The ACP entries created by copy propagation to track the implied
copies of LOAD_PAYLOAD instructions don't model the behavior of
LOAD_PAYLOAD correctly, since (as of 41868bb682) header
moves are implicitly retyped to UD and the destination of non-header
copies implicitly uses the same type as the corresponding source, even
though the ACP entries created for such copies could incorrectly
represent a type conversion, which can lead to mis-optimization of the
program.
According to Marcin, this fixes the func.mesh.ext.workgroup_id.task.q0
crucible test.
Fixes: 41868bb682 ("i965/fs: Rework the fs_visitor LOAD_PAYLOAD instruction")
Reported-by: Marcin Ślusarz <marcin.slusarz@intel.com>
Tested-by: Marcin Ślusarz <marcin.slusarz@intel.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18980>
(cherry picked from commit 7b5e933629)
`count` should not be incremented before the check, because it causes
the modifiers array to be filled starting from position 1 instead of 0.
This bug causes one less format modifier to be available than would
otherwise be expected, which could then lead to a dmabuf query failing
in situations where a supported modifier wouldn't be advertised.
It also causes garbage data to be advertised as a modifier in position 0
of the array, although this is not very likely to cause issues.
Fixes: 2a1217513 ("panfrost: Implement panfrost_query_dmabuf_modifiers")
Cc: mesa-stable
Signed-off-by: Italo Nicola <italonicola@collabora.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20879>
(cherry picked from commit 6c446377ff)
gfx11 renamed PRIM_GRP_SIZE to VERTS_PER_SUBGRP but another change was
was missed.
Update our code based on PAL's UniversalCmdBuffer::CalcGeCntl function
(especially useVgtOnchipCntlForTess being false for gfx11).
Fixes: 25a66477d0 ("radeonsi/gfx11: register changes")
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20728>
(cherry picked from commit f73cdda983)
Otherwise, the access chain you emitted last time may not dominate the
current use.
Fixes the following validation failure in
dEQP-GLES31.functional.shaders.sample_variables.sample_mask_in.bits_unique_per_sample.multisample_texture_2:
UNASSIGNED-CoreValidation-Shader-InconsistentSpirv(ERROR / SPEC):
msgNum: 7060244 - Validation Error: [
UNASSIGNED-CoreValidation-Shader-InconsistentSpirv ] Object 0: handle =
0x55cf3cea2c60, type = VK_OBJECT_TYPE_DEVICE; | MessageID = 0x6bbb14 |
SPIR-V module not valid: ID '67[%67]' defined in block '23[%23]' does
not dominate its use in block '31[%31]'
Fixes: 8899f6a198 ("zink: fix gl_SampleMaskIn spirv generation")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20756>
(cherry picked from commit 4286633eec)
Fixes validation failures:
Test case 'dEQP-GLES31.functional.android_extension_pack.shaders.es32.extension_directive.oes_sample_variables'..
MESA: error: Validation Error: [
UNASSIGNED-CoreValidation-Shader-InconsistentSpirv ] Object 0: handle =
0x563a1838b790, type = VK_OBJECT_TYPE_DEVICE; | MessageID = 0x6bbb14 |
SPIR-V module not valid: [VUID-StandaloneSpirv-Flat-04744] Fragment
OpEntryPoint operand 31 with Input interfaces with integer or float type
must have a Flat decoration for Entry Point id 4.
%gl_SampleId = OpVariable %_ptr_Input_uint Input
Test case 'KHR-GL46.shader_ballot_tests.ShaderBallotAvailability'..
MESA: error: Validation Error: [ UNASSIGNED-CoreValidation-Shader-InconsistentSpirv ] Object 0: handle = 0x5558e12f17e0, type = VK_OBJECT_TYPE_DEVICE; | MessageID = 0x6bbb14 | SPIR-V module not valid: [VUID-StandaloneSpirv-Flat-04744] Fragment OpEntryPoint operand 28 with Input interfaces with integer or float type must have a Flat decoration for Entry Point id 4.
%gl_SubgroupLocalInvocationId = OpVariable %_ptr_Input_uint Input
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20756>
(cherry picked from commit 2a33d509ca)
I've been running into failures with tests like :
dEQP-VK.robustness.robustness2.bind.notemplate.rgba32i.unroll.nonvolatile.uniform_buffer_dynamic.no_fmt_qual.len_4.samples_1.1d.frag
With the load_global_const_block_intel NIR intrinsic, you can load a
vec8/vec16 with a predicate. The predicate is correctly uniformized to
feed into the SEND instruction's flag register.
