn_nir.cpp:1010 (external/mesa/src/gallium/drivers/r600/sfn/sfn_nir.cpp:1010)
...
ld.lld: error: undefined symbol: r600::r600_merge_vec2_stores(nir_shader*)
>>> referenced by sfn_nir.cpp:1015 (external/mesa/src/gallium/drivers/r600/sfn/sfn_nir.cpp:1015)

Fixes: 165fb5117b ("r600/sfn: add lowering passes to get 64 bit ops lowered to 32 bit vec2")
Cc: 21.0 <mesa-stable>
Signed-off-by: Mauro Rossi <issor.oruam@gmail.com>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8505>

radv: do not invalidate the L2 metadata cache on compute queues

The flush VA space was only allocated for command buffers on the
graphics queue. Also, the ZPASS_DONE event should never be emitted
on compute queues because it hangs.

Invalidating the L2 metadata cache is only required for coherency
between the RBs and L2, so only on the graphics queue.

The L2 cache is invalidated at beginning of any IBs and that should
also invalidate the L2 metadata cache for compute anyways.

Fixes: 4a783a3c ("radv: Use L2 coherency on GFX9+.")
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/8494>

ci/freedreno: Remove some long-unseen a6xx known flakes.

We haven't seen any of these since June.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8510>

ci/freedreno: Drop skip list stuff from a5xx flakes.

I clearly copied the wrong file when setting this up.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8510>

ci/freedreno: Drop some long-unseen a5xx flakes.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8510>

ci/freedreno: Remove a bunch of stale flakes from a3xx.

I haven't seen these in the CI channel in many months.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8510>

ci/freedreno: Fix up the xfail/flake handling of a3xx texture functions.

There's only one that's been unstable for the last couple months, and it's
almost always a fail, so just mark that one's state.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8510>

ci/freedreno: Mark some a5xx separate_shader tests as xfails.

These almost always fail for us, so we get flake complaints on basically
every job.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8510>

gallium/ntt: Avoid referencing undefined channels of system values.

Fixes failure to compile the shader on the host with gl_NumWorkGroups (for
example) on virgl.

Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8488>

gallium/ntt: Add support for store_per_vertex_output.

Needed by virgl and r600 to use NTT.

Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8488>

gallium/ntt: Don't vectorize IBFE/UBFE/BFI.

Since our source language only allows scalar offset/bits arguments,
virglrenderer only handled scalar values too. Just refuse to vectorize
these to prevent breakage.

Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8488>

mesa/st: Dump nir-to-tgsi output when ST_DEBUG=tgsi or nir is set.

Suggested by Gert Wollny, and oh how I've wanted this.

Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8488>

mesa/st: Lower shader images before handing off to NIR-to-TGSI.

Otherwise you'll just throw unknown intrinsic errors on drivers ignorant
of this NIR option.

Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8488>

mesa/st: Make a single helper for the NIR-to-TGSI transfer.

The common path means that we get NIR IR dumping on builtin shaders, too.

Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8488>

frontend/dri: Expose RGB[AX]_SRGB as well

platform_android prefers RGB[AX] over BGR[AX] to avoid buggyness with
older android versions. But we need to include the corresponding SRGB
formats as well.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/2888
Signed-off-by: Rob Clark <robdclark@chromium.org>
Acked-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8452>

zink: decouple renderpass from framebuffer state

a framebuffer is based on the surfaces attached to it

a renderpass is based on the operations being performed

a zink_framebuffer object doesn't necessarily need to be changed
if the renderpass changes, as this might just indicate that a clear operation
is being used instead of a load, so we can cache the corresponding vk objects
into the zink_framebuffer based on the renderpass ops being used to (eventually)
reduce fb object thrashing by doing more incremental updates

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8227>

zink: use 'fb' variable name for zink_framebuffer objects in zink_framebuffer.c

let's be a bit consistent and intuitive

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8227>

zink: rename param in zink_create_framebuffer

this is consistent with other state naming

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8227>

zink: ralloc zink_framebuffer structs

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8227>

zink: remove renderpass refcounting

this is useless because we never destroy renderpasses during the lifetime
of a process and don't want to do so

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8227>

zink: add batch flag for checking renderpass state

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8227>

zink: rework framebuffer state

we don't want to actually be tracking any null surfaces, so instead use
the vk object for this purpose

also store the fb state onto the fb object so we have this info available
at all times

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8227>

ir3: add debug option to override shader assembly

IR3_SHADER_DEBUG=vs,tcs,tes... now also prints shader's sha1.

When there is a file named %sha1%.asm in IR3_SHADER_OVERRIDE_PATH
directory - ir3 assembly from file would be parsed, assembled, and
will override the shader with corresponding sha1 hash.

Parsing failure is considered unrecoverable error.
Upon successful override shader's assembly is printed with:
"Native code (overridden) for unnamed ..."

This debug option allows easier testing of small changes in
assembly without modifying the compiler or using computerator.

Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8388>

zink: only emit streamout targets during draw if we have them

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8503>

mesa: always set valid index bounds for non-indexed draws for classic drivers

This should fix crashing piglit tests on the i915 classic driver.

