Fwd: Re: [pkcs11-comment] Character set for PIN

[Valerie Fenwick <valerie.fenwick@oracle.com>]

-------- Forwarded Message --------
Subject: Re: [pkcs11-comment] Character set for PIN
Date: Tue, 04 Oct 2016 19:40:29 +0100
From: David Woodhouse <dwmw2@infradead.org>
To: Valerie Fenwick <valerie.fenwick@oracle.com>, 
pkcs11-comment@lists.oasis-open.org

On Tue, 2016-10-04 at 11:26 -0700, Valerie Fenwick wrote:
> > We should make it clearer. Can we add wording to the spec which says
> > that that functions such as C_SetPIN() and C_InitPIN() MUST return
> > CKR_ARGUMENTS_BAD if the input is not valid UTF-8. Perhaps also
> > C_Login() too?
> >
> > Or failing that, can we at *least* add such to the Usage Guide, noting
> > that applications must convert user input from their local character
> > set to UTF-8 for the purpose of calling C_Login() and similar
> > functions.
>
> As you've noted, there may have been lax programmers out there, and we certainly
> generally do not want to break compatibility with older implementations - but, a
> note in the usage guide may make sense. I will
> bring this up to our new usage guide editor.

Thanks. Just to make it explicit, although it was implied in the above:
Note that disallowing invalid UTF-8 byte sequences in C_SetPIN() and
C_InitPIN() functions is a whole lot nicer in terms of backwards
compatibility, than refusing invalid UTF-8 in C_Login().

The former would simply prevent setting a new non-ASCII PIN in broken
software (which has always been broken), and the user would have to
stick with ASCII until their software is broken.

The latter would mean that existing hardware suddenly stops working,
which is less nice.

--
dwmw2