Date: Sat, 5 Dec 2009 12:16:13 -0500 (EST)
From: Alan Stern <stern@rowland.harvard.edu>
To: Andreas Mohr <andi@lisas.de>
cc: Ondrej Zary <linux@rainbow-software.org>, <linux-usb@vger.kernel.org>,
       <linux-kernel@vger.kernel.org>
Subject: Re: debugging oops after disconnecting Nexio USB touchscreen
In-Reply-To: <20091205073652.GA30203@rhlx01.hs-esslingen.de>
Message-ID: <Pine.LNX.4.44L0.0912051211530.15558-100000@netrider.rowland.org>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1310
Lines: 34

On Sat, 5 Dec 2009, Andreas Mohr wrote:

> Hi,
> 
> > Furthermore, the patch shows that the second-to-last argument to 
> > usb_fill_bulk_urb() -- the completion function -- is NULL.  That is 
> > strictly illegal and it should have caused an oops as soon as the URB 
> > was used.
> 
> Then there's definitely a WARN_ON or so missing in
> static inline void usb_fill_bulk_urb()

No there isn't.  That inline just fills in a bunch of fields.

You could argue that there is a WARN_ON missing in usb_submit_urb().  
I don't think one is necessary, but you might disagree.  Either way,
both of us missed the fact that right at the start of usb_submit_urb()
is a check for urb->complete being NULL; if it is NULL then the
submission simply fails (and there is no oops).

> And highly likely more checks in those areas that are causing my (and
> other people's) ftdi_sio tests and USB audio (MIPS mmap) to fail.
> Followup soon.

Sometimes having too many checks is worse than having too few, 
especially if the failure modes are relatively easy to handle.

Alan Stern

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/