[linux-dvb] Re: PCMCIA docs and Nova-CI

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

On Thursday 22 January 2004 22:19, Ralph Metzler wrote:
> Holger Waechtler writes:
>  > Holger Waechtler wrote:
>  > > Andrew de Quincey wrote:
>  > >> Hi, Robert and I are getting on really well with the Nova-CI... we
>  > >> can talk to the attribute memory of a CAM module now.
>  > >>
>  > >> We're wondering if anyone has any information on PCMCIA devices;
>  > >> we're gradually working out the tuple-based attribute memory, and
>  > >> what things like the Configuration Option register do... but if
>  > >> someone could point us at some firmer docs than the plain linux
>  > >> kernel sources, it would be really handy.
>  > >
>  > > what exactly do you want to know? As soon you can read/write attribute
>  > > and i/o memory and check/set the status (module inserted or not,
>  > > enable/disable TS bypass/decryption) you are done and can hook in any
>  > > CI library, e.g. the one that comes as part of VDR or the one that has
>  > > been derived from this by the MetzlerBros...
>  >
>  > oops - sorry, this is not correct, you need for instance then implement
>  > the link layer protocol below the CI library - sorry...
>
> You also need to implement a CA device. We seperated it in its own
> file which can be used by any link layer implementation. This way, you
> do not need to duplicate the code. We use it in our driver for both
> full featured and KNC DVB cards.

Basically the Nova-CI gives you access to the raw hardware interfaces on the 
PCMCIA CAM module. You get access to attribute memory, and the 4 byte command 
interface  as defined in CENELEC EN50221 (which appears to be mapped to 
PCMCIA I/O space). Everything above that is up to me to write.

I was assuming from ca.h that the link layer protocol would be in the driver, 
and user apps would use the CA_SEND_MSG and CA_GET_MSG IOCTLS to submit the 
higher layer messages... is that correct? 

Glad to hear theres a library for the stuff above the link layer! Less work 
for me. I've not had the time to read up on the actual protocol layers 
properly yet; we've just been getting the hardware working up till now.

A problem we're having is that the IR interface at DEBI address 0x1234 is slap 
in the middle of the DEBI memory decoded by the CAM interface.


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