[linux-dvb] Re: refactoring

[Andrew de Quincey <adq_dvb@lidskialf.net>]

On Wednesday 06 Oct 2004 20:56, Gerd Knorr wrote:
> Andrew de Quincey <adq_dvb@lidskialf.net> writes:
> > [snip]
> >
> > > the ioctl wrapper is only implemented
> > > once for all drivers (the demod code could communicate with the
> > > frontend infrastructure via a thin function-pointer-in-struct
> > > interface).
> >
> > Interesting idea this one - this is the sort of thing I'm trying to sort
> > out with the refactoring.
>
> I don't think it is useful.  If I look at the cx22702 fe code for
> example I don't see how trying to replace cx22702_ioctl() with
> something generic + function pointer hooks would improve things.  Most
> of the code is just mapping dvb api enums+defines to hardware specific
> values and visa versa, that isn't something which can be shared with
> other frontend drivers ...

But it allows the other frontends which do need card specific functions to be 
implemented in a much more flexible manner. The cx22702 on the hardware we 
have doesn't need that admittedly, but virtually all the DVB-S hardware needs 
horrible hardware specific tweaks that need to override some of the DISEQC 
ioctl functions (which exactly varies from card to card). 

From other people's comments STBs are even worse for this - the dbox2 even has 
to override FE_SET_FRONTEND function to set the frequency in some odd manner.