On Tue, 2009-04-14 at 18:45 +0800, Avi Kivity wrote: > Huang Ying wrote: > >> I'm okay with an ioctl to setup MCE, but just make sure userspace has > >> all the information to know what the kernel can do rather than the > >> try-and-see-if-it-works approach. We can publish this information via > >> KVM_CAP things, or via another ioctl (see KVM_GET_SUPPORTED_CPUID2 for > >> an example). > >> > > > > Yes. MCE support should be published by KVM_CAP_MCE and other features > > can be published via reading the default value of MSR_IA32_MCG_CAP. > > > > A problem with this is that you can only read an MSR after a vcpu has > been created. But if you're writing a program to detect what features > are available (for example, when checking features common to a migration > pool), you don't want to create a vpcu (you could, but it's hacky). Yes. You are right. I will change this as you said, something like KVM_GET_SUPPORTED_CPUID2. Best Regards, Huang Ying
Attachment:
signature.asc
Description: This is a digitally signed message part
