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;
MIME-Version: 1.0
References: <Pine.LNX.4.44L0.1908061009450.1571-100000@iolanthe.rowland.org> <1565105624.8136.32.camel@suse.com>
In-Reply-To: <1565105624.8136.32.camel@suse.com>
From:   Andrey Konovalov <andreyknvl@google.com>
Date:   Wed, 7 Aug 2019 15:46:40 +0200
Message-ID: <CAAeHK+wJ_26rFeeQkfELO5nXn57-1Wo4iodLJNY=9iTeaduHyA@mail.gmail.com>
Subject: Re: KASAN: use-after-free Read in device_release_driver_internal
To:     Oliver Neukum <oneukum@suse.com>
Cc:     Alan Stern <stern@rowland.harvard.edu>,
        syzkaller-bugs <syzkaller-bugs@googlegroups.com>,
        syzbot <syzbot+1b2449b7b5dc240d107a@syzkaller.appspotmail.com>,
        LKML <linux-kernel@vger.kernel.org>,
        USB list <linux-usb@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Tue, Aug 6, 2019 at 5:34 PM Oliver Neukum <oneukum@suse.com> wrote:
>
> Am Dienstag, den 06.08.2019, 10:19 -0400 schrieb Alan Stern:
> > In any case, I don't know if this missing "get" would cause the
> > problem, but it might well.
>
> Hi,
>
> upon further thought, this should be automated. Checking for
> refcount leaks is KASAN's job. In particular, refcounts
> should not
>
> * decrease in probe()
> * increase in disconnect()
> * change in case probe() fails

It's probably a job for some other refcount debugging tool, but yes,
it would be nice to have a detector for this kind of stuff. It will
probably require some annotations in each place we want this kind of
checks to be performed.

>
>         Regards
>                 Oliver
>