May 2009 - freeswitch-dev@lists.freeswitch.org

[Freeswitch-dev] FreeSWITCH on Windows crashing when doing deflect/REFER
by Peter Olsson 12 Oct '09

12 Oct '09

I have a problem with FreeSWITCH (r13175) when using REFER. If running inside Visual Studio, everytime (during the REFER) I get a "...has triggered a breakpoint". It's possible to click continue to go on, but eventually (another 2-3 calls) FreeSWITCH crasches completely. When running outside Visual Studio I get a crasch after maybe 10 calls. Another interesting scenario is when running outside the debugger and I enable siptracing it's much more stable - even though it crasches eventually - this is what makes me wonder about mutexes... I've tracked it down to calls to malloc() and calloc() which in 100% of the cases is in the top of the call stack. My thought is that it might be related to a mutex issue somewhere. Below I've copied the call stack from two times it's happened to me. These are the most common traces, but sometimes it's a malloc being called from the sqlite lib. But so far I always think it's been related to a malloc/calloc. These things have occured in a couple of other places as well, for instance when a SIP registration fails (registering to an external server) with error 404 I also get a breakpoint triggered, but I don't think that has caused any crasches yet. I've tried to debug this on my own, but so far I didn't find anything, so that's why I'm asking you guys :) I think I'm going to try the same on Linux as well, and see if it works differently there. If you need more info, please let me know, I'll try to help out as much as possible. I thought I'd post this here first, if you have any ideas. If not I'll file a Jira for the issue. ----- ntdll.dll!DbgBreakPoint() [Frames below may be incorrect and/or missing, no symbols loaded for ntdll.dll] ntdll.dll!RtlpNtMakeTemporaryKey() + 0x3eee bytes ntdll.dll!RtlTimeToElapsedTimeFields() + 0xd8c3 bytes ntdll.dll!RtlTryEnterCriticalSection() + 0x66f bytes ntdll.dll!RtlpNtMakeTemporaryKey() + 0x4785 bytes ntdll.dll!RtlTimeToElapsedTimeFields() + 0xd6f1 bytes ntdll.dll!RtlTryEnterCriticalSection() + 0x66f bytes > msvcr90d.dll!_heap_alloc_base(unsigned int size=41) Line 105 + 0x28 bytes C msvcr90d.dll!_heap_alloc_dbg_impl(unsigned int nSize=5, int nBlockUse=1, const char * szFileName=0x00000000, int nLine=0, int * errno_tmp=0x0739f498) Line 427 + 0x9 bytes C++ msvcr90d.dll!_nh_malloc_dbg_impl(unsigned int nSize=5, int nhFlag=0, int nBlockUse=1, const char * szFileName=0x00000000, int nLine=0, int * errno_tmp=0x0739f498) Line 239 + 0x19 bytes C++ msvcr90d.dll!_nh_malloc_dbg(unsigned int nSize=5, int nhFlag=0, int nBlockUse=1, const char * szFileName=0x00000000, int nLine=0) Line 296 + 0x1d bytes C++ msvcr90d.dll!malloc(unsigned int nSize=5) Line 56 + 0x15 bytes C++ FreeSwitch.dll!my_dup(const char * s=0x035724c8) Line 95 + 0xc bytes C FreeSwitch.dll!switch_event_add_header_string(switch_event * event=0x0370fe00, switch_stack_t stack=SWITCH_STACK_BOTTOM, const char * header_name=0x0739f50c, const char * data=0x035724c8) Line 785 + 0x9 bytes C FreeSwitch.dll!switch_channel_event_set_extended_data(switch_channel * channel=0x05d86ab0, switch_event * event=0x0370fe00) Line 1365 C FreeSwitch.dll!switch_channel_event_set_data(switch_channel * channel=0x05d86ab0, switch_event * event=0x0370fe00) Line 1378 C FreeSwitch.dll!switch_core_session_exec(switch_core_session * session=0x05d82538, const switch_application_interface * application_interface=0x0357c258, const char * arg=0x05d2b8b0) Line 1472 C FreeSwitch.dll!switch_core_session_execute_application(switch_core_session * session=0x05d82538, const char * app=0x05d2b8a8, const char * arg=0x05d2b8b0) Line 1392 C FreeSwitch.dll!switch_core_standard_on_execute(switch_core_session * session=0x05d82538) Line 168 + 0x16 bytes C FreeSwitch.dll!switch_core_session_run(switch_core_session * session=0x05d82538) Line 490 + 0x204 bytes C FreeSwitch.dll!switch_core_session_thread(apr_thread_t * thread=0x05ca71f8, void * obj=0x05d82538) Line 1060 C libapr.dll!dummy_worker(void * opaque=0x05ca71f8) Line 80 C msvcr90d.dll!