Thanks, and sorry for stalling since January. This is still not
"really really" pretty, but I am still convinced that dealing with DCO
related issues inside dco*.c is a better approach :-)
I have not actually excercised the code (I do not have experienced the
underlying issue yet, which smells like a race condition between kernel
and userland), just stared at it for a bit, and relied on the BB army
to verify that it compiles and passes t_client on all DCO-enabled
platforms (it does).
Not sure why we would ever hit dco_get_peer_stats() with peer_id == -1
(in that case userland should know "the peer is gone") but the check
won't harm - though slightly unrelated. Good catch on the uint32_t.
Your patch has been applied to the master branch.
commit 6f9ba8bfd259742ee19b173898a9bfd20e22fcf3
Author: Ralf Lici
Date: Wed Mar 5 18:17:30 2025 +0100
Handle missing DCO peer by restarting the session
Signed-off-by: Ralf Lici <[email protected]>
Acked-by: Frank Lichtenheld <[email protected]>
Message-Id: <[email protected]>
URL:
https://www.mail-archive.com/[email protected]/msg31022.html
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
