r600/sfn; go back to not lowering uniforms to UBOs

Lowering uniforms to UBOs results in an aditional iadd for the
UBO buffer id evaluation, and for indirect buffers access that
results in an unnecessary op that can be avoided by not lowering
uniforms. There is some code duplication when reading the uniforms
but it saves a whole instruction group per indirect cont buffer
access.

This reverts commit 98eb00face93b9af5aac19008ecff5a2bf376745 with
some additional fixes.

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6879>

r600/sfn: Fix indirect const buffer access

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6879>

r600/sfn: Fix interpolate at sample

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6879>

r600/sfn: use fine gradient evaluation for interpolate_at_offset

Also pass in the interpolator x component (because the TGSI does this too)

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6879>

r600/sfn: save some instructions when doing multisample on sample 0

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6879>

r600/sfn: Support group memory barrier

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6879>

r600/sfn: lower to scalar for some optimizations and vectortize later

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6879>

r600/sfn: Handle nir_op_b2b32

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6879>

r600/sfn: Fix emitting shared atomics with constant sources

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6879>

nir/lower_clip: Add i/o semantics for load/store intrinsics.

ir3 looks at the .location on its inputs for handling
non-VARYING_SLOT_POS, but our uninitialized semantics matched that and
threw a compiler assertion failure.

Fixes: 502abfce7f5d ("nir: save IO semantics in lowered IO intrinsics")
Reviewed-by: Connor Abbott <cwabbott0@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6716>

ci/freedreno: Add trace tests for glxgears, 0 A.D., and xonotic.

glxgears is now trimmed to not be tons of frames, and we've picked up a
couple more traces in the public repo.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6716>

ci/freedreno: Sort the traces in the .yml of expectations

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6716>

turnip: Replace tu_log*() with mesa_log*()

This gets us logging on Android.

Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6806>

mesa: Promote Intel's simple logging façade for Android to util/

I'm bringing up freedreno Vulkan on an Android phone, and my pains are
exactly what Chad said when working on Intel's vulkan for Android in
aa716db0f64d ("intel: Add simple logging façade for Android (v2)"):

    On Android, stdio goes to /dev/null. On Android, remote gdb is even
    more painful than the usual remote gdb. On Android, nothing works like
    you expect and debugging is hell. I need logging.

    This patch introduces a small, simple logging API that can easily wrap
    Android's API. On non-Android platforms, this logger does nothing
    fancy.  It follows the time-honored Unix tradition of spewing
    everything to stderr with minimal fuss.

    My goal here is not perfection. My goal is to make a minimal, clean API,
    that people hate merely a little instead of a lot, and that's good
    enough to let me bring up Android Vulkan.  And it needs to be fast,
    which means it must be small. No one wants to their game to miss frames
    while aiming a flaming bow into the jaws of an angry robot t-rex, and
    thus become t-rex breakfast, because some fool had too much fun desiging
    a bloated, ideal logging API.

Compared to trusty fprintf, _mesa_log[ewi]() is actually usable on
Android.  Compared to os_log_message(), this has different error levels
and supports format arguments.

The only code change in the move is wrapping flockfile/funlockfile in
!DETECT_OS_WINDOWS, since mingw32 doesn't have it.  Windows likely wants
different logging code, anyway.

Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6806>

radv: Record cache flushes for RGP.

Not doing the EOP TS cacheflush event because that break wave counting
in RGP for some reason. But the rest looks to be all there.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6550>

radv: Include flushes in the barrier.

Since the flushes really happen on the next draw delay the barrier
end to include the flushes.

This fixes the barrier duration in RGP.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6550>

radv: Fix RGP Asic CU info for GFX10+.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6550>

radv: Use correct alignment for SQTT buffer sizes.

Otherwise we can miss the start of the buffer, which
results in RGP missing the initial (and only) timestamp
packet and results in RGP completely ignoring the second
SE, causing messy issues and missed events.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6550>

radv: Fix emitting SQTT userdata.

Otherwise some marker packets were missing ...

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6550>

zink: remove old extension infrastructure

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6789>

zink: use the new, generated extension infrastructure

this is just a find-and-replace from screen->... to screen->info.

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6789>

zink: use the new extension infrastructure in device creation

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6789>

zink: hook zink_device_info.py to build system

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6789>

zink: generate extension infrastructure using a python script

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6789>

nir/lower_memory_model: do not break with global atomic operations

Global atomics don't have an access flag.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6888>

nir/lower_memory_model: return progress when visiting instructions

It never returned progress=TRUE.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6888>

pan/mdg: Infer whether to disassemble shaders from info.internal

Blit shaders are now marked as internal, so remove the silent argument
from midgard_compile_shader_nir and instead use nir->info.internal to
suppress disassembling shaders.

Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6870>

panfrost: Mark blit shaders as internal

Fixes spam of blit shader nir when NIR_PRINT=1 is set.

Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6870>

nir/load_store_vectorize: improve vectorization with identical operations

We might have several identical options to vectorize an entry with, but
only one might be vectorizable because of writes interfering.

An example of this is a pattern found in some CTS tests:
a = load(0)
b = load(4)
store(0, a)
store(4, b)
a = load(0)
b = load(4)
store(0, a)
store(4, b)
...

It might have attempted to vectorize the first load(0) with the second
load(4) without attempting the second load(4) when the first fails. This
changes vectorize_entries() to continue even if the first try_vectorize()
failed.

fossil-db (Navi):
Totals from 117 (0.09% of 137413) affected shaders:
SGPRs: 7040 -> 7088 (+0.68%)
CodeSize: 276504 -> 276308 (-0.07%); split: -0.08%, +0.01%
Instrs: 51152 -> 51111 (-0.08%); split: -0.09%, +0.01%

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5415>

panfrost: Remove extra printf arguments.

Fix defects reported by Coverity Scan.

Extra argument to printf format specifier (PRINTF_ARGS)
extra_argument: This argument was not used by the format string: job +
32UL.

Fixes: 89fafe9e926b ("panfrost: XML-ify the fragment job descriptor")
Fixes: d289209ea68f ("panfrost: XML-ify the compute job descriptor")
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6829>

glsl/linker: Add support for XFB varying lowering in geometry shader

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6723>

docs: recognisable -> recognizable

Most of the docs is US English, let's stick to that.

Reviewed-by: Eric Engestrom <eric@engestrom.ch>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6864>

docs: url -> URL

Reviewed-by: Eric Engestrom <eric@engestrom.ch>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6864>

docs: gitlab -> GitLab

Reviewed-by: Eric Engestrom <eric@engestrom.ch>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6864>

docs: initialisation -> initialization

Most of the docs is US English, let's stick to that.

Reviewed-by: Eric Engestrom <eric@engestrom.ch>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6864>

docs: timeplan -> time plan

Reviewed-by: Eric Engestrom <eric@engestrom.ch>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6864>

docs: behaviour -> behavior

Most of our documentation is US English, so let's stick to that.

Reviewed-by: Eric Engestrom <eric@engestrom.ch>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6864>

docs: unecessarily -> unnecessarily

Reviewed-by: Eric Engestrom <eric@engestrom.ch>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6864>

docs: make two acronyms upper-case

Reviewed-by: Eric Engestrom <eric@engestrom.ch>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6864>

docs: frambuffer -> framebuffer

Reviewed-by: Eric Engestrom <eric@engestrom.ch>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6864>

docs: scons -> SCons

Reviewed-by: Eric Engestrom <eric@engestrom.ch>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6864>

docs: apis -> APIs

Reviewed-by: Eric Engestrom <eric@engestrom.ch>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6864>

docs: everytime -> every time

Reviewed-by: Eric Engestrom <eric@engestrom.ch>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6864>

docs: remove webmaster article

This article is out-of-date, so let's instead forward to the "About
Mesa3D.org" article on the main website. This explains where to submit
merge requests for wanted changes.

Reviewed-by: Eric Engestrom <eric@engestrom.ch>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6862>

docs: store prefixes in redirects

This allows us to link to articles that are using pretty-URLs without
adding an ugly "/index.html"-suffix to them.

