> On April 6, 2015, 4:16 p.m., Joshua Colp wrote: > > /certified/branches/13.1/res/res_pjsip_t38.c, lines 493-498 > > <https://reviewboard.asterisk.org/r/4577/diff/3/?file=73632#file73632line493> > > > > How could this occur? > > Jonathan Rose wrote: > As far as I know, it should never occur. I'm probably being paranoid.
You are. The only way that could happen is if the threading model for the PJSIP architecture was ignored. ^_^ - Joshua ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviewboard.asterisk.org/r/4577/#review15075 ----------------------------------------------------------- On April 6, 2015, 4:14 p.m., Jonathan Rose wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviewboard.asterisk.org/r/4577/ > ----------------------------------------------------------- > > (Updated April 6, 2015, 4:14 p.m.) > > > Review request for Asterisk Developers and Joshua Colp. > > > Bugs: ASTERISK-24933 > https://issues.asterisk.org/jira/browse/ASTERISK-24933 > > > Repository: Asterisk > > > Description > ------- > > Description: > Two Asterisk boxes each have the other as endpoints with authentication set. > First Asterisk box originates a call to the second using the PJSIP endpoint. > The first Asterisk box uses an extension with sendfax, the second uses > an extension with receivefax. > > The session starts fairly normally, but resolution never appears in fax show > session output. After a while (~25 seconds) the call drops and the fax fails. > Error messages shown are as follows: > Sender: The call dropped prematurely > Receiver: Disconnected after permitted retries > > Note that when not using authentication, the FAX will complete as expected. > When using chan_sip as the sender to a receiver of chan_pjsip, the FAX will > also complete as expected with authentication, but if chan_pjsip is the sender > it will fail regardless of whether the recipient is chan_sip or chan_pjsip. > > The problem is caused by duplication of a framehook in res_pjsip_t38 which > occurs on the second invite sent out when responding to the auth challenge. > > Fix: > In order to fix this, I added a simple flag to the pjsip session struct that > would > be raised when the framehook is first attached to prevent duplication. I > wouldn't > be surprised if there were a better way to do this. > > > Diffs > ----- > > /certified/branches/13.1/res/res_pjsip_t38.c 433316 > > Diff: https://reviewboard.asterisk.org/r/4577/diff/ > > > Testing > ------- > > I've duplicated and modified the t38 PJSIP fax test in the testsuite to > include authentication. It fails without the patch and passes with the patch. > I also tested this locally with my two Asterisk machines in the above > scenario. I'll be linking the test after putting it on Gerrit. > > > Thanks, > > Jonathan Rose > >
-- _____________________________________________________________________ -- Bandwidth and Colocation Provided by http://www.api-digital.com -- asterisk-dev mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-dev
