11 Nov
2013
11 Nov
'13
10:18 p.m.
Hi!
Looking for some help as to why the SCA module is crashing randomly on
4.0.4. It has crashed several times with a backtrace that shows these
stack calls:
(gdb) backtrace
#0 0x00000000 in ?? ()
#1 0x4dceca11 in sca_hash_table_slot_kv_delete_unsafe () from
/usr/lib/kamailio/modules/sca.so
#2 0x4dcecaaa in sca_hash_table_slot_kv_delete () from
/usr/lib/kamailio/modules/sca.so
#3 0x4dcecb1e in sca_hash_table_index_kv_delete () from
/usr/lib/kamailio/modules/sca.so
#4 0x4dcecb99 in sca_hash_table_kv_delete () from
/usr/lib/kamailio/modules/sca.so
#5 0x4dcf30bc in sca_appearance_unregister () from
/usr/lib/kamailio/modules/sca.so
#6 0x4dd0d9b9 in sca_call_info_update () from
/usr/lib/kamailio/modules/sca.so
#7 0x126a6a9a in do_action ()
#8 0x126af88e in run_actions ()
#9 0x126a4731 in do_action ()
#10 0x126af88e in run_actions ()
#11 0x126b007f in run_top_route ()
#12 0x1269dd7e in receive_msg ()
#13 0x12734f01 in udp_rcv_loop ()
#14 0x126d3f56 in main_loop ()
#15 0x126d76de in main ()
The message that seemed to trigger the crash in this particular case was
the following INVITE:
INVITE sip:35705@10.250.16.71;user=phone SIP/2.0
*Record-Route:*<sip:10.250.16.70;lr=on;ftag=D8B14B68-93D21A25;did=513b.a3741>
*Via:* SIP/2.0/UDP 10.250.16.70;branch=z9hG4bKc43b.14d180e6.0
*Via:* SIP/2.0/UDP 10.250.88.116;branch=z9hG4bK8d3ec33e632D8AB3
*From:* PIN-1 WKL USA <sip:15862@*obscured*>;tag=D8B14B68-93D21A25
*To:* <sip:0@*obscured*;user=phone>
*CSeq:* 1 INVITE
*Call-ID:* c69594c-ff17a5d9-d6ed4f6a(a)10.250.88.116
*Contact:* <sip:15862@10.250.88.116>
*Allow:* INVITE, ACK, BYE, CANCEL, OPTIONS, INFO, MESSAGE, SUBSCRIBE,
NOTIFY, PRACK, UPDATE, REFER
*User-Agent:* PolycomSoundPointIP-SPIP_335-UA/4.0.3.7562
*Accept-Language:* en
*Supported:* 100rel,replaces
*Allow-Events:* conference,talk,hold
*Max-Forwards:* 32
*Content-Type:* application/sdp
*Content-Length:* 296
v=0
o=- 1383792680 1383792680 IN IP4 10.250.88.116
s=Polycom IP Phone
c=IN IP4 10.250.88.116
t=0 0
a=sendrecv
m=audio 2230 RTP/AVP 9 18 0 8 101
a=rtpmap:9 G722/8000
a=rtpmap:18 G729/8000
a=fmtp:18 annexb=no
a=rtpmap:0 PCMU/8000
a=rtpmap:8 PCMA/8000
a=rtpmap:101 telephone-event/8000
I have a pcap trace that shows all the calls from 3 hours prior and this is
the first time this device appeared.
I followed the code down to the sca_hash_table_slot_kv_delete_unsafe
()routine and it appears to have crashed when it tried to use
e->free_entry which seems to have been NULL at that point. I'm not sure
how it could be NULL since it should have been initialized in
sca_hash_table_slot_kv_insert_unsafe
(). I'm guessing that either e->free_entry was never initialized or
somehow it was overwritten. Since several backtraces end in the same way
the bug seems to have be specific to some sequence of events I have not
been able to identify.
Any help tracking this down would be appreciated.
Thanks,
Bob