Re: [PATCH v7 31/36] KVM: x86: Add KVM_[GS]ET_CLOCK_GUEST for accurate KVM clock migration

David Woodhouse <[email protected]> Sat, 01 Aug 2026 09:07:46 +0100
Newsgroups org.kernel.vger.linux-kselftest,org.kernel.vger.kvm,org.kernel.vger.linux-doc,org.kernel.vger.linux-kernel,org.xenproject.lists.xen-devel
Message-ID <[email protected]>
On Fri, 2026-07-31 at 16:24 -0700, Sean Christopherson wrote:
> 
> > +	/*
> > +	 * Allow for a discrepancy of 1 kHz either way between the TSC
> > +	 * frequency used to generate the user's pvclock and the current
> > +	 * host's measured frequency, since they may not precisely match.
> > +	 */
> > +	if (user_tsc_hz < curr_tsc_hz - 1000 ||
> > +	    user_tsc_hz > curr_tsc_hz + 1000) {
> 
> I don't follow, why is KVM restricting what frequency userspace can set?

Userspace actually sets the frequency with KVM_SET_TSC_KHZ. What KVM is
insisting upon here is that the input to KVM_SET_CLOCK_GUEST is
*consistent* with the guest's TSC frequency (within a little slop
caused by different host TSCs).

I think the comment above it is fairly badly worded though ("host's
measured frequency"), so I'll fix that.
smime.p7s (application/pkcs7-signature, 6 KB) - not displayed