On Mon, Jan 20, 2025, Dapeng Mi wrote:
> On 1/18/2025 1:11 AM, Sean Christopherson wrote:
> > @@ -98,14 +149,12 @@ static uint8_t guest_get_pmu_version(void)
> > * Sanity check that in all cases, the event doesn't count when it's disabled,
> > * and that KVM correctly emulates the write of an arbitrary value.
> > */
> > -static void guest_assert_event_count(uint8_t idx,
> > - struct kvm_x86_pmu_feature event,
> > - uint32_t pmc, uint32_t pmc_msr)
> > +static void guest_assert_event_count(uint8_t idx, uint32_t pmc, uint32_t pmc_msr)
> > {
> > uint64_t count;
> >
> > count = _rdpmc(pmc);
> > - if (!this_pmu_has(event))
> > + if (!(hardware_pmu_arch_events & BIT(idx)))
> > goto sanity_checks;
> >
> > switch (idx) {
> > @@ -126,7 +175,9 @@ static void guest_assert_event_count(uint8_t idx,
> > GUEST_ASSERT_NE(count, 0);
> > break;
> > case INTEL_ARCH_TOPDOWN_SLOTS_INDEX:
> > - GUEST_ASSERT(count >= NUM_INSNS_RETIRED);
> > + __GUEST_ASSERT(count < NUM_INSNS_RETIRED,
>
> shouldn't be "__GUEST_ASSERT(count >= NUM_INSNS_RETIRED," ?
Yes. I had intentionally inverted the check to verify the assert message and
forgot to flip it back before hitting "send". Thankfully, I didn't forget before
posting formally[*]. Ugh, but I did forget to Cc you on that series, sorry :-/
[*] https://lore.kernel.org/all/20250117234204.2600624-6-seanjc@xxxxxxxxxx
