Dne 08. 03. 22 v 21:45 Richard Cochran napsal(a):
Is there anything that prevents the internal variable to be exposed? Isn't the "s2" or "s3" output in ptp4l the thing people are eyeing for when they start up sync and want to make sure it has already happened? At least in the more static systems like automotive, getting the sync once means it should be stable for the rest of the run. So the servo state pretty well corresponds to what an observer wants to know.The servo state is an internal variable has no relevance to outside observers. It does not track whether the servo is actually stable in the controls theory sense.
Instead, you can and should simply monitor the estimated offsets and feed them into your own algorithm.
I'm not a friend of duplicating already written algorithms. Moreover, ptp4l itself will have access to all the estimated offsets, while any outside observer will either not have them all, or will overload the ptp4l process by very frequent queries. Maybe the "subscription" feature of pmc would help here? Still, implementing a pmc subscriber seems like much more work than just querying one management id from time to time.
Martin
smime.p7s
Description: Elektronicky podpis S/MIME
_______________________________________________ Linuxptp-devel mailing list Linuxptp-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/linuxptp-devel
