[vdr] Re: Seperation of sat clock code

[NoSpam-@zocki.toppoint.de (Rainer Zocholl)]

Sergei.Haller@math.uni-giessen.de(Sergei Haller)  17.04.02 11:00

Once upon a time Sergei Haller shaped the electrons to say...

>On Wed, 17 Apr 2002, bernd_muc@gmx.de () wrote:


>The reason I never sent it is the following: the ntp clients use the
>Network Time Protocol (RFCs 958, 1059, 1119, 1305), which is _MUCH_
>more complicated than the Time Protocol (RFC 868). So I don't know if
>it is possible to tell an ntp client to use the Time Service at Port
>37. (but probably we could just ask the authors)

Maybe speak to the XNTP-guys to implement a "satelite clock device"
to the (X)NTP-Package?

AFAIK it's not sooooo complicate to add a new "clock device".
(No sorry, i have no time to do it ;-))
Adding the full NTP-server into VDR would be possible but of cause overkill.

But there is still the problem:
How to share the hardware resources?

Maybe it would be sufficient to "offer" the time on a named pipe or at 
port 37 so the special "ntp-vdr-clock-device" can read it and offer 
it via XNTP at port 123?

The ntp-protokoll should allow (AFAIK) to mark the time delivered as 
"not reliable yet" if the trusted channel is not selected.



>And again, this would only be available while one of the dvb cards is
>tuned to the trusted transponder.

>another advantage would be that the time client could run with root
>rights.