Re: [Qemu-devel] [PATCH 3/5] x86: fill high bits of mtrr mask

[Paolo Bonzini]

On 17/06/2016 15:51, Dr. David Alan Gilbert wrote:
> 
>> > Note that the failure mode is pretty brutal since KVM reports an
>> > internal error right after restarting on the destination, and who knows
>> > what used to happen before the assertion was introduced.  All MSRs after
>> > MTRRs would be ignored by KVM!
> I think we got lucky because the MTRRs were close to the last MSRs
> in the list, there was only some MCG/MCE stuff after it which probably
> wouldn't have been missed;  if the MTRR wasn't set I'm guessing someone
> would have just had a behaviour that looked like dreadful performance
> after migration perhaps?

No, MTRRs are only used by KVM if there's an assigned device, and that's
incompatible with migration.

Paolo