Udo van den Heuvel wrote: > I now have a patch for the NMEA refclock that comes with ntp to make it > work with linuxPPS. (thanks to Folkert and Rodolfo)
Some bits: Info about the reference NMEA ntp driver: http://www.eecis.udel.edu/~mills/ntp/html/drivers/driver20.html Original LinuxPPS snapshot: http://ftp.enneenne.com/pub/projects/linuxpps/Snapshot/ The patch for the nmea refclock that we made and the 2.6.16 kernel patches for LinuxPPS, adapted from the linuxpps-current snapshot: http://time.qnan.org/linux2.6-pps/ (alternative patch has slightly different pps source registration, is that needed?) nmea3.patch is the ntpd refclock patch. On my FC5 box I built the stock ntp 4.2.0.a.20050816-14 with all clocks disabled in the SPEC: %configure --sysconfdir=%{_sysconfdir}/ntp --bindir=%{_sbindir} --disable-all-clocks --disable-parse-clocks --with-openssl-libdir=%{_libdir} --enable-linuxcaps --enable-NMEA --enable-LOCAL-CLOCK (except of course NMEA and the local clock) This is in ntp.conf: server 127.127.20.0 prefer minpoll 4 fudge 127.127.20.0 flag3 1 flag2 0 time1 0.0 See http://time.qnan.org/ and/or the Garmin documentation at garmin.com/manuals/425_TechnicalSpecification.pdf for connecting the GPS 18 to the PC. Please review the patches before using them. > Things is that I read that the NMEA driver expects PPS after the NMEA > data. Now it appears that ntpd, using the patched NMEA driver, syncs up > very nicely with no offset, just as when using shmpps. > That is why I would like some comments on the driver and/or feedback on > the patch. > Anybody? This is still a bit vague to me. Maybe also see https://ntp.isc.org/bugs/show_bug.cgi?id=610. _______________________________________________ timekeepers mailing list [email protected] https://fortytwo.ch/mailman/cgi-bin/listinfo/timekeepers
