From mboxrd@z Thu Jan  1 00:00:00 1970
From: Hannes Frederic Sowa <hannes@stressinduktion.org>
Subject: Re: Potential out-of-bounds access in ip6_finish_output2
Date: Sat, 21 Sep 2013 17:14:57 +0200
Message-ID: <20130921151457.GA5748@order.stressinduktion.org>
References: <CACT4Y+ZXqoixvCyN5703+sVhCV4ND-vNwash2Z1vSDoLgKNyVg@mail.gmail.com> <20130917224851.GB8947@order.stressinduktion.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8
To: Dmitry Vyukov <dvyukov@google.com>, yoshfuji@linux-ipv6.org,
	netdev@vger.kernel.org, Paul Turner <pjt@google.com>,
	Andrey Konovalov <andreyknvl@google.com>,
	Kostya Serebryany <kcc@google.com>,
	Tom Herbert <therbert@google.com>
Return-path: <netdev-owner@vger.kernel.org>
Received: from order.stressinduktion.org ([87.106.68.36]:42125 "EHLO
	order.stressinduktion.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1752482Ab3IUPO7 (ORCPT
	<rfc822;netdev@vger.kernel.org>); Sat, 21 Sep 2013 11:14:59 -0400
Content-Disposition: inline
In-Reply-To: <20130917224851.GB8947@order.stressinduktion.org>
Sender: netdev-owner@vger.kernel.org
List-ID: <netdev.vger.kernel.org>

On Wed, Sep 18, 2013 at 12:48:51AM +0200, Hannes Frederic Sowa wrote:
> On Mon, Sep 16, 2013 at 10:13:10PM -0700, Dmitry Vyukov wrote:
> > I am working on AddressSanitizer -- a tool that detects use-after-free
> > and out-of-bounds bugs
> > (https://code.google.com/p/address-sanitizer/wiki/AddressSanitizerForKernel).
> > 
> > I've got a dozen of reports in ip6_finish_output2. Below are 2 of
> > them. They are always followed by kernel crash. Unfortunately I don't
> > have a reproducer because I am using trinity fuzzer. I would
> > appreciate if somebody familiar with the code look at sources and
> > maybe spot the bug.
> 
> Thanks for the report!
> 
> I tried reproducing the bug and hit some other bugs nearby. I'll try to fix
> them, but this could take some time.

I fixed the first bug I encountered with trinity here:
http://patchwork.ozlabs.org/patch/276835/

The main cause of this bug has nothing to do with raw sockets, so I first
thought they are not related. But I left my machine run trinity while
I was sleeping and did not see any other splats (I added some manually
range checks in ip6_append_data). So maybe your bug happend because the
premature exit in the dontfrag check without resetting cork->length. Maybe
you could give this patch a try? I'll have a second look later today.

Thanks,

  Hannes