Re: Serverside SNI support in libpq

Hi!

On Mon, Nov 24, 2025 at 6:53 AM Daniel Gustafsson <daniel(at)yesql(dot)se> wrote:
> The attached incorporates your tests, fixes them to make them pass. The
> culprit seemed to be a combination of a bug in the code (the verify callback
> need to be defined in the default context even if there is no CA for it to be
> called in an SNI setting because OpenSSL), and that the tests were matching
> backend errors against frontend messages.

The new v12 tests still don't pass for me (they all use "certificate
verify failed", but the failure modes should be different).

> + if (host->ssl_ca && host->ssl_ca[0] != '\0')

The comment for HostsLine.ssl_ca, and the code that assigns it,
implies to me that host->ssl_ca should never be NULL. Am I missing a
case where it could be?

On Wed, Dec 3, 2025 at 1:57 AM Heikki Linnakangas <hlinnaka(at)iki(dot)fi> wrote:
> I propose that there is no GUC. In 'pg_hosts.conf', you can specify a
> wildcard '*' host that matches anything. You can also specify a "no sni"
> line which matches connections with no SNI specified. (Or something
> along those lines, I didn't think too hard about all the interactions).

That seems to position SNI as a feature that every DBA should have to
think about by default. ("learn this file. you can't turn it off.") Is
it, yet?

Web servers enable SNI implicitly because name-based hosting is a
top-level concept for users over there (hostnames are baked into the
application layer). I would argue that we don't have that here. Maybe
in the future someone will ask for that, but at that point don't you
want a very different, name-based, config system?

On Wed, Dec 3, 2025 at 3:28 PM Daniel Gustafsson <daniel(at)yesql(dot)se> wrote:
> > On 3 Dec 2025, at 22:27, Jelte Fennema-Nio <postgres(at)jeltef(dot)nl> wrote:
> > What if we make it so that if a pg_hosts.conf file exists, then the
> > ssl_cert_file/ssl_key_file configs are ignored? And by default initdb
> > would not create a file (or it would, but with the same default
> > settings that we have now).
>
> Maybe. I'm not a big fan of magic-file-exist configurations

Me neither. (I especially don't like the idea of ignoring a
certificate+key setting that a user has taken the time to put into a
config.)

Thanks,
--Jacob