Summary:	ASTERISK-22275: [patch] T.38 Passthrough broken if peer doen't report T38MaxBitRate
Reporter:	Andreas Steinmetz (anstein)	Labels:
Date Opened:	2013-08-10 08:22:41	Date Closed:	2013-08-21 10:04:01
Priority:	Major	Regression?	No
Status:	Closed/Complete	Components:	Channels/chan_sip/T.38
Versions:	11.5.0	Frequency of Occurrence	Constant
Related Issues:
Environment:	Linux	Attachments:	( 0) fax-fix.patch
Description:	Patton Smartnode <=> Asterisk <=> Cisco SPA112 Incoming Fax from Patton Smartnode. Smartnode doesn't include T38MaxBitRate in reinvite and Asterisk thus uses a default of 2400 in reinvite to SPA112. According to http://lists.digium.com/pipermail/asterisk-users/2013-July/279964.html this is probably wrong and 14400 should be used as the default. Can easily be fixed by rearranging the "enum ast_control_t38_rate" values.
Comments:	By: Andreas Steinmetz (anstein) 2013-08-10 08:25:05.279-0500 Patch against 11.5.0 By: Richard Mudgett (rmudgett) 2013-08-13 10:55:31.598-0500 The attached patch file is not marked as a contribution.  Patches need a contributor's license to be accepted. By: Rusty Newton (rnewton) 2013-08-14 17:56:48.301-0500 I deleted the patch. Please read and follow [the instructions linked|https://wiki.asterisk.org/wiki/display/AST/Asterisk+Issue+Guidelines#AsteriskIssueGuidelines-DigiumSubmissionLicenseAgreement] Basically sign the contributor's license agreement and when approved re-upload the patch. Thanks! By: Andreas Steinmetz (anstein) 2013-08-15 11:19:54.483-0500 Patch against 11.5.0 By: Rusty Newton (rnewton) 2013-08-18 12:47:58.317-0500 Thanks Andreas! By: Walter Doekes (wdoekes) 2013-08-19 03:11:36.024-0500 I'd say ship-it. But I'd like a little comment why the value is 0. So those viewing the code won't have to guess. Something like this should suffice: {noformat} /* Set to 0 so it's taken as default when unspecified. * See ITU-T T.38 Implementors' Guide (11 May 2012), * Table H.2: if the T38MaxBitRate attribute is omitted * it should use a default of 14400. */ {noformat} By: Matt Jordan (mjordan) 2013-08-21 09:16:07.986-0500 Ugh. I do hate re-ordering enum values like that, but given the way SDP parsing occurs in chan_sip, it's definitely the easiest approach. I concur on the comment.