[ntp:hackers] Leap stuff

Kurt Roeckx kurt at roeckx.be
Tue Jul 10 18:12:04 UTC 2007


On Mon, Jul 09, 2007 at 10:14:43PM +0000, David L. Mills wrote:
> Kurt,
> 
> There is no change in the interpretation of the leap bits; it's been 
> that way for many years. The LEAP_NOWARNING means exactly what is says; 
> no leap second is anticipated in future.

Dave,

Terje suggested that it might be useful to know the difference between:
- I know that there is not going to be a leap second.
- I don't know if there is going to be a leap second or not.

You could then for instance only take those servers into account that
are sure about the leap seconds.

And as far as I know, as ntp client there is no way to tell the
difference between those cases.  Maybe something changed and you can
now actually tell the difference between those?

> If no leapseconds values are available, either from file or server, the 
> TAI is ordinaril zero. However, if a leap is signaled from a server and 
> passes the majority test, the kernel will leap as directed. In the 
> current design the kernel TAI is incremented, so an application program 
> will see the incremented value. Since leaps occur so rarely, the 
> application should assume no valid TAI information is available if the 
> TAI value is less than 10, which was the intial offset in 1972,

This part seems to be about knowing the correct offset between TAI and UTC,
which can also be very useful.  But this has little to do with deciding
if there is going to be a leap second or not if you don't have valid 
leapseconds values yourself.


Kurt

> Kurt Roeckx wrote:
> 
> > On Mon, Jul 02, 2007 at 05:00:01PM +0000, David L. Mills wrote:
> >
> >> Terje Mathisen wrote:
> >>
> >>> Here it would have been good to have effectively four possible values:
> >>>
> >>> 0) Unknown, i.e. no authoritative source available.
> >>
> >> TAI = 0 and no leap bits
> >>
> >>> 1) Known, no leapsecond. (with tracing back to NIST)
> >>
> >> Leapsec = 0 and current time less than file expiration time.
> >>
> >>> 2) Known, positive leapsecond. (ditto)
> >>
> >> Leapsec > 0. This is the number of seconds until the leap.
> >>
> >>> 3) Known, negative leapsecond. (ditto)
> >>
> >> No way to tell, unless using DUT1 trend.
> >>
> >>> The main idea being the difference between the first two, which for
> >>> many users would be quite significant.
> >>>
> >>> Or is this effectively what the new code does?
> >>
> >
> > If I understood Terje, he was talking about the Leap Indicator
> > in the protocol, which is now defined as:
> > | 0 | no warning |
> > | 1 | last minute of the day has 61 seconds |
> > | 2 | last minute of the day has 59 seconds |
> > | 3 | alarm condition (the clock is not synchronized) |
> >
> > And that 3 maybe should say: "I have no idea if there is going to be a
> > leap second or not".
> >
> > The way I understand things is that we can't just change this.
> >
> >
> > Kurt
> >
> 
> _______________________________________________
> hackers mailing list
> hackers at lists.ntp.org
> https://lists.ntp.org/mailman/listinfo/hackers


More information about the hackers mailing list