Messages in this thread | | | Date | Fri, 29 Jan 1999 09:03:06 +1100 | From | Richard Gooch <> | Subject | Re: [PATCH] msr v3 available |
| |
Albert D. Cahalan writes: > > Richard Gooch writes: > > > Hi, all. Version 3 of my msr patch is now available from: > ... > > - Included patch which adds an ioctl() on /proc/<pid>/mem for > > translating virtual to physical addresses > > As you've been told before, this is a security hole.
This has been asserted by a few. It has not yet been demonstrated (either experimentally or theoretically) to be true. I've seen various hypotheses float by, but all have proved bogus when considered. In the end, all the security concerns are based on faith that exposing physical pages must be "bad".
The same goes for denial of service attacks.
> Users have no business knowing what physical addresses are being > used.
It's a powerful debugging tool. Even on a kernel with page colouring, that doesn't mean you *get* page colouring. An application that is not behaving as expected can start probing the foundations upon which it resides.
> While many people are struggling to close information leaks and DoS > attacks, you open more.
See above. So far every proposed "example" of security/DoS attacks have been bogus.
> This gets very serious when Linux allows a limited number of > locked pages for each normal user.
Nope. Think about it. These attacks are possible anyway.
Regards,
Richard....
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |