for pipelines, we know enabled features. for classic shader objects, we do not.
therefore, we want to plumb this through explicitly for drivers using common
pipelines, rather than making drivers guess whether they can use the device
features.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35169>
This was mostly wired up, but we currently require an alignment of 64
for uniform texel buffers, because we're currently using
plane-descriptors for this.
We could lift that limitation by switching to buffer descriptors and use
LD_CVT for the format-conversion, but that's a bigger change.
Let's just fix up the aligntment and enable the extension for now.
Reviewed-by: Olivia Lee <benjamin.lee@collabora.com>
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/34826>
Vulkan is supposed to operate in explicit synchronization mode. However,
for legacy compositors that only support implicit fencing, we have to
extract the compositor implicit fence (release fence) and resolve it
properly. Since we used to rely on renderer side drivers being able to
handle implicit in-fence, here we only opt-in the new behavior for those
known to have issues with that.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/34516>
We weren't setting the accum src (which should be pT for now) and we put
ftz in the wrong place. Bit 5 is actually .bf which we want to set all
the time in order to get a float output. (Otherwise it gives an integer
output.)
Fixes: a3330f1d46 ("nak/sm20: Add float ops")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35170>
Use the new shared presentation feedback code in the loader to implement
perfetto frame delivery tracing similar to the wayland vulkan wsi code.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32757>
This is a potentially slow process, so let's break it out into its own
function so we can have more meaningful profiling data from perfetto later.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32757>
The throttling code is potentially long running. Move it to a separate
function so we can have better perfetto tracing later.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32757>
We only use the acquisition time for calculating latency for perfetto
tracks later, and the acquisition time should ideally be the start of the
perfetto flow.
This has been more or less true with very small error margin for vk wsi,
but the wayland EGL buffer handling is a lot more complicated. Moving the
time check into the flow start will make re-using this code for EGL much
simpler.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32757>
Push the presentation feedback code into the common code in the loader,
so we're one step closer to using the perfetto instrumentation here in
the EGL code.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32757>
For now just pull the tiny bit that looks up the wayland buffer id for
profiling. The end goal is to promote more code sharing between vk and egl
and improving wayland egl's perfetto profiling.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32757>
Count the outstanding feedback requests instead of relying on list
length calculations.
When moving presentation feedback bits into common code shortly, the list
will no longer be exposed.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32757>
Instead of testing the protocol object pointer, track this with a bool.
We're going to wrap the protocol object in some common code later, so
testing it directly will become inconvenient.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32757>
Now that we're no longer using multiple artifacts, we can drop the
LAVA_ prefix from the S3_ARTIFACT_NAME variable name for simplicity.
Signed-off-by: Valentine Burley <valentine.burley@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/34980>
Compared to the existing Debian-based x86_64_pyutils container, this
Alpine-based variant reduces the image size by approximately 83%.
Include all the necessary python artifacts, including lava_job_submitter
in the container to avoid having to download them at the start of each
test job.
Signed-off-by: Valentine Burley <valentine.burley@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/34980>
Kepler cards do not support shared atomic operations directly, but they
have special ldslk and stsul that can implement mutex locks on
addresses. Shared atomics can be lowered into operations in mutexes.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35028>
On a3xx-a5xx information from vertex program is used to emit state
related to vertex buffer. However this state is not re-emitted if only
vertex program changes which causes rendering artifacts on gnome-shell
where there are 2 draws with same vertex buffer but different shaders
(due to changed vertex attribute destination regid).
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35164>