The problem is that a series of optimization first remove the
find_live_channel and then changes the broadcast into a simple MOV
instruction, on the assumption that the first channel is always active
if there is not control flow. This is correct.
But after that the cmod optimzation will remove this instruction :
mov.nz.f0.0(16) null:D, vgrf16+0.0<0>:D NoMask
because it seems to be equivalent to :
cmp.g.f0.0(16) vgrf16:D, vgrf12:D, 63d
In this case vgrf16 is the predicate to the load block SEND
instruction. Since the execution mask is different between both, some
of the channels of the SEND instruction end up not being loaded or
loaded with the wrong predication and we end up with incorrect UBO
data.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: mesa-stable
Reviewed-by: Marcin Ślusarz <marcin.slusarz@intel.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20852>
(cherry picked from commit a50d2fdb46)
unlocks were incorrectly added to paths using dri2_egl_display() as
well as those using dri2_egl_display_lock()
pthread_mutex_unlock() when unlocked is documented by posix as
being undefined behaviour. On OpenBSD pthread_mutex_unlock() will call
abort(3) if this happens.
Fixes: f1efe037df ("egl/dri2: Add display lock")
Reviewed-by: Rob Clark <robclark@freedesktop.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20712>
(cherry picked from commit 0594b3c143)
It was wrong. For example, if the draw vertex count was 10 and the upload
vertex count was 150, u_vbuf wouldn't unroll the draw and would instead
memcpy 150 vertices. This fixes that case.
Fixes: 068a3bf0d7 - util: move and adjust the vertex upload heuristic equation from u_vbuf
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20824>
(cherry picked from commit 4f6e785876)
this is an artifact of very old code before the dynamic state was set
for all graphics pipelines
now the checks only cause blend constants to not be updated, which triggers
bugs and validation failures
cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20799>
(cherry picked from commit b4d18f2ad1)
offset field in kgsl_command_object is NOT used by KGSL, so
we should offset directly to iova.
Fixes weird hangs on KGSL. E.g. fixes the hang in:
dEQP-VK.memory.pipeline_barrier.transfer_dst_storage_texel_buffer.1024
cc: mesa-stable
Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20795>
(cherry picked from commit 926f626b95)
VK_MEMORY_PROPERTY_DEVICE_LOCAL_BIT = 0x01 has been incidently the
correct memory type index, but isn't guaranteed to be, which is why it
hasn't caused issues yet
Fixes: f9515d93 ("zink: allocate/place memory using memoryTypeIndex directly")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20264>
(cherry picked from commit c71287e70c)
internal_format is the "real" format of a resource, and the "real" format
of imported resources is the external-facing format, not the pipe format
this ensures the correct format is available for internal ops, such as nplanes queries
Fixes: 2e2775c11b ("zink: fix PIPE_RESOURCE_PARAM_NPLANES with format modifier")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20753>
(cherry picked from commit 072e29a22e)
When the base array layer of the image view is > 0, addrlib computes
the offset (in HwlComputeSubResourceOffsetForSwizzlePattern) which is
then added to the base VA in RADV. But if the driver doesn't reset
the base array layer, the hw will compute incorrect addressing
(ie. base array will be added twice). This also matches AMDVLK.
This fixes a VM fault followed by a GPU hang on RDNA2 when trying
to join a multiplayer game with medium settings in Halo Infinite.
Fixes: 98ba1e0d81 ("radv: Fix mipmap views on GFX10+")
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20761>
(cherry picked from commit 8d191b2cfb)
The ideal case for performance is to have a single buffer for
all display list. The caveat is that large buffers are less
likely to be freed because they're refcounted: it only takes
1 user (diplay list) to keep it in VRAM.
This lowers VRAM usage when replaying the trace attached
of the trace attached to !6140 from 5.5 GB to about 1.8 GB.
Viewperf snx performance isn't affected.
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6140
Cc: mesa-stable
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20748>
(cherry picked from commit 0f5c8c3dc3)
grow_vertex_storage may call wrap_filled_vertex, which will
trigger the assert incorrectly because the new size will be
smaller than 'new_size' but it's correct because
'vertex_store->used' has been reset to 0.
Fixes: a08baaff97 ("vbo/dlist: fix indentation in vbo_save_api.c")
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20748>
(cherry picked from commit 491f6b138e)
This avoids a violation of the Vulkan memory model that was leading to
intermittent failures of at least 8k test-cases of the Vulkan CTS
(within the group dEQP-VK.memory_model.*) on TGL and DG2 platforms.
In theory the issue may be reproducible on earlier platforms like IVB
and ICL, but the SYNC.ALLWR instruction is not available on those
platforms so a different (likely costlier) fix will be needed.
