nir/algebraic: Make some notes about comparison rearrangements versus infinity

The original comment was a little terse and a little incorrect.  The
rearrangements are fine w.r.t. NaN.  However, they produce incorrect
results if one operand is +Inf and the other is -Inf.

A later commit, "nir/algebraic: Add some compare-with-zero optimizations
that are exact", will add some more patterns here.  It may be reasonable
to squash this commit (forward) into that commit.

v2: Fix some incorrect comparisons operators in the comment (<= vs >=).
Add commentary that subtraction works like addition w.r.t. NaN.  Both
noticed / suggested by Caio.

Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6358>

nir: Make some notes about fsign versus NaN

This commit only documents the current behavior, even if that behavior
is not the behavior preferred by the relevant specs.

In SPIR-V, there are two flavors of the sign instruction, and each lives
in an extended instruction set.  The GLSL.std.450 FSign instruction is
defined as:

    Result is 1.0 if x > 0, 0.0 if x = 0, or -1.0 if x < 0.

This also matches the GLSL 4.60 definition.

However, the OpenCL.ExtendedInstructionSet.100 sign instruction is
defined as:

    Returns 1.0 if x > 0, -0.0 if x = -0.0, +0.0 if x = +0.0, or -1.0 if
    x < 0. Returns 0.0 if x is a NaN.

There are two differences.  Each treats -0.0 differently, and each also
treats NaN differently.  Specifically, GLSL.std.450 FSign does not
define any specific behavior for NaN.

There has been some discussion in Khronos about the NaN behavior of
GLSL.std.450 FSign.  As part of that discussion, I did some research
into how we treat NaN for nir_op_fsign, and this commit just captures
some of those notes.

v2: Document the expected behavior of nir_op_fsign more thoroughly.
Suggested by Rhys.  Note that the current implementation of constant
folding does not produce the expected result for NaN.  Suggested by
Caio.

Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com> [v1]
Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6358>

st/mesa: don't affect original st_CompressedTexSubImage parameters

The fallback path is still possible here so let keep them as is.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/3952
Fixes: 4b02f165 ("st/mesa: implement PBO upload for glCompressedTex(Sub)Image")
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Signed-off-by: Andrii Simiklit <andrii.simiklit@globallogic.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8258>

gallium: remove PIPE_CAP_INFO_START_WITH_USER_INDICES and fix all drivers

Drivers aren't allowed to ignore start with user index buffers anymore.
This is required by the new fast path where mesa/main is using pipe_draw_info.

Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

st/mesa: implement Driver.DrawGallium callbacks

This is the new fast path replacing the _mesa_prim path.

Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

virgl: fix handling draw info

index_bias is undefined if index_size == 0.
index bounds are undefined if index_bounds_valid == false.

Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

v3d: don't use index_bias if not indexed

index_bias is undefined if index_size == 0.

Reviewed-by: Jose Maria Casanova Crespo <jmcasanova@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

vc4: don't use index_bias if indexed

index_bias is undefined if index_size == 0.

Reviewed-by: Jose Maria Casanova Crespo <jmcasanova@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

svga: fix handling draw info

index_bias is undefined if index_size == 0.
index bounds are undefined if index_bounds_valid == false.

Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

swr: fix handling draw info

index_bias is undefined if index_size == 0.
index bounds are undefined if index_bounds_valid == false.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

r600: fix handling draw info

has_user_indices is undefined if index_size == 0.

Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

panfrost: don't use index_bias if not indexed

index_bias is undefined if index_size == 0.

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

nouveau: fix handling draw info

index_bias is undefined if index_size == 0.
index bounds are undefined if index_bounds_valid == false.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

iris: don't use index_bias if not indexed

index_bias is undefined if index_size == 0.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

freedreno: fixes handling draw info

index_bias is undefined if index_size == 0.
index bounds are undefined if index_bounds_valid == false.

Reviewed-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

gallium: fix draw info setup in draw and utilities

index_bias is undefined if index_size == 0.
index bounds are undefined if index_bounds_valid == false.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

gallium: remove and emulate PIPE_CAP_MULTI_DRAW

To remove PIPE_CAP checking in the common code.

It's better if drivers lower multi draws even if the hardware doesn't
support it beause the multi draw loop can be moved deeper into the driver
to remove more overhead.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

gallium/u_threaded: clear vertices_per_patch if prim type != PATCHES

We use memcmp when combining back-to-back draws into a multi draw.
Clear vertices_per_patch if it's irrelevant to help memcmp return
a match.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

vbo: switch immediate Begin/End to DrawGallium

This makes gallium faster because st/mesa doesn't have to translate
_mesa_prim.

