Don't setup inputs and outputs if we aren't using
load_input/store_output intrinsics. While it's mostly harmless, there
may be more outputs than expected which would lead to an oob write of
the outputs array when setting the register id to INVALID_REG.
Also be more paranoid with asserts to catch this.
Fixes: a6291b1 ("freedreno/ir3: rework setup_{input,output} to make struct varyings work")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7274>
This makes the vertex order of TRISTRIP and TRISTRIP_ADJ primitves
consistent between XFB output and GS input. Technically, the Vulkan
spec allows us to XFB out in whatever order we want but being consistent
with GS inputs is probably nicer to apps.
Fixes: 36ee2fd61c "anv: Implement the basic form of VK_EXT_transform_feedback"
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10460>
PIPE_CAP_IMAGE_LOAD_FORMATTED doesn't depend on
shaderStorageImageExtendedFormats or
shaderStorageImageWriteWithoutFormat.
PIPE_SHADER_CAP_MAX_SHADER_IMAGES enables
GL_EXT_shader_image_load_store, which *does* require
shaderStorageImageExtendedFormats. Having
shaderStorageImageWriteWithoutFormat and
shaderStorageImageReadWithoutFormat isn't enough to support this.
It *might* be possible to lower extended formats to format-less
reads or writes, but we don't currently do that, so we should
just correct the test for now.
Fixes: 3f9a6d333b ("zink: export shader image caps using features")
Fixes: 5282210c0b ("zink: check correct caps for PIPE_CAP_IMAGE_LOAD_FORMATTED")
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10456>
We do not require the extended formats feature to enable shader-images,
so we can end up accidentally enabling the cap even when we don't need
to, which can lead to validation issues.
This also prepares us for the SPV_EXT_shader_image_int64 extension, if
we ever want to support that.
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10370>
For non-CL, intrinsic access isn't set, because the image type doesn't
have access qualifier. Instead, the access qualifier is set on the variable.
So, add a mode to this pass which can chase back to the variable in addition
to the intrinsic access. Also, update the variable type and the deref chain
types so everything is consistent, that the tex is accessing a sampler. Note
we can't do this for CL, because void-typed samplers don't exist.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10356>
Modern shader APIs, like DX10 and GLSL 1.30, want min() and max() to
"cleanse" NaN. If one source is NaN, the other value should be chosen.
If both sources are NaN, the result may be either.
There are many cases where TGSI is generate from NIR, and many
optimizations in NIR expect this behavior. Not meeting these
expectations can lead to unexpected results.
Reviewed-by: Eric Anholt <eric@anholt.net>
Fixes: ffe58739da ("Softpipe: import TGSI tree. Not hooked-up yet.")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10419>
Modern shader APIs, like DX10 and GLSL 1.30, want saturate or
clamp(..., 0.0, 1.0) to "cleanse" NaN. If the source is NaN, the
result should be zero.
There are many cases where TGSI is generate from NIR, and many
optimizations in NIR expect this behavior. Not meeting these
expectations can lead to unexpected results.
Reviewed-by: Eric Anholt <eric@anholt.net>
Fixes: 56c30bf17b ("tgsi: Saturate modifier obeys ExecMask. Implement NVIDIA [-1;+1] saturate mode.")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10419>
I've seen this one happen at least twice today. Log shows something like:
Wallclock time between MSCs 16982.888889us does not match
glXGetMscRateOML 16668.071966us
or
Wallclock time between MSCs 16500.333333us does not match
glXGetMscRateOML 16668.071966us
Incidentally, both runs I've looked into had one run too fast and one run
too slow.
Acked-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10414>
Instead of using num_textures, determine the texture count from the
last bit set in textures_used.
Fixes ADDR_RANGE_FAULTs when draw_textured_quad writes only stencil.
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10422>
If no slices were valid (the resource is uninitialized), we were reading
a bogus PIPE_FORMAT_NONE format from the blit.dst.format. Fix this --
and guard against scope similar issues -- by hoisting the blit
assignments out of the loop so they're valid.
Fixes: 9d0ad7fd2e ("panfrost: Patch the gallium driver to use pan_image_layout_init()")
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10415>
Connor Abbott