_callthreadstartex() Line 348 + 0xf bytes C msvcr90d.dll!_threadstartex(void * ptd=0x03574318) Line 331 C kernel32.dll!BaseThreadInitThunk() + 0x12 bytes ntdll.dll!RtlInitializeExceptionChain() + 0x63 bytes ntdll.dll!RtlInitializeExceptionChain() + 0x36 bytes ---------------- ntdll.dll!DbgBreakPoint() [Frames below may be incorrect and/or missing, no symbols loaded for ntdll.dll] ntdll.dll!RtlpNtMakeTemporaryKey() + 0x3eee bytes ntdll.dll!RtlTimeToElapsedTimeFields() + 0xd8c3 bytes ntdll.dll!RtlTryEnterCriticalSection() + 0x66f bytes ntdll.dll!RtlpNtMakeTemporaryKey() + 0x4785 bytes ntdll.dll!RtlTimeToElapsedTimeFields() + 0xd6f1 bytes ntdll.dll!RtlTryEnterCriticalSection() + 0x66f bytes > msvcr90d.dll!_heap_alloc_base(unsigned int size=544) Line 105 + 0x28 bytes C msvcr90d.dll!_heap_alloc_dbg_impl(unsigned int nSize=508, int nBlockUse=1, const char * szFileName=0x00000000, int nLine=0, int * errno_tmp=0x042df664) Line 427 + 0x9 bytes C++ msvcr90d.dll!_nh_malloc_dbg_impl(unsigned int nSize=508, int nhFlag=0, int nBlockUse=1, const char * szFileName=0x00000000, int nLine=0, int * errno_tmp=0x042df664) Line 239 + 0x19 bytes C++ msvcr90d.dll!_calloc_dbg_impl(unsigned int nNum=1, unsigned int nSize=508, int nBlockUse=1, const char * szFileName=0x00000000, int nLine=0, int * errno_tmp=0x042df664) Line 593 + 0x20 bytes C++ msvcr90d.dll!_calloc_dbg(unsigned int nNum=1, unsigned int nSize=508, int nBlockUse=1, const char * szFileName=0x00000000, int nLine=0) Line 644 + 0x1d bytes C++ msvcr90d.dll!calloc(unsigned int nNum=1, unsigned int nSize=508) Line 56 + 0x13 bytes C++ mod_sofia.dll!su_home_new(unsigned int size=508) Line 562 + 0xe bytes C mod_sofia.dll!msg_create(const msg_mclass_s * mc=0x036f89e8, int flags=2) Line 61 + 0x12 bytes C mod_sofia.dll!nta_msg_create(nta_agent_s * agent=0x037c5430, int flags=0) Line 3305 + 0x16 bytes C mod_sofia.dll!nua_client_request_template(nua_client_request * cr=0x038c2568) Line 613 + 0x11 bytes C mod_sofia.dll!nua_client_init_request0(nua_client_request * cr=0x038c2568) Line 493 + 0x9 bytes C mod_sofia.dll!nua_client_init_request(nua_client_request * cr=0x038c2568) Line 441 + 0x9 bytes C mod_sofia.dll!nua_client_create(nua_handle_s * nh=0x038607f0, int event=33, const nua_client_methods_t * methods=0x03ca2268, const tagi_t * const tags=0x00000000) Line 199 + 0x9 bytes C mod_sofia.dll!nua_session_usage_shutdown(nua_handle_s * nh=0x038607f0, nua_dialog_state * ds=0x03860828, nua_dialog_usage * du=0x0386adc0) Line 1622 + 0x12 bytes C mod_sofia.dll!nua_dialog_usage_shutdown(su_home_s * owner=0x038607f0, nua_dialog_state * ds=0x03860828, nua_dialog_usage * du=0x0386adc0) Line 610 + 0x19 bytes C mod_sofia.dll!nua_dialog_shutdown(su_home_s * owner=0x038607f0, nua_dialog_state * ds=0x03860828) Line 587 + 0x11 bytes C mod_sofia.dll!nua_stack_destroy_handle(nua_s * nua=0x037ca4a8, nua_handle_s * nh=0x038607f0, const tagi_t * tags=0x03872944) Line 918 + 0x10 bytes C mod_sofia.dll!nua_stack_signal(nua_s * nua=0x037ca4a8, su_msg_s * * msg=0x042df868, nua_ee_data * ee=0x0387292c) Line 641 + 0x11 bytes C mod_sofia.dll!su_base_port_execute_msgs(su_msg_s * queue=0x00000000) Line 280 + 0x2c bytes C mod_sofia.dll!su_base_port_getmsgs(su_port_s * self=0x037cac68) Line 202 + 0x9 bytes C mod_sofia.dll!su_base_port_run(su_port_s * self=0x037cac68) Line 335 + 0x11 bytes C mod_sofia.dll!su_port_run(su_port_s * self=0x037cac68) Line 326 + 0x11 bytes C mod_sofia.dll!su_root_run(su_root_s * self=0x037cf1c0) Line 819 + 0xc bytes C mod_sofia.dll!su_pthread_port_clone_main(void * varg=0x03e3f96c) Line 324 + 0x9 bytes C pthreadVC2.dll!ptw32_threadStart(void * vthreadParms=0x03771cb8) Line 219 + 0x9 bytes C msvcr90d.dll!_callthreadstartex() Line 348 + 0xf bytes C msvcr90d.dll!_threadstartex(void * ptd=0x0378a3e8) Line 331 C kernel32.dll!BaseThreadInitThunk() + 0x12 bytes ntdll.dll!RtlInitializeExceptionChain() + 0x63 bytes ntdll.dll!RtlInitializeExceptionChain() + 0x36 bytes

