Re: I-D ACTION:draft-ietf-secsh-agent-02.txt

[Simon Tatham <anakin%pobox.com@localhost>]

<Internet-Drafts%ietf.org@localhost> wrote:
> http://www.ietf.org/internet-drafts/draft-ietf-secsh-agent-02.txt

... which contains ...

> The "ssh-rsa" private key format has the following specific encoding:
> 
>   string    "ssh-rsa"
>   mpint     e
>   mpint     d
>   mpint     n
>   mpint     u
>   mpint     p
>   mpint     q

Is there a standard definition of `u', `p' and `q' that I don't know
about? I have a feeling I've seen different implementations use p
and q opposite ways round.

It might be good to make it very clear which of p and q is the
larger prime, and whether u is the inverse of q mod p or vice versa.

> These messages can be sent to either direction. However, the agent MUST
> send these messages only as responses to the client's requests.

Wording nit: is this intended to indicate that the agent must send
extension messages only in response to a client's _extension_
message? If so, it isn't clear; this only seems to say that an agent
can't suddenly volunteer an extension message when no message at all
has been sent from the client.

However, I'm generally happy with this; all my suggestions have
happened, and I think I could usefully get started on some
implementation with this draft.

Cheers,
Simon
-- 
Simon Tatham         "Selfless? I'm so selfless I
<anakin%pobox.com@localhost>    don't even know who I am."