On Sat, Jun 09, 2018 at 04:07:26PM +0200, Greg Kroah-Hartman wrote: > Here's the full callstack, but yeah, it's not very obvious as to what > device is having the problem, which isn't good. I don't know what to > suggest here.
I think it is related to amdgpu -- when the script that unbinds the second GPU from host is not executed, there are no issues during reboot. I found out that amdgpu leaves i2c-dev and hwmon device nodes after unbind, which seems to be a bug in that driver. In 4.16, it only left stale hwmon node. Best, Mateusz -- mlen
signature.asc
Description: PGP signature
