So, we discussed the background of this on IRC for quite a bit - with
the multisocket patch, there is a disconnect between ce->proto
("reflects the global 'proto' setting in the config") and the
ce->local_list->array[*]->proto values ("one for each 'local' statement").
For servers, it depends a bit on how the server config is set up, but
if you do
--proto udp (default)
--local * 443 tcp
the "tcp" part will not be visible in ce->proto.
If there is *no* "local" statement (just lport + bind), the code will
synthesize local_list->array[0] from the config settings provided - so
whatever happens, for a server config, array[0]->proto contains the
protocol for the first listening socket requested.
Lev has actually tested this, while I've discussed it at length to
understand the subtle differences - written up for the benefit of
the archives.
Again, mail-archive.org refuses to acknowledge that this mail was ever
sent, so referencing sf.net + gerrit.
Your patch has been applied to the master branch.
commit 864bb498ef7a15eb256cbe6013350e76ae96c32b
Author: Gianmarco De Gregori
Date: Fri Sep 12 15:10:56 2025 +0200
Fix multi-socket and dco-win interaction
Signed-off-by: Gianmarco De Gregori <[email protected]>
Acked-by: Lev Stipakov <[email protected]>
Message-Id: <[email protected]>
URL: https://sourceforge.net/p/openvpn/mailman/message/59232442/
URL: https://gerrit.openvpn.net/c/openvpn/+/1173
Signed-off-by: Gert Doering <[email protected]>
--
kind regards,
Gert Doering
_______________________________________________
Openvpn-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/openvpn-devel