3 3

[Freeswitch-dev] How to unbridge callers on detection of DTMF * key and send them back to where they were
by echu＠vseinc.com 26 May '09

26 May '09

Hi, I've been developing an application module in C/C++ for our online personals type of application. All are well until I need to bridge two existing sessions together and need to find a way to monitor DTMF key (*) so I can un-bridge them and hopefully put them back to where they were. Call flow is like: Caller 1 sends recorded message to Caller 2 asking for live connection Caller 1 in a loop of listening to wait music and checking Caller 2 response (via database) (is there a better way like putting caller 1 on hold and somehow signal it from Caller 2's thread?) Caller 2 hears message and accepts (update database) *bridge Caller 1 and Caller 2 *monitor DTMF for asterisk key *un-bridge and send them back to where they were so they can continue to exchange messages with other callers I tried switch_ivr_uuid_bridge function like below: Originator side: switch_ivr_uuid_bridge //set up args so it stops on any dtmf while (switch_channel_ready(channel)) { status = switch_ivr_sleep(session, 60000, SWITCH_TRUE, &args); if (status == SWITCH_BREAK && dtmf_key == '*') break; } Originatee side: //set up args so it stops on any dtmf while (switch_channel_ready(channel)) { status = switch_ivr_sleep(session, 60000, SWITCH_TRUE, &args); if (status == SWITCH_BREAK && dtmf_key == '*') break; } The problem is after switch_ivr_uuid_bridge call, the channel state has changed and switch_channel_ready becomes false and I immediately got out of the loop and eventually exit from the thread. How do I make this work? Any suggestions are appreciated. Thanks. Eddy Chu

