Hi, On Fri, Oct 19, 2018 at 11:33:51AM +0200, Valentin Vidic wrote: > Booting the kernel with nouveau.runpm=0 seems to be a workaround for > this issue. Not sure why the lockup is trigered by libvirtd.
Can you reproduce this bug when you disable libvirtd to start at boot and then start libvirtd later on? Is this maybe tied to a specific VM being auto started (that gets the GPU passed in or similar)? -- Guido
