Benny Prijono | 1 Jun 2008 17:03
Favicon

Re: Odd Crash

On Sat, May 31, 2008 at 6:43 PM, Norman W. Franke <norman <at> myasd.com> wrote:
> On May 31, 2008, at 1:40 PM, Norman Franke wrote:
>
> On our production systems, we are getting some crashes like this:
> #0 0x00295108 in pjsua_call_on_media_update
> #1 0x002c79f4 in inv_negotiate_sdp
> #2 0x002c9810 in inv_check_sdp_in_incoming_msg
> #3 0x002ca128 in inv_on_state_early
> #4 0x002c7770 in mod_inv_on_tsx_state
> #5 0x002b4b14 in pjsip_dlg_on_tsx_state
> #6 0x002b9218 in tsx_set_state
> #7 0x002ba22c in tsx_on_state_proceeding_uac
> #8 0x002b9374 in pjsip_tsx_recv_msg
> #9 0x002bb334 in mod_tsx_layer_on_rx_response
> #10 0x00285d8c in endpt_on_rx_msg
> #11 0x002b10f0 in pjsip_tpmgr_receive_packe
> #12 0x002d2198 in udp_on_read_complete
> #13 0x002ada30 in ioqueue_dispatch_read_event
> #14 0x002ae398 in pj_ioqueue_poll
> #15 0x00285958 in pjsip_endpt_handle_events2
> #16 0x0028e56c in pjsua_handle_events
> #17 0x0028e5c8 in worker_thread
> #18 0x0028372c in thread_main
> #19 0x9002bd08 in _pthread_body
> I'll attempt to do both -O2 and -g to see if I can get some better stack
> traces. (I have to do -O2, or I get some glitches on some rather slow
> iMacs.)

All I can say for now is this looks to have something to do with
processing SDP in the response. If you have the message logging that
(Continue reading)

Norman Franke | 3 Jun 2008 19:57
Favicon

Re: Odd Crash

I didn't really enable logging (it's set to 2) and nothing was printed out. It happens a few times a day, but compiling with debug (-g) results in many other problems (at least under OS X), so I had to turn that off.

I wonder if it's related to another issue. If I kill my PJSUA application without terminating the call (i.e. kill -9 it), when I restart, Asterisk will send packets there (since it didn't get a terminate request) and PJSUA will sometimes crash. This hasn't happened in a while, though. 

Norman Franke
Answering Service for Directors, Inc.
www.myasd.com

On Jun 1, 2008, at 11:03 AM, Benny Prijono wrote:

On Sat, May 31, 2008 at 6:43 PM, Norman W. Franke <norman <at> myasd.com> wrote:
On May 31, 2008, at 1:40 PM, Norman Franke wrote:

On our production systems, we are getting some crashes like this:
#0 0x00295108 in pjsua_call_on_media_update
#1 0x002c79f4 in inv_negotiate_sdp
#2 0x002c9810 in inv_check_sdp_in_incoming_msg
#3 0x002ca128 in inv_on_state_early
#4 0x002c7770 in mod_inv_on_tsx_state
#5 0x002b4b14 in pjsip_dlg_on_tsx_state
#6 0x002b9218 in tsx_set_state
#7 0x002ba22c in tsx_on_state_proceeding_uac
#8 0x002b9374 in pjsip_tsx_recv_msg
#9 0x002bb334 in mod_tsx_layer_on_rx_response
#10 0x00285d8c in endpt_on_rx_msg
#11 0x002b10f0 in pjsip_tpmgr_receive_packe
#12 0x002d2198 in udp_on_read_complete
#13 0x002ada30 in ioqueue_dispatch_read_event
#14 0x002ae398 in pj_ioqueue_poll
#15 0x00285958 in pjsip_endpt_handle_events2
#16 0x0028e56c in pjsua_handle_events
#17 0x0028e5c8 in worker_thread
#18 0x0028372c in thread_main
#19 0x9002bd08 in _pthread_body
I'll attempt to do both -O2 and -g to see if I can get some better stack
traces. (I have to do -O2, or I get some glitches on some rather slow
iMacs.)

All I can say for now is this looks to have something to do with
processing SDP in the response. If you have the message logging that
could be helpful.

