On Fri, Nov 07, 2014 at 10:06:21AM -0500, Wietse Venema wrote:
> If real people have a need for SNI, what right do we have to tell
> them to fuck off because they live in an imperfect world?
The server-side SNI support in OpenSSL is currently a mess, it
muddles along, but enabling SNI leads to somewhat incorrect TLS
extension handling in some cases. I'd like to avoid this for 2.12
(or 3.0 if that's the name for the next release) and re-examine
SNI support next year, once the OpenSSL server-side SNI implementation
has been cleaned-up a bit.
In Postfix side, we'd probably need a key + chain database to
support SNI, I don't think it is wise to expose keypairs to
post-chroot privilege-reduced SMTP servers, or to have tlsmgr(8)
proxy access to keypairs for such servers.
The database driver in question has to support large objects, in
the form of private key + certificate + chain. Users would "compile"
their PEM chain files to add them to the database, signally which
domains and ".example.com" parent domains the chain should apply
to.
See also
http://www.metzdowd.com/pipermail/cryptography/2008-September/014752.html
for questions on the IDNA interaction of SNI, which I still don't
know the answer to.
--
Viktor.
