zink: assert precise queries are occlusion queries

this should always be the case

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12583>

zink: require occlusionQueryPrecise for occlusion queries

ensure this is present to avoid driver explosions

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12583>

zink: always init bordercolor value for sampler

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12574>

zink: implement PIPE_QUERY_GPU_FINISHED

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12573>

v3dv: Implement VK_EXT_pipeline_creation_feedback

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12371>

microsoft/compiler: trivial fixes to error-handling

We're really bad at making sure we report errors when we fail to
allocate memory. This makes us a bit better...

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12541>

microsoft/compiler: return errors from get_n_src

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12541>

microsoft/compiler: remove needless error-returns

There's no root error-conditions in this code, just code that assumes
they exist and tries to handle them.

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12541>

nir: fix ifind_msb_rev by using appropriate type

As you can see comparion "x < 0" doesn't make
sense if x is unsigned.

Fixes: a5747f8a ("nir: add opcodes for *find_msb_rev and lowering ")
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12548>

nir: fix shadowed variable in nir_lower_bit_size.c

Fixes: 6d792989924 ("nir/lower_bit_size: fix lowering of {imul,umul}_high")
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12527>

venus: conditionally enable async descriptor set allocation

When VK_DESCRIPTOR_POOL_CREATE_FREE_DESCRIPTOR_SET_BIT is not used to
create the pool, set allocation is guaranteed to not return
VK_ERROR_FRAGMENTED_POOL, and we can safely move set allocation to async
after doing resource tracking in the driver.

Enable after fully tested with assert(false) in the failure case.

Tested with:
- dEQP-VK.api.descriptor*
- dEQP-VK.api.object_management.*
- dEQP-VK.binding_model.descriptor*
- dEQP-VK.descriptor_indexing.*

Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12501>

venus: check descriptor allocations against pool resource

Only kick in when async_set_allocation is enabled.

Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12501>

venus: descriptor set to track descriptor count of last binding

Track the descriptor count to be used instead of the variable descriptor
count to avoid duplicate checks in later accounting.

Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12501>

venus: descriptor pool to track pool state

It also tracks whether async set allocation is enabled.

Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12501>

venus: layout to track variable descriptor count binding info

Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12501>

venus: descriptor layout to track more binding infos

Rename existing max_binding to last_binding to be consistent.

1. layout to track last binding index
2. binding to track descriptor type
3. binding to track descriptor count

Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12501>

venus: add vn_descriptor_set_layout_init

Just a refactoring without functional changes.

Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12501>

venus: refactor failure path for sets allocation

Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12501>

dri2: Fix Null pointer dereferences

Fix defect reported by Coverity Scan CID-1490794

Fixes: 0d42033b26 ("glx/dri2: Require the driver to support v4 of __DRI_DRI2")
Signed-off-by: Sergii Melikhov <sergii.v.melikhov@globallogic.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12563>

etnaviv: remove primconvert

This is no longer used.

Signed-off-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Acked-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12516>

etnaviv: export supported prim types

This is now handled by gallium, so the related codepaths can be dropped.

Signed-off-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Acked-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12516>

broadcom/qpu: remove duplicated opcode variable

`opcode` is wrote twice.

Fixes CID 1490798:  Incorrect expression  (EVALUATION_ORDER).

Fixes: 8a5f2228dbb ("broadcom/qpu: add new lookup opcode description helper")
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12554>

aco: Use workgroup size from input shader info.

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12321>

radv: Remove superfluous workgroup size calculations.

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12321>

radv: Calculate workgroup sizes in radv_pipeline.

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12321>

ac: Calculate workgroup sizes of HW stages that operate in workgroups.

