[asterisk-bugs] [Asterisk 0016677]: deadlock in app_queue with use_weight during reload
Asterisk Bug Tracker
noreply at bugs.digium.com
Fri Jan 22 17:47:51 CST 2010
The following issue has been SUBMITTED.
======================================================================
https://issues.asterisk.org/view.php?id=16677
======================================================================
Reported By: tim_ringenbach
Assigned To:
======================================================================
Project: Asterisk
Issue ID: 16677
Category: Applications/app_queue
Reproducibility: random
Severity: minor
Priority: normal
Status: new
Asterisk Version: 1.4.29
JIRA:
Regression: No
Reviewboard Link:
SVN Branch (only for SVN checkouts, not tarball releases): N/A
SVN Revision (number only!):
Request Review:
======================================================================
Date Submitted: 2010-01-22 17:47 CST
Last Modified: 2010-01-22 17:47 CST
======================================================================
Summary: deadlock in app_queue with use_weight during reload
Description:
There's a deadlock possible in app_queue if the use_weight global variable
is set, and someone reloads app_queue at the right point in time while a
call is coming in. The call's thread locks the queue list lock, but doesn't
unlock it, resulting in a dead lock when other threads try to do anything
with the queue list.
To reproduce easily, add a sleep() between the first if (use_weight) and
grabbing the lock.
I'll attach a patch that fixes the problem by using a local variable to
store a copy of use_weight.
======================================================================
Issue History
Date Modified Username Field Change
======================================================================
2010-01-22 17:47 tim_ringenbach New Issue
2010-01-22 17:47 tim_ringenbach Asterisk Version => 1.4.29
2010-01-22 17:47 tim_ringenbach Regression => No
2010-01-22 17:47 tim_ringenbach SVN Branch (only for SVN checkouts, not tarball
releases) => N/A
======================================================================
More information about the asterisk-bugs
mailing list