On Tue, 2022-03-15 at 22:42 +0800, Chao Gao wrote:
> On Fri, Mar 11, 2022 at 04:35:17AM +0000, Sean Christopherson wrote:
> > --- a/arch/x86/kvm/x86.c
> > +++ b/arch/x86/kvm/x86.c
> > @@ -9053,15 +9053,29 @@ bool kvm_apicv_activated(struct kvm *kvm)
> > }
> > EXPORT_SYMBOL_GPL(kvm_apicv_activated);
> >
> > +
>
> stray newline.
>
> > +static void set_or_clear_apicv_inhibit(unsigned long *inhibits,
> > + enum kvm_apicv_inhibit reason, bool set)
> > +{
> > + if (set)
> > + __set_bit(reason, inhibits);
> > + else
> > + __clear_bit(reason, inhibits);
> > +
> > + trace_kvm_apicv_inhibit_changed(reason, set, *inhibits);
>
> Note that some calls may not toggle any bit. Do you want to log them?
> I am afraid that a VM with many vCPUs may get a lot of traces that actually
> doesn't change inhibits.
I also think so.
Best regards,
Maxim Levitsky
>
> Anyway, this series looks good to me.
>
