glthread: disallow glthread if buffer uploads are unsupported

to remove fail paths that defeat the purpose of glthread.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20824>

glthread: do vertex uploads if an index buffer is present for MultiDrawElements

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20824>

glthread: remove the vbo_upload_ratio_too_large fallback for glMultiDrawElements

This would be possible to implement, but we don't know of any app where
it would help.

Now glthread fully handles all non-VBO uploads except glDrawIndirect.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20824>

glthread: make marshal functions for glBegin/End attribs non-static

for a future commit

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20824>

glthread: pack and name the type of glthread_vao::Attrib

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20824>

mesa: move gl_vertex_format_user definition into glthread.h

glthread.h needs it and it can't include mtypes.h because mtypes.h
includes it.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20824>

glthread: do vertex uploads if an index buffer is present for glDrawElements

glthread didn't implement uploading non-VBO vertices if indices were
in a buffer. This implements that.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20824>

glthread: change multi_draw_elements_async() to never fail due to large size

Some callers (not visible here) silently ignore the return value. Remove
the return value and handle the failure in multi_draw_elements_async.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20824>

glthread: execute glMultiDrawArrays(draw_count < 0) asynchronously

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20824>

glthread: set GL_OUT_OF_MEMORY if we fail to upload vertices

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20824>

glthread: set GL_OUT_OF_MEMORY if we fail to upload indices

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20824>

glthread: handle GL_*_ARRAY in glEnable/Disable

Surprisingly, the GL compatibility profile allows these in both
glEnableClientState and glEnable.

Fixes: 0b1dd185913 - glthread: track which vertex array attribs are enabled

Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20824>

glapi: autogenerate function parameters with no space between * and variable

print "int *v" instead of "int * v".

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20824>

mesa: allow GL_UNSIGNED_INT64_ARB as vertex format for ARB_bindless_texture

This wasn't implemented, but the spec requires it.

Fixes: 1fe7b1f9724 - mesa: implement ARB_bindless_texture

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20824>

util: fix util_is_vbo_upload_ratio_too_large

It was wrong. For example, if the draw vertex count was 10 and the upload
vertex count was 150, u_vbuf wouldn't unroll the draw and would instead
memcpy 150 vertices. This fixes that case.

Fixes: 068a3bf0d7c - util: move and adjust the vertex upload heuristic equation from u_vbuf

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20824>

glthread: fix an upload buffer leak

Fixes: befbd54864d29 - glthread: don't use atomics for refcounting to decrease overhead on AMD Zen

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20804>

anv: fix preemption enable emission in gpu_memcpy

This has to be before the MI_BATCH_BUFFER_END otherwise it has no
effect.

This also was messing around with you batch length alignment.

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Fixes: b9aa66d5d0 ("anv: disable preemption for 3DPRIMITIVE during streamout")
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20802>

radv: remove redundant type sizing

this is already 8 bits

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20751>

radv: add some graphics pipeline hints to optimize pipeline bind

this is a costly function, and we want to avoid loading random struct data
as much as possible

these struct members aren't accessed anywhere else in the function, so eliminating
access avoids some cpu overhead

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20751>

radv: simplify depth aspect check in radv_handle_image_transition()

this info is already available, so reduce cpu overhead

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20751>

radv: reorder dynamic state checks during bind

this avoids potential out-of-order reads from a struct that spans
18 CPU cachelines

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20751>

radv: repack radv_graphics_pipeline struct

this reduces the number of cachelines used by the struct and allows for
improved memory access

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20751>

dzn: Support basic subgroups

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20777>

spirv2dxil: Support basic subgroups

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20777>

spirv2dxil: Use 32-bit shared offsets

nir_build_deref_var assumes that you're going to do so, and
there's no reason to use 64-bit types for Vulkan compute.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20777>

microsoft/compiler: Add lowering passes for basic subgroup vars

DXIL doesn't have a "subgroup ID" or "num subgroups" construct,
so add lowering to construct them. Subgroup ID is done using
once-per-subgroup atomics on a workgroup-shared variable, and
then broadcasting that (using read_first_invocation) to the other
threads. Num subgroups is just a division with the workgroup size.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20777>

microsoft/compiler: Implement a few basic wave/subgroup intrinsics

These are the ones that map perfectly between SPIR-V and DXIL that
are in the "basic" extension group (except for read-lane-first,
but we'll use with some lowering shortly).

