Re: [PATCH v9 11/15] perf stat: implement control commands handling

From: Jiri Olsa
Date: Mon Jul 06 2020 - 08:37:52 EST


On Fri, Jul 03, 2020 at 10:47:22AM +0300, Alexey Budankov wrote:

SNIP

>
> while (1) {
> if (forks)
> @@ -574,11 +610,22 @@ static int dispatch_events(bool forks, int timeout, int interval, int *times, st
> if (done || stop || child_exited)
> break;
>
> - nanosleep(ts, NULL);
> - if (timeout)
> - stop = true;
> - else
> - stop = handle_interval(interval, times);
> + clock_gettime(CLOCK_MONOTONIC, &time_start);
> + if (!(evlist__poll(evsel_list, time_to_sleep) > 0)) { /* poll timeout or EINTR */
> + if (timeout)
> + stop = true;
> + else
> + stop = handle_interval(interval, times);
> + time_to_sleep = sleep_time;
> + } else { /* fd revent */
> + stop = process_evlist(evsel_list, interval, times);
> + clock_gettime(CLOCK_MONOTONIC, &time_stop);
> + diff_timespec(&time_diff, &time_stop, &time_start);
> + time_to_sleep -= time_diff.tv_sec * MSEC_PER_SEC +
> + time_diff.tv_nsec / NSEC_PER_MSEC;
> + if (time_to_sleep < 0)
> + time_to_sleep = 0;

could this computation go to a separate function? something like:

time_to_sleep = compute_tts(time_start, time_stop);

thanks,
jirka