turnip: move up LRZ invalidate in CmdClearAttachments

There is an early return if cmd->state.predication_active is true, so do
the LRZ invalidate before that.

Fixes: 2f79e0066405 ("turnip: disable LRZ on vkCmdClearAttachments()")

Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7899>

turnip: do not emit draw states in draw_cs outside of renderpass

This avoids a possible issue with MSAA sysmem clears, which use a 3D clear
path which assumes draw states are disabled, and are emitted in draw_cs in
BeginRenderPass.

(checking for TU_CMD_DIRTY_DRAW_STATE also allows not emitting the draw
states if they will be re-emitted on the next draw anyway. the previous
patch makes it so TU_CMD_DIRTY_DRAW_STATE is always set outside of
renderpasses)

Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7899>

turnip: correctly disable draw states outside of renderpasses

* do the disable in EndRenderPass2 to fix the missing disable for sysmem
* we don't need a disable at the end of every tile, or between binning pass
  and gmem pass (the first draw in draw_cs emits all the draw states)

Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7899>

turnip: always emit LRZ draw state in DIRTY_DRAW_STATE path

The packet size is constant and assumes all states, except for the 2 input
attachment states. (this means we get an invalid packet if DIRTY_LRZ isn't
set when DIRTY_DRAW_STATE is set).

Fixes: 3c07a1499863 ("turnip: enable LRZ")

Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7899>

turnip: do not include compute stage in pipeline_builder

This avoids emitting compute-related state in the graphics pipeline
(tu6_emit_xs_config was being called for compute stage).

Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7899>

turnip: no linear_to_srgb for alpha channel for gmem clear value packing

Alpha channel is always linear (oops).

Fixes: ddac5933f8f3 ("turnip: call packing functions directly for pack_gmem_clear_value")

Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7899>

gallium/u_threaded: optimize set_constant_buffer

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7719>

gallium/u_threaded: don't make a local copy of pipe_draw_start_count

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7719>

gallium/u_threaded: don't copy the indexbuf pointer if we overwrite it

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7719>

gallium/u_threaded: set has_user_indices = false in the driver thread

to remove some overhead from the main thread.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7719>

gallium/u_threaded: don't pass index bounds to the driver to decrease overhead

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7719>

glthread: count batch space in units of uint64_t elements

This removes one x86 shr instruction from _mesa_glthread_allocate_command.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7719>

glthread: change sizes to unsigned or size_t where needed

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7719>

glthread: use uint64_t to declare the batch buffer instead of align(8)

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7719>

glthread: use glthread->used instead of glthread->next_batch->used

remove one pointer dereference by having "used" in glthread too

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7719>

zink: use shader keys for samplemask

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7801>

zink: change a memcmp==0 to !memcmp

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7801>

zink: initial shader key implementation

this is a squashed version of a previously-reviewed series
ref https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7193

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7801>

vulkan/device_select: Store Vulkan vendorID and deviceID as uint32_t

Vulkan uses 32-bit IDs, compared to PCI/USB 16-bit ones.

Some driver vendorIDs do exceed 0xffff, including MESA (used by lavapipe).

Without this, the value will be truncated, and device select layer
will not match expected (or any) device.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/3935
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7977>

gallium/aux: Update scons build for u_tracepoints.[ch]

Signed-off-by: Rob Clark <robdclark@chromium.org>
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7974>

gallium/aux: Split u_tracepoints.[ch] generation

Allow for separate rules to be used to generate the .c and .h (partially
to make it easier for scons build, and partially because that is just a
better practice).

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7974>

gallium/aux: Avoid creating queue when traces not enabled

Avoids unnecessarily creating an extra thread when tracing is not
enabled.  Which also side-steps a chrome/ium sandbox issue with
the setscheduler call in u_queue.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7974>

nir: improve divergence analysis for loads with non-uniform resources

If ACCESS_NON_UNIFORM is not specified, we can assume the resource is
uniform. This requires nir_lower_non_uniform_access to remove that flag.

