On Sun, Mar 27, 2011 at 12:57:03PM +0200, Avi Kivity wrote:
> On 03/25/2011 10:44 AM, Joerg Roedel wrote:
>> The calculation of the tsc_delta value to ensure a
>> forward-going tsc for the guest is a function of the
>> host-tsc. This works as long as the guests tsc_khz is equal
>> to the hosts tsc_khz. With tsc-scaling hardware support this
>> is not longer true and the tsc_delta needs to be calculated
>>
>> --- a/arch/x86/kvm/x86.c
>> +++ b/arch/x86/kvm/x86.c
>> @@ -2126,8 +2126,13 @@ void kvm_arch_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
>> kvm_x86_ops->vcpu_load(vcpu, cpu);
>> if (unlikely(vcpu->cpu != cpu) || check_tsc_unstable()) {
>> /* Make sure TSC doesn't go backwards */
>> - s64 tsc_delta = !vcpu->arch.last_host_tsc ? 0 :
>> - native_read_tsc() - vcpu->arch.last_host_tsc;
>> + s64 tsc_delta;
>> + u64 tsc;
>> +
>> + kvm_get_msr(vcpu, MSR_IA32_TSC,&tsc);
>> + tsc_delta = !vcpu->arch.last_guest_tsc ? 0 :
>> + tsc - vcpu->arch.last_guest_tsc;
>> +
>> if (tsc_delta< 0)
>> mark_tsc_unstable("KVM discovered backwards TSC");
>> if (check_tsc_unstable()) {
>
>
> Patch is correct, but can this really happen? Won't all machines with
> scaled tsc also be stable?
>
> Perhaps not multi-board machines.
Yes, all machines with tsc-scaling should also have a stable tsc, so
this patch is not strictly required.
I added this patch as a response to Zachary's thoughts about the values
that need to be passed to adjust_tsc_offset. Zachary mentioned that with
this patch-series the adjust_tsc_offset function will either take the
adjustment in guest-tsc or host-tsc units (which is not problematic if
guest and host tsc have the same frequency).
But with tsc-scaling (and Zachary's upcoming patches) this might become
a problem. So I changed the semantic of adjust_tsc_offset to always take
the adjustment in guest-tsc units. This patch-set converts the one and
only call-site of adjust_tsc_offset where the unit is calculated as a
function of the host-tsc.
Joerg
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
