GitLab

Blob doesn't apply multipos optimization starting from 11 views
even on a650, however in practice, with the limit of 16 views,
tests pass on a640/a650 and fail on a630.

Fixes tests:
 dEQP-VK.multiview.draw_indexed.max_multi_view_view_count
 dEQP-VK.multiview.input_attachments.max_multi_view_view_count
 dEQP-VK.multiview.masks.max_multi_view_view_count
 dEQP-VK.multiview.multisample.max_multi_view_view_count
 dEQP-VK.multiview.queries.max_multi_view_view_count
 dEQP-VK.multiview.renderpass2.index.fragment_shader.max_multi_view_view_count
 dEQP-VK.multiview.secondary_cmd_buffer.max_multi_view_view_count

Fixes: 8d275778 ("tu: Enable multi-position output")
Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Backported from: b6b3b384

Part-of: <!10205>

When formatting the error here, we're currently casting an
ast_type_qualifier as a string.

But we don't need to use a string here at all, because we know from
context exactly what qualifier we're talking about, because the
if-statements explicitly check for the uniform-qualifier.

So let's just hard-code the format-string to reference the right
qualifier instead of the string-shenanigans. The latter cannot do the
right thing.

Fixes: 2d03f48a ("glsl: Add parsing for GLSL uniform blocks.")
Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <!9911>
(cherry picked from commit 437ed057)

This arugment is not present int the format-string, so we shouldn't pass
it to _mesa_glsl_error either.

Noticed by Coverity.

Fixes: 02dc74fb ("glsl: parse invocations layout qualifier for ARB_gpu_shader5")
Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <!9911>
(cherry picked from commit 2a984da1)

This breaks CLOn12's handling of CL CTS test_basic vector_creation for char3 (at least).
Removing this cast causes us to try to load from a deref with no alignment info.

Fixes: 99bb2a4d ("nir/opt_deref: Don't remove casts with alignment information")
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <!10165>
(cherry picked from commit 4b69ae8e)

bo_free is called on external BOs when there are no objects left which
reference them. The function unmaps the address range associated with
any maps which occured. However, if the BO is busy (not idle), it
doesn't mark the pointer to the start address as invalid. This can lead
to a segfault later on.

At the end of bo_free, these BOs are still present in the handle hash
table. If such a BO is reused (i.e., when a DMABUF with the same handle
is reimported) and the driver attempts to get another mapping, the
bufmgr will incorrectly assume that the map pointer is still valid and
reuse it. This leads to a segfault. Set the pointer to NULL to mark it
as invalid.

Enables iris to run and pass the piglit test,
ext_image_dma_buf_import-reimport-bug.

Cc: mesa-stable
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <!9230>
(cherry picked from commit 0092219c)

Previously panfrost_batch_add_bo was called MAX_MIP_LEVELS times on
the same batch.

Fixes: cbf68b21 ("panfrost: Move checksum_bo to panfrost_resource")
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <!10158>
(cherry picked from commit 1b41707b)

lima_pack_reload_plbu_cmd() can be used for either cbuf or zsbuf, so
use passed surface other using cbuf unconditionally.

Fixes: b8c31ac0 ("lima: fix glCopyTexSubImage2D")
Reviewed-by: Erico Nunes <nunes.erico@gmail.com>
Signed-off-by: Vasily Khoruzhick <anarsoul@gmail.com>
Part-of: <!10121>
(cherry picked from commit c8e80c24)

These are scaled up in the descriptor, which doesn't really know about
their MSAA-ness. So we have to shift them back down.

Cc: mesa-stable
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Part-of: <!10162>
(cherry picked from commit 1baefe41)

This was hidden originally to workaround a bug in the RK3399 display
driver. The patch resolving this issue has been merged in the upstream
kernel, and in fact...

1. The issue was visible on 21.0 even with this workaround under certain
   configurations (sway with an external monitor).
2. Even on buggy kernels, due to other platform details this is an
   obscure bug to hit (not aware of any ways to trigger it OOTB with
   current userspaces other than sway with an external monitor)

So why bother? Let's just delete the hack and let AFBC be used freely.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Cc: mesa-stable
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <!10148>
(cherry picked from commit e00d94f1)

Accidental read-before-write due to incorrect statement ordering. I love
SSA as much as anyone, but not everything is a parallel copy.