Reviewed-by: Zoltán Böszörményi <zboszor@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

vbo: remove _mesa_prim parameter from vbo_copy_vertices

glBegin/End won't use _mesa_prim, so we need to stop using it.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

vbo: remove _mesa_prim parameter from vbo_merge_draws

glBegin/End won't use _mesa_prim, so we need to stop using it.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

vbo: remove _mesa_prim parameter from vbo_try_prim_conversion

glBegin/End won't use _mesa_prim, so we need to stop using it.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

mesa: switch MultiDrawElements(BaseVertex) to DrawGallium*

This makes gallium faster because st/mesa doesn't have to translate
_mesa_prim.

Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

mesa: switch Draw(Range)Elements(BaseVertex) calls to DrawGallium

This makes gallium faster because st/mesa doesn't have to translate
_mesa_prim.

Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

mesa: switch (Multi)DrawArrays to DrawGallium

This makes gallium faster because st/mesa doesn't have to translate
_mesa_prim.

Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

vbo: add vbo_get_minmax_indices_gallium

to be used by st/mesa to get index bounds because it won't have _mesa_prim
with the new draw interface.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

mesa: add a fallback for drivers not implementing Driver.DrawGallium*

This is for classic drivers and it translates pipe_draw_info to the old
interface using _mesa_prim.

Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

gallium: add pipe_draw_info::index::gl_bo

mesa/main will fill pipe_draw_info but it has no access to pipe_resources,
so we need gl_buffer_object here.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

mesa: add Driver.DrawGallium* functions to be used by main/draw.c

This is a new Draw interface using the same parameters as draw_vbo.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

vbo: change the parameters of vbo_get_minmax_index to get rid of _mesa_prim

This will be used by a new Draw interface.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

vbo: only set count and end when closing _mesa_prim

It's already set everywhere except vbo_exec_wrap_buffers.
This is a small optimization removing superfluous code.

Reviewed-by: Zoltán Böszörményi <zboszor@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7679>

mesa: fix an overflow check for MultiDrawElements

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/7679>

CI: Enable d3d12 driver for Linux CI builds

Acked-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8251>

CI: Install DirectX-Headers package for x86 container

Acked-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8251>

drisw: Use debug_screen_wrap like everybody else

I tried to use GALLIUM_TRACE with llvmpipe and it didn't work, because
we were never letting the debug wrappers initialize.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8166>

microsoft/clc: Fix const violations from ralloc_steal

Fixes: ff05da7f8dc ("microsoft: Add CLC frontend and kernel/compute support to DXIL converter")

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7697>

virgl: Enable GL_EXT_texture_sRGB_RG8

Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8060>

zink: Enable GL_EXT_texture_sRGB_RG8

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8060>

zink: Enable GL_EXT_texture_sRGB_R8

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8060>

mesa: Implement GL_EXT_texture_sRGB_RG8 for softpipe and llvmpipe

sRGB_RG8 is not registered for big-GL yet, see this Khronos issue
for updates on that:

https://github.com/KhronosGroup/OpenGL-Registry/issues/450

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8060>

mesa: Cosmetic cleanups to GL_EXT_texture_sRGB_R8

This extension dates to 2015 not 2018, and since there's only one
texture format to check for the additional GL_TRUE is meaningless.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8060>

treewide: Disambiguate various variables named "debug_options"

Name them after what they control so 'vi -t' can take you somewhere
useful.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8165>

zink: handle gl_SampleMaskIn loading in ntv

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8316>

zink: set PIPE_CAP_MAX_VIEWPORTS

enables ARB_viewport_array

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8311>

zink: handle fs interpolation functions in ntv

basically load_deref but with different ops

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8203>

zink: only emit each cap once

This just prevents us from emitting some caps a bunch of times. Since
there's few caps in use, it seems reasonable to use this mechanism for
all caps, as it's unlikely to add up to anything worth worrying about.

Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8312>

android: pan/bi: Remove old IR opcode table

Fixes the following building error:

clang: error: no such file or directory: 'external/mesa/src/panfrost/bifrost/bi_tables.c'
clang: error: no input files

Fixes: b691aeaacf93 ("pan/bi: Remove old IR opcode table")
Signed-off-by: Mauro Rossi <issor.oruam@gmail.com>
Acked-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8294>

android: pan/bi: Remove NIR->old IR

Fixes the following building error:

clang: error: no such file or directory: 'external/mesa/src/panfrost/bifrost/bi_special.c'
clang: error: no input files

