On Wed, Jan 18, 2023 at 09:58:38PM -0800, Reiji Watanabe wrote: > Hi Ricardo, > > On Mon, Jan 9, 2023 at 1:18 PM Ricardo Koller <ricarkol@xxxxxxxxxx> wrote: > > @@ -898,12 +913,12 @@ static void test_overflow_interrupt(bool overflow_at_64bits) > > > > pmu_reset_stats(); > > This isn't directly related to the patch. > But, as bits of pmovsclr_el0 are already set (although interrupts > are disabled), I would think it's good to clear pmovsclr_el0 here. > > Thank you, > Reiji > There's no need in this case as there's this immediately before the pmu_reset_stats(); report(expect_interrupts(0), "no overflow interrupt after counting"); so pmovsclr_el0 should be clear. > > > > > - write_regn_el0(pmevcntr, 0, PRE_OVERFLOW); > > - write_regn_el0(pmevcntr, 1, PRE_OVERFLOW); > > + write_regn_el0(pmevcntr, 0, pre_overflow); > > + write_regn_el0(pmevcntr, 1, pre_overflow); > > write_sysreg(ALL_SET, pmintenset_el1); > > isb(); > >
