Re: [kernel-hardening] Re: Add overflow protection to kref

[Nick Bowler <nbowler@elliptictech.com>]

On 2012-02-24 10:52 -0800, Kees Cook wrote:
> On Fri, Feb 24, 2012 at 10:37 AM, Greg KH <gregkh@linuxfoundation.org> wrote:
> > On Fri, Feb 24, 2012 at 12:58:35PM -0500, David Windsor wrote:
[...]
> >> diff --git a/include/linux/kref.h b/include/linux/kref.h
> >> index 9c07dce..fc0756a 100644
> >> --- a/include/linux/kref.h
> >> +++ b/include/linux/kref.h
> >> @@ -38,8 +38,12 @@ static inline void kref_init(struct kref *kref)
> >>   */
> >>  static inline void kref_get(struct kref *kref)
> >>  {
> >> +   int rc = 0;
> >>     WARN_ON(!atomic_read(&kref->refcount));
> >> -   atomic_inc(&kref->refcount);
> >> +   smp_mb__before_atomic_inc();
> >> +   rc = atomic_add_unless(&kref->refcount, 1, INT_MAX);
> >> +   smp_mb__after_atomic_inc();
> >> +   BUG_ON(!rc);
> >
> > So you are guaranteeing to crash a machine here if this fails?  And you
> > were trying to say this is a "security" based fix?
> 
> This is the same principle as the stack protector. When something has
> gone horribly wrong and cannot be sensibly recovered from, crash the
> machine. Wrapping the refcount would cause all kinds of problems, so
> that certainly seems worthy of a BUG().

But in this case, the principle does not apply because we can recover.
The reason we cannot recover from the stack protector case is because
the stack protector is reacting after the fact, which is not the case
here.  Simply peg the reference count at the maximum value, neither
incrementing it nor decrementing it further.

Cheers,
-- 
Nick Bowler, Elliptic Technologies (http://www.elliptictech.com/)