Re: [Curdle] draft-ietf-curdle-ssh-kex-sha2 and diffie-hellman-group1-sha1 (1024-bit DH)

[Damien Miller <[e-mail]>]

On Thu, 20 Jul 2017, Peter Gutmann wrote:

> Damien Miller <[e-mail]> writes:
> 
> >Opinion: there's still enough old junk out there that optional support for
> >diffie-hellman-group1-sha1 is probably necessary for a while longer. IMO this
> >is probably worth an explicit note in the draft.
> 
> That's the catch-22 with something like this, on the one hand you want to make
> it MUST NOT so you've got something to hit vendors over the head with when
> they keep using RSA512 with MD5, on the other hand if they've got existing
> equipment they're just going to demand that you change your code to work with
> it.
> 
> As an aside, it would be nice if OpenSSH had some single enable-legacy-mode
> switch to re-enable all the MUST algorithms that are currently disabled,
> having to send people to https://www.openssh.com/legacy.html so they can
> figure out by trial and error which incantation they need to use each time
> they can't interop with current versions of OpenSSH is a pain.  Or at least
> change the error message to tell people what to do to make it work.

When I wrote legacy.html I tried to SEO it so it was the first search
result for the error message in question (it seems to have worked -- for
now).

The error message does include the set of algorithms that the user would
need to add, but I don't want to include cookbook instructions in the
error message because users will too-often just do what the cookbook says
without understanding the tradeoff they are making.

-d