[linux-dvb] Rationalisation of /dev/adapterX/caY devices

Manu Abraham abraham.manu at gmail.com
Tue Apr 11 15:22:23 CEST 2006


Johannes Stezenbach wrote:
> On Tue, Apr 11, 2006, Manu Abraham wrote:
>   
>> Johannes Stezenbach wrote:
>>     
>>> On Tue, Apr 11, 2006, Manu Abraham wrote:
>>>       
>>>> Johannes Stezenbach wrote:
>>>>    
>>>>         
>>>>> Basically this slot thing models the hardware, which is just one
>>>>> CI controller, el cheapo extended to support two slots.
>>>>>      
>>>>>           
>>>> On cards with multiple controllers/one controller with multiple devices, 
>>>> except for the case where it is daisy chained (some budget cards use 
>>>> this el cheapo method), when we have multiple slots, we can route the TS 
>>>> between slots (specific hardware). This is in fact multiple devices on 
>>>> one board/chip
>>>>
>>>> In such a case it would be nice if we can address the slots directly (we 
>>>> will be able to do many interesting things, like a router), since the 
>>>> slots are not directly linked (daisy chained). I mean this will enable 
>>>> us to leave options open.
>>>>         
>>> I think you can have multiple caX devices already if you want
>>> to, just for existing hardware I know of it doesn't make that
>>> much sense.
>>>       
>> The existing ones are fine, since (the only ones that are there now does 
>> chain the slots, no special controller (CPLD) also in some cases) But 
>> cards will be coming out with multiple frontends and ca devices with 
>> routable logic on one PCI adapter (a CPLD which does the stream control).
>>
>> In this case, we can route the TS from one frontend to the another CA 
>> device etc, functionality depends on the complexity of the CPLD/FPGA.
>>     
>
> So what you need is an API to route the TS between
> frontend, CI and demux.
>   

Yeah, we should wait on this till we get the hardware into our hands, 
but leave an option open now, such that when we get the hardware we 
shouldn't break our heads ..

> If you have two CI slots which can be connected independently
> then it seems natural to represent them using two caX devices.
>
>   

Yeah, you are right. This should be the right behaviour.


Manu




More information about the linux-dvb mailing list