Re: Clock drift and other open issues: Collecting information

[Johnny Billquist <bqt%softjar.se@localhost>]

On 2023-12-27 00:23, Jan-Benedict Glaw wrote:
> On Wed, 2023-12-27 10:03:46 +1100, matthew green <mrg%eterna23.net@localhost> wrote:
> > >      Argh... So it's probably just that we now _see_ that something
> > >      went backwards---we just didn't get informed about it
> > >      previously...
> >
> > yes - this is a new message.
> >
> > FWIW, after my rework of sparc timecounter some time ago it *still*
> > tripped the much older "time went backwards" issue, but just once at
> > boot-time, and i wasn't able to trigger it again.
> >
> > my point being that between boot time, clock setup, and when clocks
> > and interrupts are enabled, sometimes the timecounter source will
> > overflow in a way that isn't detected and is instead seen as a time
> > goes backwards issue (and reported.)  while this boot-time message
> > isn't great and please someone fix it, i would put it relatively low
> > on the priority list.
>
> So we may or may not have this backwards-jumping time during system
> operatiom, and we (for sure) have a NTP issue. Hoping that this
> initial time-went-backwards message is not overly relevant, I'll turn
> over to ntpd and see if I can get some insight there...
>
>    Does anybody see this message during normal operation? Only in
> conjunction with ntpd?

Started building on my 4000/90. I haven't seen it at all. However, ntpd 
is not managing to sync at all. Time is moving back and forth all the 
time in a way that looks broken. But over all, it is keeping time very 
well, even though ntp is not syncing.

  Johnny

--
Johnny Billquist                  || "I'm on a bus
                                  ||  on a psychedelic trip
email: bqt%softjar.se@localhost             ||  Reading murder books
pdp is alive!                     ||  tryin' to stay hip" - B. Idol