Re: [Curdle] Confirming a change to draft-ietf-curdle-rsa-sha2-12

[denis bider <[e-mail]>]

Well, to me, "without padding" means "don't add ornamentation to S". It
does not mean "trim leading zeros from S", since the spec says we're using
RSASSA-PKCS1-v1_5, and the output of that is S with a fixed-length
encoding. :)

But yes:

> Given that this has been ambiguous for untold years,
> it might be a good idea to mention that you shouldn't
> expect to find this fixed-length zero-padded encoding
> but may get a form, uh "without lengths or padding".

That is what we're doing. :) The reason for this discussion is that I
requested a sentence added (to draft-ietf-curdle-rsa-sha2) to clarify that
both S and leading-zero-trimmed-S are acceptable.

denis



On Tue, Mar 13, 2018 at 4:18 AM, Peter Gutmann <[e-mail]>
wrote:

> denis bider <[e-mail]> writes:
>
> >The problem with this phrasing is that it doesn't say:
> >
> >encode "S" as specified in RSASSA-PKCS1-v1_5
> >
> >It instead says:
> >
> >[...]
> >
> >- "a string containing s (which is an integer, without lengths or padding,
> >unsigned, and in network byte order)"
>
> By saying "without lengths or padding" it seems to be saying that you
> shouldn't zero-pad.  Or at least that's how I interpreted it.
>
> >Because of this ambiguity, the spec for "rsa-sha2-256" and "rsa-sha2-512"
> -
> >which is the subject of present discussion - comes out and says
> explicitly:
> >though shalt encode "S". Which is fixed-length.
>
> Given that this has been ambiguous for untold years, it might be a good
> idea
> to mention that you shouldn't expect to find this fixed-length zero-padded
> encoding but may get a form, uh "without lengths or padding".  I don't care
> either way since my code will handle both forms, but there are
> I-don't-know-
> how-many different SSH implementations out there where we can't easily tell
> what they do.
>
> >It has been like that since the first version if the draft, in November
> 2015,
> >which you reviewed. :)
>
> Well, yeah, but that text says different things to different people :-).
>
> Peter.
>