Re: [PATCH v9 02/21] KVM: x86: Improve accuracy of KVM clock when TSC scaling is in force
Sean Christopherson <[email protected]>
| Newsgroups | org.kernel.vger.kvm,org.kernel.vger.linux-kernel |
|---|---|
| Message-ID | <[email protected]> |
On Mon, Aug 10, 2026, Sean Christopherson wrote: > From: David Woodhouse <[email protected]> > > The kvm_guest_time_update() function scales the host TSC frequency to > the guest's using kvm_scale_tsc() and the v->arch.l1_tsc_scaling_ratio > scaling ratio previously calculated for that vCPU. Then calculates the > scaling factors for the KVM clock itself based on that guest TSC > frequency. > > However, it uses kHz as the unit when scaling, and then multiplies by > 1000 only at the end. > > With a host TSC frequency of 3000MHz and a guest set to 2500MHz, the > result of kvm_scale_tsc() will actually come out at 2,499,999kHz. So > the KVM clock advertised to the guest is based on a frequency of > 2,499,999,000 Hz. > > By using Hz as the unit from the beginning, the KVM clock would be based > on a more accurate frequency of 2,499,999,999 Hz in this example. > > Use u64 for the hw_tsc_hz field since an unsigned int would overflow for > TSC frequencies above 4GHz. Use div_u64() for the Xen CPUID leaf to > play nice with 32-bit kernels. > > Fixes: 78db6a503796 ("KVM: x86: rewrite handling of scaled TSC for kvmclock") > Reviewed-by: Paul Durrant <[email protected]> > Signed-off-by: David Woodhouse <[email protected]> > Tested-by: Dongli Zhang <[email protected]> > Signed-off-by: Sean Christopherson <[email protected]> > --- > arch/x86/include/asm/kvm_host.h | 2 +- > arch/x86/kvm/cpuid.c | 1 + > arch/x86/kvm/x86.c | 18 ++++++++++-------- > 3 files changed, 12 insertions(+), 9 deletions(-) > > diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h > index 283847619ff8..6ddc988576d0 100644 > --- a/arch/x86/include/asm/kvm_host.h > +++ b/arch/x86/include/asm/kvm_host.h > @@ -864,7 +864,7 @@ struct kvm_vcpu_arch { > gpa_t time; > s8 pvclock_tsc_shift; > u32 pvclock_tsc_mul; > - unsigned int hw_tsc_khz; > + u64 hw_tsc_hz; From https://sashiko.dev/#/patchset/20260810225500.869288-1-seanjc%40google.com: : > diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h : > index b517257a63153..099c3da20547d 100644 : > --- a/arch/x86/include/asm/kvm_host.h : > +++ b/arch/x86/include/asm/kvm_host.h : > @@ -1004,7 +1004,7 @@ struct kvm_vcpu_arch { : > gpa_t time; : > s8 pvclock_tsc_shift; : > u32 pvclock_tsc_mul; : > - unsigned int hw_tsc_khz; : > + u64 hw_tsc_hz; : > struct gfn_to_pfn_cache pv_time; : > /* set guest stopped flag in pvclock flags field */ : > bool pvclock_set_guest_stopped_request; : : The commit message mentions using div_u64() for the Xen CPUID leaf, but is : the actual update to kvm_cpuid() missing here? : : It appears kvm_cpuid() still references the old field name: : : arch/x86/kvm/cpuid.c:kvm_cpuid() { : ... : *eax = vcpu->arch.hw_tsc_khz; : ... : } : : Since the field was renamed to hw_tsc_hz in struct kvm_vcpu_arch, will this : cause a build failure when compiling for x86? The changelog is stale. This will be applied on top of commit 7d3bd21e457b ("KVM: x86: Remove runtime Xen TSC frequency CPUID update") that's sitting in kvm-x86/clocks. I'll drop the stale sentence from the changelog when applying.