Calling transactional update shell && zypper dup on my MicroOS results into a
system crash followed by reboot. This is pretty much 100% reproducible. I
recall there was recently an issue when if we were building against a certain
nvidia related library there was a kernel panic, not sure if it was mesa or
wayland.
I'm quirte certain that crash happens building of the kmod.
I was able to update the driver, by uninstalling all nvidia packages && reboot
followed by doing a system update and then installing drivers (from console
without Wayland or X).
This makes automatic update feature painful as any time my machine tries to
update ... the machine just reboots.
I was able to reproduce the issue also on TW without transactional-update, well
at least here I have at least one working snapshot :-)
localhost:/home/lkocman # nvidia-debugdump -v -l
Listing all GPUs.
Found 1 NVIDIA devices
Device ID: 0
Device name: NVIDIA GeForce RTX 3070 (*PrimaryCard)
GPU internal ID: GPU-21498f58-0fa4-a24c-b7c7-ee4bfbd25302
localhost:/home/lkocman # rpm -qa | grep kernel
kernel-firmware-radeon-20230210-1.1.noarch
kernel-firmware-platform-20230210-1.1.noarch
kernel-firmware-mediatek-20230210-1.1.noarch
kernel-firmware-bluetooth-20230210-1.1.noarch
kernel-firmware-amdgpu-20230210-1.1.noarch
purge-kernels-service-0-9.3.noarch
kernel-default-devel-6.2.1-1.1.x86_64
kernel-default-6.2.1-1.1.x86_64
kernel-macros-6.2.1-1.1.noarch
kernel-devel-6.2.1-1.1.noarch
localhost:/home/lkocman # uname -r
6.2.1-1-default
lkocman@localhost:~> rpm -qa | grep nvidia
nvidia-compute-utils-G06-525.89.02-7.1.x86_64
nvidia-video-G06-525.89.02-7.1.x86_64
nvidia-utils-G06-525.89.02-7.1.x86_64
nvidia-driver-G06-kmp-default-525.89.02_k6.2.1_1-7.1.x86_64
nvidia-compute-G06-525.89.02-7.1.x86_64
nvidia-gl-G06-525.89.02-7.1.x86_64
nvidia-drivers-G06-525.89.02-7.1.x86_64
libnvidia-egl-wayland1-1.1.11-1.1.x86_64
lkocman@localhost:~>
Interestingly I'm quite successful in buidling the official driver manually,
but then that one doesn't seem to have wayland support.