docs: Update freedreno features

Another manual re-sync with reality.

There are a few extension, notably GL_ARB_texture_compression_bptc,
which are exposed but mesamatrix doesn't show as supported.  I can't
really figure out why.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19100>

driconf: add a workaround for Cossacks 3

Wine issue: https://bugs.winehq.org/show_bug.cgi?id=44560

Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19147>

driconf: add a workaround for a cursor issue with kwin_wayland

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7351

Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19147>

docs: correct spelling of "freedreno"

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: correct spelling of "traveling"

While "travelling" isn't incorrect, we generally prefer US english in the
docs, where "traveling" is the common form.

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: correct spelling of "correctly"

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: ie. -> i.e.

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: fractionary -> fractional

While "fractionary" isn't incorrect language, it's much less common than
the more modern "fractional" term.

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: correct spelling of "rasterizer"

We use rasterizer all over the docs, so let's be consistent.

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: correct spelling of "explicitly"

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: correct spelling of "addition"

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: correct spelling of "average"

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: correct spelling of "laid"

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: correct spelling of "compression"

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: correct spelling of "commutativity"

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: correct spelling of "hierarchy"

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: correct spelling of "maximum"

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: correct spelling of "synced"

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: correct spelling of "simplification"

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: correct spelling of "shareable"

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: correct spelling of "Raspbian"

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: correct spelling of "subtly"

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: correct spelling of "choosing"

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: correct spelling of "Khronos"

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: correct spelling of "to"

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: correct spelling of "implementation"

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: correct spelling of "hierarchical"

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: correct spelling of "specified"

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: correct spelling of "bug fixes"

This isn't a compound word in English, and we tend to write it as two
worlds.

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: correct spelling of "attempts"

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: correct spelling of "implemented"

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19139>

docs: Remove document about XvMC

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Acked-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19133>

nouveau: Remove XVMC_VL environment variable usage

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Acked-by: David Heidelberg <david.heidelberg@collabora.com>
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19133>

docs: Remove rbug related documents

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19138>

docs: Remove graw related words

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19138>

ci/windows: Update build container image with all deps so wraps can be disabled

Reviewed-by: Daniel Stone <daniels@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18287>

radeonsi: lower hadd and fisnormal

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19151>

nir/lower_cl_images: set binding

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19151>

radeonsi: fail creating textures from user memory

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19151>

zink: clamp shader bindings for compute descriptors

this uses massively less memory for some drivers

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19140>

zink: reorder context init to handle COMPUTE_ONLY

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19140>

zink: delete some big struct members from zink_context

these were only used by descriptor caching, so now they just take up memory

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19140>

zink: clamp feedback loop layout to GENERAL if the ext isn't supported

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19140>

zink: enable tc cpu storage on some buffers

matching freedreno usage

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19142>

cso: make memcmp and hash computation use a literal key size for blend state

This results in faster lookups because memcmp and the hash computation
can be unrolled.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19129>

cso: inline more functions because some parameters like key_size are literals

One function is moved to the header file, the other two functions are inlined
manually.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19129>

cso: don't destroy CSOs that are saved

I think this can't happen in practice, but better safe than sorry.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19129>

cso: fix broken optimization for sampler state lookups

Since the key size wasn't a constant expression, all the function inlining
didn't do much. This makes it a constant expression.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Fixes: c4e18cd4dd1 - mesa/st: add PIPE_QUIRK_TEXTURE_BORDER_COLOR_SWIZZLE_FREEDRENO
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19129>

cso: constify some parameters to remove typecasts

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19129>

cso: start without u_vbuf by default if it's not always used

no functional change, but it's convenient for future work

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19129>

glthread: don't sync for glIsEnabled(GL_BLEND, GL_LIGHTING, GL_POLYGON_STIPPLE)

Acked-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18199>

glthread remove the unused *last pointer from unmarshal functions

Acked-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18199>

glthread: merge and collapse glBindBuffer calls that unbind and then bind

This is a small optimization for viewperf. See the comment in the code.

Acked-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18199>

gl_marshal.py: C style fixups

Acked-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18199>

gallium/u_threaded: don't call simplify_draw_info redundantly

It's always called for the first and second draw in a row, and if they
can't be merged, the second draw becomes the first draw in the next call,
and simplify_draw_info is called again on that. Thus, it's called twice
on any draw that isn't merged. To prevent that, call it when we add
the draws.

Acked-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18199>

glthread: rewrite CallList merging and do it in the app thread

This looks simpler and hopefully faster, but it may just move the overhead
to the app thread.

Acked-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18199>

glthread: demystify Draw function names

Since we had the freedom to write our own marshal and unmarshal Draw
functions, we turned it into a mess by doing whatever we want in those
functions. For example, DrawElementsInstancedBaseVertex actually
implemented DrawArraysInstancedBaseInstanceBaseInstance.

