[dpdk-dev] [PATCH] rte_alarm: modify it to make it not to be affected by discontinuous jumps in the system time

[Jay Rolette]

Back when this was first submitted in June, I mentioned that
CLOCK_MONOTONIC_RAW was ~10x slower than CLOCK_MONOTONIC:

http://dpdk.org/ml/archives/dev/2015-June/018687.html

It's not completely free from NTP frequency adjustments, but it won't have
any discontinuities.

That's what we've been using in our tree since then...

Jay


On Tue, Oct 13, 2015 at 7:33 PM, Stephen Hemminger <
stephen at networkplumber.org> wrote:

> On Fri,  5 Jun 2015 10:46:36 +0800
> Wen-Chi Yang <wolkayang at gmail.com> wrote:
>
> > Due to eal_alarm_callback() and rte_eal_alarm_set() use gettimeofday()
> > to get the current time, and gettimeofday() is affected by jumps.
> >
> > For example, set up a rte_alarm which will be triggerd next second (
> > current time + 1 second) by rte_eal_alarm_set(). And the callback
> > function of this rte_alarm sets up another rte_alarm which will be
> > triggered next second (current time + 2 second).
> > Once we change the system time when the callback function is triggered,
> > it is possiblb that rte alarm functionalities work out of expectation.
> >
> > Replace gettimeofday() with clock_gettime(CLOCK_MONOTONIC_RAW, &now)
> > could avoid this phenomenon.
> >
> > Signed-off-by: Wen-Chi Yang <wolkayang at gmail.com>
>
> Agreed, this should be applied.
> Does BSD version have same problem?
>
> Acked-by: Stephen Hemminger <stephen at networkplumber.org>
>
>