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:   Sat, 10 Aug 2019 11:23:18 -0700
From:   Kees Cook <keescook@chromium.org>
To:     Alan Stern <stern@rowland.harvard.edu>
Cc:     Greg KH <gregkh@linuxfoundation.org>,
        syzbot <syzbot+45b2f40f0778cfa7634e@syzkaller.appspotmail.com>,
        Michael Hund <mhund@ld-didactic.de>, akpm@linux-foundation.org,
        andreyknvl@google.com, cai@lca.pw, linux-kernel@vger.kernel.org,
        linux-mm@kvack.org, linux-usb@vger.kernel.org,
        syzkaller-bugs@googlegroups.com, tglx@linutronix.de
Subject: Re: BUG: bad usercopy in ld_usb_read
Message-ID: <201908101120.BE5034521A@keescook>
References: <20190809085545.GB21320@kroah.com>
 <Pine.LNX.4.44L0.1908091100580.1630-100000@iolanthe.rowland.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <Pine.LNX.4.44L0.1908091100580.1630-100000@iolanthe.rowland.org>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Fri, Aug 09, 2019 at 11:13:00AM -0400, Alan Stern wrote:
> In fact, I don't see why any of the computations here should overflow
> or wrap around, or even give rise to a negative value.  If syzbot had a
> reproducer we could get more debugging output -- but it doesn't.

Yeah, this is odd. The only thing I could see here with more study was
that ring_tail is used/updated outside of the rbsl lock in
ld_usb_read(). I couldn't convince myself there wasn't a race against
the interrupt, but I also couldn't think of a way it could break...

-- 
Kees Cook