On 09/02/2011 04:32 PM, Antti Palosaari wrote:
> On 09/02/2011 02:04 PM, Bjørn Mork wrote:
>> Antti Palosaari<crope@iki.fi> writes:
>>
>>> Since Anysee device itself does not have CCID interface it is needed
>>> to make virtual USB device in order to get CCID support. I have never
>>> seen virtual USB devices like that, but there is VHCI in current
>>> kernel staging that actually does something like that over IP.
>>
>> Don't know if you have seen this already, but there's a virtual CCID
>> device implementation in QEMU. See
>> http://wiki.qemu.org/Features/Smartcard
>> Should be a good starting point. Combine it withe the VHCI driver from
>> USBIP and you have your CCID device.
>
> It is first time I hear about QEMU virtual CCID. Now we have all parts
> needed for USBIP VHCI and QEMU virtual CCID, just glue those together.
>
> I wonder if it is wise to even create virtual CCID "core" to Kernel.
> There is few other readers that can use that too, actually I think all
> USB readers that have unique USB ID (blocking out those which uses
> USB-serial converters with common IDs).
>
> As I see that CCID still more complex as serial device I will still look
> implementing it as serial as now.

Here it is, patch attached. Implemented as serial device. Anysee uses 
two different smart card interfaces, CST56I01 and TDA8024. That one is 
old CST56I01, I will try to add TDA8024 later, maybe even tonight.

Anyhow, it is something like proof-of-concept currently, missing locks 
and abusing ttyUSB. Have you any idea if I should reserve own major 
device numbers for Anysee or should I reserve one like DVB common?

Any other ideas?

Antti
-- 
http://palosaari.fi/