The issue occurs within the sequence we emit for a NIR memory barrier
with acquire semantics requiring the synchronization of multiple
caches, e.g. in pseudocode for a barrier involving the TGM and UGM
caches on DG2:
x <- load.ugm // Atomic read sequenced-before the barrier
y <- fence.ugm
z <- fence.tgm
wait(y, z)
w <- load.tgm // Read sequenced-after the barrier
In the example we must provide the guarantee that the memory load for
x is completed before the one for w, however this ordering can be
reversed with the intervention of a concurrent thread, since the UGM
fence will block on the prior UGM load and potentially take a long
time, while the TGM fence may complete and invalidate the TGM cache
immediately, so a concurrent thread could pollute the TGM cache with
stale contents for the w location *before* the UGM load has completed,
leading to an inversion of the expected memory ordering.
v2: Apply the workaround regardless of whether the NIR barrier
intrinsic specifies multiple storage classes or a single one,
since an acquire barrier is required to order subsequent requests
relative to previous atomic requests of unknown storage class not
necessarily specified by the memory scope information of the
intrinsic.
Cc: mesa-stable
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20690>
(cherry picked from commit 4a2e7306dd)
Because anv_execbuf_add_bo_bitset() calls anv_execbuf_add_bo(), which
can fail if its memory allocations fail.
I have seen dEQP tests exercising memory allocation failures during
anv_execbuf_add_bo(), but I don't think the path coming from
add_bo_biset() was specifically exercised. Anyway, add the error check
just in case.
v2: Rebase.
Cc: mesa-stable
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20703>
(cherry picked from commit 3d37950fd9)
In anv_execbuf_add_syncobj(), we try to not create or use
exec->syncobj_values if we don't need to. But when we figure we're
going to need it (i.e., when timeline_value is not zero), then we
create exec->syncobj_values with vk_zalloc, which means every previous
value is set to zero, as it should be. This is all correct.
The problem starts when we add a 16th element. In this case we double
exec->syncobj_array_length and realloc the buffer by using vk_alloc
and copying the old array to the new one. After that, we write the
timeline_value to the array only if it's not zero, and that's the
problem: since we just used vkalloc and memcpy, we don't have any
guarantees that the new array will be zero after the 16th element, and
if timeline_value is zero we write nothing to that position.
Once we start using exec->syncobj_values we have to commit to using
it, so the "if (timeline_value)" check near the end of the function
has to be changed to "if (exec->syncobj_values)", so we actually set
elements after the 16th to zero when they need to be zero. Another
approach to fix this would be to memset the new elements once we
double syncobj_array_length.
In practice, I couldn't find any application or deqp test that used
more than 3 elements in exec->syncobj_array_length, and we need more
than 16 elements in order to be able to reproduce the bug, so I'm not
aware of any real-world bug that goes away with this patch. This issue
was found while reading code.
If we craft a little Vulkan program that submits a ton of timeline and
binary semaphores on vkQueueSubmit, then waits for them, we get the
following error without this patch:
MESA: error: ../../src/intel/vulkan/anv_batch_chain.c:1910: execbuf2 failed: Invalid argument (VK_ERROR_DEVICE_LOST)
v2: Rebase.
Cc: mesa-stable
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20703>
(cherry picked from commit ad6a036a68)
Commit e07c5467 ("v3dv/format: use XYZ1 swizzle for three-component formats")
removes the only code that handled the clamp_to_transparent_black_border
variable. Therefore, the variable can be deleted, as it is not currently
being used.
Fixes: e07c5467 ("v3dv/format: use XYZ1 swizzle for three-component formats")
Signed-off-by: Maíra Canal <mcanal@igalia.com>
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20746>
(cherry picked from commit 86c9bdcd9a)
ARB programs are logically separate, and Mesa will happily mix and match them.
We need to alert backends of this fact, by setting nir->info.separate_shader.
Otherwise, backends may link shaders invalidly.
Fixes fp-abs-01 on Bifrost. (We don't use separate_shader for anything on
Valhall, so the issue doesn't appear there.)
Compare 151aa19c21 ("ttn: Set nir->info.separate_shader"), which fixed a
similar issue with TGSI.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20688>
(cherry picked from commit 7e68cf91d7)
The host can't assign more than 32 locations explicitly, and we
exhaust this already when we handle patches and generics. So
drop the separable flag in cases when we have other IO that
uses generated names that will have to be matched by name.
v2: skip tests for VS input and FS outputs
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20738>
(cherry picked from commit 8084b412ca)