1 0

[Freeswitch-dev] Pre8 Release on Digg
by Brian West 26 May '09

26 May '09

Dear FreeSWITCHers, Now I'm gonna take a moment here to guilt each and everyone of you into checking out the story about Pre8 on Digg. We have all worked long and hard to get to 1.0.4 and we still have a little bit to go. So everyone out there that asks "What can I do to help the project?", this is your chance to do so. Help us to promote the project, which brings more people to help in supporting the community, the project and you the end user. Also looking for people to help manage jira, test bugs, ask the right questions and line up the bugs so we can knock them out! Please email me if you're interested. Here is the link for you to help out http://digg.com/search?s=FreeSWITCH+Pre8 Brian West brian(a)freeswitch.org -- Meet us at ClueCon! http://www.cluecon.com

1 0

[Freeswitch-dev] STUN and rport on Polycom phones
by Brian West 26 May '09

26 May '09

It has come to my attention that Polycom hasn't had a business case to support rport and STUN. If you can please kindly email Marek.Dutkiewicz(a)polycom.com and let him know you would like to see STUN and rport support in the polycom products. Its really one of the last things missing in the phone to make it easy to deploy in a NAT env. Thanks, Brian West brian(a)freeswitch.org -- Meet us at ClueCon! http://www.cluecon.com

1 0

[Freeswitch-dev] Sounds order
by Brian West 26 May '09

26 May '09

I'm getting ready for the next sound file order for FreeSWITCH. I have a rather large set of files to be recorded for the zRTP integration if anyone wants to help out. ;) Please contact me off list. I would like everyone to update and try out voicemail and nitpick anything that you feel is wrong there too and let me know so I can have them corrected in this order also. Thanks, Brian West brian(a)freeswitch.org -- Meet us at ClueCon! http://www.cluecon.com

1 0

Re: [Freeswitch-dev] double re-register problem (machine avaible)
by Diego Toro 22 May '09

22 May '09

Brian thank you for your help. I am testing again the integration between FS and Siemens pbx, I see now that setting expire-seconds on xml gateway to a value minor that 120 FS doen't lost the register (if this value is more than 120 the register is lost), but the calls are dropped after few seconds. Diego --- On Fri, 5/22/09, Brian West <brian(a)freeswitch.org> wrote: From: Brian West <brian(a)freeswitch.org> Subject: Re: [Freeswitch-dev] double re-register problem (machine avaible) To: freeswitch-dev(a)lists.freeswitch.org Date: Friday, May 22, 2009, 9:14 AM Are you 100% sure you have updated since I fixed the issue? You'll have to download and compile it again. /b On May 22, 2009, at 8:09 AM, Diego Toro wrote: Greetings I have the same problem reported in http://jira.freeswitch.org/browse/SFSIP-143. I am running FS (rev 13424) on Windows and I trying register FS on Siemens HighPath 3000. I have avaible on public IP the Siemens 3000 to debug, to reproduce and may be fix issue SFSIP-143 still open. I am begining with FreeSWITCH and i don't have skills yet to find the problem. More information, send me a email Thank you Diego _____________ Brian West brian(a)freeswitch.org -- Meet us at ClueCon! http://www.cluecon.com -----Inline Attachment Follows----- _______________________________________________ Freeswitch-dev mailing list Freeswitch-dev(a)lists.freeswitch.org http://lists.freeswitch.org/mailman/listinfo/freeswitch-dev UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-dev http://www.freeswitch.org

1 0

[Freeswitch-dev] double re-register problem (machine avaible)
by Diego Toro 22 May '09

