On Tue, 16 Mar 2010 17:48:35 -0400
Sebastien Roy <sebastien.roy at oracle.com> wrote:
> On 03/16/10 05:34 PM, Michael Hunter wrote:
> > Grrrr. How did this ever run? I onu'd this to a vbox and rebooted
> > multiple times 'cuz I was a little worried that this was just the first
> > layer of priv problems. It seems unlikely I just got lucky that the
> > junk on the stack was 0?
>
> In my experience, the uninitialized junk in the stack is very often 0
> especially early on in the lifetime of a process. SPARC and x86 appear
> to also behave differently in this respect.
That makes some sense. Would be interesting to have an option to
initialize stack pages to some pattern (0xdeadbeef) as they are faulted
in. I'm surprised SPARC and x86 are different. I would think pages
would need to be cleared before being handed to a process for security.
Michael
>
> -Seb
> _______________________________________________
> nwam-dev mailing list
> nwam-dev at opensolaris.org
> http://mail.opensolaris.org/mailman/listinfo/nwam-dev