A few Detroit: Become Human shaders use a index sourced from a fragment
input which is expected to be uniform.

shader-db (Navi):
Totals from 8 (0.01% of 127638) affected shaders:
SGPRs: 224 -> 384 (+71.43%)
VGPRs: 208 -> 112 (-46.15%)
CodeSize: 5360 -> 5344 (-0.30%); split: -1.49%, +1.19%
Instrs: 1036 -> 1028 (-0.77%); split: -1.93%, +1.16%
VMEM: 1320 -> 608 (-53.94%)
SMEM: 384 -> 336 (-12.50%); split: +14.58%, -27.08%
VClause: 24 -> 16 (-33.33%)
SClause: 48 -> 56 (+16.67%)
PreSGPRs: 124 -> 216 (+74.19%)
PreVGPRs: 168 -> 88 (-47.62%)

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/5201>

nir/lower_non_uniform: remove non_uniform flags after lowering

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/5201>

iris: fix memleak for query_buffer_uploader

In the Chrome WebGL Aquarium  stress test, 20 instances of Chrome will run
Aquarium  simultaneously over 20+ hours. That causes Chrome crash.
During the stress, glBeginQueryIndexed is called frequently.

1.Each query will only use 32 bytes from query_buffer_uploader. After the offset
exceed 4096, it will alloc new buffer for query_buffer_uploader->buffer
and release the old buffer.

2.But iris_begin_query will call u_upload_alloc when the offset changed, and it
will increase the query_buffer_uploader->buffer->reference.count every time
when it called u_upload_alloc.

3.So when u_upload_release_buffer try to release the resource of
query_buffer_uploader->buffer, its reference.count is
already equal to 129. pipe_reference_described will only decrease its reference
count to 128.So it never called old_dst->screen->resource_destroy.

4.The old resouce bo will never be freeed. And chrome will called mmap every time
when it alloc new resource bo.

5. Chrome process map too many vmas in its process. Its map count exceed the
sysctl_max_map_count which is 65530 defined in kernel.

6. When iris_begin_query want to alloc new resource bo, it will meet NULL pointer
because mmap return failed. Finally chrome crashed when it access this NULL resource
bo.

The fix is decrease the reference count in iris_destroy_query.

Patch is verified by chrome webgl Aquarium test case for more than 72 hours.

Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Signed-off-by: Yang Shi <yang.a.shi@intel.com>
Reviewed-by: Alex Zuo <alex.zuo@intel.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7890>

aco: use UINT64_C on 64 bit constant arguments

avoids errors seen when building on OpenBSD/amd64

../src/amd/compiler/aco_instruction_selection.cpp:1677:62: error: ambiguous conversion for functional-style cast from 'unsigned long' to 'aco::Operand'
            bld.vop3(aco_opcode::v_mul_f64, Definition(dst), Operand(0x3FF0000000000000lu), tmp);
                                                             ^~~~~~~~~~~~~~~~~~~~~~~~~~~
glibc uses unsigned long for uint64_t on LP64 archs and unsigned long long for
uint64_t on ILP32 archs.  On OpenBSD unsigned long long is used for uint64_t
on all archs.

The Operand constructors are uint8_t uint16_t uint32_t uint64_t
use UINT64_C so lu or llu suffix will be used as needed.

Fixes: df645fa369d ("aco: implement VK_KHR_shader_float_controls")
Signed-off-by: Jonathan Gray <jsg@jsg.id.au>
Reviewed-by: Tony Wasserka <tony.wasserka@gmx.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7944>

mesa/st: lower 64 bit ops to scalar before lowering to soft-float

The fp64 emulation is not prepared for vectorized 64 bit code, so
if the driver doesn't ask for lowering to scalar by itself, do it before
lowering to soft-fb, and run a vectorizazion pass afterwards.

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7961>

radv: disable alphaToOne feature

The feature was exposed but completely ignored by the driver. Other
AMD drivers don't expose it as well, probably because it's complicated
to implement alpha-to-coverage properly. Let's disable it.

Cc: mesa-stable.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7966>

driconf: add allow_incorrect_primitive_id option

And enable it for SPECviewperf.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7078>

vbo/dlist: only use merged primitives when it's ok to do so

Merging primitives generates incorrect gl_PrimitiveID[In] values.
So make merged primitives construction non-destructive and fallback
to drawing with original primitives if a program reads gl_PrimitiveId.

This commit adds _mesa_update_primitive_id_is_unused modeled after
_mesa_update_allow_draw_out_of_order to update ctx->_PrimitiveIDIsUnused
each time shaders are updated.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7078>

vbo/dlist: avoid splitting draw commands in multiple draws

For (Multi)DrawArrays and (Multi)DrawElements commands, the storage size
needed are known early so we can make sure that the prim_store/vertex_store
will be big enough to store the whole command.