Fixes: e5ec0dc18320 ("pan/bi: Remove NIR->old IR")
Signed-off-by: Mauro Rossi <issor.oruam@gmail.com>
Acked-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8294>

android: pan/bi: Remove old IR packs

Necessary for Android build after commit a1e150fc4d4e

Fixes: a1e150fc4d4e ("pan/bi: Remove old IR packs")
Signed-off-by: Mauro Rossi <issor.oruam@gmail.com>
Acked-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8294>

android: pan/bi: Remove combine lowering

Fixes the following building error:

clang: error: no such file or directory: 'external/mesa/src/panfrost/bifrost/bi_lower_combine.c'
clang: error: no input files

Fixes: 8b0d0a931b80 ("pan/bi: Remove combine lowering")
Signed-off-by: Mauro Rossi <issor.oruam@gmail.com>
Acked-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8294>

android: pan/bi: Generate instruction packer for new IR

Necessary for Android build after commit 82328a524519

Fixes: 82328a524519 ("pan/bi: Generate instruction packer for new IR")
Signed-off-by: Mauro Rossi <issor.oruam@gmail.com>
Acked-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8294>

android: pan/bi: Generate builder routines

Fixes the following building error:

external/mesa/src/panfrost/bifrost/bifrost_compile.c:38:10: fatal error: 'bi_builder.h' file not found
         ^~~~~~~~~~~~~~
1 error generated.

Fixes: 8ef0d411708a ("pan/bi: Generate builder routines")
Signed-off-by: Mauro Rossi <issor.oruam@gmail.com>
Acked-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8294>

android: pan/bi: Generate instruction printer

To avoid building errors in Android

Fixes: 1893a3805e41 ("pan/bi: Generate instruction printer")
Signed-off-by: Mauro Rossi <issor.oruam@gmail.com>
Acked-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8294>

radv: disable A2 SNORM/SSCALED/SINT for texel buffers & images on all gens

AMDVLK and AMDGPU-PRO also don't support these formats for texel
buffers and images.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/3386
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/8315>

aco: fix incorrect address calculation for load_barycentric_at_sample

Fix address calculation for indirect load_barycentric_at_sample on GFX6-8
with a uniform sample index.

A non-zero uniform sample index does not seem to be tested by CTS.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Gitlab: https://gitlab.freedesktop.org/mesa/mesa/-/issues/3966
Fixes: 93c8ebfa780 ("aco: Initial commit of independent AMD compiler")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8302>

pan/bi: Fix TEXS register counts

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Fixes: cffc1d90da0 ("pan/bi: Add staging register counts to ISA.xml")
Reported-by: Icecream95
Tested-by: macc24
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8310>

panfrost: Allow 3D AFBC on Bifrost v7

This feature doesn't seem to work properly on Midgard, and is flagged
as unsupported on Bifrost v6.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: Fix calculation of body/header pointers for 3D AFBC

When using 3D AFBC, all headers are placed at the beginning instead of
being interleaved with each surface body, which forces us to adjust
the calculation in that case.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: Allow AFBC on 2D arrays

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: Fix panfrost_should_linear_convert()

Take 3D/array textures into accound when checking for entire overwrite.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: Initialize AFBC headers to zero

This way we can load uninitialized AFBC surfaces without causing GPU
faults.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: Use panfrost_get_layer_stride() instead of open-coding it

If we don't do that we don't account for CRC buffers.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: Enable AFBC support on Bifrost

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: Fix AFBC support on Bifrost

Bifrost and Midgard render target and ZS extension descriptors are a bit
different, adjust the code to take those differences into account.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: Use proper format for Z16_UNORM

R16_UNORM isn't compatible with AFBC, we need to use the native Z16
format if we want to allow AFBC on those resources.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: Fix ZS block format v7 definition

This field takes 4 bits, not 2.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: Adjust the format for AFBC textures on Bifrost v7

On Bifrost v7, AFBC textures can't be used with a non-identity component
order. Let's patch the format so the component order is always RGB[A].
That means we're lying about the internal format, but that shouldn't be
a problem as long as we don't share the resource.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: Pass a pipe-like swizzle to panfrost_new_texture()

And let the function translate it to a mali swizzle. This way we will
be able to adjust the swizzle if needed.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: Add R5G6B5_UNORM entries to the format tables

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: Adjust surface stride calculation to take AFBC into account

AFBC strides are different from tiled/linear stride and we need to
use the value defined in slice.afbc.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: Add AFBC slice.body_size and slice.{row,surface}_stride fields

Those are needed for render target and texture descriptors and can't be
easily extracted from the other fields present in panfrost_slice

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: Fix AFBC header_size and slice size calculation

3D AFBC textures have their AFBC headers grouped together at the
beginning of the buffer which means the header_size should be multiplied
by the depth. 2D arrays have their AFBC headers placed at the beginning
of each slice, meaning that the slice size should take them into account.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: Move AFBC header_size to a sub-struct

