Hi-- On Apr 14, 2010, at 9:08 AM, David J Taylor wrote: > I think you are correct in your final statement, but my understanding (albeit > limited) is that kernel PPS provides better performance, and is used for a > very limited set of operations (possibly just timestamping the PPS signal). > All the filtering /is/ done in user-land. It isn't a pain to do, once you > have found the appropriate two or three command-lines for your system.
The main point seems to be that a PPS signal should have much more reliable latency and less jitter, so when you examine the current clock and any adjustment, you can use the PPS signal offset instead of the offset being provided to adjtime() and get better results, assuming the clock is already relatively close to accurate time. [1] Search for PPS_SYNC in: http://www.freebsd.org/cgi/cvsweb.cgi/src/sys/kern/kern_ntptime.c?rev=1.64.2.1.6.1 ...especially hardpps(). -- -Chuck [1]: If not, ntpd or whatever needs to get close to synchronized using other time sources like the GPS NMEA timestamps or other NTP timesources. _______________________________________________ questions mailing list [email protected] http://lists.ntp.org/listinfo/questions
