[LinuxPPS] Problems with ppstest and NTP

Paul paul at lavender-fam.net
Wed Jan 9 14:29:04 CET 2008


I compiled with p3 so it just possble that something has changed. When
you look in config.h which options are defined? Is the HAVE_PPSAPI
there?

Paul

On Wed, 2008-01-09 at 13:54 +0100, Heiko Gerstung wrote:
> Paul schrieb:
> > I am using the Atom driver with a GPS receiver (there are issues with
> > the time output on some GPS modules, but the PPS is rock solid). Are you
> > compiling the ntp with the atom driver or have you patched in some way?
> > When the 'new' kernel came out I tried the nmea patch out of curiosity,
> > had problems, but the dropped it because it was of no interest to me.
> >   
> 
> NTP (I use a vanilla 4.2.4p4)  has been definitely compiled with atom 
> support, you can tell from the error message the atom refclock driver 
> gives me:
> 
> Jan  9 10:46:07 xxx daemon.err ntpd[5675]: refclock_atom: 
> time_pps_create failed: Invalid argument
> 
> 
> If NTPD would not have been built with the atom driver, it says 
> something like "unknown refclock 22". Additionally, the driver for our 
> own GPS clocks (parse driver, #8) shows the same error message when 
> trying to access the PPS and this driver works fine (without PPS support 
> of course) ...
> 
> Best Regards,
>  Heiko
> 
> 
> >
> > Paul
> >
> > On Wed, 2008-01-09 at 12:55 +0100, Heiko Gerstung wrote:
> >   
> >> Paul,
> >>
> >> thanks for your comments. The clear/assert events are working fine and I 
> >> would have no problem letting ppstest go, but my problem is that NTP 
> >> cannot use the pps source as well despite the fact that Rodolfo states 
> >> that at least the Atom driver of NTP should work without any modification.
> >>
> >> Best Regards,
> >> Heiko
> >>
> >>
> >> Paul schrieb:
> >>     
> >>> If the clear/asserts change in /sys/class/pps/pps0/assert
> >>> or /sys/class/pps/pps0/clear all is well. I think the debugging output
> >>> is more a developer tool.
> >>>
> >>> Paul
> >>>
> >>> On Wed, 2008-01-09 at 12:08 +0100, Heiko Gerstung wrote:
> >>>   
> >>>       
> >>>> Hi there!
> >>>>
> >>>> I just managed to get LinuxPPS (ntp-pps-2.6.23-quater.diff) to run on 
> >>>> one of our embedded systems (kernel 2.6.23.12, i386 arch). I tried to 
> >>>> follow the instructions on
> >>>> http://wiki.enneenne.com/index.php/LinuxPPS_support
> >>>>
> >>>> but I have a few problems/comments:
> >>>>
> >>>> - when I try to fire up ppstest, I get:
> >>>> /ppstest /dev/pps0
> >>>> trying PPS source "/dev/pps0"
> >>>> cannot create a PPS source from device "/dev/pps0" (Operation not supported)
> >>>>
> >>>> My pps0 device looks OK to me:
> >>>>  ls -l /dev/pps0
> >>>> crwxrwxrwx    1 root     root     253,   0 Jan  9 10:48 /dev/pps0
> >>>>
> >>>>
> >>>> And the debugging output looks fine, too:
> >>>> PPS event on source 0 at 1199876811.203122984
> >>>> capture clear seq #1260 for source 0
> >>>>
> >>>> It seems that ntpd has the same problems here:
> >>>> Jan  9 10:46:07 xxx daemon.err ntpd[5675]: refclock_atom: 
> >>>> time_pps_create failed: Invalid argument
> >>>> Jan  9 10:49:50 xxx daemon.notice ntpd[6239]: PARSE receiver #0: 
> >>>> parse_start: could not set up PPS: Invalid argument
> >>>>
> >>>>
> >>>> Any hint what I could check?
> >>>>
> >>>> Thanks in advance,
> >>>>  Heiko
> >>>>
> >>>>
> >>>> _______________________________________________
> >>>> LinuxPPS mailing list
> >>>> LinuxPPS at ml.enneenne.com
> >>>> http://ml.enneenne.com/cgi-bin/mailman/listinfo/linuxpps
> >>>> Wiki: http://wiki.enneenne.com/index.php/LinuxPPS_support
> >>>>
> >>>>     
> >>>>         
> >>> _______________________________________________
> >>> LinuxPPS mailing list
> >>> LinuxPPS at ml.enneenne.com
> >>> http://ml.enneenne.com/cgi-bin/mailman/listinfo/linuxpps
> >>> Wiki: http://wiki.enneenne.com/index.php/LinuxPPS_support
> >>>   
> >>>       
> >>     
> >
> >   
> 
> 




More information about the LinuxPPS mailing list