Re: [PATCH v8 09/13] perf stat: implement control commands handling

[Alexey Budankov]

On 25.06.2020 15:14, Jiri Olsa wrote:
> On Wed, Jun 24, 2020 at 05:10:10PM +0300, Alexey Budankov wrote:
>>
>> On 23.06.2020 17:54, Jiri Olsa wrote:
>>> On Wed, Jun 17, 2020 at 11:41:30AM +0300, Alexey Budankov wrote:
>>>
>>> SNIP
>>>
>>>>  
>>>>  	while (1) {
>>>>  		if (forks)
>>>> @@ -581,8 +617,17 @@ static int dispatch_events(bool forks, int timeout, int interval, int *times, st
>>>>  		if (done || stop || child)
>>>>  			break;
>>>>  
>>>> -		nanosleep(ts, NULL);
>>>> -		stop = process_timeout(timeout, interval, times);
>>>> +		clock_gettime(CLOCK_MONOTONIC, &time_start);
>>>> +		if (!(evlist__poll(evsel_list, time_to_sleep) > 0)) { /* poll timeout or EINTR */
>>>> +			stop = process_timeout(timeout, 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;
>>>
>>> should we check time_to_sleep > time_diff first?
>>
>> Probably and if time_diff > time_to_sleep then time_to_sleep = 0 ?
> 
> or extra call to process_timeout? if we dont want to call evlist_poll
> with 0 timeout

poll() man page says it is ok to call poll with 0 timeout so 
process_timeout() and initialization of time_to_sleep will be
done in common flow.

~Alexey