Reviewed-by: Eric Engestrom <eric@engestrom.ch>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6862>

docs: show 'Edit on GitLab'-link

It's easier to contribute to the documentation if we have links to the
document on GitLab. This will allow people to easily edit docs, or to
realize where in the source-tree they are without having to search.

Reviewed-by: Eric Engestrom <eric@engestrom.ch>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6742>

docs: escape backquote character

This unescaped backquote charcter confuses Sphinx, making it emit a
warning and slightly confused HTML in the end.

Let's escape it instead, to keep things sane.

Fixes: fdff24d930d ("docs: add release notes for 20.1.8")
Reviewed-by: Eric Engestrom <eric@engestrom.ch>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6863>

lima: set clear depth value to 0x00ffffff as default

According to the specs, the initial depth value for a
depth buffer clear is 1. Use 0x00ffffff like the blob does.

We can remove setting this value in lima_clear, because it's
set during job creation now.

Fixes the following dEQP tests:

dEQP-GLES2.functional.fbo.render.shared_depthbuffer.rbo_rgb565_depth_component16
dEQP-GLES2.functional.fbo.render.shared_depthbuffer.rbo_rgb5_a1_depth_component16
dEQP-GLES2.functional.fbo.render.shared_depthbuffer.rbo_rgb4_depth_component16
dEQP-GLES2.functional.fbo.render.shared_depthbuffer.tex2d_rgba_depth_component16
dEQP-GLES2.functional.fbo.render.shared_depthbuffer.tex2d_rgb_depth_component16

Reviewed-by: Vasily Khoruzhick <anarsoul@gmail.com>
Signed-off-by: Andreas Baierl <ichgeh@imkreisrum.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6842>

android: pan/bi: fix typo in bifrost_gen_disasm.c gen rules

Gen rules were erroneusly named as bifrost_gen_disasm*s*
This patch fixes them

Fixes: 792b51713bf4 ("android: pan/bi: Use new disassembler")
Signed-off-by: Mauro Rossi <issor.oruam@gmail.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6873>

panfrost: Fix bifrost tiler job emission

The tiler descriptor pointer and padding sections were not filled.

Fixes: d289209ea68f ("panfrost: XML-ify the compute job descriptor")
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Tested-by: Christian Hewitt <christianshewitt@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6859>

panfrost: Fix bifrost tiler descriptor definition

The bifrost tiler descriptor definition was wrong, leading to a buffer
overflow.

Fixes: efce73d99d0e ("panfrost: XML-ify the bifrost tiler descriptors")
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Tested-by: Christian Hewitt <christianshewitt@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6859>

ci: add dEQP-VK.info.device_extensions to the list of skipped tests

It regularly fails when a new KHR extension is implemented.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6843>

clover: set LLVM min version to 8.0.1

It also bump from 8.0.0 to .1 for AMD gallium and VK

Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6336>

clover/nir: Call the memcpy lowering pass

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6713>

spirv: Add support for OpCopyMemorySized

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6713>

nir: Add a lowering pass to lower memcpy

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6713>

nir: Add a new memcpy intrinsic

This matches SPIR-V's OpCopyMemorySized

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6713>

spirv: Use the new types helpers

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6713>

compiler/types: Add glsl_baseN_t_type(bit_size) helpers

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6713>

gallivm: add support for lowered IO in vertex shaders

for Selection/Feedback/RasterPos in st/mesa.

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6698>

meson/anv: Use variable that checks for --build-id

fixes: d1992255bb29054fa51763376d125183a9f602f3
       ("meson: Add build Intel "anv" vulkan driver")

Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6819>

vtn/opencl: Switch fma to conditionally use libclc for 32bit floats

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>

vtn/opencl: Only use libclc ldexp when lower_ldexp is set

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>

vtn/opencl: Switch some nir-sequence ops to use libclc

All of these are pretty well-defined. Rather than implementing them
as a sequence of nir ops, we can just use the libclc implementation.

v2 (idr): Delete functions that are now unused.

Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>

vtn/opencl: Switch division-related ops to use libclc

Specifically, fmod only uses libclc if it was going to be lowered.
Also, add missing half_divide and half_recip handling.

Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>

vtn/opencl: Switch exp/pow/log to use libclc

Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>

vtn/opencl: Switch non-native trig to use libclc

Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>

vtn/opencl: Hook up OpenCL async copy and group wait opcodes via libclc

Adds an additional hook for spirv_to_nir to handle a core opcode via
the OpenCL libclc infrastructure, and adds handling for SpvOpGroupAsyncCopy and
SpvOpGroupWaitEvents.

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>

vtn/opencl: Rework handle_instr to be able to handle core SPIR-V opcodes via libclc

The OpenCL async copy/wait opcodes are core SPIR-V, rather than OpenCL extension opcodes.

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>

vtn/opencl: Implement a lot of opcodes via libclc

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>

clover: handle libclc shader (v3)

This works by taking the spirv produced by libclc which contains
a lot of mangled function entrypoints identified with LinkageAttribute decorations.

This patch just sets up clover to load the libclc blob and convert it to
library nir, and support inlining application nir with calls to libclc.

v2: Add a disk cache support for this object, to avoid the spirv parsing
overheads each time. move spirv->nir to lazy instantiation to avoid
the mess with glsl types and constructor ordering.
v3: make disk cache optional

v1-Reviewed-by: Jesse Natalie <jenatali@microsoft.com>

Reviewed-by: Karol Herbst <kherbst@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>

util: add missing extern C

This code is included in c++ code via disk_cache in theory,
in practice it never has been.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>

vtn/opencl: Add infrastructure for calling out to libclc

This patch adds a function remap table with name mangling, which
can convert a SPIR-V OpenCL extension opcode to a call to the external
libclc shader, which will be lowered/inlined after conversion.

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>

clover/nir: add libclc lowering pass

This pass goes through all the functions in the shader, checks
if the matching function is in the clc spir-v library and inlines
the replacement from there it is.

v2 (daniels): Also copy variables from the libclc shader
v3 (jekstrand): Fix progress return, only run variable inlining once
v4 (jekstrand): Have function inlining also copy vars for us

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>

vtn/opencl: Rework type handling for CL extension opcodes

 - Change dest type information to be vtn_type rather than glsl_type
 - Add source type information, also in vtn_type form

Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>

nir/vtn: Add handling for SPIR-V event variables

Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>

vtn: add an option to create a nir library from spirv

This adds an options to turn a spir-v library into a set of nir
functions. There is no entry point and all function implentations
are emitted.

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>

nir: Add glsl_base_type unsigned -> signed version helper

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>

nir: Add an internal flag to shader_info

Don't print the shader if it's marked internal, unless NIR_PRINT
has been explicitly set to 2 (or higher).

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>

driconf: Use nesting macros for defining options.

Manually balancing the BEGIN/ENDs is a recipe for xml validation failures,
just make the macros do the balancing.  The only ugly bit I think is that
enums take a list of DRI_CONF_ENUM() without ','s in between them.

Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6753>

virgl: Clean up the driconf definition of GLES_SAMPLES_PASSED_VALUE.

The right values ended up being passed to DRI_CONF_OPT_BEGIN_V, they were just
named wrong.

Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6753>

util/xmlconfig: Add a unit test of the code.

I want to build a non-XML-based alternative for Android, and to do that I
want to know that my equivalent code still works.

Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6753>

blorp: Ensure aligned HIZ_CCS_WT partial clears

Fixes: 5425fcf2cb3 ("intel/blorp: Satisfy HIZ_CCS fast-clear alignments")
Reported-by: Sagar Ghuge <sagar.ghuge@intel.com>
Tested-by: Ivan Briano <ivan.briano@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6854>

meson: add support for loongson's mips/mips64 arch.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6836>

freedreno/layout: add tile_all flag to the layout

Added a new tile_all flag which is used to set the TILE_ALL flag of
the texture. Enabled tile_all to depth/stencil images are they are
non-linear.

