GPF Crypto Stick vs OpenPGP Card

Hauke Laging mailinglisten at hauke-laging.de
Fri Dec 3 13:21:20 CET 2010


Am Freitag 03 Dezember 2010 09:47:27 schrieb Nils Faerber:

> The non-obvious content of the transaction, what you say as "you do not
> see what you sign even on the PIN-pad" is an issue that has been
> discussed a lot of times already - yes, it is definitely an issue but
> very hard to solve. IMHO this would require a card terminal that
> understands the data to be signed and present the user with a meaningful
> summary.
> But it strictly assumes again that this terminal cannot be compromised
> too. And being more intelligent (in order to display complex data) means
> to be a more complex device containing more complex device software
> which again opens new possible security holes.

A first improvement would be to show the hash to be signed. Of course, you 
cannot trust the hash calculation on a potentially compromised PC but this 
would be a start for further protection (e.g. by sending the file to someone 
else and comparing the hashes).

If I understand the process correctly then not the file hash is signed but the 
hash for a combination of the file hash and some metadata (timestamp, signer 
ID). For a security progress the card reader would have to see both hash 
components which would require a protocol change. IMHO it makes sense to plan 
this for the future. Ask the card reader whether it has a display and can do 
the hash calculation itself. If so then send the data in a new format.


Hauke
-- 
PGP: D44C 6A5B 71B0 427C CED3 025C BD7D 6D27 ECCB 5814
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 555 bytes
Desc: This is a digitally signed message part.
URL: </pipermail/attachments/20101203/01bf3ec2/attachment.pgp>


More information about the Gnupg-users mailing list