Re: [PATCH 03/10] KVM: x86: Add KVM_[GS]ET_CLOCK_GUEST for accurate KVM clock migration

From: David Woodhouse
Date: Fri Apr 19 2024 - 11:49:32 EST


On Fri, 2024-04-19 at 16:40 +0100, Paul Durrant wrote:
>
> > +        * If KVM_REQ_CLOCK_UPDATE is already pending, or if the
> > hv_clock has
> > +        * never been generated at all, call
> > kvm_guest_time_update() to do so.
> > +        * Might as well use the PVCLOCK_TSC_STABLE_BIT as the
> > check for ever
> > +        * having been written.
> > +        */
> > +       if (kvm_check_request(KVM_REQ_CLOCK_UPDATE, v) ||
> > +           !(hv_clock->flags & PVCLOCK_TSC_STABLE_BIT)) {
> > +               if (kvm_guest_time_update(v))
> > +                       return -EINVAL;
>
> nit: simple nested if, so you could use &&

Yeah, I frowned at that a bit, and decided I preferred it this way just
to highlight the fact that kvm_guest_time_update() is doing a *thing*.
And then we bail with -EINVAL if that thing fails.

If you stick it all in the if() statement, the code flow is logically
the same but it's just a bit less obvious that there are side-effects
here in the middle of the conditions, and that those side-effects are
actually the *main* point of the statement.

Attachment: smime.p7s
Description: S/MIME cryptographic signature