So we can add more AFBC related fields under this struct.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: Add a pan_image_layout object

Group the slices, dimension, modifier and array stride in a an object
representing the image layout. This way we shrink the number of
arguments passed to various pan_texture helpers and simplifies some
of the logic along the way.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: Stop mixing depth and number of samples

Texture depth and MSAA are two different concepts even if they are
exclusive on Mali GPUs (depth field is repurposed for sample index
there). Let's not mix them and adjust the slice_full_size calculation
to take both into account.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: Use PAN_V6_SWIZZLE() in pan_blit.c

Suggested-by: Icecream95
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: s/panfrost_slice.size0/panfrost_slice.surface_stride/

This field is encoding the stride between two consecutive surface
in a 3D texture. Let's use a name reflecting that.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: Merge panfrost_new_texture() and panfrost_new_texture_bifrost()

There's no reason to keep both now that the internal logic has been
deduplicated.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: Adjust the compression tag creation for Bifrost

Bifrost has a few more compression flags that are worth specifying.
Extend panfrost_compression_tag() to deal with those too.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: Define AFBC surface flags

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: Pass a dev object to panfrost_needs_explicit_stride()

So we can soon use the same path for Bifrost and Midgard.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: Merge emit_texture_payload() and emit_texture_payload_v7()

The only reason we have two different functions to prepare the texture
payload is the different ordering between pre-v7 and v7+ GPUs. Abstract
the surface iteration so we can merge panfrost_emit_texture_payload_v7()
and panfrost_emit_texture_payload().

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: Pass a device object to panfrost_new_texture()

We will need it to merge some of the Bifrost and Midgard logic.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: Group CRC fields in a struct

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

panfrost: Move checksum_bo to panfrost_resource

There's no reason to have the checksum_bo at the slice level since there
can only be one external CRC BO per resource. Move this field to the
panfrost_resource struct.

Suggested-by: Icecream95
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8125>

features: mark off textureGather for zink

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8151>

zink: enable PIPE_CAP_MAX_TEXTURE_GATHER_COMPONENTS

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8151>

zink: handle textureGather with Shadow-type samplers

these have to go through OpImageDrefGather without composite construction
on the result

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8151>

zink: implement ARB_texture_gather

again pretty straightforward, just hooking up tg4 tex op in ntv

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8151>

zink: use Offset param for txf ops

there's an image operand for handling offsets, so we can use this when it
hasn't previously been lowered away

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8151>

zink: run nir_lower_tex for offsets if shaderImageGatherExtended is missing

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8151>

nir: fix missing nir_lower_pntc_ytransform.c in the makefile

Fixes: 33fd9e5d "nir: account for point-coord origin when lowering it"
Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8308>

zink: handle dynamic sampler array indexing for arb_gpu_shader5

this requires that arrays of samplers be declared as single variables with
a single binding point, which is then propagated through to the descriptor
set updates

constant sampler array indexing is now un-lowered during access so we can
construct an access chain for both constant and dynamic offset paths

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8155>

radv: avoid oob read during clear

when clearing a depth/stencil image the passed colorvalue pointer is
smaller than the VkClearValue struct size

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8288>

nir: account for point-coord origin when lowering it

The resulting point-coord origin not only depends on whether
the draw buffer is flipped but also on GL_POINT_SPRITE_COORD_ORIGIN
state. Which makes its transform differ from a transform of wpos.

On freedreno fixes:
 gl-3.2-pointsprite-origin
 gl-3.2-pointsprite-origin -fbo

Fixes: d934d320 "nir: Add flipping of gl_PointCoord.y in nir_lower_wpos_ytransform."
Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Acked-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Jose Maria Casanova Crespo <jmcasanova@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8200>

radv: Use VRAM for the initial gfx cmdbuffer.

Not expect it to make any real difference, but lets be consistent.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7979>

radv: Put commandbuffers in VRAM if all VRAM is CPU visible.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7979>

radv: Use VRAM for upload buffers if entire VRAM is CPU-visible.

Not doing this for APUs because spilling is quite likely, due to
overall VRAM pressure.

Also adding a flag to disable for performance debugging.

Finally adds some memset for places where we depended on the memory
being initialized to zero, which we won't get with VRAM anymore.
(I think these places should stop depending on it since it hides
 issues with executing the cmdbuffer multiple times, but this
  preserves behavior)

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7979>

microsoft/compiler: Fix clang fallthrough warnings

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8248>

microsoft/resource_state_manager: Remove unused private variable

Clang detects that this variable is unused. We can re-add it if
we end up needing it but for now let's just remove it.

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8248>