Re: [PATCH] perf: riscv: Fix counter_idx_base in PMU snapshot restart

[Anup Patel]

On Sat, Aug 22, 2026 at 9:43 PM liutong <liutong@xxxxxxxxxxx> wrote:
>
> In pmu_sbi_start_ovf_ctrs_snapshot(), after for_each_set_bit() finishes
> iterating used_hw_ctrs[i], idx is left at BITS_PER_LONG. The subsequent
> sbi_ecall() uses idx * BITS_PER_LONG as counter_idx_base, passing an
> out-of-range value (4096 on 64-bit) to SBI_EXT_PMU_COUNTER_START.
>
> This causes the firmware to reject the call and overflowed counters are
> never restarted.
>
> Use the outer loop index 'i' instead of the exhausted iterator 'idx'.
>
> Fixes: a8625217a054 ("drivers/perf: riscv: Implement SBI PMU snapshot function")
> Signed-off-by: liutong <liutong@xxxxxxxxxxx>

LGTM.

Reviewed-by: Anup Patel <anup@xxxxxxxxxxxxxx>

Thanks,
Anup

> ---
>  drivers/perf/riscv_pmu_sbi.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/perf/riscv_pmu_sbi.c b/drivers/perf/riscv_pmu_sbi.c
> index dfc886dee..099ba974a 100644
> --- a/drivers/perf/riscv_pmu_sbi.c
> +++ b/drivers/perf/riscv_pmu_sbi.c
> @@ -1021,7 +1021,7 @@ static inline void pmu_sbi_start_ovf_ctrs_snapshot(struct cpu_hw_events *cpu_hw_
>                         sdata->ctr_values[idx] =
>                                         cpu_hw_evt->snapshot_cval_shcopy[idx + i * BITS_PER_LONG];
>                 /* Start all the counters in a single shot */
> -               sbi_ecall(SBI_EXT_PMU, SBI_EXT_PMU_COUNTER_START, idx * BITS_PER_LONG,
> +               sbi_ecall(SBI_EXT_PMU, SBI_EXT_PMU_COUNTER_START, i * BITS_PER_LONG,
>                           cpu_hw_evt->used_hw_ctrs[i], flag, 0, 0, 0);
>         }
>  }
> --
> 2.34.1
>