num_syncs was being incremented by one if 'utrace_submit != NULL' but
a sync was only being set if also
'util_dynarray_num_elements(&utrace_submit->batch_bos) == 0'.
This mismatch could cause application to abort due to
'assert(count == num_syncs)'.
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27244>
This causes lavapipe to use the split code and fixes accuracy
for CTS.
Fixes dEQP-VK.glsl.builtin.precision_fconvert.f64_to_f16*
Cc: mesa-stable
Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27228>
With regards to implicit masking of the shift counts for 8- and 16-bit
types, the PRMs are incorrect. They falsely state that on Gen9+ only the
low bits of src1 matching the size of src0 (e.g., 4-bits for W or UW
src0) are used. The Bspec (backed by data from experimentation) state
that 0x3f is used for Q and UQ types, and 0x1f is used for **all** other
types.
To match the behavior expected for the NIR opcodes, explicit masks for
8- and 16-bit types must be added.
This fixes (the updated version, see crucible!138) of
func.shader.shift.int16_t on all Intel platforms. According to Karol,
this also fixes "integer_ops integer_rotate" tests in OpenCL CTS.
No shader-db or fossil-db changes on any Intel platform.
Tested-by: Karol Herbst <kherbst@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23001>
During the development of this fix, I utilized the `test_full_yaml_log`
test, which is marked as slow. This test is excellent for reproducing
past job submissions. It can be executed using the following commands:
```
lavacli jobs logs --raw 12496073 > /tmp/log.yaml
pytest .gitlab-ci/tests/test_lava_job_submitter.py -m slow -s
```
Here, `12496073` is the LAVA job ID from this specific job:
https://gitlab.freedesktop.org/mesa/mesa/-/jobs/53546660
The logs were not functioning as expected due to a few mistakes I made
with generators, such as:
- Not reusing the `time_traveller_gen` generator to call `next` more
than once.
- Forgetting to parse the YAML inside `time_travel_from_log_chunk`.
Additionally:
- Added some statistics at the end of the test to aid in diagnosing
whether the test was reproduced accurately.
- Renamed some variables for clarity.
Signed-off-by: Guilherme Gallo <guilherme.gallo@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/26996>
In our process of monitoring LAVA logs, we typically skip numerous
messages to enhance log clarity. We already exclude `feedback` messages
from display. These messages were just used as a heartbeat signal,
indicating that if we are receiving them, the Device Under Test (DUT) is
active.
Practically, if we continuously receive feedback messages without any
other message level (either `debug` or `target`) for several minutes,
this could be a cause for concern, as it likely indicates the device is
in a kind of livelock state.
Therefore, it is more prudent to ignore feedback messages, as they tend
to occur frequently in unstable scenarios. However, it's important to
note that any other message level will still be considered as a
heartbeat signal.
Real case where several minutes of feedback messages indicate a hang:
https://gitlab.freedesktop.org/mesa/mesa/-/jobs/53546660
Signed-off-by: Guilherme Gallo <guilherme.gallo@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/26996>
Fixes RADV compilation of a Cyberpunk 2077 RT pipeline with
PIPELINE_CREATE_DISABLE_OPTIMIZATION_BIT.
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Fixes: dfaa3c0af6 ("aco: Flip s_cbranch / s_cselect to optimize out an s_not if possible.")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27194>
The GFX11 ISA docs description of this hazard says it's about v_permlane in
general, not just v_permlane(x)16.
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27118>
The GFX11 ISA doc explicitly recommends using v_nop in
7.2.8. PERMLANE Specific Rules.
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27118>
Xe2+ changed the msaa mapping for 2D/3D Tile64 surfaces, introduce a
Xe2+ specific enum to handle this change.
Signed-off-by: Rohan Garg <rohan.garg@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27113>
The render area clamping was lost during the transition to the FB
helpers. Restore the original logic so we can benefit from
EGL_KHR_partial_update on v4, and on v5 when only one damage
rectangle is passed.
Fixes: ff3eada7eb ("panfrost: Use the generic preload and FB helpers in the gallium driver")
Reported-by: Sjoerd Simons <sjoerd.simons@collabora.com>
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Tested-by: Sjoerd Simons <sjoerd.simons@collabora.com>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27215>
On my machine, this folder gets created by .gitlab-ci/bin/ci_run_n_monitor.sh.
Signed-off-by: Christian Gmeiner <cgmeiner@igalia.com>
Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27227>
Vivante NPUs can contain systolic arrays that can be used to perform
several kinds of tensor transformations.
Use these to offload the CPU.
Acked-by: Christian Gmeiner <cgmeiner@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25714>
Vivante NPUs contain MAC arrays that can be used to perform
convolution operations on tensors.
This commit implements the Gallium functions that accept convolutions
and executes them on these MAC arrays.
Acked-by: Christian Gmeiner <cgmeiner@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25714>
Teflon is a Gallium frontend that TensorFlow Lite can load to delegate
the execution of operations in a neural network model.
See docs for more.
Acked-by: Christian Gmeiner <cgmeiner@igalia.com>
Acked-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25714>