Add 4 internal GL functions that pass user buffers through marshal
/unmarshal functions, and clean up the other names to match their
behavior.

The new functions don't need any parameters in their definitions because
we don't call them directly.

Acked-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18199>

glapi: remove EXT and ARB suffixes from Draw functions

This swaps the function names with aliased names that don't have those
suffixes.

Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Acked-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18199>

glthread: add more DrawArrays/Elements variants with fewer fields

The idea is to save 8 bytes per call by removing 2 fields that are not
specified by the user such as instance_count=1, basevertex=0, and
baseinstance=0.

Acked-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18199>

glthread: use a constant expression instead of cmd_size in custom functions

cmd_size is constant in these cases

Acked-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18199>

gl_marshal.py: simplify print_sync_call and rename to print_call

no spaces around = is a python convention for implicit arguments only

Acked-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18199>

gl_marshal.py: inline functions for readability

All functions have only one use. Jumping constantly between functions
in the file is confusing.

No spaces around = is a python convention for implicit arguments only.

Acked-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18199>

gl_marshal.py: remove the -O1 hack and manual SET_* inlining

The compile times aren't so bad anymore now that the initialization is
split among all 8 files.

Also add one assertion.

Acked-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18199>

gl_marshal.py: rework how the marshal dispatch table is initialized

Instead of setting all function pointers in marshal_generated0.c,
set the function pointers in the file that contains the functions,
and remove all the forward declarations of marshal functions
in marshal_generated.h.

Acked-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18199>

gl_marshal.py: move the unmarshal table into a separately generated file

It's unrelated to the rest of the script and it's in the way of bigger
changes.

Acked-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18199>

gl_marshal.py: remove/simplify parameters

Acked-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18199>

glsl: move lower instructions logic inside that pass

There is now only a single called of this pass so tidy things up
and move all this logic inside the pass.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19112>

glsl: always do {CARRY,BORROW}_TO_ARITH lowering

The only caller always sets these so here we just remove the
option to disable it.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19112>

glsl: drop sub to add neg lowering in GLSL IR

NIR opt algebraic does this for us so no need to have it
implemented here also.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19112>

glsl: move rule inside lower_packing_builtins()

We only have a single user of this pass so lets tidy things up and
move all the rules in the pass itself.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19112>

anv: work around improper buffer usage in hitman3

Avoid memory errors by returning an error immediately instead of
attempting to allocate the buffer view with an unsupported format.

Fixes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6998

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17972>

anv: add a layer for hitman3

This title is trying to create a buffer view with usage not supported by Anv.

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17972>

anv: compile anv_acceleration_structure.c

This file was unintentionally excluded from meson build support.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17972>

aco: add ACO_DEBUG=force-waitdeps

GFX11 has a lot of complicated data dependency hazards.

For debugging GFX10+ data dependency hazards. This creates an excessive
amount of s_waitcnt_depctr.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18273>

aco/gfx11: workaround VALUMaskWriteHazard

fossil-db (gfx1100):
Totals from 62812 (46.52% of 135032) affected shaders:
Instrs: 43971580 -> 44069887 (+0.22%)
CodeSize: 233473420 -> 233866648 (+0.17%)
Latency: 463487489 -> 463520688 (+0.01%); split: -0.00%, +0.01%
InvThroughput: 86505748 -> 86509679 (+0.00%); split: -0.00%, +0.00%

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18273>

aco/gfx11: workaround VALUPartialForwardingHazard

fossil-db (gfx1100):
Totals from 18121 (13.42% of 135032) affected shaders:
Instrs: 28272367 -> 28323468 (+0.18%)
CodeSize: 152581916 -> 152786320 (+0.13%)
Latency: 366477785 -> 366482318 (+0.00%); split: -0.00%, +0.00%
InvThroughput: 74475188 -> 74475758 (+0.00%); split: -0.00%, +0.00%

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18273>

aco/gfx11: workaround VALUTransUseHazard

fossil-db (gfx1100):
Totals from 116990 (86.64% of 135032) affected shaders:
Instrs: 67942325 -> 69493991 (+2.28%)
CodeSize: 366448984 -> 372655648 (+1.69%)
Latency: 673236871 -> 673269808 (+0.00%); split: -0.00%, +0.00%
InvThroughput: 128266905 -> 128270175 (+0.00%); split: -0.00%, +0.00%

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18273>

aco/gfx11: workaround LdsDirectVMEMHazard

fossil-db (gfx1100):
Totals from 27217 (20.16% of 135032) affected shaders:
Instrs: 18010853 -> 18047277 (+0.20%)
CodeSize: 99369568 -> 99515264 (+0.15%)
Latency: 207454040 -> 207464932 (+0.01%); split: -0.00%, +0.01%
InvThroughput: 39810158 -> 39810628 (+0.00%); split: -0.00%, +0.00%

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18273>

aco/gfx11: workaround LdsDirectVALUHazard

fossil-db (gfx1100):
Totals from 57858 (42.85% of 135032) affected shaders:

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18273>

aco/gfx11: workaround VcmpxPermlaneHazard

Same as GFX10, but in a separate pass because it's the only hazard that's
shared.

