Summary:ASTERISK-05798: random crash in pri_fixup_principle
Reporter:Keiron Liddle (keiron)Labels:
Date Opened:2005-12-07 16:26:07.000-0600Date Closed:2006-01-12 00:35:24.000-0600
Versions:Frequency of
Environment:Attachments:( 0) output.txt
( 1) output2.txt
Description:We have observed asterisk to crash on the line on two occasions:
7712                                    if (pri->pvts[principle]->owner) {

It would appear to be due to some other thread changing something and maybe having a reasonable number of active calls (30 - 40).

Comments:By: Clod Patry (junky) 2005-12-07 19:53:49.000-0600

could u add that to your backtrace please?

"bt full" and "thread apply all bt"

By: Mark Spencer (markster) 2005-12-20 02:29:06.000-0600

You may need to build without optimizations in order to get a useful backtrace since the stack in your second file is clearly not valid...  Can you please rebuild without optimizations and get us another backtrace?  Thanks!

By: Mark Spencer (markster) 2005-12-26 16:18:50.000-0600

Closing due to inactivity from the bug placer.  If you continue to have the problem please feel free to reopen and place the new backtrace as requested.  Thanks!

By: Keiron Liddle (keiron) 2006-01-12 00:13:05.000-0600

I found out what the problem is.
I was getting a principle of -1 in the pri_fixup_principle which was due to a call of that method on the last line of the case of PRI_EVENT_RING in the pri_dchannel method where the else statement did not properly have brackets around the code.

Anyway this appears to be fixed/changed in the current SVN code.

By: Jason Parker (jparker) 2006-01-12 00:35:15.000-0600

Appears to have been fixed in svn trunk.