Hi Viktor,
Sorry for the late answer for "type" and "ef_structure" question. In
that case, the type is SC_FILE_TYPE_WORKING_EF and the ef_structure is
SC_FILE_EF_TRANSPARENT.
On 05/05/2012 07:29 PM, Viktor Tarasov wrote:
> Le 05/05/2012 07:14, Nguyễn Hồng Quân a écrit :
>> Thanks Viktor,
>>
>> I found the defect at the function pgp_get_blob() in card-openpgp.c.
>> There are lines:
>>
>> if (child->id == id) {
>> (void) pgp_read_blob(card, child);
>> *ret = child;
>> return SC_SUCCESS;
>>
>> The problem is either:
>> 1. The child blob does not exist, but there still exists its ID.
>> 2. The result of pgp_read_blob(card, child) is not checked.
>>
>> This function is called by file_select and because it returns SUCCESS,
>> it makes file_select SUCCESS although the blob does not exist.
>>
>> I think fixing 1 is better. What do you think? (Or the ID is
>> pre-defined?)
>>
>> I'm new (and this driver was not written by me), so I'm grateful to
>> receive your guidance.
> First of all,
> I do not know openpgp card and do not have this card to make the tests.
>
> Afaiu, the 'child' and it's 'ID' are the openpgp specific features that do
> not have any relation to the 7816 standards.
> They has to be hidden from the common OpenSC library part by the openpgp
> card's driver.
>
> The authors of openpgp driver could explain better,
> by from my point of view,
> if the blob cannot be read, pgp_select() has to return 'file-not-found' or
> other error.
> In any case with such openpgp internal errors there is no possibility to
> return a valid FCP/FCI and valid file length.
>
> Looking onto the code I suppose that the FCP returned by
> pgp_select(public-key) belongs in fact to MF (or intermediate DF).
> That's for, in my previous mail, I asked you
> what are the 'type' and 'ef-structure' of the sc_file data returned by
> 'successful' pgp_select() ?
>
--
Regards,
Quân
_______________________________________________
opensc-devel mailing list
[email protected]
http://www.opensc-project.org/mailman/listinfo/opensc-devel
