[Linuxptp-users] Incorrect time with ts2phc tod_source nmea and extts_polarity both

[James C. <jj...@jc...>]

I have a u-blox M8T connected to an i210-T1 and am trying this config
with ts2phc 4.0

[global]
ts2phc.pulsewidth 100000000
ts2phc.tod_source nmea
ts2phc.nmea_serialport /dev/ttyUSB0
ts2phc.nmea_baudrate 9600
leapfile /usr/share/zoneinfo/leap-seconds.list

[enp1s0]
ts2phc.extts_polarity both

I have chrony controlling the system clock using a GPS-connected local
NTP server and after the offsets have settled down I compare the PHC
and system clock using:

phc_ctl enp1s0 cmp

I get results that are close to 36.9 seconds (e.g.  -36899918303ns).

If I remove the tod_source line (so it is using the generic source), I
get results that are close to 37s (e.g. -36999949290ns), which is what
I would expect.

James