radv: remove radv_translate_vertex_format

This is not used anymore.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17894>

radv,aco: use pipe_format for dynamic vertex input state

Also prepare for 64-bit and R8G8B8/R16G16B16 with the addition of
radv_vs_input_state::nontrivial_formats.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/5021
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17894>

radv,aco: use pipe_format for static vertex input state

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17894>

radv: add radv_is_vertex_buffer_format_supported

We will want to support formats which cannot used for texel buffers.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17894>

ac: add ac_vtx_format_info

This will be used by RADV and ACO.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17894>

util/format,vulkan/format: update p_format for Vulkan vertex buffers

We need to move some earlier (so they are <= 255 like the comment says)
and add a few 64-bit ones.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17894>

nir: use a GC context for instructions

Gives an roughly -15% change in compile-time for RADV/ACO.

Memory usage increase seems to be 5-6%.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/5034
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12910>

nir: adjust nir_src_copy signature to take a nir_instr *

This is almost always a nir_instr and updating the src of a nir_if will
have to work slightly differently in the future.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Acked-by: Alyssa Rosenzweig <alyssa@collabora.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12910>

Revert "nir: Drop the unused instr arg for src/dest copy functions."

This reverts commit c3a01841184ee8303c0c5ebe58491301622c5ad6.

Acked-by: Alyssa Rosenzweig <alyssa@collabora.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12910>

nir/serialize: remove unused parameter from read_src()

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12910>

util: add freelist allocator with mark/sweep

Based on the allocator https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/524,
but modified a bit (in particular, it's now separate from ralloc).

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12910>

util/ralloc: add HEADER_ALIGN macro

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12910>

nir: Free instructions more often

Soon we'll be allocating instructions out of a per-shader pool, which
means that if we don't free too many instructions during the main
optimization loop, the final nir_sweep() call will create holes which
can't be filled. By freeing instructions more aggressively, we can
allocate more instructions from the freelist which will reduce the final
memory usage.

Modified from Connor Abbott's original patch to rebase on top of
refactored DCE and so that the use-after-free in nir_algebraic_impl() is
fixed.

Co-authored-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12910>

gallium/u_threaded: buffer sharedness tracking

This fixes TC's buffer invalidation code for buffers that are shared
between contexts.

TC is unable to notify other TCs in case it replaces a shared buffer's
underlying storage when invalidating, causing those other TCs to use
wrong buffer IDs for busyness tracking, which leads to corruption due
to invalidation fast-paths being triggered when they shouldn't be.

This patch addresses this issue by tracking if a buffer is shared, and
if it is, disabling buffer storage replacement for the affected buffer.
This is achieved by tracking which TC instance first accessed a certain
buffer. If a second instance then accesses it as well, it will realize
that it isn't the only one working on the buffer and mark the buffer
accordingly.

If TC needs to invalidate a buffer for the correctness of an operation
at any point, it will fall back to doing the operation in a synchronous
fashion with this patch if the buffer is shared and currently busy.

It might be possible to later detect that a buffer has become un-shared;
however, this is outside of the scope of this bugfix patch.

v2: Do not disable buffer busyness tracking for shared buffers.

Fixes: e9c41b32 ("gallium/u_threaded: add buffer lists - tracking of buffers referenced by tc")

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17338>

gallium/u_threaded: Prepare for buffer sharedness tracking

tc_add_to_buffer_list and tc_bind_buffer need the tc instance too.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17338>

u_atomic: Add a helper for pointer compare-exchange

Notably this helps with MSVC, which complains about compiling the not-taken
branches of the ternary when called with pointer args. Using a version that
doesn't have "runtime" sizeof checks eliminates the warnings.

Reviewed-by: Jonathan Strobl <jonathan.strobl@gmx.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17338>

Revert "ci: disable the freedreno farm."

It is back again.  Hopefully it stays that way.

This reverts commit ab473a4f486146dcd474d81addc3bfcb136efd9f.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18321>

aco/live_var_analysis: implement faster merging of live_out sets for some cases

If we know that logical and linear predecessors are the same,
we don't need to check for the register type of each variable.

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18105>

aco: use std::vector::reserve() more often

This removes the majority of vector re-allocations.

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18105>

lima/ci: Make use of .lava-piglit

This brings back the old CI behaviour.

Signed-off-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18299>

gallium/u_transfer_helper: hide deinterleave helpers

There's no more callers of these, so let's make them an implementation
detail instead of public API.

Reviewed-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17959>

zink: use full transfer-helper api

This is going to make it easier to maintain u_transfer_helper in the
future.

Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17959>

gallium/u_transfer_helper: add in-place interleave flag

This flag should allow Zink to stop using the interleave-specific API,
and instead just use the normal/full u_transfer_helper API, and start
using some more of these helpers.

