gpsd-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Gpsd sending SKY message twice


From: Gary E. Miller
Subject: Re: Gpsd sending SKY message twice
Date: Thu, 1 Dec 2022 19:12:43 -0800

Yo Tomáš!

On Thu, 1 Dec 2022 08:04:10 +0100
Tomáš Vostřel <vostrelt@gmail.com> wrote:

> OK, gpsdebuginfo output attached.  

Interesting.

You never mentioned MIPS64 before.  Not a well tested arch.  Big endian
or little endian?

gpsd, and friends are not in the PATH, why not?  That severely limited
the usefulness of gpsdebuginfo.  Can you try again with gpsd, and
clients, in the PATH?

Busy Box has a really stupid "ps".  I pushed an update to gpsdebuginfo
to handle that.  Pleas retest wit the new one:

    https://gpsd.io/gpsdebuginfo

I'm pretty sure this is not a good idea:

lrwxrwxrwx    1 root     root             4 Nov 29 07:19 /dev/ttyACM0 -> gps0

ttyACM0 is reserved for drivers.

> > #if defined(CONTROL_SOCKET_ENABLE) && defined(SOCKET_EXPORT_ENABLE)
> >
> > tell me you did not compile so that this part did not get built?
> >    
> 
> As I want the gpsd binary to be as low in size as possible and no
> clients to be build, I used following flags.
> 
> scons -j 8 minimal=True manbuild=False gpsdclients=False ublox=True
> socket_export=True nostrip=True prefix=''  

You want it minimal, and then don't strip the binaries.  That is your
own foot you just shot.

Pretty hard to debug without the clients.  You can always delete them in
production.  Since you are interested in PPS, you at least should have
ntpshmmon.

> I thought the control socket is not needed to compile if not used.  

True.  But you need to use it.

> But then the PPS messages were not be shipped. After adding
> control_socket=True they started coming.  

Yup.  Because PPS uses control sockets.

How do you expect to use the PPS from gpsd without SHM or control sockets?

Another reason to have Python and sockets:

"GPS regression tests suppressed because socket_export or python is off."

Not running the regression tests on the target is a bad idea.  Especially
on MIPS64.

> There is the TCP socket for clients to connect. And SHM blocks for NTP
> synchronization.  

Yes, and more.

> Yes, the build is definitely not standard. I cross-compile in a
> proprietary build system. I attach the build log.  

Looks pretty normal to me.

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

Attachment: pgpl4O7BKeKkY.pgp
Description: OpenPGP digital signature


reply via email to

[Prev in Thread] Current Thread [Next in Thread]