Quoting the documentation :
"The returned range is a half-open interval where all of the
addresses within the subimage are < end_tile_B."
This is obviously not true with images smaller than a logical tile.
Currently the code return 1.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: mesa-stable
Reviewed-by: Nanley Chery <nanley.g.chery@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24276>
lower_scan_reduce only worked when ballot_components equals one. This
commit adds support for arbitrary ballot_components.
Signed-off-by: Job Noorman <jnoorman@igalia.com>
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31587>
This functionality will become more complex in the next commit so
separate it into a helper function.
Signed-off-by: Job Noorman <jnoorman@igalia.com>
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31587>
build_subgroup_mask and build_ballot_imm_ishl will be needed by other
functions higher-up the file.
Signed-off-by: Job Noorman <jnoorman@igalia.com>
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31587>
[WHY]
- was hardcoded to store 16 configs only
- as the config descriptor usage grows, more is needed
- in bypass case, we also generate a new config which is a waste
[HOW]
- change to use vector to store configs
- don't force new config desc if in bypass
- revise the vector API, reduce the parameter passing
[TESTING]
- Tested with corresponding test cases
Reviewed-by: Brendan Leder <breleder@amd.com>
Acked-by: Chih-Wei Chien <Chih-Wei.Chien@amd.com>
Signed-off-by: Roy Chan <roy.chan@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31693>
Generalize the visual confirm handling through helper function with
minor coding style changes
Reviewed-by: Evan Damphousse <evan.damphousse@amd.com>
Reviewed-by: Roy Chan <Roy.Chan@amd.com>
Acked-by: Chih-Wei Chien <Chih-Wei.Chien@amd.com>
Signed-off-by: Brendan Leder <breleder@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31693>
Fix potential duplicated config if config_writer_complete is called
multiple times
Reviewed-by: Roy Chan <Roy.Chan@amd.com>
Reviewed-by: Evan Damphousse <evan.damphousse@amd.com>
Acked-by: Chih-Wei Chien <Chih-Wei.Chien@amd.com>
Signed-off-by: Brendan Leder <BrendanSteve.Leder@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31693>
Refactor 3D LUT code and tested with corresponding test cases
Reviewed-by: Roy Chan <Roy.Chan@amd.com>
Acked-by: Chih-Wei Chien <Chih-Wei.Chien@amd.com>
Signed-off-by: Mike Hsieh <Mike.Hsieh@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31693>
[WHY]
Config writer duplicates packets when switching pipes.
[HOW]
Update config writer to only write new packets if new type is required.
[TESTING]
Tested with corresponding test cases.
Reviewed-by: Roy Chan <Roy.Chan@amd.com>
Acked-by: Chih-Wei Chien <Chih-Wei.Chien@amd.com>
Signed-off-by: Brendan Leder <breleder@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31693>
- Generalize the p2b programming to support chroma viewport if any
- Turn vpe_get_element_size_in_bytes to a global helper function
Reviewed-by: Jesse Agate <Jesse.Agate@amd.com>
Acked-by: Chih-Wei Chien <Chih-Wei.Chien@amd.com>
Signed-off-by: Jiali <jiali.zhao@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31693>
We jump out of the loops whenever result is not VK_SUCCESS, there is
no need to check for it there. I guess I missed this detail in the
most recent rework for this function.
Reviewed-by: Iván Briano <ivan.briano@intel.com>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31698>
When the VkBuffer is of size 2^32 (which matches maxBufferSize), we
have vm_bind->size set to 2^32, which is fine because it fits in an
uint64_t. What is not fine is the 'i' variable being size_t, because
on 32bit systems it will loop forever since it will always be smaller
than 2^32.
Credits to Iván for not only reporting it, but also coming up with the
solution at the same time as I did, then testing it.
Cc: mesa-stable
Reported-by: Iván Briano <ivan.briano@intel.com>
Reviewed-by: Iván Briano <ivan.briano@intel.com>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31698>
For each pipeline, we launch 13 jobs for kingoftown, let's decrement it
to match the number of 12 trogdor-kingoftown available in Collabora's
farm atm.
Let's decrement the parallel no., and increment the dEQP fraction for
a618_vk job.
Signed-off-by: Guilherme Gallo <guilherme.gallo@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31699>
We have 6 DUTs but we are running a total of 7 jobs per pipeline,
increasing the DUT pending time.
- Job: parallel no.
- 'a618_gl: 2'
- 'a618_egl: 1'
- 'a618_piglit: 1'
- 'zink-tu-a618: 1'
So, reduce the `a618_gl` parallel from 2 to 1 and increase dEQP fraction
from 2 to 5 to reach the 10-min mark.
Signed-off-by: Guilherme Gallo <guilherme.gallo@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31699>
We have one spare meson-g12b-a311d-khadas-vim3 DUT, and this job is
taking 14 minutes to run on average and retries are rare, so let's
increment the parallel to reach the 10 min-mark.
Signed-off-by: Guilherme Gallo <guilherme.gallo@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31699>
With the addition of the radeonsi-raven-cdna-lower-image:x86_64 job,
zork devices are being overutilized.
To balance the load, we're moving the radv-raven-traces job to dalboz,
where one DUT per pipeline remains available.
Signed-off-by: Guilherme Gallo <guilherme.gallo@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31699>
These skip files contain tests that exceed one minute in execution time,
which are better suited for nightly jobs. This change aims to optimize
execution time and prevent delays in merge pipelines under time
constraints.
Signed-off-by: Guilherme Gallo <guilherme.gallo@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31699>
- Increase the fraction for radv-stoney-vkcts from 15 to 18.
- Same for radv-angle, from 2 to 3
The last one should take around 7 minutes, but this new 3 minute gap can
be filled in by the ~95% DUT availability and eventual infra retries.
Signed-off-by: Guilherme Gallo <guilherme.gallo@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31699>
This change has two main advantages. First is that the previous code
had a subtle bug where state updates would drop the previous state
pushbuf without first guaranteeing that the queue was idle, potentially
leading to a fault. This worked when I first wrote the code because we
were using the old nouveau UAPI which took a list of BOs on each submit
and the kernel would interanlly reference count them. With the modern
UAPI, there is no such safety net.
The second effect of this commit is that we're now only submitting the
pushbuf when one of the pointers or ranges actually changes. This
should hopefully reduce GPU stalling as setting those pointers is a
pretty heavy-weight operation.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30596>
There are more driver params than what is contained in the indirect
buffer. So switch things around to always allocate a buffer into which
we copy the indirect params.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31534>
Before the gallium driver can support load_shader_consts_via_preamble it
needs a way to differentiate between # of API level UBOs and the number
of UBOs after the driver inserts ones for driver-params, etc. This is
because the driver UBOs are emit in a different place from user UBOs.
If we didn't have this distinction, and the app had more UBOs bound than
the shader used, we would otherwise emit some conflicting UBO
descriptors (ie. ones for unused const slots conflicting with the driver
UBO slots).
This also moves the consts_ubo emit into the PROG state.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31534>
Move this to where the rest of the hs/ds params are emit, because for
a7xx with load_shader_consts_via_preamble we'll need this all to be a
single UBO. So we won't be able to piecemeal upload consts.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31534>
