[LinuxPPS] New PPS test patch

[William S. Brasher]

On Wed, 20 May 2009, Rodolfo Giometti wrote:

> Here a new test patch that implements PPS header file split up:
> 
>    http://ftp.enneenne.com/pub/misc/linuxpps/patches/tests/ntp-pps-v2.6.30-rc5-bis.diff
> 
> Please, test it and report results.
> 
> Ciao,
> 
> Rodolfo
> 
> -- 
> 
> GNU/Linux Solutions                  e-mail: giometti at enneenne.com
> Linux Device Driver                          giometti at linux.it
> Embedded Systems                     phone:  +39 349 2432127
> UNIX programming                     skype:  rodolfo.giometti
> 
> _______________________________________________
> 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
> 
> 

I patched linux-2.6.30-rc6 with ntp-pps-v2.6.30-rc5-bis.diff, built, 
installed and booted into the new kernel.  Ntp started without a fuss, and 
was stable with 0 offset within 10 minutes.

I have had no problem starting ppsldisc, and have been unable to trigger 
the kernel warning I received earlier.



Finally, I found one reason I had to start ntp manually on some of the 
pps-enabled machines, running both 2.4 and 2.6 kernels:  I had enabled the 
crypto feature in order to have ntp track TAI offset and leapseconds.

The crypto library, openssl, looks for a .rnd file in the user's home 
directory on startup, but my startup scripts start daemons in the root of 
the file system and sanitize the environment, leaving openssl to look for 
.rnd at the root of the file system.  When .rnd is not available, ntp 
exits and logs a complaint.

Of course, most of my clocks run on machines with solid state scan disk 
drives, and I've turned off logging on those machines to minimize writing 
and to hide complaints from ntp.  

This issue is discussed in the NTP documentation, of course.


Anyway, good job, guys, it does look good right now.