"rcs0 reset for CS error" regression when using OCL on Gemini Lake
I found a regression when using the intel-compute-runtime on a Gemini Lake laptop.
I bisected it down to change: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=d5ef86b38e4c2a65d5c1d64d8d0f3fcf58aa0884
drm/i915: Add pci ids and uapi for DG1
DG1 has support for local memory, which requires the usage of the
lmem placement extension for creating bo's, and memregion queries
to obtain the size. Because of this, those parts of the uapi are
no longer guarded behind FAKE_LMEM.
According to the pull request referenced below, mesa should be mostly
ready for DG1. VK_EXT_memory_budget is not hooked up yet, but we
should definitely just enable the uapi parts by default.
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
References: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11584
Cc: Jordan Justen <jordan.l.justen@intel.com>
Cc: Jason Ekstrand <jason@jlekstrand.net>
Link: https://patchwork.freedesktop.org/patch/msgid/20210812124452.622233-2-maarten.lankhorst@linux.intel.com
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
System details:
- Lenovo Duet 3i - Pentium N5030 (dmidecode attached)
- Gentoo with: intel-compute-runtime (22.20.23198), libdrm (2.4.110), kernel @ drm-tip (ecc81cedc1a4b62b69c748fd07ca880bd4c1cd3f), mesa (22.1.0)
It reproduces for me every time darktable wants to copy data to OCL device. (Not always a GPU reset happens,sometimes the OCL process just freezes).
Logs:
I have reverted the aforementioned commit on 5.17.8 and it worked again, I will also verify on drm-tip