Re: UBLOX and NTPD

Follow Up with a few answers and the requested logs. If a different debug level is needed then please let me know.

Running gpsd in two configurations:

running gpsd directly attached to the gps device (debug output in attached file gps_direct.log)
running custom application that attaches directly to the gps device and forwards all messages to another device (debug output in attached file gps_indirect.log)

In the direct configuration gpsd is connected directly to device "/dev/TTPS1" (Ublox).

gpsd -n -N -D8 /dev/ttyPS1 2>&1 | tee gpsd_direct.log

For the indirect configuration, we use openpty() to create a pseudo-terminal. We only use the FD returned for the master (ignore slave). We then create a symbolic link to the pseudo terminal called "/dev/ttyGPS0". In the indirect configuration we starrt up our application and then gpsd using device "/dev/ttyGPS0".

gpsd -n -N -D8 /dev/ttyGPS0 2>&1 | tee gpsd_indirect.log

Output of "ipcr -m" shows the same for both configurations running. Looks like the segments are correctly created. I experimented with stopping/starting ntpd and see the number of "nattch" change from 2 to 1 when stopped and 1 to 2 when started. Output is:

# ipcs -m

------ Shared Memory Segments --------
key shmid owner perms bytes nattch status
0x4e545030 0 root 600 80 2
0x4e545031 32769 root 600 80 1
0x4e545032 65538 root 666 80 1
0x4e545033 98307 root 666 80 1
0x4e545034 131076 root 666 80 1
0x4e545035 163845 root 666 80 1
0x4e545036 196614 root 666 80 1
0x4e545037 229383 root 666 80 1

0x47505344 262152 root 666 8928 1

On Tue, Mar 19, 2024 at 4:50 PM Gary E. Miller <gem@rellim.com> wrote:

Yo Joseph!

On Tue, 19 Mar 2024 16:31:36 -0500
Joseph Beissel <joebeis64@gmail.com> wrote:

> - Using a UBLOX M8

Which submodel? M8N? M9T?

> - Using GPSD 3.17 due to system built on Yocto Warrior (I know,
> sucks for me...)

Yup.

> I can run gpsd on and directly connect to our gps device using the
> following command:
>
> gpsd -b -nN8 /dev/ttyPS1 2>&1 | tee /tmp/gpsd_direct.log

Running with -b is discouraged.

> The command "ipcs -m" shows the shared memory segments. The segment
> at 0x4e545030 has two processes attached to it (nattch=2). I assume
> one is gpsd and the other is ntdp. I can run ntpshmmon and see
> samples being written to the shared memory segment:

Got some output of that?

> sample NTP0 aaaa.aaaa bbbb.bbbb cccc.ccc (repeats with different
> values

Pretend data is not useful.

> Running the "date" command on the box shows the correct date and time.

What does ntpmon say?

> Now I change things up a bit. I stop gpsd and start our custom
> application that attaches to the gps device (/dev/ttyPS1) and
> forwards all UBX and NEMA messages to another device (/dev/ttyGPS0).
> I know there's better ways to do this but other software relies on
> some of the things that our cucstom aplication does and we can't
> change that for now (down the road most likely). So having our
> application forward all messages is the best we can do for now.

Not rare.

> I then startup gpsd as follows:
>
> gpsd -b -nN8 /dev/ttyGPS0 2>&1 | tee /tmp/gpsd_indirect.log

What is /dev/ttyGPS0? A real device? A loopback device? A pipe?

> The command "ipcs -m" shows the shared memory segments. The segment
> at 0x4e545030 has two attached to it (nattch=2). I assume gpsd and
> ntpd are the same as above.

You are saving the logs. Reading the logs will allow you to now
make assumptions.

> When I run ntpshmmon I do not see any
> samples being written to the shared memory segment.

The answer will be in the log.

> Since I used "N8" I compared the 2 logs created. One thing I noticed
> is that the variable ntpshm_link_activate is set differently for the
> direct connection (/dev/ttyPS1) then it is for the indirect connection
> (/dev/ttyGPS0):
>
> - direct(/dev/ttyPS1): ntpshm_link_activate=1
> - indirect(/dev/ttyGPS0): ntpshm_link_activate=0

Not much I can do woth log fragments. Please sned the FULL logs.

> I am not sure if this is causing my problem.

No link, means no SHM output from gpsd. That would be your problem.

> And I am not sure why
> it is not set to "1". Would this keep the shared memory segment from
> being written to?

Yup.

> If so, what keeps it from getting set to "1"? Or
> is there something else that I am missing. I used the
> not-supported-terrible-tool gpsmon when running in the indirect
> configuration and things look good,

We never care what gpsmon says. We care what the logs say.

> I also compared the 2 logs to see what UBX and NMEA messages are
> being sent and they appear to be the same for the direct and indirect
> casses.

I woud like to do the same, but...

> In the direct case I captured the UBX raw configuration
> messages that gpsd sends to the device and made sure that our custom
> application was doing the same. In addition I compared the NMEA
> messages and they appear to be identical as well.

You are now totally lost.

> Also, one more question about sending messages to this group. When I
> get a reply to a message that I posted, how do I reply to it?

Without knowing yur email client, we can't know. I use claws-mail. I just
type 'R'.

Ansking questions, without the data to answer them, is not useful.

> Do I
> simply use "reply all" or is there another way?

"Reply Sender". Assuming your email client is not brain dead and knows
that "gpsd-users" is the sender.

> If the reply comes
> from gem, will a "reply-all" be sent to gem directly?

NOTHING comes from me, only from the list. Please do not do Reply-all.

> I thought I
> read that it is taboo since I am not a paying customer. Want to make
> sure I am doing the right thing by the user group.

"Reply-sender" is the way.

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

From:	Joseph Beissel
Subject:	Re: UBLOX and NTPD
Date:	Wed, 20 Mar 2024 12:39:02 -0500