[vdr] Re: Two cards, two recordings, bad signal on one = two broken recordings

[Rene Bartsch <ml@bartschnet.de>]

Am Dienstag, 14. Oktober 2003 22:47 schrieben Sie:
> Hi Rene,
>
> > > So I'd say everybody should change this in whatever way they
> > > consider best fitting.
> >
> > Yes, but bad reception shouldn't crash a system.
> >
> > What about just checking two values?
> >
> > 1.) FE_HAS_LOCK
> > 2.) BER
> >
> > If there's no FE_HAS_LOCK it might be a crash, but if there's a BER You
> > receive something and the driver hasn't crashed.
>
> The simple answer to that one is not every frontend delivers all values.
> Frontends which return a *real* BER are hard to find. I never understood
> why Convergence didn't update all the frontends, so that they deliver
> all four values (BER, SNR, SS, UNC) of interest
>

Because they do programming by hacking. Most stability problems we have in VDR 
are caused by ugly workarounds for driver problems.

My server used to crash for half a year with strange segfaults and IDE-errors. 
I screwed it into parts twice a week and changed hardware for 1000,- EUR, but 
nothing improved. I even lost all file-systems three times and my BIOS got 
corrupted several times (--> only pulling power-cord and removing 
CMOS-battery helped to get it working correctly).

After switching to DVB-driver version 1.0.1 (in which the DMA-bug has been 
fixed) it works like a charm for three days, now.

Rene



-- 
Info:
To unsubscribe send a mail to ecartis@linuxtv.org with "unsubscribe vdr" as subject.