"Simon J. Gerraty" <sjg@quick.com.au> writes:

> Since we are talking clued bad guys... open(2) detects text busy but
> is that infallible?  Is there no way that they can update pages of a
> running binary?  If the answer to either is not provably "yes", then
> in the presence of demand paging, you'd really need to compute 
> (and re-compute) the hash of each page as (and when ever) you paged it in, 
> otherwise how do you know that your cached hash is still valid?

How about the additional requirement that hashed binaries are
immutable?

--Michael