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12321>

ci: update the list of skipped tests for Fiji/RADV

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12553>

panvk: Initialize timestamp for disk cache

Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12552>

panfrost: Make pan_blit() return the tiler job pointer

The Vulkan driver needs to patch job headers when re-issueing a command
buffer. Return the tiler job pointer to allow that.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12552>

panfrost: Fix pan_blit_ctx_init() when start > end

This can happen when one wants to obtain a mirrored view. We need
to make sure the min <= max before emitting the viewport.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12552>

panfrost: Prepare indirect draw helpers to per-gen XML

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12551>

panfrost: Prepare indirect dispatch helpers to per-gen XML

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12551>

panvk: Prepare per-gen split

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12551>

panfrost: Only emit special attribute buffer entries on pre-v6 hardware

There's no special attributes on Bifrost.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12551>

panfrost: Move panfrost_modifier_to_layout() to pan_texture.c

This function is not used outside pan_texture.c.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12551>

panfrost: Get rid of all _packed structs in pan_context.h

Such that pan_context.h remains generic even after the per-gen XML
switch.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12551>

pan/gen_pack: Add parens around packed1/2 vars in pan_merge()

This way we can pass *desc instead of (*desc).

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12551>

panfrost: Add generic mappings for the gen-specific tiler descriptor macros

This way the transition to per-gen XML files gets simpler.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12551>

panfrost: Get rid of the mali_xxx enum redefinitions

The gen_macros.h header should include the common header file when
PAN_ARCH is undefined, thus making those redefinitions useless.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12551>

panfrost: s/[idep_]midgard_pack/[idep_]pan_packers/

We will soon have per-gen pack headers. Let's rename the meson variables
to reflect that.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12551>

panfrost: Stop including midgard_pack.h directly

Include gen_macros.h instead so we can smoothly transition to per-gen
XMLs.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12551>

pan/gen_macros: Include midgard_pack.h from gen_macros.h

This way we will only have a single entry point, thus making the
transition to per-gen XML files easier.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12551>

pan/gen_macros: Map {TEXTURE,SAMPLER} to the arch-specific descriptor

This will ease the transition to per-arch XML files. While at it, patch
pan_cmdstream.c to use this new macro and avoid an #if section.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12551>

pan/gen_macros: Move the TEXTURE definition to gen_macros.h

So we can use the same trick in panvk.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12551>

panfrost: Move the polygon list init logic to pan_cmdstream.c

This logic is Midgard specific and we need move the
panfrost_scoreboard_initialize_tiler() call to a per-arch file so
we can make panfrost_scoreboard_initialize_tiler a per-arch function.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12551>

pan/gen_pack: Add pan_size() and pan_align() macros

And replace all references to MALI_xxx_{LENGTH,ALIGN} by
pan_{size,alignment}(xxx) calls so we can give generic alias to
midgard/bifrost specific descriptors.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12551>

panvk: Add missing midgard_pack dependency

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12551>

pan/gen_pack: Generalize the PREFIX() trick

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12551>

intel/compiler: INT DIV function does not support source modifiers

BSpec says that for all generations.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/5281
CC: mesa-stable
Signed-off-by: Marcin Ślusarz <marcin.slusarz@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Sagar Ghuge <sagar.ghuge@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12518>

radeonsi: Change GS vertex offset arguments to use gs_vtx_offset array.

This makes RadeonSI behave like RADV, and hence makes it possible to
use the NIR based ESGS I/O lowering and NGG lowering.

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12511>

radv, ac, aco: Use indices 0-2 of gs_vtx_offset argument array on GFX9+.

Previously, indices 0, 2, 4 were used.
This worked, but it was somewhat unintuitive.
This commit changes it to use indices 0, 1, 2 instead, which
makes the code easier to understand.

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12511>

glsl: fix explicit-location ifc matching in presence of array types

We were treating each field as if it took up a single slot. However
that's not the case. And with strict matching (GLSL 4.20+ / ES 3.1+) we
would end up not matching identical interfaces.

