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:   Tue, 25 Sep 2018 13:54:38 -0400 (EDT)
From:   Alan Stern <stern@rowland.harvard.edu>
To:     Andrey Konovalov <andreyknvl@google.com>
cc:     Vladis Dronov <vdronov@redhat.com>,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        Oliver Neukum <oneukum@suse.com>,
        Hans de Goede <hdegoede@redhat.com>,
        syzkaller <syzkaller@googlegroups.com>,
        USB list <linux-usb@vger.kernel.org>,
        LKML <linux-kernel@vger.kernel.org>,
        stable <stable@vger.kernel.org>
Subject: Re: [PATCH] usb: usbfs: fix crash in check_ctrlrecip()->usb_find_alt_setting()
In-Reply-To: <CAAeHK+weYa=d=dSgCJ0o6dZBAxBakeAy2hodHhuPSeTr3iesuA@mail.gmail.com>
Message-ID: <Pine.LNX.4.44L0.1809251352540.1517-100000@iolanthe.rowland.org>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Tue, 25 Sep 2018, Andrey Konovalov wrote:

> On Tue, Sep 25, 2018 at 5:15 PM, Alan Stern <stern@rowland.harvard.edu> wrote:
> > On Tue, 25 Sep 2018, Vladis Dronov wrote:
> >
> >> > What reason is there for having two different fixes for the same bug?
> >> > This one isn't going to get into any mainline trees that don't already
> >> > have c9a4cb204e9e.
> >>
> >> I believe this is the right thing to do, so usb_find_alt_setting()
> >> is not called with a known-bad argument.
> >>
> >> Honestly, I would change "if (!config)" in usb_find_alt_setting() to
> >> "BUG_ON(!config)" so we know when its callers do smth wrong and go
> >
> > (You'll be lucky if Linus doesn't see that.  He yells at anybody who
> > suggests adding BUG_ON for anything that doesn't completely crash the
> > kernel.  The basic problem is that "BUG_ON" is not a good name: That
> > routine doesn't really report bugs; instead it brings everything to a
> > halt in situations where the kernel is unable to proceed.  In practice
> > this tends to make actual debugging more difficult.)
> 
> What about adding a WARN_ON()? It doesn't crash the kernel and it will
> be detected and reported by syzbot.

Sure, we could do that.  But would be the point?  After c9a4cb204e9e, 
calling usb_find_alt_setting() with a NULL config is no more of a bug 
than calling kfree() with a NULL pointer.  You wouldn't want to put a 
WARN_ON in kfree(), would you?

Alan Stern