Re: Flush+Reload: a High Resolution, Low Noise, L3 Cache Side-Channel Attack

[Jeffrey Walton <noloader () gmail com>]

On Fri, Jul 26, 2013 at 3:37 PM,  <Valdis.Kletnieks () vt edu> wrote:
> On Fri, 26 Jul 2013 07:31:09 +0100, Hurgel Bumpf said:
> > Just found this online.. might be of interest
>
> > Direct PDF: http://eprint.iacr.org/2013/448.pdf
>
> From the fine PDF:
>
> "The Flush+Reload attack is a variant of the Prime+Probe attack that relies on
> sharing pages between the spy and the victim programs. With shared pages, the
> spy program can ensure that a specic memory line is evicted from the whole
> cache hierarchy. The spy uses this to monitor access to the memory line."
>
> The fact you need to get gnupg to share the pages in question with you
> does mean that this isn't, by itself, a knockout blow.
>
> Still quite the interesting attack.  And attacks always improve.  Maybe
> somebody will find a way to do better...
Dr. Bernstein puts a lot of effort into defending against timing
attacks and other side channels in his NaCl library. I'm not aware of
any other libraries which go to the same depths. On the downside, NaCl
is not easy to work with (for example, change compilers or
cross-compile for iOS or Android); its not really portable (lots of C
language violations); nor is it easy to get analysis tools on it.

Recently, he presented an OWASP talk that included the subject matter
(including lots of other practical crypto failures).
  * Slides: http://www.secappdev.org/handouts/2012/Dan%20J.%20Bernstein/worst%20practices.pdf
  * Talk: https://www.owasp.org/download/jmanico/owasp_podcast_95.mp3
  * Video: http://secappdev.org/lectures/144

For DNSSEC fans, he beats the hell out of DNSSEC for its amplification
attacks and other info leaks.

Jeff

_______________________________________________
Full-Disclosure - We believe in it.
Charter: http://lists.grok.org.uk/full-disclosure-charter.html
Hosted and sponsored by Secunia - http://secunia.com/