[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <d1117d48-2e3e-4e8e-bc47-d5f182dac928@linux.intel.com>
Date: Tue, 21 Nov 2023 10:57:54 -0500
From: "Liang, Kan" <kan.liang@...ux.intel.com>
To: Namhyung Kim <namhyung@...nel.org>,
Peter Zijlstra <peterz@...radead.org>,
Ingo Molnar <mingo@...nel.org>
Cc: Mark Rutland <mark.rutland@....com>,
Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
Arnaldo Carvalho de Melo <acme@...nel.org>,
LKML <linux-kernel@...r.kernel.org>,
Ian Rogers <irogers@...gle.com>,
Mingwei Zhang <mizhang@...gle.com>
Subject: Re: [PATCH 2/3] perf/core: Reduce PMU access to adjust sample freq
On 2023-11-20 5:19 p.m., Namhyung Kim wrote:
> For throttled events, it first starts the event and then stop
> unnecessarily. As it's already stopped, it can directly adjust
> the frequency and then move on.
>
> Signed-off-by: Namhyung Kim <namhyung@...nel.org>
Reviewed-by: Kan Liang <kan.liang@...ux.intel.com>
Thanks,
Kan
> ---
> kernel/events/core.c | 13 ++++++++++---
> 1 file changed, 10 insertions(+), 3 deletions(-)
>
> diff --git a/kernel/events/core.c b/kernel/events/core.c
> index 53e2ad73102d..fd3449e4d081 100644
> --- a/kernel/events/core.c
> +++ b/kernel/events/core.c
> @@ -4119,10 +4119,15 @@ static void perf_adjust_freq_unthr_events(struct list_head *event_list)
> if (hwc->interrupts == MAX_INTERRUPTS) {
> hwc->interrupts = 0;
> perf_log_throttle(event, 1);
> - event->pmu->start(event, 0);
> - }
>
> - if (!event->attr.freq || !event->attr.sample_freq)
> + if (!event->attr.freq || !event->attr.sample_freq) {
> + delta = 0;
> + goto next;
> + }
> +
> + if (event->hw.state & PERF_HES_STOPPED)
> + goto adjust;
> + } else if (!event->attr.freq || !event->attr.sample_freq)
> continue;
>
> /*
> @@ -4130,6 +4135,7 @@ static void perf_adjust_freq_unthr_events(struct list_head *event_list)
> */
> event->pmu->stop(event, PERF_EF_UPDATE);
>
> +adjust:
> now = local64_read(&event->count);
> delta = now - hwc->freq_count_stamp;
> hwc->freq_count_stamp = now;
> @@ -4144,6 +4150,7 @@ static void perf_adjust_freq_unthr_events(struct list_head *event_list)
> if (delta > 0)
> perf_adjust_period(event, period, delta, false);
>
> +next:
> event->pmu->start(event, delta > 0 ? PERF_EF_RELOAD : 0);
> }
> }
Powered by blists - more mailing lists