[Oisf-devel] Suricata 1.2.1 + OpenBSD 5.1 = segmentation fault

[Steve Grubb]

On Friday, May 04, 2012 07:12:49 AM Henri Wahl wrote:
> Is it of any value?

A little. Someone might be able to sleuth it out. My guess is its a pointer 
deref problem without a NULL check. I have a couple guesses about which one. But 
it looks like a check was missing way up the call stack somewhere. But the 
quickest and most certain thing might be to rebuild the daemon and shared 
objects so that the debug info is still there and re-run it. Mkae sure the -g 
flag is being passed to gcc and there is no stripping of the symbol table. You 
should be able to use the file command to see that the daemon is unstripped.

In cases like this, I sometimes leave the daemon in a build directory and point 
the init script to the build location of the daemon. If you got a crash once, 
you'll probably get it again.

-Steve