ACK.
I can't easily reproduce this issue here (it's timing critical and you
need a slow resolver to be able to "hit" that window), but I think the
bug is clear enough - without the patch, the code will destroy "res",
then decide to ignore the signal (= not go to "done") and then fall over
ASSERT(res)... so with the new code, it won't destry "res" for SIGUSR1.
Patch has been applied to the master and release/2.3 branch, as this
is just a bug, not really related to dual-stack support.
commit 282788a835f6c9dfb85e8f9a3bd45f5841271b06 (master)
commit 47a963681a6ed18d32405ec26b74d56925425093 (release/2.3)
Author: Arne Schwabe
List-Post: openvpn-devel@lists.sourceforge.net
Date: Thu Nov 21 21:08:05 2013 +0100
Fix assertion when SIGUSR1 is received while getaddrinfo is successful
Acked-by: Gert Doering <g...@greenie.muc.de>
Message-Id: <1385064495-25877-4-git-send-email-a...@rfc2549.org>
URL: http://article.gmane.org/gmane.network.openvpn.devel/8019
Signed-off-by: Gert Doering <g...@greenie.muc.de>
--
kind regards,
Gert Doering
