On Mon, Nov 27, 2006 at 07:46:17PM +0200, Alon Bar-Lev wrote:
> On 11/27/06, Joe Orton <[EMAIL PROTECTED]> wrote:
> >Hi Alon,
> >
> >I'm definitely interested in seeing the neon SSL API extended to be able
> >to handle hardware tokens.  The intent is that new interfaces could be
> >added to support this stuff, certainly.
> 
> I am glad!
> BTW: Something wrong with the list, I cannot subscribe.

The box running mailman currently has a large backlog of mail it's 
chewing through - it should catch up eventually :(

> >How would this work at OpenSSL level?  (i.e. what is the interface to
> >OpenSSL to do this; a callback is registered?)
> 
> pkcs11-helper can extract the X509 certificate and RSA of private key.
> It registers the RSA callbacks and perform private key operations.

OK, thanks for explaining that.

> And if you wish to be nice to users, you can enumerate existing
> certificates, I don't know how much you wish the API to be flexible...
> In order to display the user a set of available certificates you need
> to enumerate objects:
> 
> list = pkcs11h_certificate_enumCertificateIds ()

It sounds like it would be useful to expose that too.

> Well... The callbacks should be common to sessions... And I think you
> should add such callback to PKCS#12 as well, when user should be
> prompted for passphrase.

Do you mean common to *all* sessions; does the library require that to 
be so?  This means process-global state!

> Also there should be a way to specify somekind of configuration, for
> example which providers to load, provider specific parameters etc...

This is where supporting hardware tokens generally gets really awkward - 
providers will be process-global state, is that correct?

Regards,

joe
_______________________________________________
neon mailing list
[email protected]
http://mailman.webdav.org/mailman/listinfo/neon

Reply via email to