Signed-off-by: Samuel Iglesias Gonsálvez <siglesias@igalia.com>
Reviewed-by: Jonathan Marek <jonathan@marek.ca>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6848>

nir/liveness: Consider if uses in nir_ssa_defs_interfere

Fixes: f86902e75d9 "nir: Add an SSA-based liveness analysis pass"
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/3428
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Yevhenii Kharchenko <yevhenii.kharchenko@globallogic.com>
Reviewed-by: Connor Abbott <cwabbott0@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6824>

turnip: implement VK_EXT_extended_dynamic_state

Passes dEQP-VK.pipeline.extended_dynamic_state.*

Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5641>

turnip: move A6XX_RB_ALPHA_CONTROL write to init_hw

Its always 0.

Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5641>

turnip: fix wrong indentation in tu6_draw_common

Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5641>

nir/instr_set: hash intrinsic sources

ministat (CSE only):
Difference at 95.0% confidence
-9.80325 +/- 0.173089
-41.4434% +/- 0.461972%
(Student's t, pooled s = 0.0763653)

ministat (entire run):
Difference at 95.0% confidence
-3.13667 +/- 0.61519
-5.11107% +/- 0.990737%
(Student's t, pooled s = 0.271416)

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6860>

lima: fix glCopyTexSubImage2D

The reload texture descriptor needs to take care of the mipmap level
and the layer in case of GL_TEXTURE_CUBE_MAP.

glCopyTexSubImage2D triggers the lima_blit function which ends in a draw.
A reload is necessary. The reload texture descriptor is always built with
just one mipmap level, but this needs to be the level we want to reload,
not just 0. We also have to take care of the cubemap face.

This fixes the following dEQP tests:

dEQP-GLES2.functional.texture.specification.basic_copytexsubimage2d.2d_rgb
dEQP-GLES2.functional.texture.specification.basic_copytexsubimage2d.2d_rgba
dEQP-GLES2.functional.texture.specification.basic_copytexsubimage2d.cube_rgb
dEQP-GLES2.functional.texture.specification.basic_copytexsubimage2d.cube_rgba

Reviewed-by: Erico Nunes <nunes.erico@gmail.com>
Signed-off-by: Andreas Baierl <ichgeh@imkreisrum.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6816>

kmsro: Extend to include imx-dcss

This allows using the dcss display controllers (with the dcss drm
modesetting driver) along with the Etnaviv render-only drivers. DCSS is
found on i.MX SoCs.

Signed-off-by: Guido Günther <agx@sigxcpu.org>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3449>

intel/fs: Disable sample mask predication for scratch stores

Scratch stores are being lowered to the instructions with side-effects,
however they should be enabled in fs helper invocations, since they
are produced from operations which don't imply side-effects.

To fix this - we move the decision of whether the sample mask predication
is enable to the point where logical brw instructions are created.

GLSL example of the issue:

 int tmp[1024];
 ...
 do {
   // changes to tmp
 } while (some_condition(tmp))

If `tmp` is lowered to scrach memory, `some_condition` would be
undefined if scratch write is predicated on sample mask, making
possible for the while loop to become infinite and hang the GPU.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/3256
Fixes: 53bfcdeecf4c9632e09ee641d2ca02dd9ec25e34
Signed-off-by: Danylo Piliaiev <danylo.piliaiev@globallogic.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6056>

zink: only set stencil-ref for back if two-sided

Otherwise, we want to set both front and back to the same state.

Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6853>

zink: store base-object of DSA-state

This is useful in the next commit, where we need to inspect the
base-state.

Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6853>

zink: correct typo in stencil-setup

Without this, we end up using two-sided stencil when one-sided stencil
should be used. Whoops.

Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6853>

radeonsi: remove redundant info.uses_fbfetch

Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6782>

radeonsi: remove redundant variables from struct si_compute

Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6782>

radeonsi: assume that constant load_local_group_size has been optimized out

Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6782>

radeonsi: run NIR optimizations that glsl_to_nir runs but other places might not

Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6782>