Re: [PATCH v11 08/11] KVM: x86/pmu: Emulate LBR feature via guest LBR event

[Xu, Like]

On 2020/5/19 19:00, Peter Zijlstra wrote:
> On Thu, May 14, 2020 at 04:30:51PM +0800, Like Xu wrote:
> > +static inline bool event_is_oncpu(struct perf_event *event)
> > +{
> > +	return event && event->oncpu != -1;
> > +}
>
> > +/*
> > + * It's safe to access LBR msrs from guest when they have not
> > + * been passthrough since the host would help restore or reset
> > + * the LBR msrs records when the guest LBR event is scheduled in.
> > + */
> > +static bool intel_pmu_access_lbr_msr(struct kvm_vcpu *vcpu,
> > +				     struct msr_data *msr_info, bool read)
> > +{
> > +	struct kvm_pmu *pmu = vcpu_to_pmu(vcpu);
> > +	u32 index = msr_info->index;
> > +
> > +	if (!intel_is_valid_lbr_msr(vcpu, index))
> > +		return false;
> > +
> > +	if (!msr_info->host_initiated && !pmu->lbr_event)
> > +		intel_pmu_create_lbr_event(vcpu);
> > +
> > +	/*
> > +	 * Disable irq to ensure the LBR feature doesn't get reclaimed by the
> > +	 * host at the time the value is read from the msr, and this avoids the
> > +	 * host LBR value to be leaked to the guest. If LBR has been reclaimed,
> > +	 * return 0 on guest reads.
> > +	 */
> > +	local_irq_disable();
> > +	if (event_is_oncpu(pmu->lbr_event)) {
> > +		if (read)
> > +			rdmsrl(index, msr_info->data);
> > +		else
> > +			wrmsrl(index, msr_info->data);
> > +	} else if (read)
> > +		msr_info->data = 0;
> > +	local_irq_enable();
> > +
> > +	return true;
> > +}
> So this runs in the vCPU thread in host context to emulate the MSR
> access, right?
Yes, it's called to emulate MSR accesses when the guest LBR event is
scheduled on while the LBR stack MSRs have not been passthrough to the vCPU.

Thanks,
Like Xu
>