This is only needed for GLES where floats have been lowered to
half floats internally. We don't do that in desktop GL and we
need to skip it to return the correct enums for the
AMD_gpu_shader_half_float extension.
Acked-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18540>
Commit a52c1994aa ("intel/dev: generate declarations for struct
intel_device_info") started generating and including
intel/dev/intel_device_info_gen.h, but it's not present in the
documentation build, leading to errors:
docs/isl/formats.rst:222: ERROR: src/intel/dev/intel_device_info.h:39: 'intel/dev/intel_device_info_gen.h' file not found
docs/isl/tiling.rst:87: ERROR: src/intel/dev/intel_device_info.h:39: 'intel/dev/intel_device_info_gen.h' file not found
Add an empty header stub under docs/header-stubs to fix the build.
Fixes: a52c1994aa ("intel/dev: generate declarations for struct intel_device_info")
Signed-off-by: Jani Nikula <jani@nikula.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27573>
Allows to use zink with drivers that do not expose this extension.
Backport-to: 23.3 24.0
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Signed-off-by: Corentin Noël <corentin.noel@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27516>
If a program does two blits in a row, we internally do a sequence of
operations that involves binding vb0.
Previously, the vb0 state after each operation would look something like:
| operation | cmd->state.gfx.vb0 | hardware | save->vb0 |
| ---------------------------- | ------------------ | --------- | --------- |
| | user | user | |
| nvk_meta_begin() | user | user | user |
| BindVertexBuffers(internal0) | internal0 | internal0 | user |
| nvk_meta_end() | internal0 | user | |
| nvk_meta_begin() | internal0 | user | internal0 |
| BindVertexBuffers(internal1) | internal1 | internal1 | internal0 |
| nvk_meta_end() | internal1 | internal0 | |
That is, CmdBindVertexBuffers() would update cmd->state.gfx.vb0, but
nvk_meta_end() would not. This meant that the last operation would bind a
driver-internal buffer instead of the original value that the user set.
This change fixes the issue by tracking cmd->state.gfx.vb0 in
nvk_cmd_bind_vertex_buffer(), which both CmdBindVertexBuffers() and
nvk_meta_end() call into.
After this commit, the state looks like:
| operation | cmd->state.gfx.vb0 | hardware | save->vb0 |
| ---------------------------- | ------------------ | --------- | --------- |
| | user | user | |
| nvk_meta_begin() | user | user | user |
| BindVertexBuffers(internal0) | internal0 | internal0 | user |
| nvk_meta_end() | user | user | |
| nvk_meta_begin() | user | user | user |
| BindVertexBuffers(internal1) | internal1 | internal1 | user |
| nvk_meta_end() | user | user | |
To test this commit, build gtk4 commit 87b66de1, run:
GSK_RENDERER=vulkan gtk4-demo --run=image_scaling
then select trilinear filtering in the dropdown and check for rendering
artifacts.
Fixes: e1c66501 ("nvk: Use vk_meta for CmdClearAttachments")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27559>
Should lead to slightly better reliability + eventual performance boost
on Intel HW.
Also reverted:
- `ae795abe7b3e ("drm/msm/dsi: Use pm_runtime_resume_and_get to prevent refcnt leaks")`
- `78a1eb102c7f ("drm/msm/dsi: Enable runtime PM")`
so a306 testing won't break.
Acked-by: Guilherme Gallo <guilherme.gallo@collabora.com>
Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27532>
MAILBOX reduced from 5 to 4 to be more inline with WL.
With xwayland_wait_ready = false now being the default,
add some extra consideration for non-FIFO present modes.
Add a lengthy rationale for why minImageCount = 3 is exposed on Xwl.
Signed-off-by: Hans-Kristian Arntzen <post@arntzen-software.no>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27074>
It prevents reading the invalid out-of-bounds offset if exec_mask
is 0 without building the 'if' condition. We can do it only when
reading the memory. It's faster than run time 'if' condition
Reviewed-by: Konstantin Seurer <konstantin.seurer@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27302>
