[ntpwg] KISS codes and timestamps

[David L. Mills]

Danny,

It shouldn't have to say anything. By definition, if leap bits show 
never synchronized and stratum 16, the client should assume the 
timestamps are not valid. The dirty little secret in the reference 
implementation is that the stamps are valid if the server is in fact 
properly synchronized.

We have talked on occasion whether sufficient mischief could be achieved 
if the stamps were set to zero or produced a sinusoidal variation of 
fifty years over the day.

Dave

Danny Mayer wrote:

> [I managed to send this to the wrong mailing list. My apologies if you
> see this twice - PDM]
>
> I don't want to open a can of worms but should we add language to the
> draft that says that if you receive DENY, RSTR, or RATE KISS codes you
> should not rely on the contents of the NTP Timestamps in the NTP packet
> or should we say that the contents of the Timestamps will always be valid?
>
> Danny
>
> _______________________________________________
> ntpwg mailing list
> ntpwg at lists.ntp.org
> https://lists.ntp.org/mailman/listinfo/ntpwg