On 03/08/2016 08:46, Michael Ellerman wrote: >> Considering the h in "hmi" is for hypervisor, > > Well hypervisor != KVM. > > Though in this case hmi.c was pretty safe because it was new code. But > if I'd received a powerpc patch to hmi.c I wouldn't have thought to > check if it conflicted with the KVM tree.
This, together with the fact that hmi.c does nothing if KVM is inactive, suggests to me that hmi.c should be in arch/powerpc/kvm. I'll send a patch soo; I don't care who merges it, but a topic branch would be appropriate. ;) Paolo
