directly accessing the pipe_resource::screen pointer crashes trace,
so it's necessary to have an unwrapped pointer available if such things
are needed
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11017>
without enough nir optimizing, some shaders fail to reduce as they should,
resulting in the wrong output
fixes dEQP-VK.graphicsfuzz.spv-stable-mergesort-flatten-selection-dead-continues
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11019>
this was added in 315f8daab1 without much detail, then refactored
a bunch of times, and there's even now been a comment added questioning
whether it's correct: it does not appear to be.
fixes lavapipe cts cases:
dEQP-VK.draw.inverted_depth_ranges.depthclamp_deltaone_bias_clamp_neg
dEQP-VK.draw.inverted_depth_ranges.nodepthclamp_deltaone_bias_clamp_neg
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11012>
The quick_gl set is too unstable -- even when I switched to a consistent
set of tests, and added lots of flakes, I keep getting new ones as some
test (unclear which, but it's like 7-8 minutes into the run) kills other
innocent ones. Until we get per process pagetables, disable this testing
by default.
However, now that we've freed up a board that was doing quick_gl, we have
time to do all of quick_shader so that piglit uprevs don't reshuffle the
test list and expose new failures.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11018>
Besides advertising AHB support, for Android, we do not advertise
KHR_external_memory_fd, EXT_external_memory_dma_buf or
EXT_image_drm_format_modifier so that AHB remains the sole channel for
external resource sharing.
Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10960>
This patch refactors the struct look up logic for memory allocation, and
it prepares the necessary info for ahb allocation and import.
Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10960>
This patch wraps up fd import logic, which will be used later by ahb
import. This patch also removes a redundant check in need_bo.
Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10960>
1. bail early if there's no compatiable AHB format
2. check against the corresponding drm format modifier
Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10960>
Started happening after disabling cpufreq, devfreq and runtime PM.
At least one of these fail in each run, so it's blocking MRs.
Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7987>
Use Piglit's replay profile to measure and store the time that frames
take to render in the GPU.
This job won't run automatically in regular pipelines, but will be
triggered automatically by a script for every successful pre-merge
pipeline.
This is because we want to generate performance data for every relevant
commit merged in main, but we don't want to keep a device busy during
the pre-merge run.
Signed-off-by: Antonio Caggiano <antonio.caggiano@collabora.com>
Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-By: Rohan Garg <rohan.garg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7987>
Lock CPU and GPU frequency scaling on devices so to speed up test
execution and lower the variation of frame times from performance replay
jobs.
Also disable autosuspend of the GPU device.
Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-By: Rohan Garg <rohan.garg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7987>
Now that the compile entrypoints don't touch the context, it's clear
that we can support this. Note, even though the pools for shaders and
descriptors are referenced from particular contexts, they are unowned
pools -- once uploaded, any thread can use the results.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10976>
DFSM has never been enabled by default because it was slower.
RadeonSI is also dropping support for this because they discovered
that's actually not efficient in practice.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10968>
ACO is the default compiler for almost a year from now, so it should
be fine to replace RADV/ACO by just RADV. LLVM is still added
when RADV_DEBUG=llvm is used for convenience.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10972>
For fragment shaders that only contain a discard, the exec mask has
to be zero'd and everything discarded.
It seems unnecessary to emit an export here because if the FS has no
exports, the compiler already emits a null export at the end.
Fixes incorrect hair rendering in Detroit: Become Human.
fossil-db (Sienna Cichlid):
Totals from 3 (0.00% of 149839) affected shaders:
CodeSize: 2896 -> 2872 (-0.83%)
Instrs: 556 -> 553 (-0.54%)
Latency: 29266 -> 29214 (-0.18%)
InvThroughput: 3374 -> 3372 (-0.06%)
Cc: 21.1 mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10955>