Fixes: c4545676d7 ("glsl/linker: fix location aliasing checks for interface variables")
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12479>

meson: stop searching for execinfo

Signed-off-by: Simon Zeni <simon@bl4ckb0ne.ca>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12108>

util/u_debug_symbol: remove debug_symbol_name_glibc and execinfo dependency

Signed-off-by: Simon Zeni <simon@bl4ckb0ne.ca>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12108>

glapi/gl_gentable.py: drop call to backtrace on no op

The function uses glibc specific functions

Signed-off-by: Simon Zeni <simon@bl4ckb0ne.ca>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12108>

i915: remove use of backtrace and backtrace_symbols

The function `debug_backtrace` relied on glibc specific functions, and is
unused unless someone manually patches the source to set DEBUG_BACKTRACE_SIZE

Signed-off-by: Simon Zeni <simon@bl4ckb0ne.ca>
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12108>

anv/image: Don't assert that HiZ can be added

HiZ isn't yet enabled for Tile4/64.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12132>

iris: Disable tiled memcpy for Tile4

ISL's tiled memcpy functions don't support it yet.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12132>

iris: Disable the Y-tiled modifiers on XeHP+

I915_FORMAT_MOD_Y_TILED_CCS was already disabled for Xe+ due to a change
in the CCS layout. Disable the remaining Y-tiled modifiers since XeHP
lacks support for Y-tiling.

Rework:
* Nanley: Include Anuj's fix for the non-CCS modifiers.
* Nanley: Split out Anuj's fix into a separate if statement.
* Nanley: Rewrite commit message.

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12132>

intel/isl: Disable I915_FORMAT_MOD_Y_TILED on XeHP+

XeHP lacks support for Y-tiling.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12132>

intel: Add underscores to HALIGN and VALIGN enums

The HALIGN enums for XeHP already have underscores. Make the other
HALIGN and VALIGN enums conform.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12132>

intel: Update surface states for XeHP alignments

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12132>

intel/isl: Use a switch for HALIGN/VALIGN encoding

Avoid using a sparse and relatively large array for HALIGN encoding.
Additionally, this provides validation of the input alignment values.

Suggested-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12132>

intel/isl: Fix halign/valign of uncompressed views

We're going to start asserting for valid halign/valign values during
surface state emission. Pre-SKL, isl_surf_get_uncompressed_surf creates
uncompressed surfaces with invalid halign/valign values - 1x1. Fix this
by replacing the call to isl_surf_get_image_surf with isl_surf_init,
passing in the uncompressed format up-front.

As we're no longer using isl_surf_get_image_surf, we also need to get
the x and y offset of the image ourselves. Instead of getting a
sample-based offset, then converting to elements later on, we use
isl_surf_get_image_offset_B_tile_el to get the offset in elements
up-front.

With the above two changes, the generic code after the else-block is no
longer needed for the single-layer-view code path. We move it and
specialize it to the if-block (which is executed SKL+ and handles
multi-layer views).

Cc: mesa-stable
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12132>

intel/blorp: Fix Gfx7 stencil surface state valign

Stencil on Gfx7 has a vertical alignment element of 8, but the largest
its surface state can express is 4. Apply the Gfx6 solution of changing
the alignment in blorp_surf_retile_w_to_y.

Cc: mesa-stable
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12132>

intel/blorp: Fix faked RGB image alignment on XeHP

On XeHP, NPOT and POT formatted surfaces will use different image
alignment units when emitting surface states. When BLORP fakes an RGB
image as RED, update the image alignment to prevent assert failures when
emitting surface states.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12132>

intel: Support Tile4/64 in surface states

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12132>

intel: Support Tile4/64 in depth/stencil state

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12132>

intel/isl: Update tiling filter functions for XeHP

Enable the XeHP-specific tilings and restrict them to that platform.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12132>

intel/isl: Use an allow-list in gfx6_filter_tiling

Try to avoid having to update isl_gfx6_filter_tiling when new tilings
are added for new platforms. Note that the allow-list uses
ISL_TILING_ANY_Y_MASK and thus assumes that no new Y-tilings will be
added in the future.

Suggested-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12132>

intel/isl: Drop ISL_SURF_USAGE_DISPLAY_*_BIT

We haven't used these since their introduction.

Suggested-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12132>

intel/isl: Drop extra assert on array_pitch_el_rows

ISL already asserts that the variable is a multiple of the tile height
via isl_assert_div.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12132>

intel/isl: Size Tile64 surfaces with 4 dimensions

In order to size Tile64 surfaces correctly, make sure that the total
physical extent is arrayed. The code should handle 3D surfaces as well,
but is untested for now.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12132>

intel/isl: Update image alignments on XeHP

Implement the new XeHP alignment rules for surface layout.
RENDER_SURFACE_STATE objects still need updating, but that's left for a
separate commit.