This reduces the amount of drawcalls in snx03 tests. For instance in test10:

      | Num draw calls |     GPU-load    |
------|----------------|-----------------|
      | Before | After |  Before | After |
------|--------|-------|---------|-------|
test10|  35k   |   8k  |   58%   |  80%  |

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7078>

vbo/dlist: refactor prim_store/vertex_store allocations

This will be used in the next commit.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7078>

mesa: optimize _mesa_program_resource_location

- xxhash is faster than sha1.
- remove superfluous calls to strlen

Using SPECviewperf13 snx-03 first subtest and "perf -e cycles -g", perf report says:

 Before  | After  | Function
---------|--------|---------------
  47.39% | 47.36% | _mesa_CallList
   5.00% |  3.03% | _mesa_program_resource_location

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7078>

util/hash_table: add _mesa_hash_data_with_seed function

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7078>

vbo/dlist: implement primitive merging

Merge consecutive primitives using the same mode while constructing the index buffer.

This improves performance a lot (x3 - x10) SPECviewperf13 snx-03 test by reducing the
number of draw calls per frame.
Here are some numbers for 4 of the tests:

      | Num draw calls |     GPU-load    |
------|----------------|-----------------|
      | Before | After |  Before | After |
------|--------|-------|---------|-------|
test1 |  390k  |  16k  |   68%   |  90%  |
test2 |  370k  |  16k  |   40%   |  90%  |
test3 |  1.2M  |  35k  |   38%   |  78%  |
test10|  3.5M  |  35k  |   36%   |  58%  |

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7078>

vbo/dlist: convert LINE_STRIPS to LINES

Less primitive modes allows for better primitive merging.

Lines are always used (instead of picking dynamically lines or line
strips for instance) because:
- they don't need primitive restarts to be merged
- they perform better (at least on radeonsi) - SPECviewperf13 snx subtests
  with lines (like 4 or 10) are 1.5x-2x faster.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7078>

vbo/dlist: create an index buffer in compile_vertex_list

This will be useful to implement other optimizations.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7078>

dlist: do not call _mesa_lookup_list twice

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7078>

android: radv: add libcutils shared dependency

Fixes the following building error:

    FAILED: out/target/product/x86_64/obj/SHARED_LIBRARIES/vulkan.android-x86_intermediates/LINKED/vulkan.android-x86.so
    ...
    ld.lld: error: undefined symbol: property_get
    >>> referenced by os_misc.c:193 (external/mesa/src/util/os_misc.c:193)
    >>>               os_misc.o:(os_get_option) in archive out/target/product/x86_64/obj/STATIC_LIBRARIES/libmesa_util_intermediates/libmesa_util.a

Fixes: eeecc21d ("util: Add property_get() fallback for android")
Reviewed-by: Marijn Suijten <marijn.suijten@somainline.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7861>

mesa/math: Fix address of array always returning true

    external/mesa3d/src/mesa/math/m_matrix.c:1403:13: error: address of array 'mat->inv' will always evaluate to 'true' [-Werror,-Wpointer-bool-conversion]
       if (mat->inv && (mat->flags & MAT_DIRTY_INVERSE)) {
           ~~~~~^~~ ~~

Fixes: 3175b63a0df ("mesa: don't allocate matrices with malloc")
Reviewed-by: Rob Clark <robclark@freedesktop.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7861>

android: util: Add libcutils to Android.mk shared libs

Otherwise:

    external/mesa3d/src/util/os_misc.c:59:12: fatal error: 'cutils/properties.h' file not found
    #  include <cutils/properties.h>
               ^~~~~~~~~~~~~~~~~~~~~

And:

    ld.lld: error: undefined symbol: property_get
    >>> referenced by os_misc.c:193 (external/mesa3d/src/util/os_misc.c:193)
    >>>               os_misc.o:(os_get_option) in archive out/target/product/discovery/obj/STATIC_LIBRARIES/libmesa_util_intermediates/libmesa_util.a

Fixes: eeecc21d935 ("util: Add property_get() fallback for android")
Acked-by: Rob Clark <robclark@freedesktop.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7861>

zink: Cap PIPE_SHADER_CAP_MAX_CONST_BUFFERS to 32

PIPE_MAX_CONSTANT_BUFFERS is 32, however many Vulkan implementations
has maxPerStageDescriptorUniformBuffers that exceeds it, for example:

radv 8388606,
anv 64
nvidia 1048580 for RTX 2000 and up.

and, together with the current zink logic, the returned value
will exceed the maximum allowed value for the cap.

