RE: [chrony-dev] pps source is still marked as synchronised when no absolute clock sources are available

["Hattink, Tjalling (FINT)" <T.Hattink@xxxxxxxx>]

>>>
>>> Ideally, this would be configurable. Perhaps a refclock option to
bind it
>>> with another one. This would also allowed to use PPS signal
immediately
>>> without waiting for first sync.
>>
>> This would be a really nice solution. It gives the user full control
in
>> how the PPS should be treated. I can bind the PPS signal to the SHM
>> reference clock, which makes it clear to chrony that the PPS signal
>> should only be trusted when the SHM clock is trusted. If no binding
is
>> made chrony should always trust the PPS signal.
>
>I think this adds more complexity to chrony, when it belongs in the
refclock
>software. In your case both the pps and the nmea come from the same
device,
>and both are good or bad. That single device should be delivering a
single
>time, not two times, to chrony. Ie, the software which reads the gps
should be
>reading both the pps and the nmea and delivering one time to for
example the
>shm driver. One should not complicate chrony to make up for
deficiencies in
>the driver. One job per program.
>

I can live with this (I have to add PPS support in my SHM driver now),
but
I think a warning about this behaviour of PPS clocks should be put in
the
chrony manual. The PPS driver behaves differently than used in ntpd. It
should mention that the PPS signal is always trusted (in ntpd it is only
trusted when an absolute time source or reference clock is trusted).

--
Tjalling Hattink

---
To unsubscribe email chrony-dev-request@xxxxxxxxxxxxxxxxxxxx with "unsubscribe" in the subject.
For help email chrony-dev-request@xxxxxxxxxxxxxxxxxxxx with "help" in the subject.
Trouble?  Email listmaster@xxxxxxxxxxxxxxxxxxxx.