[LinuxPPS] LinuxPPS new functionalities

clemens at dwf.com clemens at dwf.com
Mon Feb 22 20:06:04 CET 2010


> On 2010-02-21 20:40, clemens at dwf.com wrote:
> >> Do we know *WHY* this patch works?
> >> I got this question from the person that I sent the previous ldattach
> >> patches to. The answer might help add the needed functionality.
> >>
> > I dont know anything about the patch.
> > I am using the ldattach as distributed in Fedora12, and am having no
> > problems with it at all,- so  it would seem that in at least my case
> > that the patch is not necessary.
> 
> Hmm, OK.
> What exactly is the hardware that you are using?
> And/or how do you start ntpd et all? (this might help explaining)
> 

Well, I think the most important thing is that this is ldattach from Fedora12,
It is part of the glibc yum package, which shows a version of
    glibc-2.10.1-2
for x86_64 and i586 and i686.  Previous versions (Fedora11) complain that
PPS is not defined.

It is also possible (as you surmise) that its the way I use it.
My startup code does the following 
       ldattach PPS /dev/ttyS0
        loop, up to 10 times with 5 sec pause, looking for the /dev/pps<x>
            that belongs to the device on the ldattach line.
        set up the symbolic link from /dev/oncore.pps.0 to that device
        daemon ntpd 

The only really interesting thing here is the wait for /dev/pps<x> to
appear.  This was important with a previous version of the ldisc code, and
the search for the correct /dev/pps<x> is important on my machine
since I keep changing things around and /dev/pps0 is not guaranteed.

As for hardware, this is working on everything from a Pentium II, thru
a  couple of Pentium4's, and a Core 2 duo.

If anyone is interested, I can provide my startup script (from 
/etc/rc.d/init.d)
and/or the little job that searches for the right /dev/pps<x> .

-- 
                                        Reg.Clemens
                                        reg at dwf.com





More information about the LinuxPPS mailing list