Fixes: 2358da81d26 - mesa: switch (Multi)DrawArrays to DrawGallium

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8453>

iris: Fix android build due to missing link to libmesa_iris_gen125

Reported-by: Yugang Fan
Fixes: cd3251d6baf ("intel/iris: Build gen 12.5")
Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8426>

turnip: make GS use correct varyings size from previous stage

Fixes:
dEQP-VK.tessellation.invariance.primitive_set.triangles_fractional_even_spacing_ccw
dEQP-VK.tessellation.invariance.outer_edge_division.triangles_fractional_even_spacing
dEQP-VK.tessellation.invariance.outer_edge_symmetry.triangles_fractional_odd_spacing_cw
dEQP-VK.tessellation.invariance.outer_edge_symmetry.quads_fractional_odd_spacing_ccw
dEQP-VK.tessellation.invariance.outer_edge_symmetry.isolines_equal_spacing_cw
dEQP-VK.tessellation.invariance.outer_edge_index_independence.triangles_equal_spacing_ccw
dEQP-VK.tessellation.invariance.outer_edge_index_independence.triangles_fractional_even_spacing_cw
dEQP-VK.tessellation.invariance.inner_triangle_set.triangles_equal_spacing

Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8497>

iris: only set point sprite overrides if actually using points

Fixes black screen in some FNA games.

Cc: <mesa-stable@lists.freedesktop.org>
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/3431
Signed-off-by: Yevhenii Kolesnikov <yevhenii.kolesnikov@globallogic.com>
Reviewed-by: Nanley Chery <nanley.g.chery@intel.com>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7218>

zink: partially enable SSBO pipe cap

this enables atomic counters

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8330>

zink: add more usage bits for buffer types

these things can be anything they want to be when they grow up

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8330>

zink: flag ssbo buffer resources as having pending writes on batch

ssbos are the only descriptor type we support (so far) that allows writes
during the draw, so we have to ensure that we set the right flag on the batch
reference to handle sync if the buffer is later read from

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8330>

zink: split UBOs and samplers into 'read' batch references during draw

we have the technology, so now we can just flag these resources as being
read from instead of also being written to, freeing them up for concurrent reads
without requiring fencing

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8330>

zink: start supporting atomic shader ops

this handles atomic_add, which is what atomic counters use

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8330>

zink: modify ubo loading in ntv to work for ssbos

if ssbo variables are declared using the same structure as ubos, then
the mechanics here are identical, and we can just add a couple extra lines
to handle the atomic ops and different array usage

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8330>

zink: emit ssbo variables in ntv

we're structuring these the same as ubos so we can reuse the same load
function later

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8330>

zink: hook up ssbo shader bindings

this is just the struct zink_shader binding point stuff, but nothing is
using it yet

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8330>

zink: add set_shader_buffers pipe_context method

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8330>

zink: add util function for emitting ntv atomic ops

these all need specific memory params, but otherwise they're the same
as normal ops, so we can just make a helper function here to add in the memory
params

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8330>

zink: add spirv builder function for runtime array type

we'll need this for ssbos

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8330>

nir/lower_uniforms_to_ubo: set explicit_binding on uniform_0

this variable is always bound to buffer index 0, so the binding info
here is actually useful

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7935>

zink: optimize renderpass hash table

only the existing render targets need to be used for table entries

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8487>

util/set: split off create() into an init() function

this brings parity with the matching hash_table api

Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8450>

util/set: add the found param to search_or_add

this brings parity with the internal api

Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8450>

util/set: optimize rehash for empty table and no-func clears

if the table is filled with deleted entries, we don't need to rzalloc+free an identical
block of memory for the table, we can just memset the existing one

the same applies to table clears without a function passed in that the table
doesn't need to be iterated and can just be memset

Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8450>

util/hash_table: optimize rehash for empty table and no-func clears

if the table is filled with deleted entries, we don't need to rzalloc+free an identical
block of memory for the table, we can just memset the existing one

the same applies to table clears without a function passed in that the table
doesn't need to be iterated and can just be memset

Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8450>

radeonsi: invalidate compute sgprs in si_rebind_buffer

If we don't tag compute sgpr as dirty they will point to the
ol buffer location.
This fixes arb_compute_shader-dlist with mcbp enabled.

Fixes: 85a6bcca615 ("radeonsi: pass at most 3 images and/or shader buffers via user SGPRs for compute")
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8433>

ac: add ifdef __cplusplus guard to header

ac_shadowed_regs.h can be included from si_state_draw.cpp so this commit
adds the needed guards.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8433>

ci: Use meson test directly instead of ninja test

The former allows specifying how many processes to spawn for tests. The
latter seems to spawn (up to) as many test processes as there are CPU
cores.

Acked-by: Daniel Stone <daniels@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8451>

ci: Remove .gitlab-ci/meson-build.bat

Unused since 07885cbcdb0b "CI: Add native Windows VS2019 build"

Acked-by: Daniel Stone <daniels@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8451>

anv: add transfer usage for color/depth/stencil attachments

We sometimes use anv_layout_to_aux_state() to compute the aux state of
an image during the resolve operations at the end of a render
(sub)pass.

