Add pipe format to sized internal format mappings for R32F,
RG32F and RGBA32F in the DRIImage format table. This fixes shader
image store operations on EGL images imported from DMA-BUF with
respective DRM formats.
Without these mappings, DMA-BUF imported textures received base internal
formats instead of sized internal formats, causing shader image unit
validation to fail and DispatchCompute write zeroes on compute shader
imageStore operation.
Fixes: 6d3f266406 ("dri: Add additional 16/32b float/int formats")
Signed-off-by: Lakshman Chandu Kondreddy <lkondred@qti.qualcomm.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39004>
ANV can return different memory requirements with and w/o the ALIAS bit.
See https://gitlab.freedesktop.org/mesa/mesa/-/issues/14671 for details.
Meanwhile, venus has a driver side cache for image memory requirements.
As blessed per spec for memory aliasing, venus strips the ALIAS bit when
populating the cache key. Because of the use of imageless mem req query,
the ALIAS mem req now can hit the cache first, leaving a smaller/relaxed
requirement in the cache...busted.
Venus is unable to fix ANV behavior behind the scene, so this workaround
is only to align Venus behavior with ANV to not suffer from Venus-only
rendering artifacts.
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39368>
Fixes the following building error happening with clang:
FAILED: src/amd/vulkan/libvulkan_radeon.so.p/radv_pipeline_rt.c.o
...
../src/amd/vulkan/radv_pipeline_rt.c:934:38: error: use of GNU empty initializer extension [-Werror,-Wgnu-empty-initializer]
struct nir_function raygen_stub = {};
^
1 error generated.
Fixes: 0a1911b2 ("radv,aco: Use function call structure for RT programs")
Reviewed-by: Konstantin Seurer <konstantin.seurer@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39370>
The Function glInternalReleaseBufferMESA is called whenever a
intermediate buffer is used by MESA internal and being released, however
it's only exported for GLES3, which leads to drawing w/o VBO not
operational for GLES1/2.
Fixes invalid operation warning in `glmark2-es2 -b build:use-vbo=false` for
GLES2 HW / any HW with MESA_GLES_VERSION_OVERRIDE=2.0 .
VBO-less drawing is possible for GLES 1.0 too, so this function is also
exported with GLES1. Tested on GLES1 with mesa-demos gears demo with
VBO allocation commented out.
Fixes: ae75b59cb5 ("glthread, tc: Fix buffer release with glthread and tc")
Signed-off-by: Icenowy Zheng <uwu@icenowy.me>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39092>
Kopper draw codepath calls an internal helper called
InternalInvalidateFramebufferAncillaryMESA, which is currently only
exported for GLES3 contexts.
This leads to a warning when the underlying Vulkan driver capability
isn't enough for GLES3 or GLES version number is forced to 2.0 or the
program is a GLES1 one.
Export this internal helper for GLES1/2 too.
Fixes: 0c7994bb2d ("mapi: add InternalInvalidateFramebufferAncillaryMESA")
Signed-off-by: Icenowy Zheng <uwu@icenowy.me>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39092>
Mesa internally converts some draw calls to DrawArraysInstancedBaseInstance
or DrawElementsInstancedBaseVertexBaseInstance, which are only available
in GLES contexts via GL_EXT_base_instance, a GLES3-only extension.
Export these two functions into the internal namespace for GLES 2.0
contexts too.
This fixes black screen for `glmark2-es2 -b build:use-vbo=false` (which is
the first test item for glmark2 auto test sequence) on GLES 2 HW / GLES
3 HW with MESA_GLES_VERSION_OVERRIDE=2.0 .
Fixes: 7c1f8b63ad ("glthread: work around GL_INVALID_OPERATION with OpenGL ES 1.x draws")
Signed-off-by: Icenowy Zheng <uwu@icenowy.me>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39092>
We need to support this if we advertise KHR_swapchain_maintenance1
Cc: mesa-stable
Signed-off-by: llyyr <llyyr.public@gmail.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39218>
Similar to last MR, since we advertise support for
KHR_swapchain_maintenance1, we need to advertise this
Cc: mesa-stable
Signed-off-by: llyyr <llyyr.public@gmail.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39218>
This adds support for cooperative matrix to lavapipe.
It uses 8x8 matricies as the size, and loads the first row
of 8 values into a the subgroup.
It stores the B matrix transposed to make the matmul operation
a lot more subgroup friendly.
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/38935>
Now that we are setting COMPUTE_WALKER_BODY once makes more sense to remove
iris_emit_execute_indirect_dispatch() and emit EXECUTE_INDIRECT_DISPATCH in
iris_upload_compute_walker().
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39326>
Both functions setting COMPUTE_WALKER_BODY had almost the same code, the only
difference is that iris_emit_execute_indirect_dispatch() keeps
ThreadGroupIDX/Y/ZDimension set to 0.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39326>
Fixes a crash with:
dEQP-VK.api.external.semaphore.opaque_fd.signal_export_import_wait_temporary
when driver calls genX(CmdSetEvent2) -> emit_apply_pipe_flushes with
having NULL in emitted_flush_bits.
Fixes: 8834ef8bcd ("anv: use flushing PIPE_CONTROL for event signaling")
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39343>
Combine the `radeonsi-raven-va` and `radeonsi-raven-vaapi-fluster` jobs
into one deqp-runner suite.
OOM kills no longer appear to be an issue.
Signed-off-by: Valentine Burley <valentine.burley@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39015>
The only difference was that these rules incorrectly added a second
-fluster suffix to the expectation files after a refactoring.
Signed-off-by: Valentine Burley <valentine.burley@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39015>
This version notably also brings in vkd3d-proton support.
Also remove the redundant DEBIAN_TEST_ANDROID_TAG mention.
Signed-off-by: Valentine Burley <valentine.burley@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39015>
This fixes situations where the variable being split is a parallelcopy
definition.
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/38811>
For example, v6b at byte=2 should be split into v2b,v1 instead of v1,v2b.
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/38811>
