Commit graph

178152 commits

Author SHA1 Message Date
Martin Roukala (né Peres)
cf3fb0fee7 radv/ci: add more tests to the navi21 vkcts flake list
Signed-off-by: Martin Roukala (né Peres) <martin.roukala@mupuf.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25325>
2023-09-25 11:27:31 +00:00
Martin Roukala (né Peres)
0a065b5cf5 radv/ci: increase the parallelism of the vkcts-navi21 job
The vkcts-navi21 job currently take ~13 minutes each, but we are about
to start running the VK_EXT_shader_object tests which should
significantly increase our runtime and thus would exceed the 15 minutes.

Signed-off-by: Martin Roukala (né Peres) <martin.roukala@mupuf.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25325>
2023-09-25 11:27:31 +00:00
Juan A. Suarez Romero
bf32199386 v3d/shim: include new ioctl parameters
Some new parameters added recently were missed in the DRM shim.

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/9880
Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25370>
2023-09-25 10:47:58 +00:00
Samuel Pitoiset
219a14d093 radv: enable cache counters for RGP on GFX11
This was the last missing RGP feature on GFX11 compared to older chips.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25271>
2023-09-25 09:05:58 +02:00
Samuel Pitoiset
70a4290e28 radv: add SPM support for GFX11
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25271>
2023-09-25 09:05:58 +02:00
Samuel Pitoiset
301c402bd8 ac/spm: add support for GFX11
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25271>
2023-09-25 09:05:58 +02:00
Samuel Pitoiset
26c710b7c4 ac/spm: add new segment types for GFX11
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25271>
2023-09-25 08:34:31 +02:00
Samuel Pitoiset
97d87a431a ac/gpu_info: define AMD_MAX_WGP
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25271>
2023-09-25 08:34:31 +02:00
Dave Airlie
129e27a46e llvmpipe/linear: move to nir inputs
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24436>
2023-09-25 13:23:07 +10:00
Dave Airlie
90c4468ed0 llvmpipe/linear: convert to using nir for output.
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24436>
2023-09-25 13:23:07 +10:00
Dave Airlie
8477b97f1b llvmpipe/fs: start converting dervied state to nir based.
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24436>
2023-09-25 13:23:07 +10:00
Dave Airlie
b72e77bd97 llvmpipe/fs: start converting interp/input paths to nir.
This input_usage_mask is going to take a bit more thinking

Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24436>
2023-09-25 13:23:07 +10:00
Dave Airlie
2b09d2d715 llvmpipe/nir: call gather info to update inputs read properly
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24436>
2023-09-25 12:06:39 +10:00
Dave Airlie
5827034dfc gallivm: drop tgsi aos paths.
These were only used by the linear code, and that doesn't use
tgsi anymore

Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24436>
2023-09-25 12:06:37 +10:00
Dave Airlie
e9a41875da llvmpipe/linear: drop tgsi path.
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24436>
2023-09-25 12:05:50 +10:00
David Heidelberg
1f3446499f ci/freedreno: disable broke cheza (Adreno 630) runners
Revert ASAP as `google-freedreno-cheza` gets running again.

Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25365>
2023-09-25 02:04:28 +05:30
Gert Wollny
a57cc0ae0c r600/sfn: Fix use of scheduled_shader vs shader
Thanks to Vitaliy Kuzmin for pointing this out.

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25364>
2023-09-24 19:52:36 +02:00
Gert Wollny
34a068de6f r600/sfn: sort the uniforms of the right shader
Thanks to Vitaly Kuzmin for pointing this out.

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25364>
2023-09-24 19:52:36 +02:00
Faith Ekstrand
c04f313cc5 r600: Use nir_builder helpers for load/store_shared_r600
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24762>
2023-09-24 17:30:03 +00:00
Faith Ekstrand
28137a187b r600: Use more auto-generated nir_builder helpers
Use nir_umad24 instead of rolling our own.  Also, all system values
already have NIR builder helpers and we don't need to be manually
building intrinsics or rolling our own helpers.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24762>
2023-09-24 17:30:03 +00:00
Karmjit Mahil
feafb8a256 pvr: Fix dynamic desc offset storage
The index at which the dynamic descriptor offsets were being
stored was incorrect, leading to some offsets not being stored and
thus `0` being applied as the offset to the descriptors instead.

dEQP test fixed:
  dEQP-VK.binding_model.shader_access.{primary,secondary}_cmd_buf
    .uniform_buffer_dynamic
    .{vertex,fragment,compute,vertex_fragment}
    .multiple_discontiguous_descriptor_sets
    .*_descriptor.offset_view_{,non}zero_dynamic_nonzero

Fixes: aa791961a8 ("pvr: Add support for dynamic buffers descriptors")
Signed-off-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25320>
2023-09-24 14:00:57 +00:00
Karmjit Mahil
bf17e4fe33 pvr: Fix for loop itarator usage
The `i` iteration variable was being used instead of `k` when
appending new mappings, and getting the current source rect.