Rework:
* Nanley: Include Sagar's VALIGN fix for D16_UNORM.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12132>

intel/isl: Define ISL_TILING_4/64 for XeHP

XeHP defines new tiling formats, Tile4 and Tile64. They are needed in
order to support depth/stencil surfaces and multisampling. Create new
ISL enums and define some initial tiling information in order to enable
them later on.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12132>

intel/isl: Add msaa_layout param to isl_tiling_get_info

The additional parameter will be used by Tile64.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12132>

intel/isl: Add more parameters to isl_tiling_get_info

They are not used yet but the layout of Yf and Ys tiles are dependent on
these parameters.  While we're here, better document the function.

Rework:
* Nanley: Update crocus.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Nanley Chery <nanley.g.chery@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12132>

zink: only remove programs from hash tables on shader deletion if needed

Reviewed-by: Hoe Hao Cheng <haochengho12907@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12532>

zink: set inlinable_uniforms_mask first when binding a shader

Reviewed-by: Hoe Hao Cheng <haochengho12907@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12532>

zink: add some null checks for shader variant key generation

Reviewed-by: Hoe Hao Cheng <haochengho12907@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12532>

zink: stop using dirty_shader_stages for shader binds

by only using this mask for variant updates, we can begin to do some
neat things

Reviewed-by: Hoe Hao Cheng <haochengho12907@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12532>

zink: simplify a bitmask init

Reviewed-by: Hoe Hao Cheng <haochengho12907@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12532>

zink: create compute programs on bind

this simplifies the launch_grid codepath and gets shader binds out
the shader update bitmask

Reviewed-by: Hoe Hao Cheng <haochengho12907@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12532>

zink: avoid hashing shader stages multiple times for new gfx programs

Reviewed-by: Hoe Hao Cheng <haochengho12907@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12532>

zink: split gfx shader cache based on stages present

this makes the pool of progs in a given hash smaller and also minimizes
the hashing required for a given program

Reviewed-by: Hoe Hao Cheng <haochengho12907@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12532>

zink: track mask of bound gfx shader stages

Reviewed-by: Hoe Hao Cheng <haochengho12907@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12532>

zink: inline program cache structs

derefs--

Reviewed-by: Hoe Hao Cheng <haochengho12907@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12532>

zink: return early when getting resource modifer if no modifier is used

Reviewed-by: Hoe Hao Cheng <haochengho12907@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12534>

zink: verify program key sizes before checking for default variant

currently keys for shaders are always the same size, but this will change
in the future, at which point the keysize becomes relevant

Reviewed-by: Hoe Hao Cheng <haochengho12907@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12547>

microsoft/compiler: Miscellaneous fixes from running clang-format

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12513>

microsoft/compiler: Emit a flat interpolation method for SV_SampleIndex

We do not want to set an interpolation method for vertex shader inputs.

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12513>

microsoft/compiler: Set the SampleFrequency runtime metadata

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12513>

microsoft/compiler: More robustly handle setting Register=-1

This is the 'N/A mask' case in the DXIL disassembly.

This logic is taken from: https://github.com/microsoft/DirectXShaderCompiler/blob/7c9e487afd17b7726229d8ceb3242a412134afe7/tools/clang/tools/dxcompiler/dxcdisassembler.cpp#L106

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12513>

microsoft/compiler: Add support for SV_SampleIndex intrinsic

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12513>

pan/bi: Fix format specifiers in disassembler

Flagged by cppcheck.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12530>

pan/bi: Remove unused clause_start field

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12530>

pan/bi: Assert l != NULL in bi_ra

Confuses cppcheck; indeed, the proof is confusing.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12530>

pan/bi: Simplify condition

Flagged by cppcheck.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12530>

panfrost: Remove unused functions

Flagged by cppcheck.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12530>

panfrost: Remove stale TODOs and XXXs

These don't meaningfully apply but their comments never got updated.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12530>