I caught a bug which seems to occur when a nonblocking connect() is
interrupted by a connection to the management port which causes the
standalone management event loop to be called via openvpn_sleep().
When polling the management port for connection status at an interval
of less than a second the management loop never exited and hung
forever. This patch fixes it. It didn't seem to cause any other
issues. Does this patch look ok?
>From 8f19f1088be171336840110cdd4365e18ea59d7c Mon Sep 17 00:00:00 2001
From: coderrr <coderrr.cont...@gmail.com>
List-Post: openvpn-devel@lists.sourceforge.net
Date: Tue, 17 Apr 2012 22:12:24 +0700
Subject: [PATCH] don't stay in management event loop forever
---
src/openvpn/manage.c | 3 ++-
1 files changed, 2 insertions(+), 1 deletions(-)
diff --git a/src/openvpn/manage.c b/src/openvpn/manage.c
index 1dddd41..574b111 100644
--- a/src/openvpn/manage.c
+++ b/src/openvpn/manage.c
@@ -2871,7 +2871,8 @@ management_event_loop_n_seconds (struct
management *man, int sec)
man_check_for_signals (&signal_received);
if (signal_received)
return;
- } while (expire);
+ update_time();
+ } while (now >= expire);
/* revert state */
man->persist.standalone_disabled = standalone_disabled_save;
--
1.7.4.1
- coderrr
https://www.privateinternetaccess.com
