[Home]

Summary:ASTERISK-20184: Asterisk crash in Dial
Reporter:David Cunningham (dcunningham)Labels:
Date Opened:2012-07-30 18:20:36Date Closed:2012-08-30 12:49:06
Priority:MajorRegression?
Status:Closed/CompleteComponents:Applications/app_dial
Versions:1.8.11.1 Frequency of
Occurrence		Occasional
Related
Issues:
Environment:CentOS release 5.6 (Final) Attachments:( 0) backtrace.txt
( 1) full.477.gz
Description:Asterisk crashed when performing a Dial(). Backtrace is attached, and the Asterisk log from the time. Could you please advise what you need? Thanks.
Comments:By: David Cunningham (dcunningham) 2012-07-30 18:21:01.465-0500

Backtrace.
By: David Cunningham (dcunningham) 2012-07-30 18:21:21.812-0500

Asterisk log attached.
By: Richard Mudgett (rmudgett) 2012-07-30 19:05:17.759-0500

This is likely a duplicate of ASTERISK-19923 memory corruption from voicemail.
By: David Cunningham (dcunningham) 2012-07-30 19:08:13.873-0500

I forgot to mention - the server is running the patch from issue 19923.

By: David Cunningham (dcunningham) 2012-07-30 19:11:53.479-0500

It's also running the patch recommending in this issue:
ASTERISK-20013

By: Richard Mudgett (rmudgett) 2012-07-30 21:40:21.959-0500

The patch that was committed for ASTERISK-18404 or the original patch for that issue?
By: David Cunningham (dcunningham) 2012-07-30 22:02:04.408-0500

I don't see it as the same patch as ASTERISK-18404. Here's the actual patch applied:

{noformat}
Index: include/asterisk/time.h
===================================================================
--- include/asterisk/time.h (revision 363307)
+++ include/asterisk/time.h (working copy)
@@ -83,8 +83,9 @@
  is handled for positive and negative numbers, by ensuring
  that the divisor is always positive
*/
- return  ((end.tv_sec - start.tv_sec) * 1000) +
- (((1000000 + end.tv_usec - start.tv_usec) / 1000) - 1000);
+ int64_t sec_dif = (int64_t)(end.tv_sec - start.tv_sec) * 1000;
+ int64_t usec_dif = (1000000 + end.tv_usec - start.tv_usec) / 1000 - 1000;
+ return  sec_dif + usec_dif;
}
)
{noformat}

By: Rusty Newton (rnewton) 2012-08-02 17:17:29.190-0500

Are you able to reproduce it? Can you get it again but with a full log that also contains DEBUG level 5  ?  Also, a "bt full" on the backtrace.


By: David Cunningham (dcunningham) 2012-08-02 18:00:55.283-0500

I will enable debug and wait for another crash. We don't know what causes it yet, so it might take a while.

By: Leif Madsen (lmadsen) 2012-08-29 10:47:46.737-0500

Have you been able to reproduce this, or should we close as "Unable to reproduce" for now?
By: David Cunningham (dcunningham) 2012-08-29 17:56:29.977-0500

Let me check with the customer as we were waiting to hear from them. Thanks.
By: David Cunningham (dcunningham) 2012-08-30 00:21:09.623-0500

The customer advises it hasn't reoccurred, so I guess we can't reproduce this and it can be closed. Thanks.