Reviewed-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17959>

gallium/u_transfer_helper: clean up boolean flags

Whenever we add or remove a flag here, we need to update a bunch of
drivers in a fragile way. Moving to flags here instead should make this
a bit easier to maintain in the future.

Reviewed-by: Alyssa Rosenzweig <alyssa@collabora.com>
Reviewed-by: Vasily Khoruzhick <anarsoul@gmail.com>
Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Juan A. Suarez <jasuarez@igalia.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17959>

nir/opt_algebraic: a & ~a -> 0

Also re-ordered some optimizations for better readability.

Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18250>

radv/ci: uprev boot2container

This brings in massively-tidier logs.

Signed-off-by: Martin Roukala (né Peres) <martin.roukala@mupuf.org>
Acked-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18267>

v3dv: expose VK_KHR_maintenance4

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18246>

v3dv: implement vkGetDeviceImageSparseMemoryRequirementsKHR

Added with KHR_maintenance4.

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18246>

v3dv: implement vkGetDeviceBufferMemoryRequirementsKHR

Added with VK_KHR_maintenance4.

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18246>

v3dv: add a buffer_init helper

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18246>

v3dv: add a get_buffer_memory_requirements helper

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18246>

v3dv: implement vkGetDeviceImageMemoryRequirementsKHR

Added with KHR_maintenance4.

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18246>

v3dv: add a get_image_memory_requirements helper

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18246>

v3dv: add a v3dv_image_init helper

This is different from the internal create_image in that it doesn't
allocate memory for an image object, instead it expects the object
to be created and it initializes it from a VkImageCreateInfo struct.

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18246>

v3dv: set maxBufferSize property

Included since KHR_maintenance4 and Vulkan 1.3.

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18246>

v3dv: ref/unref pipeline layout objects

Since KHR_maintenance4 it is possible to destroy pipeline layouts
immediately after creating pipelines.

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18246>

v3dv: add a helper to destoy pipeline layouts

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18246>

radeonsi: use LOAD_CONTEXT_REG_INDEX for VGT_STRMOUT_DRAW_OPAQUE

Based on PAL's UniversalCmdBuffer::CmdDrawOpaque.

We don't need to use PFP_SYNC_ME because it's done in emit_cache_flush.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18129>

mesa: fall-back to fp16 for bptc-textures

FP16 is way more commonly supported than FP32, and requires half the
memory bandwidth. Because the internal format of BPTC is really FP16,
let's fall back to that instead.

Acked-by: Marek Olšák <marek.olsak@amd.com>
Acked-by: Soroush Kashani <soroush.kashani@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18012>

mesa/main: add fp16-versions of bptc decompression functions

These are essential copies of the float-versions, with minor
adjustments.

Acked-by: Marek Olšák <marek.olsak@amd.com>
Acked-by: Soroush Kashani <soroush.kashani@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18012>

mesa/main: remove always-true define

We don't include this header anwhere where BPTC_BLOCK_DECODE isn't
already true, so let's remove it.

Acked-by: Marek Olšák <marek.olsak@amd.com>
Acked-by: Soroush Kashani <soroush.kashani@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18012>

zink: allow X32 -> A32

The 32-bit per component pipe_formats are a bit different than the 8 and
16 bit formats, in that there's no UNORM or SNORM variants of them. So
let's just omit those variants.

While we're at it, update a comment that was already out-of-date anyway.

Acked-by: Marek Olšák <marek.olsak@amd.com>
Acked-by: Soroush Kashani <soroush.kashani@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18012>

mesa/st: enable bptc extension with fallback

This implements emulation of BPTC for all drivers that don't natively
support it in HW, similar to what we have for ETC and ASTC already. The
only significant difference is that we don't implement any transcoding
DRI-conf for it.

Acked-by: Marek Olšák <marek.olsak@amd.com>
Acked-by: Soroush Kashani <soroush.kashani@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18012>

mesa/st: implement fallback for bptc

Acked-by: Marek Olšák <marek.olsak@amd.com>
Acked-by: Soroush Kashani <soroush.kashani@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18012>

mesa: add _mesa_unpack_bptc-function

Acked-by: Marek Olšák <marek.olsak@amd.com>
Acked-by: Soroush Kashani <soroush.kashani@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18012>

mesa: add format-helper for bptc

Acked-by: Marek Olšák <marek.olsak@amd.com>
Acked-by: Soroush Kashani <soroush.kashani@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18012>

mesa/st: add context-flag for bptc-support

Acked-by: Marek Olšák <marek.olsak@amd.com>
Acked-by: Soroush Kashani <soroush.kashani@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18012>

mesa/st: do not use memcpy when using compressed fallback

If we're using the memcpy-path, we don't update the compressed image,
which will be needed if an application try to download the texture-image
again afterwards.