dEQP test fixed:
  dEQP-VK.pipeline.monolithic.image.suballocation.sampling_type
    .combined.view_type.3d.format.*.count_1.
    .{3x3x3,5x5x5,...(odd dimensions)}

Fixes: 060c3db4ef ("pvr: Complete pvr_generate_custom_mapping()")

Reported-by: James Glanville <james.glanville@imgtec.com>
Signed-off-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25321>
2023-09-24 13:44:19 +00:00
Faith Ekstrand
47b481cace nvk: Set GUARDBAND_Z_SCALE_1 when Z-clipping
This fixes most of the remaining dEQP-VK.*.inverted_depth_range.* tests.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25359>
2023-09-23 13:01:52 -05:00
Faith Ekstrand
d7f3fdc262 nvk: Enable dynamic clip/clamp enable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25356>
2023-09-23 14:58:22 +00:00
Faith Ekstrand
07091f0f94 nvk: Fix depth clipping parameters
This is a pretty complete overhaul which I think may actually be correct
and there are now comments explaining why various parameters were
chosen.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25356>
2023-09-23 14:58:22 +00:00
Faith Ekstrand
e317ba373b nvk: Clean up viewport math
This should make everything a bit easier to read

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25356>
2023-09-23 14:58:22 +00:00
David Heidelberg
17dfbc25d2 iris: do not mention specifically clover for OpenCL support
It's confusing, when rusticl is used.

Acked-by: Karol Herbst <kherbst@redhat.com>
Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25287>
2023-09-23 05:04:29 +00:00
David Heidelberg
a4e4bf8ccd ci/venus: add fragment.32B_in_memory_with_vec4_s32 flake
Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25349>
2023-09-23 04:44:37 +00:00
Felix DeGrood
b9c9fb7259 anv: fix frame count reporting in INTEL_MEASURE
Report frame count at CB submit time, instead of CB build time.

Reviewed-by: Mark Janes <markjanes@swizzler.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25235>
2023-09-22 22:41:53 +00:00
Faith Ekstrand
6f85e6b06b nvk: Disable statistics around meta ops
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25353>
2023-09-22 21:23:11 +00:00
Faith Ekstrand
169835da39 nvk: Lower interp_at_sample to interp_at_offset
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25353>
2023-09-22 21:23:10 +00:00
Frank Binns
53ed2ff742 pvr: remove pvr_pbe_get_src_pos()
This should have been removed as part of f59c6eef16 ("pvr: Move PBE START_POS
into csb enum helpers header").

Signed-off-by: Frank Binns <frank.binns@imgtec.com>
Reviewed-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25328>
2023-09-22 21:03:43 +00:00
Frank Binns
6619802446 pvr: treat VK_IMAGE_CREATE_MUTABLE_FORMAT_BIT as not supported
The required compiler support is missing for mutable format support, so return a
not supported error from vkGetPhysicalDeviceImageFormatProperties2() when the
VK_IMAGE_CREATE_MUTABLE_FORMAT_BIT flag is set in the image format info
structure.

This results in the dEQP-VK.image.mutable.* conformance tests being treated as
Not Supported.

Signed-off-by: Frank Binns <frank.binns@imgtec.com>
Reviewed-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25241>
2023-09-22 20:47:17 +00:00
Mike Blumenkrantz
865bdc3da2 lavapipe: set default min sample shading to 1
this should only affect shobj handling and even then probably no effect

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25354>
2023-09-22 20:03:06 +00:00
Mike Blumenkrantz
a2a848cb19 lavapipe: set default viewport and scissor count for cmdbufs
this otherwise breaks shobj tests

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25354>
2023-09-22 20:03:05 +00:00
Mike Blumenkrantz
0d975219ad lavapipe: more vertex stride fixups
for handling cases where bind2 and dynamic vi are used interchangeably

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25354>
2023-09-22 20:03:05 +00:00
Mike Blumenkrantz
94debd4b4c lavapipe: always set independent blend
no functional changes, but this is just how vulkan works

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25354>
2023-09-22 20:03:05 +00:00
Mike Blumenkrantz
74ac66c275 lavapipe: fix variable descriptor count support handling
if no variable counts are present this must be zero, otherwise yolo