This causes cso_destroy_context to pass big values back to zink
(via zink_set_constant_buffer), resulting in access beyond end of
allocated buffer for all UBOs.

Cap the cap to PIPE_MAX_CONSTANT_BUFFERS (32), not INT_MAX.

Add an assert to verify future drivers.

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Fixes: daaf5f1d186 ("gallium: Fix leak of currently bound UBOs at CSO context destruction.")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7976>

gallium: fix the PIPE_SHADER_CAP_SUPPORTED_IRS value for all drivers

Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7826>

st/mesa: remove less useful debug options in hot paths

Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7826>

st/mesa: don't generate TGSI for the draw VS because it now supports NIR too

Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7826>

draw: add NIR support to draw_create_vertex_shader

Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7826>

zink: fix channel ordering in format-mapping

This looks like a typo. Packed vulkan formats should always map to the
inverse order of the corresponding gallium notation. Besides, it makes
no sense that unsigned and signed formats have different ordering.

Fixes: cdfb1d925f3 ("zink: add last few format maps for ARB_vertex_type_2_10_10_10_rev")
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7964>

r600/sfn: Add support for shader_clock

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7973>

r600/sfn: Fix dest-swizzle for GS vertex loads

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7962>

gallium: Fix VAAPI postproc blit

Fixes the VAAPI postproc issue mentioned in this comment
(https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6736#note_626808)
without changing the height of the underlying resource when doing the
blit.

This commit removes the 0.5 pixel center offset from the compute blit - VAAPI postproc is the only function that uses this compute blit.

Fixes: 49465babdb3 ("frontends/va/postproc: Use the actual image height when blitting")
Signed-off-by: Thong Thai <thong.thai@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7970>

nir/lower_non_uniform: Better handle non-derefs

In particular, if we have an index or bindless handle we were passing
the original handle which, technically, is uniform within the context of
the if.  However, we can save the back-end compiler some effort if we
pass it the result of the read_first_invocation().

(Rebased by Kenneth Graunke and Rhys Perry.)

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7592>

nir/lower_non_uniform: Refactor for better code organization

In theory, I don't think this is a functional change.  We should
generate the same code before and after.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7592>

nir/lower_non_uniform: Use nir_read_first_invocation helper.

We now have a general helper for this, and don't need to roll our own.

Suggested by Rhys Perry in the review for MR !7592.

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7592>

gallium: do not reset buffers for unsupported stages

There's no good reason why drivers that doesn't grok geometry,
tesselation or compute shaders needs to deal with them.

This fixes a crash on a lot of Piglit tests for Zink.

Fixes: daaf5f1d186 ("gallium: Fix leak of currently bound UBOs at CSO context destruction.")
Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7971>

llvmpipe: work around mingw compiler optimization bug

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/3906

Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7914>

radv: fix clearing FMASK for layered MSAA images on GFX9+

If we always clear the whole FMASK buffer, layers can be corrupted.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/3710
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7924>

ac/surface: initialize the FMASK slice size for GFX9+

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7924>

radv: advertise VK_EXT_sample_locations on GFX10+

Only MSAA2x and MSAA4X sample locations can be used.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7850>

radv: enable using MSAA2x and MSAA4x sample locations on GFX10+

These failures are really weird but MSAA2x and MSAA4x work fine.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7850>

radv: Implement VK_VALVE_mutable_descriptor_type.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7967>

vulkan: Update to 1.2.164.

Acked-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7967>

android: freedreno/ir3: use python3 in gen rules

Completes freedreno gen rules migration to python3 as per meson.build
With this change all freedreno gen rules use $(MESA_PYTHON3)

Acked-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7942>

android: freedreno: Remove fd_log()

Fixes the following building error:

FAILED: ninja: 'external/mesa/src/gallium/drivers/freedreno/freedreno_log.c',
needed by 'out/target/product/x86_64/obj_x86/STATIC_LIBRARIES/libmesa_pipe_freedreno_intermediates/freedreno_log.o',
missing and no known rule to make it

Fixes: 03e7c93b828a ("freedreno: Remove fd_log()")
Acked-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7942>

android: freedreno: Add GPU tracepoints

Changelog:
- add freedreno_tracepoints.c.{c,h} gen rules for Android $(MESA_PYTHON3)
- update Makefile.sources with the required generated sources

Fixes the following building errors:

external/mesa/src/gallium/drivers/freedreno/freedreno_gmem.c:35:10:
fatal error: 'u_tracepoints.h' file not found
         ^~~~~~~~~~~~~~~~~
1 error generated.

