On Mon, 10 Oct 2005, Brett Lymn wrote:
>>  Introduce per-page fingerprints in Veriexec.
>
> This has been backed out for the moment due to a bad assumption I made
> in the implementation.  Technically, the assumption I made was that
> all file systems used the genfs getpages() routine - this is not true
> in all cases so the code has been disabled for the moment.  Elad and I
> are testing a version that is file system independent, expect a diff
> on tech-kern sometime soon.

Thanks for working on this, it sounds like a very nice addition.
Can you give any numbers on the speed impact of this?
E.g. for startup times of (say) firefox?


  - Hubert