pinpad entry support in Git repository
gniibe at fsij.org
Thu Jan 12 02:12:02 CET 2012
On 2012-01-11 at 11:01 +0200, Martin Paljak wrote:
> Did some testing with three readers that were not mentioned, which I
> had available. Attached a small "report".
Thanks a lot.
> Reader 1: ACS non-CCID reader ACR83, with the vnedor-provided modified
> CCID driver. Did not work at all.
> Reader 2: Gemalto Ezio Shield (variant): PIN commands worked as
> expected (with pinmax up to 32, I did not type 32 digits though),
> plaintext PIN commands were disallowed with 6d00
> Reader 3: Omnikey 3821: worked as expected with pinpad.
> Also a small patch against pinpadtest.py as I have several readers I
> can't disconnect.
I added your patch to the script.
> It might make sense to make a "probing script" that would discover
> deficiencies in reader firmwares (like require certain message bits
> (some of them are fixed in the CCID driver) or require fixed PIN
> lengths etc)
Pinpad entry for vaiable length PIN is a corner area. If there is a
way to identify card readers with PC/SC API, we could have white list
or black list, but I don't know portable way to distinguish readers.
It would be good if we can do probing the feature with no user
intervention, but I don't have idea for that.
More information about the Gnupg-devel