Fixes faults in glamor in 21.0 when using GIMP on v5.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Fixes: e8b997e1 ("panfrost: Add AFBC slice.body_size and slice.{row,surface}_stride fields")
Closes: #4389Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <!10148>
(cherry picked from commit 23b060bb)

This fixes the following dEQP tests:

- dEQP-GLES3.functional.shaders.operator.unary_operator.minus.lowp_uvec2_fragment
- dEQP-GLES3.functional.shaders.operator.unary_operator.minus.lowp_uvec4_fragment
- dEQP-GLES3.functional.shaders.operator.unary_operator.minus.mediump_uvec2_fragment
- dEQP-GLES3.functional.shaders.operator.unary_operator.minus.mediump_uvec4_fragment

Fixes: 1af8fe4e ("gallium: add shader caps INT16 and FP16_DERIVATIVES")
Reviewed-by: Alyssa Rosenzweig <alyssa@collabora.com>
Tested-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <!10128>
(cherry picked from commit 6a2ee3b1)

Fixes: b83f4b9f ("glsl: Add an IR lowering pass to convert mediump operations to 16-bit")
Reviewed-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <!10128>
(cherry picked from commit 07a02a33)

This shouldn't be unsigned, or else we'll trigger asserts.

Fixes: 7f00d4da ("glsl: Handle 16-bit types in loop analysis")
Reviewed-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <!10125>
(cherry picked from commit c64e2bf2)

This shouldn't check for 64-bit...

Fixes: a052a9c2 ("glsl: handle int16 and uint16 types and add instructions for mediump")
Reviewed-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <!10125>
(cherry picked from commit b718de74)

Fixes: 682c81bd "vulkan,anv: Add a base object struct type"
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <!10123>
(cherry picked from commit 2b19b7a1)

Fixes: 682c81bd "vulkan,anv: Add a base object struct type"
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <!10123>
(cherry picked from commit 46749afe)

Fixes: 682c81bd "vulkan,anv: Add a base object struct type"
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <!10123>
(cherry picked from commit e20c6537)

PIPE_BUFFER layout is incompatible with the 3D pipe, so don't try to
blit it via a 3D engine blit, but fall back to the software copy.

Fixes crashes in piglit tests arb_copy_buffer and arb_map_buffer_range.

Fixes: c9e8b49b (etnaviv: gallium driver for Vivante GPUs)
Cc: mesa-stable@lists.freedesktop.org
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Part-of: <!9310>
(cherry picked from commit 171d7a30)

Multidraws with large draw count will try to call tc_add_slot_based_call
with a payload larger than TC_CALLS_PER_BATCH.

This commit split these draws in multiple commands to avoid this issue.

Cc: mesa-stable
Closes: #4534Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <!10170>

texture_index is meaningless when a tex_instr has deref src.
Use var->data.binding instead.

CC: mesa-stable
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Acked-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <!10170>

It ignored non-harvested chips with a non-power-of-two memory bus.

Fixes: abed921c - amd: add support for Navy Flounder
Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <!10170>

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <!10170>

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <!10170>

Enabling this fast path, while making CPU side simpler
(and thus reducing driver's CPU overhead), forces us to
generate multiple VERTEX_BUFFER_STATE packets pointing to
the same buffer, but with slightly shifted BufferStartingAddress.
This is fine on GFX version >= 11 and < 8, but on version
8 and 9, thanks to internal details of the VF cache, vertex
data from each VERTEX_BUFFER_STATE is cached separately
and this results in a lot of cache misses.

Disabling this fast path restores previous performance levels.

On my SKL GT2 machine this improves performance in:
- GLB27 Egypt offscreen by 37%
- GLB27 TRex offscreen by 22%
- gfxbench5 Manhattan offscreen by 1.75%
- gfxbench5 Manhattan31 offscreen by 1.9%
- gfxbench5 Aztec Ruins high by 2.3%

In Intel performance CI on GFX version 9 it improves performance in:
- gfxbench5 Manhattan offscreen by 1.65%
- gfxbench5 Aztec Ruins normal offscreen by 1.33%

Closes: #3277
Fixes: 42842306 ("mesa,st/mesa: add a fast path for non-static VAOs")
Signed-off-by: Marcin Ślusarz <marcin.slusarz@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <!10175>

Fixes: 42842306 ("mesa,st/mesa: add a fast path for non-static VAOs")
Signed-off-by: Marcin Ślusarz <marcin.slusarz@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <!10175>

The heap usage/budget needs to be tracked per process instead
of per device.

Fixes: 7bef1920 ("radv: add support for VK_EXT_memory_budget")
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <!10174>

Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Part-of: <!10135>