Re: Help needed to configure and compile gpsd to run on a Raspberry Pi 4 and NOT use systemd

[Gary E. Miller]

Yo Mick!

On Fri, 15 Nov 2024 20:17:16 +0000
Mick Durkin <mickdurkinuk@gmail.com> wrote:

> OK, we cloned the latest version from git and compiled/installed it.
> Attached is a log of the compile and install (Compile.log).

Sadly you started compile.log after a configure pass.  So you only got
part oi the build process.  Best to do "scons -c" to ensure a fresh and
complete build for the Compile.log

Thses are handy, but you can run them remotely:

WARNING: gtk+-3.0 not found.
WARNING: xgps and xgpsspeed are missing runtime dependencies

Ditto:

WARNING: AsciiDoctor not found.
WARNING: Some documentation and html will not be built.

Otherwise looks fine for a minimal build.

> We then ran gpsdebuginfo under the assumption that gpsd should not be
> started and running as a daemon. Attached is the log (GPSD Debug
> Info.log).

Wrong assumption.  I see your Control.log has a few of the things we
expect from gpsdebuginfo.

> We then ran gpsd with "gpsd -n /dev/ttyS0", as this is the serial
> device it is attached to, as opposed to "gpsd -n /dev/ttyAMA0".

That makes no sense.  I thought you were on a RasPi??

Is your ttyS0 a simlink to ttyAMA0?  ttyS0 is supposed to be a UART,
but ttyAMA0 is only TX/RX.  The difference is that  ttyS0 support DCD
and automatci PPS.  ttyAMA0 has no UART, had no DCD, thus has no PPS.

> "cgps" gave sensible output.

Good.

> However, when we tried "ntpshmmon" we were only
> seeing NTP0 data, no NTP1 data at all.

Which is expected, assuming you are using ttyS0 on a RasPi.  Since the
RasPi has no DCD for gpsd to autodetect.

> We then ran "ipcs" and shmid 0
> - 11 showed, with shm 0 & 1 having 600 perms, as expected and the rest
> with 666 perms.

As long as you start gpsd as root, almost impossible to break that.

> We then ran "ppscheck /dev/ttyS0" which showed the
> attached pps device was pps1, not the expected pps0. We then ran "ls
> -al /dev/pps*" and saw both pps0 and pps1.

See comments on DCD above.

> When we killed gpsd, the same command still showed pps0.

Means nothing.

> However, at that point we noticed the
> creation date was not today.

Means nothing.

> These tests are attached (General.log).

Not much of use in there.

> As a verification, we rebooted the server and before doing anything to
> do with gpsd, we again checked for /dev/pps* and found pps0 created.

Can you supply your /boot/config.txt file contents?  You did modify per
the directions?

> This simple test is not in the log, but the pps0 device had a creation
> date and time appropriate to the system reboot time. This seems to
> indicate something in the startup of the system is grabbing this pps
> device.

Yes, on all UNIX, /dev is populated at boot time for devices not hot
plugged.

> We suspect this may be some remnant of previous systemd
> operations but have no way of telling.

The system log would tell you, but I seriously doubt systemd(umb) did itt.

> Perhaps you could point us to
> finding why this device is being created?

Send us your: /boot/config.txt, dmesg output just after boot, a gpsd -D4
log, and the output of gpsdebuginfo.

RGDS
GARY
---------------------------------------------------------------------------
Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
        gem@rellim.com  Tel:+1 541 382 8588

            Veritas liberabit vos. -- Quid est veritas?
    "If you can't measure it, you can't improve it." - Lord Kelvin