No fossil-db changes.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18273>

aco: add search_backwards helper

This will be useful for VALUPartialForwardingHazard.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18273>

nir2dxil: Lower texture projections

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19115>

mesa/st: remove translate_mode

This no longer does anything useful, so let's just drop it.

Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19034>

mesa/st: move static asserts out of translate_prim

Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19034>

CI: Add gallium-va and video-codecs in windows-vs2019 and debian-mingw32-x86_64

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19063>

d3d12: resource_from_handle to validate importing resource ptr from same d3d12 device

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19063>

frontends/va: Add windows VA frontend support via vl_winsys_win32 and libva-win32

Acked-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19063>

gallium/vl: Add vl_winsys_win32 support

Acked-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19063>

hasvk: use DX rules for point rasterization

It seems that's what the reference renderer in the CTS expects for
Vulkan. This mostly matters if the edges of a point primitive fall
exactly on a pixel sampling point.

Fixes some upcoming tests under
dEQP-VK.pipeline.monolithic.depth.format.*.point_list*

Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19026>

anv: use DX rules for point rasterization

It seems that's what the reference renderer in the CTS expects for
Vulkan. This mostly matters if the edges of a point primitive fall
exactly on a pixel sampling point.

Fixes some upcoming tests under
dEQP-VK.pipeline.monolithic.depth.format.*.point_list*

Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19026>

anv: compile_upload_rt_shader expects a valid pointer

Fixes crashes on almost every CTS test that uses raytracing pipelines.

Fixes: ff91c5ca42b ("anv: add analysis for push descriptor uses and store it in shader cache")

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19130>

util: Move the implementation of futex_wake and futex_wait from futex.h to futex.c

Doing this is to avoid include bloat on Windows that pulls windows.h,
for other unix platforms, it's also benefit include the used symbols only when use
futex_wake and futex_wait.
No functional change, there is no performance incur because function call is consume
much less time than syscall

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19120>

freedreno: Fix fence unref race

Destroying a fence that was never flushed could race with the submit
queue, which could cause either an fd leak or closing an fd that the
retire queue would want to use.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19080>

radeonsi: fix max_dw computation for CS preambles

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19131>

winsys/amdgpu: fix (enable) preemption for chained IBs

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19131>

radeonsi: rename si_create_multi_fence -> si_alloc_fence

We no longer have any multi fences (gfx+sdma fence).

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19131>

v3dv: drop error overwrite

Let the error returned be bubbled up.

Fixes: dEQP-VK.api.device_init.create_instance_device_intentional_alloc_fail.basic
Fixes: 591103d04d5f18ae30c7 ("v3dv: don't return incompatible driver if GPU is not present")
Signed-off-by: Eric Engestrom <eric@igalia.com>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18901>

asahi: Lower point sprites in driver

mesa/st can't save us now!

Fixes: 310959d9fe1 ("mesa/st: rip out point-sprite cap")
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18651>

u_transfer_helper: Handle Z24S8 with z24_in_z32f but no interleaving

First, the separate Z/S condition kicks in, splitting up Z24S8 into
Z24X8 + S8. But then the driver is asked to create a Z24X8 resource,
which it may not be able to do. We need to further lower at create
time to Z32F + S8, so the driver is only asked to create supported
formats. We can do this with an extra condition at create time.

The alternate approach would be recursing into
transfer_helper_resource_create, but this seemed more obvious.

Fixes assertion failure running neverball on Asahi.

Assertion failed: (templ->format != PIPE_FORMAT_Z24X8_UNORM &&
      templ->format != PIPE_FORMAT_Z24_UNORM_S8_UINT &&
      "u_transfer_helper should have lowered"), function
agx_resource_create, file agx_pipe.c, line 174.

Fixes: 45a37ace287 ("u_transfer_helper: Handle Z24X8 for drivers that don't use the interleaved transfer_map")
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Acked-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18651>

nir/lower_io: Set interpolated_input dest_type

...even for non-pixel interpolation, for consistency. Otherwise backends get
funny intrinsics with interpolateAt:

   vec4 32 ssa_4 = intrinsic load_interpolated_input (ssa_3, ssa_2) (base=1, component=0, dest_type=invalid /*0*/, io location=33 slots=1 /*161*/)

We know it'll be a float, but backends shouldn't need to special case this. (Or
maybe interpolated_input shouldn't have a dest_type index. I'd be ok with that
resolution too. But having one and not setting it consistently is wrong.)

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19085>