"Andy Ruhl" wrote:

> On 9/16/07, James Hartley <jjhartley@gmail.com> wrote:
> > If you see the same behavior after (re)compiling everything with 4.0's
> > gcc, I would consider this as being ominous.
> 
> Nope, a recompile fixed it.
> 
> But it shouldn't crash the box. That might be interpreted as a
> localhost exploit type security problem.

It looks like lsof is only setgid kmem (and not setuid root), so
crashing the box isn't great.  You couldn't get in to DDB after
it became unresponsive?

Cheers,
Simon.