On 02/05/2018 18:56, Jim Mattson wrote: > On Wed, May 2, 2018 at 9:02 AM, Paolo Bonzini <pbonzini@xxxxxxxxxx> wrote: >> On 02/05/2018 17:52, Jim Mattson wrote: >>>>> vmcs_field_readonly is now unused. With that removed, >>>>> >>>>> Reviewed-by: Paolo Bonzini <pbonzini@xxxxxxxxxx> >>>> Actually, no. The error must be kept if the host has disabled the >>>> feature with a KVM_SET_MSR ioctl for MSR_IA32_VMX_MISC. >>>> >>> Coming back to this patch set... >>> >>> Is there a good reason to allow userspace to clear >>> MSR_IA32_VMX_MISC_VMWRITE_SHADOW_RO_FIELDS? It's not yet possible to >>> migrate a VMX-capable VM (though I think that ability is imminent), so >>> there are no compatibility issues. Why not just force this on? >> >> I suppose that can be done, yes. The only reason is that silicon ties >> that bit to the availability of shadow VMCS, but we can certainly do >> things differently than actual processors. >> >> Technically it would be a small userspace API break, but we can live >> with it. > > Does qemu explicitly set IA32_VMX_MISC today? It doesn't use KVM_SET_MSR for VMX MSRs at all. "We" is KVM, not QEMU. Paolo
