On 15 March 2012 15:42, Avi Kivity <[email protected]> wrote: > On 03/15/2012 05:37 PM, Peter Maydell wrote: >> It would be cleaner to cause the guest to take an exception >> IMHO. (This is typically what ought to happen anyway if the >> guest accesses an unimplemented cp15 register, although our >> current mess has a habit of underdecoding these things.) > > At the very least, print something out so it's clear that qemu may be at > fault, not the guest. This is what kvm does in such cases (which we > discourage, but still exist).
We really need better logging support to do this properly, so you can enable or disable it with some sensible granularity. (=> I agree in principle but I don't think we have the facilities and with what we have at the moment I prefer silence.) -- PMM
