[vdr] Re: final solution to the signalless operation problem found!! -> testers+RFC required

[Guido Fiala <gfiala@s.netic.de>]

Am Wednesday 24 March 2004 21:14 schrieb Andreas Oberritter:
> On Wed, 2004-03-24 at 19:46, Guido Fiala wrote:
> > Am Wednesday 24 March 2004 15:17 schrieb Andreas Oberritter:
> > > The only remaining problem is that the VES is in an undefined state
> > > after insmod. So how about this patch instead?
> >
> > You _should_ be right - unfortunately my test showed, that the ARM
> > crashes instantly if using your patch instead of calling ioctl(FE_SLEEP)
> > in vdr just after opening the frontend (this way my ARM run now 24hours
> > without any outcommand error or reset or crash - it's rock-solid!).
>
> Yes, of course. My patch addressed the undefined state before opening
> the device. VDR must not open the device if it should stay powered down.
> open() calls FE_INIT and therefore wakes up the VES1893.

Yes - and i would have thought your patch should work if vdr would not open 
the device (commented out the open) - yet it did not work ! It was unexpected 
to me, i really wished it worked that way for simplicity.

thanks for your help anyway, 
Guido


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