Horton

Re: Direct SSL connection with ALPN and HBA rules

Jacob Champion <jacob.champion@enterprisedb.com>

View thread View on postgr.es
From: Jacob Champion <jacob.champion@enterprisedb.com>
To: Heikki Linnakangas <hlinnaka@iki.fi>
Cc: Robert Haas <robertmhaas@gmail.com>, Daniel Gustafsson <daniel@yesql.se>, Michael Paquier <michael@paquier.xyz>, Postgres hackers <pgsql-hackers@lists.postgresql.org>
Date: 2024-05-15T18:24:00Z
Lists: pgsql-hackers 
On Wed, May 15, 2024 at 6:33 AM Heikki Linnakangas <hlinnaka@iki.fi> wrote:
> Ok, yeah, I can see that now. Here's a new version to address that. I
> merged ENC_SSL_NEGOTIATED_SSL and ENC_SSL_DIRECT_SSL to a single method,
> ENC_SSL. The places that need to distinguish between them now check
> conn-sslnegotiation. That seems more clear now that there is no fallback.

That change and the new comment that were added seem a lot clearer to
me, too; +1. And I like that this potentially preps for
encryption=gss/ssl/none or similar.

This assertion seems a little strange to me:

>                   if (conn->sslnegotiation[0] == 'p')
>                   {
>                       ProtocolVersion pv;
>
>                       Assert(conn->sslnegotiation[0] == 'p');

But other than that nitpick, nothing else jumps out at me at the moment.

Thanks,
--Jacob

Commits

  1. Remove option to fall back from direct to postgres SSL negotiation

  2. Reject SSL connection if ALPN is used but there's no common protocol

  3. libpq: Enforce ALPN in direct SSL connections

  4. libpq: If ALPN is not used, make PQsslAttribute(conn, "alpn") == ""

  5. Fix documentation and comments on what happens after GSS rejection

  6. doc: Add note to prevent server spoofing with SCRAM