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20777>

microsoft/compiler: Add an overload param to unary function helpers

Subgroup ops aren't i32, they use none.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20777>

zink: don't use ds3 blend states without color attachments

this is illegal and causes validation errors

cc: mesa-stable

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20799>

zink: delete need_blend_constants

this is an artifact of very old code before the dynamic state was set
for all graphics pipelines

now the checks only cause blend constants to not be updated, which triggers
bugs and validation failures

cc: mesa-stable

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20799>

intel/dev: Split hwconfig i915 specific code

Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Marcin Ślusarz <marcin.slusarz@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20437>

intel/dev: Move i915 code to i915/intel_device_info.c

No behavior changes here.

Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Marcin Ślusarz <marcin.slusarz@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20437>

intel/dev: Export functions that will be used by different kernel drivers

No behavior changes here.

Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Marcin Ślusarz <marcin.slusarz@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20437>

radv: remove an old FIXME about a possible bug with TC-compat HTILE

I added this FIXME 2 years ago because it was unclear if it was
broken or not. Since, CTS coverage improved and the number of tests
with depth/stencil on the compute queue increased a lot. vkd3d-proton
also widely uses depth/stencil with GENERAL on GFX10+ and likely with
async compute as well. No issues so far.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/4048
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20785>

freedreno: Add driconf to disable conservative LRZ

The problematic sequence of draws is pretty rare.  But there are a small
handful of games which do not exhibit the problematic sequence and for
which invalidating LRZ on draws with blend plus depthwrite enabled hurts
performance slightly.  This driconf option enables opting in to the
previous behavior.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20708>

util/xmlconfig: Use os_get_option()

Enable the property_get() fallback on android.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20708>

turnip: Invalidate LRZ on blend+depthwrite

See the previous commit for details.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20708>

turnip: Rename lrz force_disable_mask

Prep to use it to communicate more than just force-disable cases.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20708>

freedreno/a6xx: Invalidate LRZ on blend+depthwrite

Writing depth with blend enabled means we need to invalidate LRZ,
because the written depth value could mean that a later draw with
depth enabled (where we would otherwise write LRZ) could have
fragments which don't pass the depth test due to this draw.  For
example, consider this sequence of draws, with depth mode GREATER:

  draw A:
    z=0.1, fragments pass
  draw B:
    z=0.4, fragments pass
    blend enabled (LRZ write disabled)
    depth write enabled
  draw C:
    z=0.2, fragments don't pass
    blend disabled
    depth write enabled

Normally looking at the state in draw C, we'd assume we could
enable LRZ write.  But this would cause early-z/lrz to discard
fragments from draw A which should be visible due to draw B.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/8065
Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20708>

freedreno/a6xx: Add LRZ perf warn for ztest direction changes

Add a perf warning when LRZ invalidate happens due to depth test
direction reversal.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20708>

gallium/util: Add util_writes_depth() helper

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20708>

radeonsi: report 0 block size for Polaris HEVC encoding

makes encoded videos resemble the input again :)

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7992
Fixes: c4482a3c1a9 ("radeonsi/vcn: enable multi-slice encoding")
Reviewed-by: Ruijing Dong <ruijing.dong@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20714>

radv: Make NGG query emission a dirty flag.

Don't emit the NGG query user SGPR if its state doesn't change.
Based on original work by Mike Blumenkrantz.

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18819>

iris: add restrictions for 3DSTATE_RASTER::AntiAliasingEnable

Field must be disabled if any render targets have integer
format, additionally for Gfx12+ field must be disabled when
num multisamples > 1 or forced multisample count > 1.

Cc: mesa-stable
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7892
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-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/20671>

hasvk: add restrictions for 3DSTATE_RASTER::AntiAliasingEnable

Field must be disabled if any render targets have integer format.

Cc: mesa-stable
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20671>

anv: add restrictions for 3DSTATE_RASTER::AntiAliasingEnable

Field must be disabled if any render targets have integer
format, additionally for Gfx12+ field must be disabled when
num multisamples > 1 or forced multisample count > 1.

Cc: mesa-stable
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20671>

radv: advertise extendedDynamicState3ColorBlendEquation

This enables full ds3 support with Zink!

