Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Mon, 26 Sep 2022 13:15:02 -0700
From:   Kees Cook <keescook@chromium.org>
To:     Andrey Konovalov <andreyknvl@gmail.com>
Cc:     Feng Tang <feng.tang@intel.com>,
        Andrew Morton <akpm@linux-foundation.org>,
        Vlastimil Babka <vbabka@suse.cz>,
        Christoph Lameter <cl@linux.com>,
        Pekka Enberg <penberg@kernel.org>,
        David Rientjes <rientjes@google.com>,
        Joonsoo Kim <iamjoonsoo.kim@lge.com>,
        Roman Gushchin <roman.gushchin@linux.dev>,
        Hyeonggon Yoo <42.hyeyoo@gmail.com>,
        Dmitry Vyukov <dvyukov@google.com>,
        Jonathan Corbet <corbet@lwn.net>,
        Dave Hansen <dave.hansen@intel.com>,
        Linux Memory Management List <linux-mm@kvack.org>,
        LKML <linux-kernel@vger.kernel.org>,
        kasan-dev <kasan-dev@googlegroups.com>
Subject: Re: [PATCH v6 2/4] mm/slub: only zero the requested size of buffer
 for kzalloc
Message-ID: <202209261305.CF6ED6EEC@keescook>
References: <20220913065423.520159-1-feng.tang@intel.com>
 <20220913065423.520159-3-feng.tang@intel.com>
 <CA+fCnZfSv98uvxop7YN_L-F=WNVkb5rcwa6Nmf5yN-59p8Sr4Q@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <CA+fCnZfSv98uvxop7YN_L-F=WNVkb5rcwa6Nmf5yN-59p8Sr4Q@mail.gmail.com>
Precedence: bulk

On Mon, Sep 26, 2022 at 09:11:24PM +0200, Andrey Konovalov wrote:
> On Tue, Sep 13, 2022 at 8:54 AM Feng Tang <feng.tang@intel.com> wrote:
> >
> 
> Hi Feng,
> 
> > kzalloc/kmalloc will round up the request size to a fixed size
> > (mostly power of 2), so the allocated memory could be more than
> > requested. Currently kzalloc family APIs will zero all the
> > allocated memory.
> >
> > To detect out-of-bound usage of the extra allocated memory, only
> > zero the requested part, so that sanity check could be added to
> > the extra space later.
> 
> I still don't like the idea of only zeroing the requested memory and
> not the whole object. Considering potential info-leak vulnerabilities.

I really really do not like reducing the zeroing size. We're trying to
be proactive against _flaws_, which means that when there's a memory
over-read (or uninitialized use), suddenly the scope of the exposure (or
control) is wider/looser.

Imagine the (unfortunately very common) case of use-after-free attacks,
which leverage type confusion: some object is located in kmalloc-128
because it's 126 bytes. That slot gets freed and reallocated to, say, a
97 byte object going through kzalloc() or zero-on-init. With this patch
the bytes above the 97 don't get zeroed, and the stale data from the
prior 126 byte object say there happily to be used again later through
a dangling pointer, or whatever. Without the proposed patch, the entire
128 bytes is wiped, which makes stale data re-use more difficult.

> > Performance wise, smaller zeroing length also brings shorter
> > execution time, as shown from test data on various server/desktop
> > platforms.

For these cases, I think a much better solution is to provide those
sensitive allocations their own dedicated kmem_cache.

> >
> > For kzalloc users who will call ksize() later and utilize this
> > extra space, please be aware that the space is not zeroed any
> > more.
> 
> CC Kees

Thanks! Well, the good news is that ksize() side-effects is hopefully
going to vanish soon, but my objections about stale memory remain.

-Kees

-- 
Kees Cook