[Home]

Summary:ASTERISK-08140: Chanspy application in asterisk 1.4 ver crash the asterisk-segmentation fault (core dumped)
Reporter:mottano (mottano)Labels:
Date Opened:2006-11-16 00:49:15.000-0600Date Closed:2007-01-10 10:24:42.000-0600
Priority:CriticalRegression?No
Status:Closed/CompleteComponents:Core/General
Versions:Frequency of
Occurrence
Related
Issues:
Environment:Attachments:( 0) bigpbx_core.txt
Description:Consuder i have 3 channels like 6002,6004,6006.
I need to monitor the 6006 ,when 6004 and 6006 have conversation thro. chanspy (asterisk 1.4, beta3)
In my  chanspy dial plan is like..
exten => 6008,1,chanspy(SIP/6006 |wbq)

when i dial 6008, i got the following message in console and immed. asterisk terminated ..


------------------------------------------------------------


*CLI> -- Executing [6006@from-sip:1] Dial("SIP/6002-08590248", "SIP/6006|15|tr") in new stack
-- Called 6006
-- SIP/6006-08594188 is ringing
-- SIP/6006-08594188 answered SIP/6002-08590248
-- Executing [6008@from-sip:1] ChanSpy("SIP/6004-08589040", "SIP/6006|wq") in new stack
== Spying on channel SIP/6006-08594188
[Nov 11 15:19:37] NOTICE[8974]: app_chanspy.c:202 start_spying: Attaching SIP/6004-08589040 to SIP/6006-08594188
Segmentation fault (core dumped)
linux:~ #

---------------------------------------------------------------------------

****** ADDITIONAL INFORMATION ******

i need to clarify, wethere the new future in chanspy,i.e,"Whispering mode" have functioning correctly.
Comments:By: Joshua C. Colp (jcolp) 2006-11-16 09:50:22.000-0600

Please try the latest 1.4 branch, I believe there were further improvements. If that also fails then please get a backtrace from the core dump. Instructions can be found in the doc directory, backtrace.txt - Thanks!

By: Jason Parker (jparker) 2006-11-21 12:26:01.000-0600

Were you able to test the latest 1.4 branch?

By: Mitch Sharp (bluecrow76) 2006-11-22 15:03:08.000-0600

Just posted bigpbx core.txt.  Got back to the office today and found the system down.  Investigated the core dump and found the segfault occured when someone used ChanSpy on a SIP channel.  Hope the dump helps.

By: Joshua C. Colp (jcolp) 2006-11-27 15:35:10.000-0600

Just to confirm - that is a relatively recent 1.4 checkout or still the beta?

By: Mitch Sharp (bluecrow76) 2006-11-27 16:16:36.000-0600

This occured on a 1.4-beta3 install.  The bigpbx_core.txt file is from that core dump.

I've also had this happen recently on a 1.2.12.1 system with the following trace:

(gdb) bt
#0  0x00617fa9 in free () from /lib/tls/libc.so.6
#1  0x0805904d in ast_frfree (fr=0x6dbff4) at frame.c:277
#2  0x08061b8e in ast_channel_spy_free (spy=0xb7566910) at channel.c:1115
#3  0x00241693 in channel_spy (chan=0xb7a06200, spyee=0xb7a00ac0, volfactor=0xb7566a74, fd=0) at app_chanspy.c:341
#4  0x0024275d in chanspy_exec (chan=0xb7a06200, data=0xb756b0a0) at app_chanspy.c:508
ASTERISK-1  0x0809141d in pbx_extension_helper (c=0xb7a06200, con=Variable "con" is not available.
) at pbx.c:553
ASTERISK-2  0x080926e6 in __ast_pbx_run (c=0xb7a06200) at pbx.c:2230
ASTERISK-3  0x0809424c in pbx_thread (data=0x6dd804) at pbx.c:2517
ASTERISK-4  0x00962371 in start_thread () from /lib/tls/libpthread.so.0
ASTERISK-5  0x0067d9be in clone () from /lib/tls/libc.so.6
(gdb) bt full
#0  0x00617fa9 in free () from /lib/tls/libc.so.6
No symbol table info available.
#1  0x0805904d in ast_frfree (fr=0x6dbff4) at frame.c:277
No locals.
#2  0x08061b8e in ast_channel_spy_free (spy=0xb7566910) at channel.c:1115
       f = Variable "f" is not available.
(gdb)

By: Serge Vecher (serge-v) 2006-11-29 15:50:41.000-0600

bluecrow76: as suggest by file, please try the latest svn branch checkout, at r > 48000 while waiting for 1.4-beta4. Thanks.

By: Joshua C. Colp (jcolp) 2006-12-18 21:07:32.000-0600

If this is still not fixed please report back and we can go from there, which will probably requiring getting access to a box where this is happening so I can examine the core dump.

By: jlimb (jlimb) 2006-12-22 02:05:47.000-0600

this may or may not be relative:
on my 1.2.12.1 box I run chanspy with the (q) quiet option because if I don't I can crash asterisk every time by hanging up while chanspy is playing the name of the channel I am monitoring.

By: Joshua C. Colp (jcolp) 2006-12-22 10:54:43.000-0600

Can you provide backtraces of these crashes? I attempted to recreate them using both 1.2.12.1 and the latest 1.2 but was unable to on my box.

By: jlimb (jlimb) 2006-12-22 19:43:08.000-0600

Here is my crash on 1.2.12.1

Reading symbols from /lib/libgcc_s.so.1...done.
Loaded symbols for /lib/libgcc_s.so.1
#0  0x00a65fa9 in free () from /lib/tls/libc.so.6
(gdb) bt
#0  0x00a65fa9 in free () from /lib/tls/libc.so.6
#1  0x0805904d in ast_frfree (fr=0xb29ff4) at frame.c:277
#2  0x08061b8e in ast_channel_spy_free (spy=0xb7a27910) at channel.c:1115
#3  0x00ff5693 in channel_spy (chan=0x9f54138, spyee=0x9f56b08,
   volfactor=0xb7a27a74, fd=0) at app_chanspy.c:341
#4  0x00ff675d in chanspy_exec (chan=0x9f54138, data=0xb7a2c0a0)
   at app_chanspy.c:508
ASTERISK-1  0x0809141d in pbx_extension_helper (c=0x9f54138, con=Variable "con" is not available.
) at pbx.c:553
ASTERISK-2  0x080926e6 in __ast_pbx_run (c=0x9f54138) at pbx.c:2230
ASTERISK-3  0x0809424c in pbx_thread (data=0xb2b804) at pbx.c:2517
ASTERISK-4  0x00bca371 in start_thread () from /lib/tls/libpthread.so.0
ASTERISK-5  0x00acb9be in clone () from /lib/tls/libc.so.6
(gdb) bt full
#0  0x00a65fa9 in free () from /lib/tls/libc.so.6
No symbol table info available.
#1  0x0805904d in ast_frfree (fr=0xb29ff4) at frame.c:277
No locals.
#2  0x08061b8e in ast_channel_spy_free (spy=0xb7a27910) at channel.c:1115
       f = Variable "f" is not available.
(gdb)

By: Serge Vecher (serge-v) 2006-12-27 13:43:03.000-0600

what about 1.2.14?

By: jlimb (jlimb) 2007-01-09 23:44:46.000-0600

As for me and my problem, It does not happen on 1.2.14.  But it doesn't sound like my problem is the same as the initial reporter's problem.

By: Serge Vecher (serge-v) 2007-01-10 10:24:30.000-0600

ok, op (mottano) hasn't responded for over a month. If the problem still persists with 1.4.0 release, please reopen the bug with a new backtrace attached.