@danielcassidy see that doesn't even make sense either. The GPU for the host has a stub driver on the host. Its then passed through to the VM directly. So the host doesn't actually touch the graphics driver being used.
@roxie I mean obviously it *shouldnt* be that way but the gpu is pretty much directly attached to the pci-x bus, which is also more or less the main system bus, and it wouldn't take much of a bug for something attached to the system bus to trigger a reboot.
@roxie still, if the vm has direct access to the gpu it is not at all surprising to me that it's possible to leverage that access to crash or reboot the host, intentionally or otherwise.