On Tue, Jul 02, 2019 at 02:19:04PM +0200, Armin HAMAR wrote: > So it's also the cable that needs to be compensated?
Only when calculating the timestamping corrections. If you have 938ns over 3m cable, that should be about 923ns over 0m cable. > Will be a lot of tweaking does somebody know how intel got those offsets? > And howlong one needs to measure to get a more or less good value? Most accurate should be a scope with Ethernet analyzer. You could also try to find a calibrated GM, but you would have to trust the vendor they did it right. Few hours worth of data should be enough. Note that the timestamping errors may depend on the network load and configuration of the NIC, so make sure your testing environment is similar to the production environment. -- Miroslav Lichvar _______________________________________________ Linuxptp-users mailing list Linuxptp-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/linuxptp-users