This isn't a problem yet, because none of the current compressed formats
with fallback are copyable as per the OpenGL 4.6 Core spec, table 8.14.
However, this is about to change, when we introduce fallback for BPTC
formats.

Acked-by: Marek Olšák <marek.olsak@amd.com>
Acked-by: Soroush Kashani <soroush.kashani@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18012>

mesa/st: avoid double-mapping if both images are the same

If both src_image and dst_image are pointing to the same image, we'll
end up double mapping the image, which throws an assert in
st_texture_image_map.

So let's check if these images are the same, and avoid the extra mapping
in that case.

Acked-by: Marek Olšák <marek.olsak@amd.com>
Acked-by: Soroush Kashani <soroush.kashani@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18012>

mesa/st: simplify st_compressed_format_fallback

By looking at the format-layout, we can do switch-cases here instead of
serial calls to _mesa_is_format_foo(). This isn't a huge deal yet, but
this function is going to learn about a lot more formats soon...

Acked-by: Marek Olšák <marek.olsak@amd.com>
Acked-by: Soroush Kashani <soroush.kashani@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18012>

mesa/main: simplify implementation of _mesa_is_format_astc_2d

Reviewed-by: Alyssa Rosenzweig <alyssa@collabora.com>
Acked-by: Marek Olšák <marek.olsak@amd.com>
Acked-by: Soroush Kashani <soroush.kashani@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18012>

mesa/main: simplify implmementation of _mesa_is_format_etc2

Reviewed-by: Alyssa Rosenzweig <alyssa@collabora.com>
Acked-by: Marek Olšák <marek.olsak@amd.com>
Acked-by: Soroush Kashani <soroush.kashani@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18012>

venus: Document VK_EXT_texture_compression_astc_hdr support

Signed-off-by: Chad Versace <chadversary@chromium.org>
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18000>

venus: Enable VK_EXT_tooling_info

Implement it locally with the common implementation.

Signed-off-by: Chad Versace <chadversary@chromium.org>
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18000>

venus: Enable VK_EXT_private_data

Implement it as hybrid native/passthrough as a workaround until venus
learns how to deep surgery on pNext chains. Eventually, we want to
implement it purely natively. For details, see the big code comment.

See: https://gitlab.freedesktop.org/virgl/virglrenderer/-/merge_requests/908
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Signed-off-by: Chad Versace <chadversary@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18000>

venus: Enable VK_KHR_format_feature_flags2

Signed-off-by: Chad Versace <chadversary@chromium.org>
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18000>

venus: Enable VK_EXT_ycbcr_2plane_444_formats

Signed-off-by: Chad Versace <chadversary@chromium.org>
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18000>

venus: Enable VK_EXT_texel_buffer_alignment

The patch looks asymmetic because the extension was promoted to 1.3 but
not its feature struct.

Signed-off-by: Chad Versace <chadversary@chromium.org>
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18000>

venus: Enable VK_EXT_subgroup_size_control

Signed-off-by: Chad Versace <chadversary@chromium.org>
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18000>

venus: Enable VK_KHR_zero_initialize_workgroup_memory

Signed-off-by: Chad Versace <chadversary@chromium.org>
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18000>

venus: Enable VK_KHR_shader_terminate_invocation

Signed-off-by: Chad Versace <chadversary@chromium.org>
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18000>

venus: Enable VK_KHR_shader_non_semantic_info

This extension has no Vulkan API, only SPIR-V API.

Signed-off-by: Chad Versace <chadversary@chromium.org>
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18000>

venus: Enable VK_KHR_shader_integer_dot_product

Signed-off-by: Chad Versace <chadversary@chromium.org>
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18000>

venus: Simplify vn_GetPhysicalDeviceProperties2

Reduce approximately 220 lines to 80.

Signed-off-by: Chad Versace <chadversary@chromium.org>
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18000>

venus: Simplify vn_GetPhysicalDeviceFeatures2

Reduce approximately 300 lines to 60.

Signed-off-by: Chad Versace <chadversary@chromium.org>
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18000>

vulkan: Make vk_copy_struct_guts() public

Venus wants to copy struct guts too.

Signed-off-by: Chad Versace <chadversary@chromium.org>
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18000>

venus: Sort passthrough extensions

Only the Vulkan 1.3 extensions were unsorted.

