Re: [gentoo-hardened] Assessing the Tux Strength: Part 2 - Into the Kernel

On 3 Sep 2010 at 11:56, Daniel Kuehn wrote:

> The randomisation bit was particularily interesting because as far as I
> understand that is one of the better security measures we can use.

actually, if you ask me, ASLR is the least useful security feature :P. it's not
even really security, it's mere obfuscation, and it's great when it's works but
it'll never provide guarantees (which is what we prefer in security).

> Shame on fedora for only 3-bits randomisation for shared libs :P

a note here: fedora uses exec-shield which maps libraries in two different
regions: ascii-armor (lower 16MB) and the rest. i think what paxtest measured
there is the former where the usable entropy is necessarily less than elsewhere
and may not be representative of real life apps and their address spaces (not
saying the whole ascii-armor region is worth anything for security though ;).

PS: when discussing null deref protections, it's worth mentioning UDEREF which
is a tad bit more general and useful than mmap_min_addr ;).