FAILED: out/target/product/x86_64/obj/SHARED_LIBRARIES/gallium_dri_intermediates/LINKED/gallium_dri.so
...
ld.lld: error: undefined symbol: __trace_end_clear_restore
>>> referenced by freedreno_tracepoints.h:38 (out/target/product/x86_64/obj/STATIC_LIBRARIES/libmesa_pipe_freedreno_intermediates/
freedreno_tracepoints.h:38)
...
ld.lld: error: undefined symbol: __trace_start_vsc_overflow_test
>>> referenced by freedreno_tracepoints.h:272 (out/target/product/x86_64/obj/STATIC_LIBRARIES/libmesa_pipe_freedreno_intermediates
/freedreno_tracepoints.h:272)
ld.lld: error: too many errors emitted, stopping now

Fixes: a02dcb970fff ("freedreno: Add GPU tracepoints")
Acked-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7942>

android: gallium/aux: Add GPU tracepoint mechanism

Android rules to build u_trace sources and u_tracepoints generated sources

Changelog:
- add util/u_tracepoints.{c,h} gen rules for Android using $(MESA_PYTHON3)
- update Makefile.sources with the required sources and generated sources

Fixes: 3471af9c6cfa ("gallium/aux: Add GPU tracepoint mechanism")
Acked-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7942>

android: gallium/aux: update old generated sources rules

This is in preparation for additional generated sources rules for Android
which will require ad hoc rules, so it is necessary to replace old ones

NOTE: pre-existing gen rules based on $(transform-generated-source) macro
are both obsolete and use of '%' pattern rule is incompatible with ad hoc
python commands for different targets

Changelog:
- remove util/u_format_srgb.c target
- replace obsolete indices/{u_indices,unfilled}_gen.c 'common' gen rules
  with 'per target' gen rules using $(MESA_PYTHON3) as per meson gen rules

Fixes: 3471af9c6cfa ("gallium/aux: Add GPU tracepoint mechanism")
Acked-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7942>

ci: update the list of expected failures for RADV/FIJI

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7958>

radv: fix using FS sample shading if the linker optimized inputs away

During NIR linking, constant varyings might be moved to the next
stage and the sample qualifier removed.

shader_info::uses_sample_shading remembers if the sample qualifier
was used before optimizations.

No fossils-db changes on Sienna Cichlid.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7892>

lavapipe: fix logic-op support

Lavapipe exposes support for the logicOp feature, but doesn't actually
respect the state. This is easy to fix, so let's plumb it through.

This fixes spec@!opengl 1.0@gl-1.0-logicop When running with Zink on
Lavapipe.

Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7932>

zink. Fixing vkGetPhysicalDeviceProperties2 and vkGetPhysicalDeviceFeatures2 for Vk 1.1 and VK_KHR_get_physical_device_properties2.

MoltenVK does not export the vkGetPhysical*2() functionns, even in Vulkan 1.2.154.0 where the instance version moves from 1.0 to 1.1.
If the extension is present and used the KHR versions of the functions can be used.
From the spec the vkGetPhysicalDevice*2() functions should be avaiable from Vk 1.1 loaders and devices. Which implies MoltenVK might be misbehaving.
This change allows the extension to be used, if present, before the Vk 1.1 version check.

Fixes: 752f6d80 ("zink: setup version dependent VkPhysicalDeviceVulkan*Features and VkPhysicalDeviceVulkan*Properties.")
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7960>

radv: use 32-bit predication for skipping FCE on GFX10.3+

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7897>

radv: set the predication boolean as 32-bit if necessary

CTS is missing tests.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7897>

radv: fix exporting multiviews with NGG

If a subpass uses multiview but the fragment shader doesn't load it
we still have to export it.

Cc: mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7815>

radv: mark GFX10.3 as a non-conformant Vulkan implementation

In theory, GFX10.3 is not considered to be a conformant Vulkan
implementation because we didn't submit a conformance submission
package.

Cc: mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7913>

radeonsi: fix regression on gpus using the radeon winsys.

