I have a laptop with Intel+NVIDIA graphics. Running Leap 15.5, this system has run flawlessly, with over a year between reboots (daily suspend/resume cycles). Finally, I decided to try to update to Leap 16.0 (15.5 → 15.6 → 16.0), and found that not only did all of my hardware acceleration stop working (Mesa, VA-API, etc.), but that the system would freeze after suspend, forcing a hard reset. Obviously unacceptable, but I am not asking for help in troubleshooting this.
Eventually, I restored a snapshot of my system to Leap 15.5, and it has returned to stability. Running inxi -Gxx, I see that for this installation (Leap 15.5), the system has disabled the NVIDIA card (driver: N/A), but after upgrading to 16.0, the new installation was using it (driver: nvidia), without me having enabled anything explicitly. My question is: how does my Leap 15.5 system disable the NVIDIA device? (If I decide to upgrade again, I might experience better results if I manage to disable the NVIDIA card. This isn’t the first time I’ve had problems with NVIDIA cards.)
There are sites that mention ways to disable the NVIDIA GPU, but they reference software I don’t have, or other methods that don’t apply. For example, the source is not:
- BIOS: I changed no BIOS settings during the upgrade
nvidia-smitool: I only use open-source software (OSS repo), so I don’t have this tool on my system- In fact,
zypper se -i nvidiaonly matcheskernel-firmware-nvidiaandopenSUSE-repos-Leap-NVIDIA(but I disabled that repo, since it’s non-free)
- In fact,
- Boot parameter: only have
mitigations=autoin the boot line; that doesn’t relate, does it? - Module/driver parameters: Nothing NVIDIA-specific in either
/etc/modprobe.dor/etc/X11/xorg.conf.d
Lastly, in 15.5, I use X11 (XOrg), but in 16.0, Gnome forces Wayland. I would think that this disabling of the GPU happens at a lower level than the compositor, but if not, I would still have interest in knowing the mechanism how this happens.