On Mon, Sep 27, 2021 at 10:34:30AM -0700, Richard Cochran wrote: > On Wed, Sep 15, 2021 at 12:08:01PM +0200, Miroslav Lichvar wrote: > > When the master updated the UTC offset and leap flags in an announce > > message after a leap second, only the new flags were accepted for > > synchronization of the local clock, which caused a one-second error to > > appear until a state decision event updated the UTC offset. > > The once second error appeared only in the case of ptp4l using SW > time stamping and controlling clock_realtime directly, right?
I think it impacts also ptp4l using HW timestamping in a non-TAI PTP domain. > Or does this affect ptp4l->PHC + phc2sys PHC->clock_realtime as well? It shouldn't when when PTP is in TAI as phc2sys updates the UTC offset and leap flags at the same time when it gets TIME_PROPERTIES_DATA_SET. -- Miroslav Lichvar _______________________________________________ Linuxptp-devel mailing list Linuxptp-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/linuxptp-devel