If we're dealing with a multisampled image that is created without a
transfer usage, our internal code might trigger a resolve using the
transfer layout (see genX_cmd_buffer.c:cmd_buffer_end_subpass), for
which the image doesn't the usage bit. The current code tries to AND
the 2 usages which won't have any bit in common, thus skipping all
checks below.

v2: Add the transfer usages depending on attachment usage (Lionel)

v3: Limit to samples > 1 (Jason) && DEPTH_STENCIL_ATTACHMENT_BIT (Lionel)

v4: Add transfer usage at image creation (Jason)

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Fixes: 54b525caf0aa99 ("anv: Rework anv_layout_to_aux_state")
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/4037
Reviewed-by: Reviewed-by: Tapani Pälli <tapani.palli@intel.com> (v1)
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8307>

turnip/ir3: handle image load/stores produced by AtomicLoad/Store

SpvOpAtomicLoad and SpvOpAtomicStore are translated into
nir_intrinsic_image_deref_store/load instead of some separate
atomic intrinsics, however they don't have src or dest type
specified. Turnip doesn't support shaderImageFloat32Atomics
so type is just integer.

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

Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8476>

lavapipe: Defer lavapipe warning to CreateDevice

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/4055
Fixes: b38879f8c5f57b7f1802e433e33181bdf5e72aef
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8377>

panfrost: Fix typos.

Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8400>

gallium/tgsi: garbage collect unused TGSI_UTIL_SIGN_MODE.

Noticed while git grepping for abs/neg stuff.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8369>

gallium/tgsi: Rewrite the docs on source modifiers.

Clarify what the rules are, and warn about the exceptions.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8369>

gallium/tgsi: Remove support for f64 src modifiers.

The tgsi.rst was not very clear but didn't indicate that they were
supported, and llvmpipe only supported double abs and not negate.
glsl_to_tgsi never generated them, and neither did the new nir_to_tgsi, so
just garbage collect it.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8369>

gallium/tgsi: Assert that we don't see integer abs modifiers.

tgsi.rst says it's not supported, and llvmpipe, r600, and virgl don't
support it. Make sure nobody else introduces them like I did while
testing only on softpipe.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8369>

gallium/ntt: Stop lowering integer source mods.

While tgsi_exec.c (softpipe) implemented 32b integer src mods, the
tgsi.rst documentation says only 32b negate is supported and not abs.
llvmpipe implemented 32 and 64 negate but not abs, virgl implemented
negate incorrectly, and r600 apparently doesn't do any integer src mods.
glsl_to_tgsi has apparently never generated integer src mods.

Given that r600 can't do any integer src mods, just stop trying to
generate them for TGSI.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8369>

gallium/ttn: Add support for TGSI_OPCODE_I64NEG/ABS.

Found when converting AMD's built-in TGSI shaders to not using 64-bit src
mods.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8369>

nv50,nvc0: explicitly list recently-added caps

Some of these should be implemented, but that can be done in a later
change.

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/8461>

nv50/ir: ignore FS_BLEND_EQUATION_ADVANCED

It's unsupported, but gets set unconditionally by the state tracker,
even when fbfetch is to be used.

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/8461>

zink: clamp sampler+samplerview limits

* struct shader_info provides a 32bit mask for textures_used
* samplers and samplerviews are a 1:1 mapping for shader descriptors
* also according to spec this is always 32 per stage

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8457>

intel/genxml,anv,iris: Drop the legacy compute path from gen125.xml

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8342>

iris: Add support for COMPUTE_WALKER

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8342>

anv: Enable push constants on gen12-hp

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8342>

intel/fs: Emit code for Gen12-HP indirect compute data

Reworks:
* Jordan: Apply to gen > 12
* Jordan: Adjust comment about loading constants

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8342>

anv: Add a general state pool

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8342>

anv: Use COMPUTE_WALKER for gen12-hp

Rework: Rafael Antognolli, Jason Ekstrand
Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8342>

anv: Don't use MEDIA_INTERFACE_DESCRIPTOR_LOAD for gen12-hp

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8342>

anv: Emit CFE_STATE for gen12-hp

Rework:
* make scratch TODO. (Jason)
* emit_compute_cs_state => emit_compute_state. (Jason)

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8342>

intel/fs: Allow compute dispatch without a pushed subgroup ID on Gen12-HP

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8342>

intel/compiler: Disable push constants on gen12-hp

We currently don't use push constants with the COMPUTE_WALKER command.
Make all uniforms to be pull constants.

The local group id previously was a push constant, but is now
available in R0.2[7:0].

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8342>

intel/tools: Decode COMPUTE_WALKER

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8342>

intel/genxml/gen125: Add CFE_STATE and COMPUTE_WALKER

Reworks:
* Jason: Captalize COMPUTE_WALKER::BTD mode
* Jason: Make COMPUTE_WALKER::InlineData an array
* Jason: Remove stale fields, uint=>bool, add missing enum values.

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8342>