Cheers
 Benny

_______________________________________________
Visit our blog: http://blog.pjsip.org

pjsip mailing list

<div>
<div>I didn't really enable logging (it's set to 2) and nothing was printed out. It happens a few times a day, but compiling with debug (-g) results in many other problems (at least under OS X), so I had to turn that off.</div>
<div><br></div>
<div>I wonder if it's related to another issue. If I kill my PJSUA application without terminating the call (i.e. kill -9 it), when I restart, Asterisk will send packets there (since it didn't get a terminate request) and PJSUA will sometimes crash. This hasn't happened in a while, though.&nbsp;</div>
<br><div> <span class="Apple-style-span"><span class="Apple-style-span"><span class="Apple-style-span"><div>
<div>Norman Franke</div>
<div>Answering Service for Directors, Inc.</div>
<div>www.myasd.com</div>
</div></span></span></span> </div>
<br><div>
<div>On Jun 1, 2008, at 11:03 AM, Benny Prijono wrote:</div>
<br class="Apple-interchange-newline"><blockquote type="cite">
<div>On Sat, May 31, 2008 at 6:43 PM, Norman W. Franke &lt;<a href="mailto:norman <at> myasd.com">norman <at> myasd.com</a>&gt; wrote:</div> <blockquote type="cite">
<div>On May 31, 2008, at 1:40 PM, Norman Franke wrote:</div>
<div><br></div>
<div>On our production systems, we are getting some crashes like this:</div>
<div>#0 0x00295108 in pjsua_call_on_media_update</div>
<div>#1 0x002c79f4 in inv_negotiate_sdp</div>
<div>#2 0x002c9810 in inv_check_sdp_in_incoming_msg</div>
<div>#3 0x002ca128 in inv_on_state_early</div>
<div>#4 0x002c7770 in mod_inv_on_tsx_state</div>
<div>#5 0x002b4b14 in pjsip_dlg_on_tsx_state</div>
<div>#6 0x002b9218 in tsx_set_state</div>
<div>#7 0x002ba22c in tsx_on_state_proceeding_uac</div>
<div>#8 0x002b9374 in pjsip_tsx_recv_msg</div>
<div>#9 0x002bb334 in mod_tsx_layer_on_rx_response</div>
<div>#10 0x00285d8c in endpt_on_rx_msg</div>
<div>#11 0x002b10f0 in pjsip_tpmgr_receive_packe</div>
<div>#12 0x002d2198 in udp_on_read_complete</div>
<div>#13 0x002ada30 in ioqueue_dispatch_read_event</div>
<div>#14 0x002ae398 in pj_ioqueue_poll</div>
<div>#15 0x00285958 in pjsip_endpt_handle_events2</div>
<div>#16 0x0028e56c in pjsua_handle_events</div>
<div>#17 0x0028e5c8 in worker_thread</div>
<div>#18 0x0028372c in thread_main</div>
<div>#19 0x9002bd08 in _pthread_body</div>
<div>I'll attempt to do both -O2 and -g to see if I can get some better stack</div>
<div>traces. (I have to do -O2, or I get some glitches on some rather slow</div>
<div>iMacs.)</div> </blockquote>
<div><br></div>
<div>All I can say for now is this looks to have something to do with</div>
<div>processing SDP in the response. If you have the message logging that</div>
<div>could be helpful.</div>
<div><br></div>
<div>Cheers</div>
<div>
<span class="Apple-converted-space">&nbsp;</span>Benny</div>
<div><br></div>
<div>_______________________________________________</div>
<div>Visit our blog: <a href="http://blog.pjsip.org">http://blog.pjsip.org</a>
</div>
<div><br></div>
<div>pjsip mailing list</div>
<div><a href="mailto:pjsip <at> lists.pjsip.org">pjsip <at> lists.pjsip.org</a></div>
<div><a href="http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org">http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org</a></div> </blockquote>
</div>
<br>
</div>
Benny Prijono | 3 Jun 2008 22:12
Favicon

Re: Odd Crash

On Tue, Jun 3, 2008 at 6:57 PM, Norman Franke <norman <at> myasd.com> wrote:
> I didn't really enable logging (it's set to 2) and nothing was printed out.
> It happens a few times a day, but compiling with debug (-g) results in many
> other problems (at least under OS X), so I had to turn that off.

Without any other info there's very little I can do. But we will be
releasing 0.9 in few weeks time and by then I'll be doing some QA, so
something might come up.

> I wonder if it's related to another issue. If I kill my PJSUA application
> without terminating the call (i.e. kill -9 it), when I restart, Asterisk
> will send packets there (since it didn't get a terminate request) and PJSUA
> will sometimes crash. This hasn't happened in a while, though.

I don't think that's related, as this is happening in pjmedia.

Cheers
 Benny


Gmane