On Thu, Oct 06, 2022 at 02:29:40PM +0200, Arne Schwabe wrote:
> Remove function event_timeout_clear_ret as it is unused.
>
> Cleanup event_timeout_trigger a bit. Do an instant return false if the
> timeout is not defined and inline local_now and use
> event_timeout_remaining instead of local duplicated code.
>
> Add doxygen comments for all timeout function, especially for the
> event_timeout_trigger function that is hard to understand otherwise.
>
> Patch v2: add many fixes/correction suggested by Frank
>
The new sentences have some small issues. Gert can probably fix them on
apply.
Acked-By: Frank Lichtenheld <fr...@lichtenheld.com>
[...]
> -/*
> - * Will return the time left for a timeout, this function does not check
> - * if the timeout is actually valid
> +/**
> + * Returns the time until the timeout should triggered from from now.
"should be triggered, from \c now."
> + * This function does not check if the timeout is actually valid.
> */
> static inline interval_t
> event_timeout_remaining(struct event_timeout *et)
> {
> - return (interval_t) (et->last - now + et->n);
> + return (interval_t) ((et->last + et->n) - now);
> }
>
> -/*
> - * This is the principal function for testing and triggering recurring
> - * timers and will return true on a timer signal event.
> - * If et_const_retry == ETT_DEFAULT and a signal occurs,
> - * the function will return true and *et will be armed for the
> - * next event. If et_const_retry >= 0 and a signal occurs,
> - * *et will not be touched, but *tv will be set to
> - * minimum (*tv, et_const_retry) for a future re-test,
> - * and the function will return true.
> - */
> -
> #define ETT_DEFAULT (-1)
>
> +/**
> + * This is the principal function for testing and triggering recurring
> + * timers.
> + *
> + * If *et is not triggered, *tv is set to remaining time until the timeout if
> + * not already lower:
> + *
> + * *tv = minimum(*tv, event_timeout_remaining(*et))
> + *
> + * If *et triggers and et_const_retry is negative (ETT_DEFAULT is -1):
> + * - the function will return true
> + * - *et will be armed for the next event (et->last set to now).
> + * - *tv will be lowered to the event period (n) if larger than the
> + * period of the event (set to *et's next timeout)
> + * - *et will not changed (ie also not rearmed, stays armed)
> + *
> + * If *et triggers and et_const_retry >= 0, *tv will be lowered to
> et_const_try
> + * if larger:
> + *
> + * *tv = *minimum(*tv, et_const_retry)
> + *
> + * This is mainly useful is the trigger cannot yet be triggered for other
"if the trigger"
> + * reasons and a backoff timeout should be returned if the timer is ready
> + * to trigger.
> + *
> + *
> + * @param et the timeout to check
> + * @param tv will be set to timeout for next check for this
> + * timeout unless already smaller.
> + * @param et_const_retry see above
> + * @return if the timeout has triggered and event has been
> reset
> + */
> bool event_timeout_trigger(struct event_timeout *et,
> struct timeval *tv,
> - const int et_const_retry);
> + int et_const_retry);
>
> /*
> * Measure time intervals in microseconds
> diff --git a/src/openvpn/openvpn.h b/src/openvpn/openvpn.h
> index 00cd652fa..95c7d7520 100644
> --- a/src/openvpn/openvpn.h
> +++ b/src/openvpn/openvpn.h
> @@ -386,7 +386,9 @@ struct context_2
> * Event loop info
> */
>
> - /* how long to wait on link/tun read before we will need to be serviced
> */
> + /** Time to next event of timers and similar. This is used to determine
> + * how long to wait on event wait (select/poll on link/tun read)
> + * before this context wants to be serviced */
Add period for consistency.
> struct timeval timeval;
>
> /* next wakeup for processing coarse timers (>1 sec resolution) */
Regards,
--
Frank Lichtenheld
_______________________________________________
Openvpn-devel mailing list
Openvpn-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/openvpn-devel
