Re: [sr-dev] t_suspend and t_continue in tm module

Hi I've finally got working functionality for t-suspend and t_continue for SIP responses in the tmp/tm_async_reply_support branch. I've memory and load tested the branch and have also isolated the changes so that if there are problems they will only be experienced if you use the new API functions. However keeping mind that TM can be tricky to program (as mentioned in the documentation ;)) - I've included below a summary of the changes and a git diff against the latest master. So the TM experts can have a look and see if there are any obvious problems. If there are no complaints or queries I will commit the changes to master next week Friday (July 12). Regards Richard. Change summary: 1. parser/msg_parser.h: added new message state FL_RPL_SUSPENDED 2. tm_load.c: new api functions: *t_suspend_reply*, *t_continue_reply* and *t_cancel_suspend_reply* 3. t_suspend.c/h: new methods: *t_suspend_reply*, *t_continue_reply* and *t_cancel_suspend_reply *4. tm/t_reply.c: if msg is in state FL_RPL_SUSPENDED then we skip sending the reply on 5. tm/t_reply.c/h: new methods to fake env for responses: *faked_env_resp, faked_resp, free_faked_resp * 6. API documentation to include new functions: *t_suspend_reply*, *t_continue_reply* and *t_cancel_suspend_reply * * * * * On 18 March 2013 14:13, Richard Good &lt;richard.good(a)smilecoms.com <mailto:richard.good@smilecoms.com>> wrote: Hi Daniel and other TM experts We have been working on a prototype for being able to use the t_suspend() and t_continue() methods for SIP responses. The version in branch /tmp/tm_async_reply_support is working, though still needs to be 100% tested and only works for 1 branch (i.e. no forking) Would it be possible for you to have a look at this branch code and see if we are on the right track. The code is working but TM (as mentioned in the documentation ;)) can be tricky to program and we want to make sure we are not breaking anything, as the goal is to get this into the master branch. The changes we have done: 1. parser/msg_parser.h: added new message state FL_RPL_SUSPENDED 2. tm/h_table.h: added int suspended_request to ua_server and int suspended_reply to ua_client 3. tm/t_suspend.c:t_suspend 1. Changed suspend method to check if the message is a request or response and set the suspended_request/suspended_reply variable accordingly 2. If its a request we do as the existing code 3. If its a response we get the correct uac branch, clone the msg into t->uac[branch].reply and set the msg state to FL_RPL_SUSPENDED 4. tm/t_reply.c: if msg is in state FL_RPL_SUSPENDED then we skip sending the reply on 5. tm/t_suspend.c:t_continue 1. Check if this uas was suspended - if so this is a continue for a request so continue with existing code 2. If not then we search for the branch that was suspended 3. Once branch is found unsuspend the message state 4. do pre-script, run_top_route and post scripts 5. Relay the reply using basically the same code we skipped in step 4 First next step after we confirm we are on the right track is to add forking support - this will require a new t_continue method where the calling application passes in the branch of the suspended reply when wanting to continue. Any comments/feedback would be appreciated. Regards Richard. On 11 March 2013 14:34, Daniel-Constantin Mierla &lt;miconda(a)gmail.com <mailto:miconda@gmail.com>> wrote: Hello, t_suspend() and t_continue() are only for requests at this moment. Cheers, Daniel On 3/6/13 2:49 PM, Richard Good wrote: -- Daniel-Constantin Mierla -http://www.asipto.com http://twitter.com/#!/miconda <http://twitter.com/#%21/miconda> -http://www.linkedin.com/in/miconda Kamailio World Conference, April 16-17, 2013, Berlin -http://conference.kamailio.com - _______________________________________________ sr-dev mailing list sr-dev(a)lists.sip-router.org <mailto:sr-dev@lists.sip-router.org> http://lists.sip-router.org/cgi-bin/mailman/listinfo/sr-dev This email is subject to the disclaimer of Smile Communications at http://www.smilecoms.com/disclaimer