Messages in this thread | | | Date | Fri, 4 Dec 2009 21:10:56 +0100 | From | Andi Kleen <> | Subject | Re: [PATCH] intel_txt: add s3 userspace memory integrity verification |
| |
On Fri, Dec 04, 2009 at 09:53:37AM -0800, H. Peter Anvin wrote: > On 12/04/2009 09:13 AM, Andi Kleen wrote: > >>> > >>> So no, you did not audit do_suspend_lowlevel to make sure it does not > >>> follow function pointers. Bad. > >> > >> We aren't aware of any code or data used by the resume path that is outside of the tboot-MAC'ed regions above--if you can point out any then we will gladly address them. > > > > Code coverage is not enough, you need data coverage too. If someone > > modifies kernel data it's typically easy to subvert code as a next step. > > > > The only function pointers that are invoked on the do_suspend_lowlevel > path are some paravirt_crap pointers, but those are located inside > kernel static data.
Was referring to panic(), like Pavel said.
It would be relatively easy to subvert something called by panic that just jumps back to after the MAC checks.
-Andi
-- ak@linux.intel.com -- Speaking for myself only.
| |