NMEA but no PPS in ntp, Ubuntu 18.04.5

[Steve Bourland]

Yo Gary!

>   2. NMEA but no PPS in ntp, Ubuntu 18.04.5 (Gary E. Miller)

> Date: Thu, 3 Dec 2020 16:20:56 -0800
> From: "Gary E. Miller" <gem@rellim.com>
> Subject: NMEA but no PPS in ntp, Ubuntu 18.04.5

> Sorry, I forgot the -N.  That shold be an FAQ.

No worries, I've been around gpsd long enough I knew I'd need it :)

> After looking at the code, I decided the log never tells you which SHM()
> was used.  The ntpshm_link_activated just tells you 0 == Serial, 1 ==
> PPS, not which SHM.  That logging is improved in git head.

I'll pull git head and build this morning, let you know when I have that 
done (probably sending the new log).  I'm happy to test anything you want 
on this server while it's down, least I can do to help you get more test 
points.

> Yup, cgps thinks 3D, but PPS thinks otherwise.
>
> gpsd:INFO: PPS:/dev/ttyS0 Assert hooks called clock:
> 1607005294.000156294 real:  1607005294.000000000: no fix
>
> Very odd.  Hard to debug threading remotely.
>
> Did you ever get aound to getting ntpd to listen to SHM(0) and SHM(2)?

Did try using 127.127.28.2 with no change.  The SHM(2) only has a single 
entry at startup of ntpshmmon, only SHM(0) gets per second entries on the 
sick system.  The "healthy" system (Ubuntu 18.04.4 LTS, Linux 
4.15.0-72-generic #81-Ubuntu, gpsd: 3.19.1~dev (revision 
3.19.1~dev-2019-12-16T16:01:12), ntpd ntpsec-1.1.0+419 
2018-03-14T12:03:57-0700, using 127.127.28.0 and 127.127.28.1 
successfully) reports entries for NTP2 and NTP3 at startup, then NTP0 and 
NTP1 per second (I hope that makes sense).  The "sick" system was working 
just fine prior to my upgrade from 16.04 to 18.04.  The upgrade was done 
in-situ, so the cabling and receiver didn't move (ie, I have high 
confidence my issues are software not hardware.  Given we're seeing PPS 
signals with ppstest and cgps thinks we have a fix but pps doesn't, I 
think you're on board with me?  I haven't ruled out "permissions" issues 
that may have come up with the OS move, but have shutdown apparmor to no 
avail, and can't find a difference between the two systems in that 
regard).

> Depending you your GNSS receiver that change may help, or not.  The
> benefit, depending on model is that gpsd may optimize the sentences that
> your receiver sends.

Both systems are using (hold your nose) Garmin 18x LVCs, so I suspect 
there won't be much change if I drop the -b.
                                        Thanks,
                                          Steve