fixes:
dEQP-VK.api.maintenance3_check.support_count_combined_image_sampler_extra_bindings_no_variable_size
dEQP-VK.api.maintenance3_check.support_count_combined_image_sampler_no_variable_size
dEQP-VK.api.maintenance3_check.support_count_inline_uniform_block_extra_bindings_no_variable_size
dEQP-VK.api.maintenance3_check.support_count_inline_uniform_block_no_variable_size
dEQP-VK.api.maintenance3_check.support_count_input_attachment_extra_bindings_no_variable_size
dEQP-VK.api.maintenance3_check.support_count_input_attachment_no_variable_size
dEQP-VK.api.maintenance3_check.support_count_sampled_image_extra_bindings_no_variable_size
dEQP-VK.api.maintenance3_check.support_count_sampled_image_no_variable_size
dEQP-VK.api.maintenance3_check.support_count_sampler_extra_bindings_no_variable_size
dEQP-VK.api.maintenance3_check.support_count_sampler_no_variable_size
dEQP-VK.api.maintenance3_check.support_count_storage_buffer_dynamic_extra_bindings_no_variable_size
dEQP-VK.api.maintenance3_check.support_count_storage_buffer_dynamic_no_variable_size
dEQP-VK.api.maintenance3_check.support_count_storage_buffer_extra_bindings_no_variable_size
dEQP-VK.api.maintenance3_check.support_count_storage_buffer_no_variable_size
dEQP-VK.api.maintenance3_check.support_count_storage_image_extra_bindings_no_variable_size
dEQP-VK.api.maintenance3_check.support_count_storage_image_no_variable_size
dEQP-VK.api.maintenance3_check.support_count_storage_texel_buffer_extra_bindings_no_variable_size
dEQP-VK.api.maintenance3_check.support_count_storage_texel_buffer_no_variable_size
dEQP-VK.api.maintenance3_check.support_count_uniform_buffer_dynamic_extra_bindings_no_variable_size
dEQP-VK.api.maintenance3_check.support_count_uniform_buffer_dynamic_no_variable_size
dEQP-VK.api.maintenance3_check.support_count_uniform_buffer_extra_bindings_no_variable_size
dEQP-VK.api.maintenance3_check.support_count_uniform_buffer_no_variable_size
dEQP-VK.api.maintenance3_check.support_count_uniform_texel_buffer_extra_bindings_no_variable_size
dEQP-VK.api.maintenance3_check.support_count_uniform_texel_buffer_no_variable_size

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25354>
2023-09-22 20:03:05 +00:00
Mike Blumenkrantz
207b37bfd9 zink: always trace_screen_unwrap in acquire
it's possible for acquire to be called from the frontend, in which
case the screen will still be trace-wrapped

cc: mesa-stable

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25288>
2023-09-22 19:34:21 +00:00
Mike Blumenkrantz
aaabb5b0f2 zink: avoid UAF on wayland async present with to-be-retired swapchain
wayland surfaces are likely to become unlinked in WSI implementations upon
retiring a swapchain, requiring the pending present to complete
in order to avoid invalid access

cc: mesa-stable

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25288>
2023-09-22 19:34:21 +00:00
Mike Blumenkrantz
2b7b1059e4 zink: move swapchain fence to swapchain object
this is more accurate in terms of usage/ownership and avoids potential
illegal fence usage in the scenario where multiple async presents are
in flight

cc: mesa-stable

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25288>
2023-09-22 19:34:21 +00:00
Rob Clark
e24a710c1e freedreno/decode: Remove gpu_id
Now unused.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25333>
2023-09-22 19:17:14 +00:00
Rob Clark
066f80e31c freedreno/decode: Use info->chip to decode
Use the chip generation in cases where decoding is generation specific,
rather than range testing gpu_id.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/9310
Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25333>
2023-09-22 19:17:14 +00:00
Rob Clark
ebdac80b0d freedreno/decode: Lookup device info
First step for migration to chip_id.  With newer devices we won't be
able to decode the chip_id in any meaningful way, ie. it is just a
number.  But we can use the device table to figure out things like
generation which are needed to properly decode cmdstream, etc.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25333>
2023-09-22 19:17:14 +00:00
Caio Oliveira
ae99a3d76c util: Add more PRINTFLIKE and MALLOCLIKE annotations
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25230>
2023-09-22 10:47:33 -07:00
Caio Oliveira
931253b45d microsoft/compiler: Fix printf formatting string issues
Found when adding PRINTFLIKE macro to util functions.

Suggested-by: Eric Engestrom <eric@igalia.com>
Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25230>
2023-09-22 10:47:33 -07:00
Caio Oliveira
1cdc4be14b intel/compiler: Don't allocate memory for SIMD select error handling
The position in the error array already indicate the SIMD in question,
so take off all the formatted printing from the errors -- which in some
cases were just not needed.  We lose a little bit of extra context but
it is all easily derivable from the message and the SIMD.

This also will remove the overhead when SIMD selection is being used to
just to find the selected dispatch width -- at a point where the shaders
were already compiled -- and the errors are not used at all.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/9849
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25336>
2023-09-22 16:23:02 +00:00
Caio Oliveira
e55aa87f32 compiler: Only enable mesaclc helper if we have OpenCL SPIR-V support
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25341>
2023-09-22 15:29:25 +00:00
Caio Oliveira
4e065cdaaa clover: Only compile/depend libclspirv and libclnir when using SPIR-V support
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25341>
2023-09-22 15:29:25 +00:00
Caio Oliveira
3a210932ad clover: Hide SPIR-V related code behind HAVE_CLOVER_SPIRV
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25341>
2023-09-22 15:29:25 +00:00