Re: [PATCH] x86: add missing verify_cpu to 32bit wakeup

[Kees Cook]

Hi,

On Fri, Jul 01, 2011 at 02:26:27PM -0700, H. Peter Anvin wrote:
> On 07/01/2011 02:19 PM, Kees Cook wrote:
> > Some BIOSes will reset the Intel XD_DISABLE MSR bit when resuming from S3,
> > which can interact poorly with ebba638ae723d8a8fc2f7abce5ec18b688b791d7.
> > In 32bit PAE mode, this can lead to a fault when EFER is restored by
> > the kernel wakeup routines, due to it setting the NX bit for a CPU
> > that (thanks to the BIOS reset) now incorrectly thinks it lacks the NX
> > feature. 64bit wakeup already handled this through its common call path
> > that would hit verify_cpu(). 32bit has a separate path for restoring
> > CPU state on S3 wakeup, and needed to call verify_cpu() to handle this
> > situation.
> 
> For S3, we should save/restore MISC_ENABLE instead... in fact, we
> already save it, we just restore it too late.

Given that MISC_ENABLE may not be available for a given CPU, it seems that
it's basically the same detection code as in verify_cpu() already. Since
this bit is the only part that is needed that early, I think the patch is
good the way it is (especially since it balances the 64bit path which
already calls this logic). I don't think doing the full early MISC_ENABLE
save/restore this early is worth it. Thoughts?

-Kees

-- 
Kees Cook
Ubuntu Security Team
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/