RE: rsa-sha2-256/512: handling of incorrect signature encoding

Peter Gutmann <[email protected]>
Newsgroups gmane.ietf.secsh
Message-ID <9A043F3CF02CD34C8E74AC1594475C73F4CDF9D6@uxcn10-5.UoA.auckland.ac.nz>
denis bider (Bitvise) <[email protected]> writes:

>What type of “advertising SHA-2” do you mean? If the server advertises rsa-
>sha2-256 or -512 for server authentication, then it also needs to have an RSA
>host key. But the server might not have an RSA host key, it might only have
>an ECDSA or Curve25519 host key. 

I was thinking of it in an opportunistic-upgrade sense, for example for PGP
and S/MIME the mandatory algorithm is SHA-1 but if you receive a message
signed with SHA-2 you can switch to that because the client will be able to
process it.  So if you see ecdsa-sha2... or rsa-sha2... in the keyex then you
know the other side can do SHA2, and should do the auth with SHA2 as well.

>In this situation, the server cannot advertise an rsa-sha2-XXX algorithm for
>host authentication. 

It's not so much can it do RSA-SHA2, but can it do SHA2 in general rather than
SHA1.

Peter.
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.