[Home]

Summary:ASTERISK-14736: asterisk sends no RELEASE_COMPLETE when mISDN Call gets disconnected from asterisk side in 1.6.0.13 and likely others
Reporter:Schnulli.Maus (schnullimaus)Labels:
Date Opened:2009-08-30 13:35:05Date Closed:2011-07-27 13:32:27
Priority:MajorRegression?No
Status:Closed/CompleteComponents:Channels/chan_misdn
Versions:Frequency of
Occurrence
Related
Issues:
Environment:Attachments:
Description:Trace from mISDN:

Sun Aug 30 00:36:27 2009: P[ 4] * IND : HANGUP pid:3 ctx:default dad: oad:*14 State:CONNECTED
Sun Aug 30 00:36:27 2009: P[ 4]   --> out_cause:16
Sun Aug 30 00:36:27 2009: P[ 4]   --> state:CONNECTED
Sun Aug 30 00:36:27 2009: P[ 4]  I SEND:DISCONNECT oad:29 dad:14 pid:3
Sun Aug 30 00:36:27 2009: P[ 4]   --> channel:1 mode:TE cause:16 ocause:16 rad: cad:14
Sun Aug 30 00:36:27 2009: P[ 4]   --> info_dad: onumplan:0 dnumplan:0 rnumplan:0 cpnnumplan:4
Sun Aug 30 00:36:27 2009: P[ 4]  I IND :RELEASE oad:29 dad:14 pid:3 state:CLEANING
Sun Aug 30 00:36:27 2009: P[ 4]   --> channel:1 mode:TE cause:16 ocause:16 rad: cad:14
Sun Aug 30 00:36:27 2009: P[ 4]   --> info_dad: onumplan:0 dnumplan:0 rnumplan:0 cpnnumplan:4
Sun Aug 30 00:36:27 2009: P[ 4]   --> No need to queue hangup
Sun Aug 30 00:36:27 2009: P[ 4]  Cannot hangup chan, no ast
Sun Aug 30 00:36:27 2009: P[ 4]  $$$ Cleaning up bc with stid :10010400 pid:3

looks like that all data for that call (the ast) gets voided when sending the DISCONNECT, so the RELEASE cannot be associated to a call and no RELEASE_COMPLETE is sent. According to Q931 Specs there must be a RELEASE_COMPLETE sent.

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

I tried to trace it down (i am not a programmer) and found a function in
chan_misdn.c, where i put a dirty fix:

/************************************************************/
/*  Receive Events from isdn_lib  here                     */
/************************************************************/
static enum event_response_e
cb_events(enum event_e event, struct misdn_bchannel *bc, void *user_data)

(...tons of lines removed...)

case EVENT_RELEASE:
{
bc->need_disconnect = 0;
bc->need_release = 0;

hangup_chan(ch);
/* send RLC on reception of RELEASE */
misdn_lib_send_event(bc, EVENT_RELEASE_COMPLETE);
release_chan(bc);
}
break;

I just inserted a line of code to send a RELEASE_COMPLETE message, after a
short test it seems to do the trick, but i am uncertain wether this is the
best place or the best way to fix it. (At least, i feel one should first
call release_chan and then send the RELEASE_COMPLETE, but i was afraid that
release_chan might void the bc variable.

Now the trace  looks like this:

Sun Aug 30 19:14:45 2009: P[ 4]  I SEND:DISCONNECT oad:29 dad:14 pid:4
Sun Aug 30 19:14:45 2009: P[ 4]   --> channel:1 mode:TE cause:16 ocause:16 rad: cad:14
Sun Aug 30 19:14:45 2009: P[ 4]   --> info_dad: onumplan:0 dnumplan:0 rnumplan:0 cpnnumplan:4
Sun Aug 30 19:14:45 2009: P[ 4]   --> caps:Speech pi:0 keypad: sending_complete:0
Sun Aug 30 19:14:45 2009: P[ 4]   --> Channel: mISDN/7-u3 hanguped new state:CLEANING
Sun Aug 30 19:14:45 2009: P[ 4]  I IND :RELEASE oad:29 dad:14 pid:4 state:CLEANING
Sun Aug 30 19:14:45 2009: P[ 4]   --> channel:1 mode:TE cause:16 ocause:16 rad: cad:14
Sun Aug 30 19:14:45 2009: P[ 4]   --> info_dad: onumplan:0 dnumplan:0 rnumplan:0 cpnnumplan:4
Sun Aug 30 19:14:45 2009: P[ 4]   --> caps:Speech pi:0 keypad: sending_complete:0
Sun Aug 30 19:14:45 2009: P[ 4]  ast_hangup already called, so we have no ast ptr anymore in
event(RELEASE)
Sun Aug 30 19:14:45 2009: P[ 4]   --> No need to queue hangup
Sun Aug 30 19:14:45 2009: P[ 4]  Cannot hangup chan, no ast
Sun Aug 30 19:14:45 2009: P[ 4]  I SEND:RELEASE_COMPLETE oad:29 dad:14 pid:4
Sun Aug 30 19:14:45 2009: P[ 4]   --> channel:1 mode:TE cause:16 ocause:16 rad: cad:14
Sun Aug 30 19:14:45 2009: P[ 4]   --> info_dad: onumplan:0 dnumplan:0 rnumplan:0 cpnnumplan:4
Sun Aug 30 19:14:45 2009: P[ 4]   --> caps:Speech pi:0 keypad: sending_complete:0
Sun Aug 30 19:14:45 2009: P[ 4]  $$$ CLEANUP CALLED pid:4
Sun Aug 30 19:14:45 2009: P[ 4]  $$$ Cleaning up bc with stid :10010400 pid:4
Sun Aug 30 19:14:45 2009: P[ 4]  Sending Control ECHOCAN_OFF



A RELEASE_COMPLETE is now sent as expected. Probably, because of the fix,
an unneeded RELEASE_COMPLETE is sent here and there which might cause
new problems.

I found this when troubleshooting an issue with asterisk connected to my
Ackermann Euracom 182 PBX. Sometimes the PBX goes into an undefined state
where the caller hears a busy tone and the called persons voice. I was
able to reproduce it when the SIP phone which is used to call into
the PSTN via asterisk and Euracom just was disconnected from the
network, then i started tracing. Have to observe some time if the issue
is gone, but i assume that the missing RELEASE_COMPLETE leaves the PBX
in an undefined state. But not sending the RELEASE_COMPLETE violates
the standards and should be fixed.

Best regards,

Stephan    

Comments:By: Schnulli.Maus (schnullimaus) 2009-08-31 01:08:48

same issue found in 1.6.0.14, and same workaround applies.
By: Russell Bryant (russell) 2011-07-27 13:32:22.506-0500

Per the Asterisk maintenance timeline page at http://www.asterisk.org/asterisk-versions maintenance (bug) support for the 1.4 and 1.6.x branches has ended. For continued maintenance support please move to the 1.8 branch which is a long term support (LTS) branch. For more information about branch support, please see https://wiki.asterisk.org/wiki/display/AST/Asterisk+Versions

If this is still an issue, please open a new issue so it can be re-triaged appropriately. Thanks!