On 3/31/2022 2:44 PM, Cole Walker wrote: > Thanks for the reply, Jacob. > > The kernel is 5.10.74-200.1644 > > I am going to try the 1.8.3 ICE driver first, as that is easier than a > kernel change in my environment. > > Do you have any insight as to what could be causing the nmea delay to > increase like that? Would the driver affect it? It does seem to be a > common factor in this behaviour and I wonder if it is worth > investigating further. > > Cole
Sure. Please try out the 1.8.3 ICE driver and let me see if there's any results different. I suspect that something must be going wrong with delivering the timestamps to ts2phc, either in the device or in the driver. Nothing else makes much sense. You get an incorrect timestamp, and then it ofcourse thinks there is a big delay, so it starts frequency tuning, then the next few timestamps are valid so it eventually corrects itself. I doubt this is a bug in ts2phc but perhaps it is. I am a bit suspicious of that particular line in ts2phc you brought up where that check adds to the second value. The jump in difference is almost exactly 0.5 seconds. That feels quite suspicious. Thanks, Jake _______________________________________________ Linuxptp-users mailing list Linuxptp-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/linuxptp-users
