Hi Marc, Thank you for the patch! I have already sent a patch for this [1], which was queued by Drew [2], but apparently has not landed in master yet. [1] https://www.spinics.net/lists/kvm-arm/msg44084.html [2] https://gitlab.com/rhdrjones/kvm-unit-tests/-/commit/241dac4cadfd7d7ace8c8c3f0613376311b4e785 Thanks, Alex On 3/24/21 2:38 PM, Marc Zyngier wrote: > The PMU unit tests fail with an UNDEF exception when no PMU > is available (although KVM hasn't been totally consistent > with that in the past). > > This is caused by PMCR_EL0 being read *before* ID_AA64DFR0_EL1 > is checked for the PMU version. > > Move the PMCR_EL0 access to a reasonable place, which allows the > test to soft-fail gracefully. > > Fixes: 784ee933fa5f ("arm: pmu: Introduce defines for PMU versions") > Reported-by: Andrew Walbran <qwandor@xxxxxxxxxx> > Signed-off-by: Marc Zyngier <maz@xxxxxxxxxx> > --- > arm/pmu.c | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/arm/pmu.c b/arm/pmu.c > index cc959e6..15c542a 100644 > --- a/arm/pmu.c > +++ b/arm/pmu.c > @@ -988,7 +988,7 @@ static void pmccntr64_test(void) > /* Return FALSE if no PMU found, otherwise return TRUE */ > static bool pmu_probe(void) > { > - uint32_t pmcr = get_pmcr(); > + uint32_t pmcr; > uint8_t implementer; > > pmu.version = get_pmu_version(); > @@ -997,6 +997,7 @@ static bool pmu_probe(void) > > report_info("PMU version: 0x%x", pmu.version); > > + pmcr = get_pmcr(); > implementer = (pmcr >> PMU_PMCR_IMP_SHIFT) & PMU_PMCR_IMP_MASK; > report_info("PMU implementer/ID code: %#"PRIx32"(\"%c\")/%#"PRIx32, > (pmcr >> PMU_PMCR_IMP_SHIFT) & PMU_PMCR_IMP_MASK,