Re: [PATCH v8 07/17] KVM: x86: Move "no master clock" fallback from __get_kvmclock() to get_kvmclock()

On Tue, Aug 04, 2026, [email protected] wrote:
> commit cb25993776f1881f5ff66de261984bd8992ceece
> Author: Sean Christopherson <[email protected]>
> 
> KVM: x86: Move "no master clock" fallback from __get_kvmclock() to get_kvmclock()
> 
> This commit refactors the kvmclock retrieval logic by extracting the master
> clock fallback out of the inner helper and changing it to return a boolean.
> 
> > diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> > index b520928579129..7366d4602f89f 100644
> > --- a/arch/x86/kvm/x86.c
> > +++ b/arch/x86/kvm/x86.c
> 
> [ ... ]
> 
> > @@ -1667,8 +1673,6 @@ static void __get_kvmclock(struct kvm *kvm, struct kvm_clock_data *data)
> >  				   &hv_clock.tsc_shift,
> >  				   &hv_clock.tsc_to_system_mul);
> >  		data->clock = __pvclock_read_cycles(&hv_clock, data->host_tsc);
> > -	} else {
> > -		data->clock = get_kvmclock_base_ns() + ka->kvmclock_offset;
> >  	}
> >  }
> 
> [Severity: High]
> Does this function need to return true on the success path?

Doh, yes.  I botched the splitting and incorrectly added the "return true" in
"KVM: x86: Wrap all of __get_kvmclock_master_clock() with CONFIG_X86_64=y".  I'll
fix, and properly compile test all patches (I did that before splitting the patch,
but obviously not after).