Hi Victor, On 28/05/14 07:15, Victor Kamensky wrote: > Hi Marc, Christoffer, > > While working on BE patches the following observation occurred > to me. In the same vein as Marc caught my mistake with BUG_ON > in [1] does the code have similar issue in access_vm_reg function? > In this function we have: > > BUG_ON(!p->is_write); > > My reading of this: if guest will issue read instruction for those > registers before HCR_TVM is cleared it will crash host kernel. Is > it right? I think guest should not crash host kvm kernel in similar > way as userspace KVM ioctl should not crash kernel. Or am I > missing something here? Yes, you're missing the definition of the TVM bit. "When this bit is set to 1, any valid Non-secure PL1 or PL0 write to a virtual memory control register is trapped to Hyp mode." The BUG_ON() is actually there to catch bugs. M. -- Jazz is not dead. It just smells funny... _______________________________________________ kvmarm mailing list kvmarm@xxxxxxxxxxxxxxxxxxxxx https://lists.cs.columbia.edu/mailman/listinfo/kvmarm