[asterisk-bugs] [JIRA] (ASTERISK-26616) T38 calls not being hung up correctly

[Morten Tryfoss (JIRA)]

    [ https://issues.asterisk.org/jira/browse/ASTERISK-26616?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=233854#comment-233854 ] 

Morten Tryfoss commented on ASTERISK-26616:
-------------------------------------------

After some more debugging, I think this might solve the issue (at least for me in this case):

In function chan_sip.c/interpret_t38_parameters:
transmit_reinvite_with_sdp(p, FALSE, FALSE);
was removed for AST_T38_TERMINATED and AST_T38_REQUEST_TERMINATE.

Then the BYE seems to be sent through the whole chain of channels. 
But there might be cases where T38 should be terminated and go on with audio?

> T38 calls not being hung up correctly
> -------------------------------------
>
>                 Key: ASTERISK-26616
>                 URL: https://issues.asterisk.org/jira/browse/ASTERISK-26616
>             Project: Asterisk
>          Issue Type: Bug
>      Security Level: None
>          Components: Channels/chan_sip/General
>    Affects Versions: 13.8.0
>            Reporter: Morten Tryfoss
>         Attachments: debuglog.txt
>
>
> We've got this scenario:
> There's a bridged call which negotiates T38. 
> After fax is sent the caller sends BYE. 
> Then Asterisk tries a reINVITE to the callee (back to normal audio). 
> This is rejected by a 488 not accepted here.
> No BYE is sent to the callee.
> The call stays up for several hours until dropped by the callee.
> I found this in the debug log at BYE:
> [Nov 21 16:35:14] DEBUG[15411][C-001906b6] chan_sip.c: Hangup call SIP/sipic2-00320075 (This is the callee-channel), SIP callid 754576b53f450b526e26caf705dbbf57 at 158.58.152.21:5060
> [Nov 21 16:35:14] DEBUG[18031][C-001906b6] chan_sip.c: (Provisional) Stopping retransmission (but retaining packet) on '754576b53f450b526e26caf705dbbf57 at 158.58.152.21:5060' Request 104: Found
> It seems like Asterisk is trying to CANCEL the call instead of BYE? Invitestate is incorrect  or channelstate is not up after the failed reINVITE, which in turn sets needcancel to TRUE?
> It seems to be easy to reproduce, but it takes some time for me to identify the long running calls since they're not visible in Asterisk anymore.
> The issue could also be related to ASTERISK-26609.



--
This message was sent by Atlassian JIRA
(v6.2#6252)