Re: [PATCH v2 1/8] KVM: x86/xen: improve accuracy of Xen timers

[Sean Christopherson <seanjc@google.com>]

On Tue, Mar 05, 2024, Paul Durrant wrote:
> On 04/03/2024 23:44, Sean Christopherson wrote:
> > On Tue, Feb 27, 2024, David Woodhouse wrote:
> > 	/*
> > 	 * Xen has a 'Linux workaround' in do_set_timer_op() which checks for
> > 	 * negative absolute timeout values (caused by integer overflow), and
> > 	 * for values about 13 days in the future (2^50ns) which would be
> > 	 * caused by jiffies overflow. For those cases, Xen sets the timeout
> > 	 * 100ms in the future (not *too* soon, since if a guest really did
> > 	 * set a long timeout on purpose we don't want to keep churning CPU
> > 	 * time by waking it up).  Emulate Xen's workaround when starting the
> > 	 * timer in response to __HYPERVISOR_set_timer_op.
> > 	 */
> > 	if (linux_wa &&
> > 	    unlikely((int64_t)guest_abs < 0 ||
> > 		     (delta > 0 && (uint32_t) (delta >> 50) != 0))) {
> 
> Now that I look at it again, since the last test is simply a '!= 0', I don't
> really see why the case is necessary. Perhaps lose that too. Otherwise LGTM.

Hmm, I think I'll keep it as-is purely so that the diff shows that it's a just
code movement.  There's already enough going on in in this patch, and practically
speaking I doubt anything other than checkpatch will ever care about the "!= 0".

Thanks!