22 May '09

Greetings I have the same problem reported in http://jira.freeswitch.org/browse/SFSIP-143. I am running FS (rev 13424) on Windows and I trying register FS on Siemens HighPath 3000. I have avaible on public IP the Siemens 3000 to debug, to reproduce and may be fix issue SFSIP-143 still open. I am begining with FreeSWITCH and i don't have skills yet to find the problem. More information, send me a email Thank you Diego

2 1

[Freeswitch-dev] IMPORTANT: Latest FreeSWITCH SVN Updates - Bootstrap Required
by Michael Collins 22 May '09

22 May '09

FYI, We just want to let everyone know that we have made a few updates that will require a rebootstrap. One of the key updates was a security fix for libsndfile. In this particular case it won't be possible simply to "make current" like you normally do. Here is a common set of commands for a typical Linux rebootstrap: cd /usr/src/freeswitch.trunk make clean svn up ./bootstrap.sh ./configure make install NOTE: if you've got the libzrtp file and you've already run the buildzrtp.sh script then be sure to use "./configure --enable-zrtp" in the above operation. Thank you for your continued support of the FreeSWITCH project! -Michael S Collins http://www.cluecon.com

1 0

[Freeswitch-dev] ignore_early_media and mod_conference two different cases
by Szymon Olko 21 May '09

21 May '09

Hi, I made scenario in which I originate call and put it to conference. I use events to check if call was answered or not. I do it all with "ignore_early_media" set to false, so I can record/hear what remote leg is saying, this way call center operators can help to teach my system to recognize it automatically for next calls. I wrote that having one member in conference that I put this call in. I use that command. ORIGINATE {camel_command_id=4}sofia/gateway/sip.freeconet.pl/664558242 &conference(digitalvoice_call#0@talk) '' '' ULTIMO 429955 Call is not answered, it is rejected on phone. There is one member in conference. I get following events scenario (I omit here not interesting events for me): CHANNEL_CREATE CHANNEL_OUTGOING CHANNEL_ORIGINATE CHANNEL_STATE CHANNEL_STATE CHANNEL_STATE CHANNEL_PROGRESS_MEDIA CHANNEL_STATE CHANNEL_EXECUTE CUSTOM_conference::maintenance CHANNEL_PROGRESS CUSTOM_conference::maintenance CHANNEL_EXECUTE_COMPLETE CHANNEL_HANGUP CHANNEL_DESTROY If I do all the same but conference is not existing (there is no member in that conference), I get following events in order: CHANNEL_CREATE CHANNEL_OUTGOING CHANNEL_ORIGINATE CHANNEL_STATE CHANNEL_STATE CHANNEL_STATE CHANNEL_PROGRESS_MEDIA CHANNEL_STATE CHANNEL_EXECUTE CUSTOM_conference::maintenance CHANNEL_ANSWER CHANNEL_PROGRESS CUSTOM_conference::maintenance CHANNEL_EXECUTE_COMPLETE CHANNEL_HANGUP CHANNEL_DESTROY Call was not answer so there should be no CHANNEL_ANSWER event but it is there. In other cases when I answer call I get events order: - when there is member in conference CHANNEL_PROGRESS before CHANNEL_ANSWER event, time between them is the time of phone ringing. - where there is no member in conference CHANNEL_PROGRESS event 4 seconds after CHANNEL_ANSWER event, I cannot use this what I made for empty conferences. Maybe I don't understand early_media, but I would like it to work like in NOT EMPTY conference case. Is there a way to make some 'virtual/local' member that will be in conference just to not be empty? Regards Szymon

1 0

[Freeswitch-dev] New ZDNet Article About FreeSWITCH
by Michael Collins 19 May '09

19 May '09

Gang, Please visit the main FreeSWITCH site and check out the links<http://www.freeswitch.org/node/185>to the new ZDNet story. Please spread the word with all of the link sharing sites that you have. Thanks! Michael S Collins http://www.cluecon.com

1 0