Martin Paljak wrote:
> On Mar 1, 2010, at 10:59 , Viktor TARASOV wrote:
>   
>> Hi,
>>
>> do somebody uses the card profiles with the 'pin-domains' activated?
>> Will you have any objections against the idea to abandon the support of 
>> the 'pin-domains' in pkcs15init ?
>>
>> Actually only one card driver cflex (and cyberflex), one of the first 
>> drivers, implements it .
>> Afaik, for a long time these cards are not more produced.
>> It's not going about the total elimination of the 'flex' card support, 
>> but about the support with one 'pin-domain',
>> like the others card drivers do.
>>
>> Imho, the support of 'pin-domains' is an non-justified burden for the 
>> pkcs15init core .
>>
>> Alternative to 'pin-domains' is the multi-pkcs15-applications, which the 
>> implementation can be considered.
>>     
> PKCS#15 does not talk about PIN domains, so the implementation should be 
> implemented in the card specific driver. Do you have any pointers do docs?
>   

I have no special docs, but, every card that supports the local PINs (I 
guess, practically, all the cards), can be used with the 'pin-domains'. 
It can be, but nobody (excepted 'flex') use it.


> At the same time, the way I see it, it would break (multiple, all?) PINs on 
> Cryptoflex, which is a good working driver?
>   

All my respects to this driver, the ancestor of many current drivers. It 
took me a certain time to get know how it works.
Yes, it will break multiple PINs for 'flex' cards -- for this card only 
two local PINs per DF are possibles.
But, does the multiple PINs functionality is really asked for (for this 
card)?


> I have a cryptoflex, so I can help testing. Probably the only source for PIN 
> domains and the way it is implemented would be cryptoflex doc?
> From the source I see:
> /* Some cards need to keep all their PINs in separate directories.
>          * Create a subdirectory now, and put the pin into
>          * this subdirectory
>          */
>
> At the same time, I was not able to delete cryptoflex with current trunk, it 
> seems that it tries to cache as a PIN the transport keys and crashes. Will 
> investigate.
>   

Sorry, it's currently broken for this card.
I've noticed that I have tested it with the unprotected profile.
That's why these questions.


Kind wishes,
Viktor.



-- 
Viktor Tarasov  <[email protected]>

_______________________________________________
opensc-devel mailing list
[email protected]
http://www.opensc-project.org/mailman/listinfo/opensc-devel

Reply via email to