[LinuxPPS] Recommendation / Re: Stats & architectures

[Robert Jenkins]

> From: clemens at dwf.com 12 March 2008 17:30

> Subject: Re: [LinuxPPS] Recommendation / Re: Stats & architectures 
> 
> > 
> > This is documented in the wiki, it's a 'none error'.
> > 
> > It's more of a test of availability for a particular feature, as I
> > understand it.
> > 
> > Robert.
> > 
> Fine, as long as you understand what is going on.
> In the oncore driver getting this message would be a fatal 
> error, since
> you would have asked for HARDPPS and the kernel did not support it.
> 
> 
> -- 
>                                         Reg.Clemens
>                                         reg at dwf.com
> 

To my understanding of the situation, HARDPPS is not part of the kernel code
in the present (V5.x ?) PPS API.

The kernel code stores timestamps the PPS signal edge transitions and the
NTPD pps code in whatever refclock uses these timestamps to do the rest of
the work.

The last kernel supported by the 'ppskit-lite' patch (which included Hardpps
capability) was around 2.6.17
As far as I understand it, the author of those patches stopped producing
them as kernel changes repeatedly messed the system up.

You'd have to talk to him (Ulrich Windl), Rodolfo Giometti & the NTP guys
for the technical reasons behind the changes, I'm just a user.

Robert.