Signed-off-by: Chad Versace <chadversary@chromium.org>
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18000>

aco: remove dead code for querying image size/samples/levels

ac_nir_lower_resinfo() now lowers these.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17991>

radv: enable ac_nir_lower_resinfo for ACO

fossil-db (navi21):
Totals from 4022 (2.98% of 134913) affected shaders:
VGPRs: 207824 -> 205264 (-1.23%); split: -1.25%, +0.02%
CodeSize: 18486240 -> 18778488 (+1.58%); split: -0.00%, +1.58%
MaxWaves: 84398 -> 84882 (+0.57%)
Instrs: 3408968 -> 3470936 (+1.82%); split: -0.00%, +1.82%
Latency: 43493420 -> 43168394 (-0.75%); split: -1.57%, +0.82%
InvThroughput: 6911341 -> 6816816 (-1.37%); split: -1.41%, +0.05%
VClause: 88256 -> 82728 (-6.26%); split: -6.31%, +0.04%
SClause: 130061 -> 131526 (+1.13%); split: -0.39%, +1.51%
Copies: 219849 -> 238548 (+8.51%); split: -0.05%, +8.56%
Branches: 79168 -> 79174 (+0.01%); split: -0.00%, +0.01%
PreSGPRs: 180683 -> 184952 (+2.36%); split: -0.17%, +2.53%
PreVGPRs: 179716 -> 178956 (-0.42%)

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17991>

nir/lower_tex: ignore width of cube textures

On AMD hardware, height is faster to access and we're already doing so.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17991>

aco: add SCC clobber in build_cube_select

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17991>

nir/print: support nir_texop_descriptor_amd

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Fixes: 3098000e712 ("nir: add nir_texop_descriptor_amd")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17991>

radeonsi: move set_patch_vertices into si_state_shaders.cpp

it's a better place for it

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18195>

radeonsi: move patch_vertices-related tessellation updates out of si_draw

This only depends on the patch_vertices and the TCS.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18195>

radeonsi: remove 1 draw packet order codepath, keep the first one

Multi-mode multi-draws will make it more complicated, so let's start with
simpler code.

I changed the order a little: I put the VBO update next to emit_draw_packets.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18195>

radeonsi: cosmetic changes in si_emit_rasterizer_prim_state

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18195>

radeonsi: set GS_STATE_OUTPRIM and PROVOKING_VTX_INDEX only when they change

This moves setting those registers from an unconditional place in draw_vbo
into si_set_rasterized_prim (for draw_vbo), si_update_rasterized_prim
(for bind_xx_shader), and si_bind_rs_state.

It's a little more complicated than expected.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18195>

radeonsi: precompute GS_OUT_PRIM in advance

We don't have to do it every draw now if the rasterized prim type
doesn't change.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18195>

radeonsi: unify the logic that sets rast_prim

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18195>

radeonsi: move fixing ngg_culling into si_update_shaders

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18195>

radeonsi: remove the prim_restart_tri_strips_only option

Not used enough, no difference in performance for Dirt Rally on 6800.
Move the variable down.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18195>

radeonsi: move *rs to its only use in si_draw

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18195>

radeonsi: make the primitive type constant with tessellation

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18195>

radeonsi: don't pass num_patches via derived_tess_state, pass it via si_context

This removes the parameter from si_emit_derived_tess_state and uses
si_context to pass it. This rework is needed for multi-mode draws
where num_patches will be needed much later.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18195>

radeonsi/ci: add skips of tests not built on Linux

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18195>

panfrost: Use util/bitpack_helpers.h

Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18062>

asahi: Use util/bitpack_helpers.h

Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18062>

util: Add util_bitpack_[su]fixed_clamp helpers

These clamp the value to the fixed-point range instead of asserting.

Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18062>

broadcom: Use util/bitpack_helpers.h

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18062>

util,intel: Pull the bit packing helpers from genxml to a common header

Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18062>

zink: defer all compute shader creation

with other various changes, this should enable the entirety of compute
shader creation to occur in the compile thread

it also enables (slightly) simplifying shader create by moving cube check
out to the caller

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18197>

zink: add function to check whether a shader has cube samplers

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18197>

zink: generate sha1 for compute programs

this got broken in the move to u_live_shader_cache, which does not
handle compute programs and thus does not populate the shader's sha1

cc: mesa-stable

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18197>

zink: init fbfetch/bindless ctx stuff on shader create

deferring until later will be problematic for threads

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18197>

zink: add a context pointer to zink_program struct

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18197>

zink: add locking for descriptor layout caches

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18197>

zink: add partial async shader compile support

when compute shaders can be precompiled, they can be precompiled asynchronously
which allows the implementation of the parallel shader compile hooks

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18197>

zink: add precompilation for compute shaders

for compute shaders that don't need spec constants or cube lowering,
precompiles are possible and can be performed immediately after disk
cache lookup completes

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18197>

zink: move compute descriptor binds down in execution

this has to occur after the pipeline has been updated to account for
async compile

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18197>

zink: add a param to allow zink_screen_update_pipeline_cache to run directly

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18197>