During running a sipp-scenario against an opensips2.1 doing tls<->udp
traversing we are able to easily crash this opensips within seconds. As soon as
we disable tcp_accept_aliases the crashes are gone. We are heavily generating
new tls-sessions (around 1000 per second), so we will reuse source-ports. Maybe
this forces this segfault.
gdb(bt full):
#0 0x000000000059a2bf in tcpconn_add_alias (id=<error reading variable: Cannot
access memory at address 0x7ffefb7b78b8>, port=5067,
port@entry=<error reading variable: Cannot access memory at address
0x7ffefb7b78b8>, proto=<error reading variable: Cannot access memory at address
0x7ffefb7b78b8>) at net/net_tcp.c:766
c = 0x7ff422570658
hash = 0
a = <optimized out>
__FUNCTION__ = "tcpconn_add_alias"
---
Reply to this email directly or view it on GitHub:
https://github.com/OpenSIPS/opensips/issues/589
_______________________________________________
Devel mailing list
Devel@lists.opensips.org
http://lists.opensips.org/cgi-bin/mailman/listinfo/devel