Date: Wed, 12 Aug 2009 21:19:59 +0100 From: Bruce Cran <bruce@cran.org.uk> To: current@freebsd.org Subject: sctp panic in _mtx_lock_sleep when attempting to connect to a remote machine Message-ID: <20090812211959.0000293c@unknown>
next in thread | raw e-mail | index | archive | help
I've found a way to reliably panic two machines running 8.0-BETA2. It seems that there's a problem with SCTP connection requests being made at the same time as other network traffic. The panic I see is: Fatal trap 12: page fault while in kernel mode cpuid = 0; apic id = 00 Stack trace: _mtx_lock_sleep sctp_lower_sosend sctp_sosend kern_sendit sendit sendto sycall Xfast_syscall I can trigger it by running the SCTP-enabled version of ncat from http://www.roe.ch/Nmap_SCTP . I put a few thousand lines of: cat /dev/random | ./ncat --sctp 192.168.1.80 2345 into a shell script, where 192.168.1.80 is a machine running 7.2 with SCTP enabled but no server listening - I mostly see "Connection refused" errors when I run the script. When I run the script and at the same time generate some tcp traffic by running csup for example, the box panics. -- Bruce
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20090812211959.0000293c>