For GPUs using the radeon kernel driver, num_se was never
getting initialised.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/3939
Fixes: f2977a162af4 ("ac: fix min/max_good_num_cu_per_sa on gfx10.3 with disabled SEs")
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7954>

radeonsi/gfx10: added support for gfx10 conditional rendering

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7526>

r300,r600,radeonsi: inline struct radeon_cmdbuf to remove dereferences

It's straightforward except that the amdgpu winsys had to be cleaned up
to allow this.

radeon_cmdbuf is inlined and optionally the winsys can save the pointer
to it. radeon_cmdbuf::priv points to the winsys cs structure.

Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7907>

panfrost: Only set varyings and varying_buffers when varying_count > 0

This removes the "warn: No Varying records" printed by pandecode.

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/7921>

panfrost: Fix draw descriptor definition

Move the 2 words padding out of the draw descriptor to fix Midgard tiler
job decoding.

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/7921>

panfrost: Fix texture payload decoding

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/7921>

isl: Fix android build

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Suggested-by: Mauro Rossi <issor.oruam@gmail.com>
Fixes: f08d8c849e0df3 ("intel/isl: Build gen 12.5")
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/3931
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7938>

ci: Run sanity job automatically for forked branches as well

This should make running the CI for forked branches less painful again.

Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7904>

ci: Manual test jobs don't need the Git tree

If they do need something, it should be passed via artifacts from a
build job.

Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7904>

ci: sanity job doesn't need the Git tree

This should fix the sanity job taking >= 1 minute in some cases.

Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7904>

ci: Drop x86_build_old image

Currently not used for anything.

v2:
* Drop build script as well (Eric Anholt)

Fixes: a3543adc2628 "clover: set LLVM min version to 8.0.1"
Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7902>

ci: Append $MESA_TEMPLATES_COMMIT to image tags

This ensures all images get rebuilt when we update to a newer
ci-templates commit.

v2:
* Append to WINDOWS(_UPSTREAM)_IMAGE instead of WINDOWS_TAG. The latter
  failed, apparently variables are not expanded recursively on the
  Windows runners.
* Use separate MESA_IMAGE_TAG/MESA_BASE_IMAGE variables instead of
  appending to each FDO_DISTRIBUTION_TAG/FDO_BASE_IMAGE separately for
  Linux jobs, to prevent accidentally dropping the suffix.

Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7902>

ci: Define global variable MESA_TEMPLATES_COMMIT for ci-templates commit

No functional change.

Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7902>

ci: Run git gc before creating Git cache tarball

Since the job which creates the cache tarball starts from the previous
cache, the cache kept accumulating cruft and growing bigger.

This cuts the size of the tarball in half (from almost 600M to under
300M), which can translate to significant time savings when downloading
it on some runners.

v2:
* Use git gc --aggressive (Eric Anholt)

Reviewed-by: Andres Gomez <agomez@igalia.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7920>

meson: Fix Clang microsoft-enum-value detection.

Fixes: 3aee462781a ("meson: add windows compiler checks and libraries")
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Dylan Baker <dylan@pnwbakers.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7543>

gallium/osmesa: Return cleanly for OSMesaGetDepthBuffer() with no depth.

This makes our behavior match classic.

Closes: #2034
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7886>

gallium/osmesa: Fix leak of the ST manager/api on library unload.

Closes: #883
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7886>

gallium/osmesa: Fix data race on setting up the ST API.

We're going to need it anyway, put it in the same call_once as the ST
manager.

Closes: #880
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7886>

gallium/osmesa: Remove the broken buffer-reuse scheme.

Besides leaking and a lack of thread-safety, it would also incorrectly
share front buffers if multiple contexts happened to use the same
size/format, as demonstrated by the new unit test.

Closes: #2035
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7886>

gallium/osmesa: Fix flushing and Y-flipping of the depth buffer.

We were returning a pointer to use-after-free the depth buffer, not
updating it in after future rendering, and also not y flipping it.  A
little refactor to mostly reuse the color buffer's path makes it easy to
do it all right.

Adds a unit test to check for these bugs.

Closes: #885
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7886>

zink: assert all index values in ntv OpAccessChain constructor

this is super annoying

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7933>

xmlconfig: Warn if parsing the engine/app versions fails.

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7048>

xmlconfig: Add unit tests for recent bugs in the driconf rewrite.

This covers:
7fb4ab9ec122 ("driconf: Restore the ability to override driconf with the environment.")
2b977adff805 ("xmlconfig: fix scandir_filter")

and touches a bit more of drirc logic while I'm here.

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7048>

docs: update calendar and link releases notes for 20.2.4

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7931>

relnotes: Add sha256sums for 20.2.4

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7931>

docs: add release notes for 20.2.4

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7931>