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20281>

radv: fix detecting that blend is enabled when all CB states are dynamic

It's allowed to be NULL.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20281>

radv: enable compiling PS epilogs on-demand for dynamic color blend equations

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20281>

radv: add support for dynamic blend equation

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20281>

radv: add a new helper for normalizing blend factors

It will be also used when compiling PS epilogs on-demand.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20281>

radv: move some color blend helpers to radv_private.h

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20281>

zink/ci: add spec@!opengl 1.1@line-smooth-stipple to the fails list

Fixes: a8d2b288eee3 ("ci/piglit: 2023-01-19 uprev")
Signed-off-by: Martin Roukala (né Peres) <martin.roukala@mupuf.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20807>

iris: Don't flush the render cache for a compute batch

Make sure we comply with BSpec and ensure that certain flush flags
are not set for compute batches

Signed-off-by: Rohan Garg's avatarRohan Garg <rohan.garg@intel.com>
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15664>

tu/kgsl: do not use kgsl_command_object::offset

offset field in kgsl_command_object is NOT used by KGSL, so
we should offset directly to iova.

Fixes weird hangs on KGSL. E.g. fixes the hang in:
 dEQP-VK.memory.pipeline_barrier.transfer_dst_storage_texel_buffer.1024

cc: mesa-stable

Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20795>

anv: uncompressed views of compressed 3d images are now valid

There was a VUID-VkImageViewCreateInfo-image-04739 in the Vulkan 1.3
spec that said:

    If image was created with the
    VK_IMAGE_CREATE_BLOCK_TEXEL_VIEW_COMPATIBLE_BIT flag and format is a
    non-compressed format, viewType must not be VK_IMAGE_VIEW_TYPE_3D

That VUID has since been removed, and when a view of a 3D image is
created, with put the depth into the array_len, so it won't be always 1.

Reviewed-by: Mark Janes <markjanes@swizzler.org>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20803>

radv: Fix depth-only-with-discard when epilogs are used.

For a depth-only-with-discard pipeline, spi_shader_col_format needs to be
fixed up to a single channel export, or otherwise discard will not work.

Since col_format can change depending on the dynamic state, precompute the
need for this workaround on pipeline creation and apply it when emitting
prolog states.

Fixes: eb07a11b8f4 ("radv: add support for compiling PS epilogs on-demand")
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20704>

meson: use sections in summary()

Makes the list easier to read

Reviewed-by: Dylan Baker <dylan@pnwbakers.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12161>

meson: use summary()

Make use of mesons summary() to create and align the configuration
summary.

Reviewed-by: Dylan Baker <dylan@pnwbakers.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12161>

dzn: Rework meta blit VS

Some D3D12 drivers, like my PC's AMD driver, don't like using a
dynamic index to load from a constant buffer that's bound via
root constants. Instead, just go ahead and load the full set of
vertex data and just bcsel which one to use.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20778>

dzn: Delete an unnecessary assert

As far as I can tell, having multiple levels in the view for a
color target just means to target the first one.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20778>

dzn: Enable independent blending

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20778>

ci: bump ci-fairy with session support (robust downloads)

Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20788>

ci/piglit: 2023-01-19 uprev

Emergency uprev to get robustness against http 50x errors for traces.

Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20788>

ci: migrate from wget to curl

Better error handling is more reliable.

Options:
 -L, follow location
 --retry, number of retries
 --retry-all-errors, does not fail on ALL errors, that's why there is -f
 -f, fail fast with no output at all on server errors
 --retry-delay, make curl sleep this amount of time before each retry

Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20788>

ci/piglit: explicitely define we want GLX tests

Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20788>

meson: use the same workaround for setting 'lib' on windows

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20752>

meson: use [] instead of 'lib for !windows name_prefix

This is a poorly documented value that uses the default value, but
doesn't emit a warning. In the future Meson will fix this.

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20752>

meson: use builtin support for reading version from a file

In meson 0.57 support was added for reading a version from a file to
meson natively, so we don't need this workaround anymore.

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20752>

meson: remove version checks for < 0.59

Which is now required, so these are useless

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20752>

dzn: Enable multiview

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20650>

dzn: Implement multiview queries

This is a weird way to do queries, but in multiview, each query
takes up N slots, where N is the number of views. D3D doesn't do
it that way, and only has one result, which fortunately is a valid
way to do Vulkan queries. We just need to take care to zero out
the other view results, and make sure they get "signaled" when
the cmdbuf is submitted.

