Re: [PATCH] KVM: x86: Fix perfctr WRMSR for running counters

[Paolo Bonzini]

On 27/01/20 22:22, Eric Hankland wrote:
> Correct the logic in intel_pmu_set_msr() for fixed and general purpose
> counters. This was recently changed to set pmc->counter without taking
> in to account the value of pmc_read_counter() which will be incorrect if
> the counter is currently running and non-zero; this changes back to the
> old logic which accounted for the value of currently running counters.
> 
> Signed-off-by: Eric Hankland <ehankland@xxxxxxxxxx>
> ---
>  arch/x86/kvm/vmx/pmu_intel.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/arch/x86/kvm/vmx/pmu_intel.c b/arch/x86/kvm/vmx/pmu_intel.c
> index 34a3a17bb6d7..9bdbe05b599c 100644
> --- a/arch/x86/kvm/vmx/pmu_intel.c
> +++ b/arch/x86/kvm/vmx/pmu_intel.c
> @@ -264,9 +264,10 @@ static int intel_pmu_set_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
>  				pmc->counter = data;
>  			else
>  				pmc->counter = (s32)data;
> +			pmc->counter += pmc->counter - pmc_read_counter(pmc);

I think this best written as it was before commit 2924b52117:

			if (!msr_info->host_initiated)
				data = (s64)(s32)data;
			pmc->counter += data - pmc_read_counter(pmc);

Do you have a testcase?

Paolo

>  			return 0;
>  		} else if ((pmc = get_fixed_pmc(pmu, msr))) {
> -			pmc->counter = data;
> +			pmc->counter += data - pmc_read_counter(pmc);
>  			return 0;
>  		} else if ((pmc = get_gp_pmc(pmu, msr, MSR_P6_EVNTSEL0))) {
>  			if (data == pmc->eventsel)
>