Adiletta, Andrew <[email protected]> wrote:

> As far as SSH is concerned there are ways to handle synchronization (we 
> outline them
> in the paper). The POC concept we present in the paper should be acceptable 
> to anybody
> who is fluent in the Rowhammer/microarch attack literature.

I disagree.  I believe your simulation is fake. 

> There are numerous results
> where the target is slowed down to solve synchronization. We don’t brush 
> aside or hide
> the synchronization issue in the paper but discuss it explicitly.

Then why don't you produce exactly such a synchronization without any source 
code
changes, in any actual distribution build system??

It's not that you brush or hide; it is that you overplay the seriousness by
providing an entirely synthetic example.

Unless your document is amended, non-astute readers in the future will
read this as "serious OpenSSH problem" instead of "entirely synthetic
and fake demonstration".

I think you made a mistake choosing OpenSSH as a demonstration target
for reputational benefit, that hand-waving about realistic attacks when
you didn't execute one, and I'm calling you out for what is effectively
academic malpractice.

Reply via email to