[ntp:hackers] Re: gps_vme.c and Bug #143

Harlan Stenn stenn at whimsy.udel.edu
Tue Aug 19 14:02:57 PDT 2003


> Yes and Yes (the answer is almost always yes).

Thanks!

> I wanted to change the PCI driver name on the box "time" since (a) its
> PCI not VME and (b) its not GPS (although the same driver is used for
> Brandywine cards with GPS). But we don't have a separate refclock
> number assigned.

I don't care what we call it.  I can rename it easily if somebody tells
me a better name.

I'm not sure I understand the issue of the refclock number on this.

Dave, any thoughts?

> Aug 19 20:15:16 time ntpd[8751]: synchronized to GPS_VME(1), stratum=0
> Aug 19 20:23:58 time ntpd[8751]: synchronized to GPS_VME(0), stratum=0
> Aug 19 20:32:25 time ntpd[8751]: synchronized to GPS_VME(1), stratum=0
> Aug 19 20:41:00 time ntpd[8751]: synchronized to GPS_VME(0), stratum=0

> The NTP daemon is flipping between the two cards, perhaps one has to have
> the "prefer" keyword attached in ntp.conf?

I don't know the answer to this one.  I was curious if it would make
sense to add a PPS clock to the box as well - that would be a different
way to solve the problem.

H
-
> Regards, Rich
> 
> > Guys,
> > 
> > Two questions.
> > 
> > First, do I have permission to update the gps_vme code in the public NTP
> > sources to include your changes?
> > 
> > Second, is it OK if I monitor ntpd on time.usno.navy.mil from my
> > framework at UDel?  Basically, I'll be checking the output of ntpq every
> > 10 minutes or so.
> > 
> > Harlan



More information about the hackers mailing list