Update:
this is not releated only with huntgroup but also Dial(VISDN/visdn0). I
can reproduce this by dial and immediately hangup and immediately dial.
So it seems, that there is some race condition with several fast dial at
once.
Martin Vít wrote:
Hello,
i'm testing visdn-0.15 with 5 PCI HFC-S
All cards are configured as NT P2MP
I've defined huntgroup tr1 as follows:
[huntgroup:tr1]
mode = rotary
members = visdn0,visdn1,visdn2,visdn3,visdn4
Scenario:
-----------
always reproducable after second call to huntgroup
asterisk log:
------------
-- Executing Dial("IAX2/172.24.45.1:4569-2",
"VISDN/huntgroup:tr1/xxxx||j") in new stack
-- Called huntgroup:tr1/605267610
Jun 22 14:57:10 DEBUG[11485]: channel.c:2350 set_format: Set channel
VISDN/visdn0/98.O to read format slin
Jun 22 14:57:10 DEBUG[11485]: channel.c:2350 set_format: Set channel
IAX2/172.24.45.1:4569-2 to write format slin
Jun 22 14:57:10 DEBUG[11485]: channel.c:2350 set_format: Set channel
IAX2/172.24.45.1:4569-2 to read format slin
Jun 22 14:57:10 DEBUG[11485]: channel.c:2350 set_format: Set channel
VISDN/visdn0/98.O to write format slin
Jun 22 14:57:10 DEBUG[11487]: app_queue.c:524 changethread: Device
'VISDN/visdn0/98.O' changed to state '0' (Unknown) but we don't care
because they're not a member of any queue.
Segmentation fault (core dumped)
backtrace:
-------------
#0 0xb764234c in q931_channel_select_response (call=0x81a27c0,
ies=0x8171d30, causes=0xb7634f90) at call.c:692
692 if (ci->chanset.chans[i]->state ==
(gdb) bt
#0 0xb764234c in q931_channel_select_response (call=0x81a27c0,
ies=0x8171d30, causes=0xb7634f90) at call.c:692
#1 0xb7649052 in q931_handle_call_proceeding (call=0x81a27c0,
msg=0x81b73a0) at call.c:3869
#2 0xb764ec2c in q931_dispatch_message (call=0x81a27c0,
msg=0x81b7178) at call.c:6379
#3 0xb764054a in q931_receive (dlc=0x8171da0) at input.c:1205
#4 0xb766ed39 in visdn_q931_thread_do_poll () at chan_visdn.c:2433
#5 0xb766ef15 in visdn_q931_thread_main (data=0x0) at chan_visdn.c:2495
#6 0xb7f4fb63 in start_thread () from /lib/tls/libpthread.so.0
#7 0xb7e4a18a in clone () from /lib/tls/libc.so.6
regards festr
_______________________________________________
Visdn-hackers mailing list
[email protected]
https://mailman.uli.it/mailman/listinfo/visdn-hackers
--
Martin Vít
LAM plus s.r.o.
http://www.vasesit.cz/
mobil: 605 267 610
_______________________________________________
Visdn-hackers mailing list
[email protected]
https://mailman.uli.it/mailman/listinfo/visdn-hackers
