Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Wed, 11 Mar 2020 10:44:09 -0700
From:   Kees Cook <keescook@chromium.org>
To:     'Christopher Lameter' <cl@linux.com>
Cc:     David Laight <David.Laight@ACULAB.COM>,
        Andrew Morton <akpm@linux-foundation.org>,
        Pekka Enberg <penberg@kernel.org>,
        David Rientjes <rientjes@google.com>,
        Joonsoo Kim <iamjoonsoo.kim@lge.com>,
        Daniel Micay <danielmicay@gmail.com>,
        Vitaly Nikolenko <vnik@duasynt.com>,
        Silvio Cesare <silvio.cesare@gmail.com>,
        "linux-mm@kvack.org" <linux-mm@kvack.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH] slub: Relocate freelist pointer to middle of object
Message-ID: <202003111039.24B8A0B@keescook>
References: <202003051624.AAAC9AECC@keescook>
 <alpine.DEB.2.21.2003081919290.14266@www.lameter.com>
 <6fbf67b5936a44feaf9ad5b58d39082b@AcuMS.aculab.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <6fbf67b5936a44feaf9ad5b58d39082b@AcuMS.aculab.com>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

From: Christopher Lameter
> Sent: 08 March 2020 19:21
> 
> On Thu, 5 Mar 2020, Kees Cook wrote:
> 
> > Instead of having the freelist pointer at the very beginning of an
> > allocation (offset 0) or at the very end of an allocation (effectively
> > offset -sizeof(void *) from the next allocation), move it away from
> > the edges of the allocation and into the middle. This provides some
> > protection against small-sized neighboring overflows (or underflows),
> > for which the freelist pointer is commonly the target. (Large or well
> > controlled overwrites are much more likely to attack live object contents,
> > instead of attempting freelist corruption.)
> 
> Sounds good. You could even randomize the position to avoid attacks on via
> the freelist pointer.

That's a good point. "offset" is just calculated once, and for many
slabs, the available space is quite large. I wonder what the best
practice might be for how far from the edge to stay. Hmmm. Maybe simply
carve it into thirds, and randomize the offset within the middle third?

-- 
Kees Cook