[LinuxPPS] One step backward (scratch that) but...

[Rodolfo Giometti]

On Fri, Sep 14, 2007 at 07:01:30PM -0600, clemens at dwf.com wrote:
> 
> OK, 
>         
>         PPS_HAVE_FINDSOURCE
> 
> is defined in timepps.h.new.syscall.API, so that problem goes away with the
> 'correct' timepps.h, however
> 
>         struct pps_source_data_s
> 
> is used in the definitions of time_pps_findpath and time_pps_findsource
> but never defined.  
> 
> I seem to remember this problem being found, and fixed once before, but
> its enough to cause me to stop working at the computer and go walk the dogs...

:)

> More when I actually get things working.

Define PPS_HAVE_FINDSOURCE is gone. Any file containg such define is
_wrong. The file timepps.h defined here:

    http://ftp.enneenne.com/pub/misc/linuxpps/patches/ntp-pps-2.6.23-rc2.diff

is correct.

Please, take a look at my previous letters with Udo where I explain to
him how refclock should correctly access to the PPS devices.

You can also look at PARSE refclock. It correctly allow an user to
define a GPS data source _and_ a PPS data source.

PARSE refclock first opens the GPS data source device, then it tries
if the just opened device is also a PPS one and, if not, it opens the
second device as PPS data source.

Ciao,

Rodolfo

-- 

GNU/Linux Solutions                  e-mail:    giometti at enneenne.com
Linux Device Driver                             giometti at gnudd.com
Embedded Systems                     		giometti at linux.it
UNIX programming                     phone:     +39 349 2432127