[asterisk-bugs] [JIRA] (ASTERISK-26100) [patch] Asterisk will reject, with a 488, subsequent T.38 reinvite to T.38 negotiation.

Parantido Julius De Rica (JIRA) noreply at issues.asterisk.org
Thu Jun 9 05:22:56 CDT 2016


Parantido Julius De Rica created ASTERISK-26100:
---------------------------------------------------

             Summary: [patch] Asterisk will reject, with a 488, subsequent T.38 reinvite to T.38 negotiation.
                 Key: ASTERISK-26100
                 URL: https://issues.asterisk.org/jira/browse/ASTERISK-26100
             Project: Asterisk
          Issue Type: Bug
      Security Level: None
          Components: Channels/chan_sip/T.38
    Affects Versions: 11.22.0
         Environment: Ubuntu 14.04 LTS (asterisk is always compiled from scratch)
            Reporter: Parantido Julius De Rica
            Severity: Blocker


Hi all,

I already searched bug tracker and found similar solved bug, but no one fixed my issue.

I've te following scenario:

FAX (Offerer) <--> SIP Server <--> Asterisk T.38 Passthrough <--> Oracle SBC (ex Acme Packet) <--> PSTN <--> Fax (Receiver)

Offerer starts a Fax transmission as in-audio call (g729 coded). This call get switched by network side (Oracle SBC) as T.38 with a first invite. During fax negotiation Oracle SBC send back a second T.38 re-invite and Asterisk answer with a 488 after 5 seconds due the following code statement:

p->t38id = ast_sched_add(sched, 5000, sip_t38_abort, dialog_ref(p, "passing dialog ptr into sched structure based on t38id for sip_t38_abort."));

Reading code seems that a reinvite is not allowed, during fax transmission, after channel is placed in AST_STATE_UP.

Reading all RFCs and ITU-T specifications no one describe this scenario but  at the same times no one disallow it.

I wrote a patch for chan_sip.c to handle this behaviour. All seems to works fine for all cutomers that reports to me this issue and no regressions are observed.

I attached a PCAP trace (for better scenario description) and a patch.

Regards



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



More information about the asterisk-bugs mailing list