Re: [RFC v1] mm: SLAB freelist randomization

[Kees Cook]

On Thu, Apr 7, 2016 at 2:14 PM, Jesper Dangaard Brouer
<brouer@xxxxxxxxxx> wrote:
>
> On Wed, 6 Apr 2016 14:45:30 -0700 Kees Cook <keescook@xxxxxxxxxxxx> wrote:
>
>> On Wed, Apr 6, 2016 at 12:35 PM, Thomas Garnier <thgarnie@xxxxxxxxxx> wrote:
> [...]
>> > re-used on slab creation for performance.
>>
>> I'd like to see some benchmark results for this so the Kconfig can
>> include the performance characteristics. I recommend using hackbench
>> and kernel build times with a before/after comparison.
>>
>
> It looks like it only happens on init, right? (Thus must bench tools
> might not be the right choice).

Oh! Yes, you're right. I entirely missed that detail. :) 0-cost
randomization! Sounds good to me. :)

-Kees

>
> My slab tools for benchmarking the fastpath is here:
>  https://github.com/netoptimizer/prototype-kernel/blob/master/kernel/mm/slab_bulk_test01.c
>
> And I also carry a version of Christoph's slab bench tool:
>  https://github.com/netoptimizer/prototype-kernel/blob/master/kernel/mm/slab_test.c
>
> --
> Best regards,
>   Jesper Dangaard Brouer
>   MSc.CS, Principal Kernel Engineer at Red Hat
>   Author of http://www.iptv-analyzer.org
>   LinkedIn: http://www.linkedin.com/in/brouer



-- 
Kees Cook
Chrome OS & Brillo Security