Note that it is invalid in D3D to use ResolveQueryData on query
slots that have never actually been begun/ended, so we zero out
the data by copying zeroes into the buffer. This probably could
be optimized but oh well.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20650>

dzn: Handle draws and clears for multiview rendering

For draws, when we're emulating multiview, we need to loop them
and set up the right sysval. For clears, we always need to loop.
When not emulating, we also need to set up the right view instance
mask.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20650>

dzn: Handle multiview pipeline creation

This handles both native and emulated pipeline creation

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20650>

dzn: Put nir compilation options in a struct

The function signatures are getting unwieldly...

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20650>

spirv2dxil: Claim multiview support

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20650>

spirv2dxil: Implement lowering for multiview

D3D's view instancing is an optional feature, and even when it's
supported, it only goes up to 4 views, where Vulkan requires a
minimum of 6 supported views. So, we need to have a path for handling
the cases where we can't use the native feature.

In this mode, pass the view ID as a runtime var. The caller is then
responsible for looping the draw calls and filling out the constant
buffer value correctly for each draw. When we get to the last pre-rast
stage, we'll additionally want to write out gl_Layer to select the
right RTV array slice. Lastly, for the fragment shader, if there's
any input attachments, those get loaded using the RTV slice instead
of the view ID. RTV slice input into the PS is done with a signature
entry (which must be output from the previous stage) rather than a
system value.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20650>

spirv2dxil: Pass runtime conf struct to lower_shader_system_values

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20650>

microsoft/compiler: Support view instancing

This adds support for D3D12-native view instancing to the compiler.
Essentially, it's just the ability to load SV_ViewID (dx.op.viewID),
set the right capability, and fill out some more PSV data. Note that
the PSV data is currently garbage. Ideally, we'd fill out a proper
input -> output and viewID -> output dependency table, but AFAIK
this is only used to enforce D3D API validation, and drivers ignore
it, so it's less critical to get it right.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20650>

microsoft/compiler: Delete stale TODO comment

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20650>

microsoft/compiler: Subpass textures are supposed to be arrays

When doing multiview subpass rendering, they get loaded with the
view/layer index.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20650>

microsoft/compiler: Delete incorrect implementation for load_layer_id

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20650>

zink: simplify some dynarray concat descriptor code

TIL this exists

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20681>

zink: consolidate semaphore creation where possible

all cacheable semaphores should now be using the cache

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20681>

zink: move semaphore caching to zink_reset_batch_state()

this makes semaphores available for reuse more rapidly

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20681>

zink: add a binary semaphore cache

after being waited upon, a binary semaphore can be reused, which saves
tons of present-related ioctls when fps is high

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20681>

zink: add a util function for creating semaphores

annoying to keep copy/pasting this around

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20681>

zink: preserve present resources during async presentation

ensure that these have a lifetime great enough to be presented

fixes #7781

cc: mesa-stable

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20793>

ci/bare-metal: add more timestamps to help debugging issues

Acked-by: Emma Anholt <emma@anholt.net>
Acked-by: David Heidelberg <david.heidelberg@collabora.com>
Signed-off-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20792>

zink: NV_compute_shader_derivatives

nothing complicated here

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12120>

zink: correct sparse bo mem_type_idx placement

VK_MEMORY_PROPERTY_DEVICE_LOCAL_BIT = 0x01 has been incidently the
correct memory type index, but isn't guaranteed to be, which is why it
hasn't caused issues yet

Fixes: f9515d93 ("zink: allocate/place memory using memoryTypeIndex directly")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20264>

zink: trival renames heap_idx -> memoryTypeIndex

Trival renames to correctly identify vulkan memory type indices aren't
the same as zink heaps

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20264>

zink: zink_heap isn't 1-to-1 with memoryTypeIndex

Clarify the relationship between zink heaps and vulkan memory type
indices, and resolve the issues from mixing the two up.

Closes: #7588, #7813
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20264>

meson: Use feature option methods for xmlconfig

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Acked-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20409>

meson: use a feature option for shared-llvm

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Acked-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20409>

meson: use a feature option for egl

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Acked-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20409>

meson: use a feature option for zstd

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Acked-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20409>

meson: use a feature option for xlib-lease

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Acked-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20409>

meson: use a feature option for power8

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Acked-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20409>

meson: use a feature option for lmsensors

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Acked-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20409>