Re: [TLS] TLS 1.2 draft (issue #25 about SSLv2 Hello)

<Pasi.Eronen@nokia.com> writes:

>> Cleaned up backward compatibility text [issue 25]
>
> This text (in E.2) still needs some cleaning. Currently
> it says:
>
>    challenge_length
>        The length in bytes of the client's challenge to the server to
>        authenticate itself. Historically, permissible values are between
>        16 and 32 bytes inclusive. When using the SSLv2 backward
>        compatible handshake the client MUST use a 32-byte challenge.
> [...]
>    challenge
>        Corresponds to ClientHello.random. If the challenge length is
>        less than 32, the TLS server will pad the data with leading
>        (note: not trailing) zero bytes to make it 32 bytes long.
>
> Which is not exactly consistent: if the challenge length MUST be 32
> bytes, it can't be less than 32 bytes.
>
> It was also noted that for historical and/or compatibility reasons,
> many existing browsers (e.g. Opera, IE6, FireFox 1.5) use 16-byte
> challenges:
>
> http://www1.ietf.org/mail-archive/web/tls/current/msg00985.html
>
> My proposal (in the mail linked above) was to to change this to "The
> length of the challenge field in bytes; MUST be between 16 and 32
> (inclusive)."
>
> Comments? (At the very least, the text needs to be internally
> consistent. Preferably it should also promote real-world
> interoperability and reflect what implementations are expected to do.)

So, here's what I was trying to say:

1. As a TLS 1.2 impl yuou MUST use a 32-byte challenge.
2. Downrev impls may use a 16-32 byte challenge at which point
   you would pad with zeros but in < 1.2 mode.

That may not be the right thing to say. :)

-Ekr

_______________________________________________
TLS mailing list
TLS@lists.ietf.org
https://www1.ietf.org/mailman/listinfo/tls