[ntp:questions] SIRF time output wobble, GPS or NTP fault?

Dave Hart hart at ntp.org
Thu Mar 8 09:00:27 UTC 2012


On Thu, Mar 8, 2012 at 08:17, David J Taylor
<david-taylor at blueyonder.co.uk.invalid> wrote:
> Dave, perhaps you could clarify one thing on Windows about the user-mode PPS
> timestamping.  When you say "serial I/O layer", my understanding is that
> this /would/ include virtual COM port devices such as RS-232/USB adapters.

Correct.

> Perhaps the Linux port could also benefit from a user-mode DCD timestamp, or
> is there not the justification for the effort?

It's an ugly hack, and one that fortunately will be migrating out of
the ntpd Windows port serial I/O code and into its own PPSAPI provider
DLL, thanks to a suggestion by Juergen Perlinger which I should have
considered myself.

There's no provision to enable or disable it currently, for example.
By putting it in an optional DLL, only those who want end-of-line
timestamps replaced with PPS timestamps taken in user mode would get
them.  It's particularly confusing for drivers that have different
fudge factors for PPS and serial, as NMEA does, and putting it in a
PPSAPI provider DLL will correct that confusion.

Apparently no one has had the right mix of idea and motivation to
modify the POSIX ntpd serial I/O code to watch for DCD transitions to
accomplish similar.  It's not hard to find proper PPS support in free
OSes.

Cheers,
Dave Hart


More information about the questions mailing list