Summary:	ASTERISK-09748: Libss7 and Siemens EWSD (ITU style),SS7 link crashes.
Reporter:	Zurab Rukhadze (zurab)	Labels:
Date Opened:	2007-06-25 05:44:51	Date Closed:	2007-06-29 08:58:11
Priority:	Critical	Regression?	No
Status:	Closed/Complete	Components:	Core/General
Versions:		Frequency of Occurrence
Related Issues:
Environment:		Attachments:	( 0) libss7-ewsd-dbg.txt ( 1) libss7-ewsd-trace.log
Description:	The existence of such ISDN user part parameters as:Call diversion info, Backward call indicators, etc. in the received ISDN user part  Call progress (CPG) mesaages, causes Libss7 crashes.
Comments:	By: Matthew Fredrickson (mattf) 2007-06-25 17:21:48 Ok, checking into this. By: Matthew Fredrickson (mattf) 2007-06-25 17:32:40 Hrm... can you provide me with login access to see this?  I checked your trace and I'm not sure what might be going wrong without further investigation. By: Matthew Fredrickson (mattf) 2007-06-25 17:39:30 Oh, does it still happen if you have ss7 debug disabled on asterisk? By: Zurab Rukhadze (zurab) 2007-06-26 08:45:35 libss7-trunk rev.91 test with Siemens EWSD. "libss7 crash" problem was solved. However, as soon as "call forward no response" appears, the connection fails. After that, Libss7 ignors all calls (See attached file libbs7-ewsd-dbg.txt). The above-mantioned situation does not depend on the ss7 debug status. By: Matthew Fredrickson (mattf) 2007-06-26 09:13:51 This looks like it might be related to a bug that I have seen only in certain scenarios.  Something is overwriting the network indicator on the linkset, or it's changing for some reason.  Just to check, if you can revert back to the version that made it crash and get a core dump (start asterisk with the -g option) and give me a backtrace of the crash, that might help track it down. By: Matthew Fredrickson (mattf) 2007-06-26 09:16:13 Or if you can reproduce this using valgrind (valgrind asterisk) that might give the information I need. By: Mike Anikienko (mike anikienko) 2007-06-26 10:26:18 I have too problem with CPG with Huawei M800. This problem only where CPG present. When call come with CPG (client set hold on his phone), after this call I see ERROR[16262] chan_zap.c: [1] Received message destined for point code 0x1234 but we're 0x3f9f.  Dropping But 0x3f9f - is not my point code, my point code is 0x1234 ! 0x3f9f - random time to time. After this SS7 link is down. And no more calls served. I need FULL restart asterisk. unload/load chan_zap - not solve. With this problem I live above 1 month. Before I not see this problem. Sorry, but now i have not more information. By: Matthew Fredrickson (mattf) 2007-06-26 10:32:03 Thanks for the report.  So you're sure, it's only when we receive CPG that this happens?  If you can try to reproduce this with valgrind, that might help to give more information about what is happening. By: Mike Anikienko (mike anikienko) 2007-06-26 14:58:48 Yes, I see problem only where incoming call have CPG. Tomorrow, I try valgrind By: Matthew Fredrickson (mattf) 2007-06-27 17:23:55 Do you mean it's when we receive a CPG message, or when we transmit a CPG message? By: Mike Anikienko (mike anikienko) 2007-06-28 00:37:37 when asterisk receive a CPG message (when incoming call have CPG - when clent set his phone to hold) [1] Len = 15 [ 8e 95 0e c5 18 37 c5 4d 04 00 2c 02 01 2c 01 ] [1] [ 8e 95 0e ] [1] FSN: 21 FIB 1 [1] BSN: 14 BIB 1 [1] <[0] MSU [1]  [ c5 ] [1]  Network Indicator: 3 Priority: 0 User Part: ISUP [1]  [ 18 37 c5 4d ] [1]  OPC 14100 DPC 14104 SLS 4 [1]          [ 04 00 ] [1]          CIC: 4 [1]          [ 2c ] [1]          Message Type: CPG [1] [1] Unhandled optional parameter 0x2c 'Generic Notification Indication' [1] [ [1] 0xf9 [1] ] DEBUG[16262] chan_zap.c: Queuing frame PROGRESS on CIC 4 WARNING[16262] file.c: Unexpected control subclass '14' By: Matthew Fredrickson (mattf) 2007-06-28 09:59:57 If I give you a code change to try, can you reproduce it quickly enough to test it for me?  Also, is it every time you receive a CPG, or just sometimes? By: Matthew Fredrickson (mattf) 2007-06-28 10:11:40 Also, if you can reproduce it, have you been able to do so with valgrind running? By: Matthew Fredrickson (mattf) 2007-06-28 11:20:45 I think I just found the fix!!!! I made a mistake in which ss7 I was passing to the zap_queue_frame routine.  Update asterisk to the latest and I think it will work now. By: Mike Anikienko (mike anikienko) 2007-06-29 01:10:12 Thanks, problem with imcomming call with CPG solved.