We are running ptp4l on our grandmaster clock. It has a ptp device, which the kernel identifies, which is not attached to a particular interface device. We use an upstream ICD-GPS-153 receiver to provide 1PPS and timestamps. I wrote some code for ts2phc to interface with this device. It behaves as expected, and we have checked that the PTP messages [announce, sync, follow up, delay request, and delay response] using wireshark. The timestamping in the messages all looks reasonable, given the situation.
Our GMC is a PPC-based computer running Gentoo linux. For integration reasons, we upgraded to ptp4l 3.1 to serve our PTP time. Our clients are running Red Hawk linux with ptpd2. The GPS receiver is plugged into an INS, and the INS puts out a high frequency nav message that I can use to roughly check when the top of the UTC second occurs. When I run solely my ts2phc application on the GMC, the PTP messages look sensible, but the clients add a second to their system time every 32 seconds. System time steps away from UTC by 1s every 32s. It ran for 12 hours with that behavior. When I add phc2sys on my GMC, then the client system times converge on UTC [which is what I expect a call to get system time to return in this case]. I have no idea where to look to figure out why slave behavior is affected by applications that should only have effects on the master. Do you have any recommendations for where to start looking? Thanks! -- David ==============
_______________________________________________ Linuxptp-users mailing list Linuxptp-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/linuxptp-users
