[ntp:questions] Fwd: Re: NetBSD GPS/PPS using 4.2.6p3

A C agcarver+ntp at acarver.net
Wed Aug 24 15:01:23 UTC 2011

On 8/24/2011 02:36, Dave Hart wrote:
> On Wed, Aug 24, 2011 at 07:52, A C<agcarver+ntp at acarver.net>  wrote:
>> So it seems like a port contention issue though I don't understand why
> It may be specific to NetBSD.  The Windows port of ntpd has special
> code in refclock_open() and tty_open() to duplicate rather than
> re-open multiple references to the same underlying COMx device.  Given
> the POSIX platform code doesn't do anything of the kind, I assume most
> POSIX systems allow multiple opens by ntpd to operate correctly.
>> Any thoughts about all of this?  I certainly have a little less hair than I
>> did when I started.  However, I'd still like to get GPS_NMEA working or at
>> least understand why GPS_NMEA and PPS_ATOM kill each other.
> I don't see code that's writing to the PPS tty in refclock_atom.c, so
> I'm curious what you find about that.  You might try using the NMEA
> driver's built-in PPSAPI support (flag 1 1 flag2 1) without the atom
> driver.  refclock_nmea.c can either use /dev/gpsppsX for the PPSAPI,
> or it can reuse the same descriptor opened on /dev/gpsX, if
> /dev/gpsppsX can't be opened.  Given your experiences with atom and
> NMEA both opening the GPS tty, you should ensure /dev/gpsppsX does not
> exist.
> Congratulations and good luck,
> Dave Hart

I will try to capture the outgoing serial data sometime either this 
weekend or the next when I have a bit more time.  I'll send it along to 
the list as soon as I do.  For now I'll leave things using PPS only and 
see how it works.  But I still want to eventually get the GPS refclock 
working along with PPS just to have a backup clock for network outages.

More information about the questions mailing list