Thread: Re: pg_stat_statements: Avoid holding excessive lock

> +             /*
> +              * There is no need to hold entry->mutex when reading stats_since and
> +              * minmax_stats_since for (unlike counters) they are always written
> +              * while holding pgss->lock exclusively. We are holding pgss->lock
> +              * shared so there should be no race here.
> +              */
>               stats_since = entry->stats_since;
>               minmax_stats_since = entry->minmax_stats_since;
> -             SpinLockRelease(&entry->mutex);

>> The comment could be simpler, say a "The spinlock is not required when
>> reading these two as they are always updated when holding pgss->lock
>> exclusively.".  Or something like that.

Agree , It reduces the lock time , The new comment are short and concise, It  sounds good .