gmane.linux.kernel.input

Oliver Neukum | 26 Jun 15:30

[patch]fix race between disconnect and open in usbtouch

Hi,

if you kill the URB before you unregister, open can come in between
and resubmit the urb. This patch reverses the order.

	Regards
		Oliver

Signed-off-by: Oliver Neukum <oneukum <at> suse.de>

---

--- linux-2.6.26-sierra/drivers/input/touchscreen/usbtouchscreen.c.alt	2008-06-26
15:23:34.000000000 +0200
+++ linux-2.6.26-sierra/drivers/input/touchscreen/usbtouchscreen.c	2008-06-26
15:23:38.000000000 +0200
@@ -973,8 +973,8 @@ static void usbtouch_disconnect(struct u

 	dbg("%s - usbtouch is initialized, cleaning up", __FUNCTION__);
 	usb_set_intfdata(intf, NULL);
-	usb_kill_urb(usbtouch->irq);
 	input_unregister_device(usbtouch->input);
+	usb_kill_urb(usbtouch->irq);
 	usb_free_urb(usbtouch->irq);
 	usbtouch_free_buffers(interface_to_usbdev(intf), usbtouch);
 	kfree(usbtouch);
--
To unsubscribe from this list: send the line "unsubscribe linux-input" in
the body of a message to majordomo <at> vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

(Continue reading)

headers

Ville Syrjälä | 26 Jun 15:41

Re: [patch]fix race between disconnect and open in usbtouch

On Thu, Jun 26, 2008 at 03:32:05PM +0200, Oliver Neukum wrote:
> Hi,
> 
> if you kill the URB before you unregister, open can come in between
> and resubmit the urb. This patch reverses the order.
> 
> 	Regards
> 		Oliver
> 
> Signed-off-by: Oliver Neukum <oneukum@...>
> 
> ---
> 
> --- linux-2.6.26-sierra/drivers/input/touchscreen/usbtouchscreen.c.alt	2008-06-26
15:23:34.000000000 +0200
> +++ linux-2.6.26-sierra/drivers/input/touchscreen/usbtouchscreen.c	2008-06-26
15:23:38.000000000 +0200
> @@ -973,8 +973,8 @@ static void usbtouch_disconnect(struct u
>  
>  	dbg("%s - usbtouch is initialized, cleaning up", __FUNCTION__);
>  	usb_set_intfdata(intf, NULL);
> -	usb_kill_urb(usbtouch->irq);
>  	input_unregister_device(usbtouch->input);
> +	usb_kill_urb(usbtouch->irq);
>  	usb_free_urb(usbtouch->irq);
>  	usbtouch_free_buffers(interface_to_usbdev(intf), usbtouch);
>  	kfree(usbtouch);

I think you can just remove that usb_kill_urb() call since
usbtouch_close() has already called it by that time.

(Continue reading)

headers

Oliver Neukum | 26 Jun 15:49

Re: [patch]fix race between disconnect and open in usbtouch

Am Donnerstag 26 Juni 2008 15:41:10 schrieb Ville Syrjälä:
> > ---
linux-2.6.26-sierra/drivers/input/touchscreen/usbtouchscreen.c.alt        2008-06-26
15:23:34.000000000 +0200
> > +++ linux-2.6.26-sierra/drivers/input/touchscreen/usbtouchscreen.c    2008-06-26
15:23:38.000000000 +0200
> > @@ -973,8 +973,8 @@ static void usbtouch_disconnect(struct u
> >  
> >       dbg("%s - usbtouch is initialized, cleaning up", __FUNCTION__);
> >       usb_set_intfdata(intf, NULL);
> > -     usb_kill_urb(usbtouch->irq);
> >       input_unregister_device(usbtouch->input);
> > +     usb_kill_urb(usbtouch->irq);
> >       usb_free_urb(usbtouch->irq);
> >       usbtouch_free_buffers(interface_to_usbdev(intf), usbtouch);
> >       kfree(usbtouch);
> 
> I think you can just remove that usb_kill_urb() call since
> usbtouch_close() has already called it by that time.

Are you sure? Would you prepare a patch?

	Regards
		Oliver
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@...
More majordomo info at  http://vger.kernel.org/majordomo-info.html

(Continue reading)

headers

Dmitry Torokhov | 26 Jun 16:30

Re: [patch]fix race between disconnect and open in usbtouch

On Thu, Jun 26, 2008 at 03:49:26PM +0200, Oliver Neukum wrote:
> Am Donnerstag 26 Juni 2008 15:41:10 schrieb Ville Syrjälä:
> > > ---
linux-2.6.26-sierra/drivers/input/touchscreen/usbtouchscreen.c.alt        2008-06-26
15:23:34.000000000 +0200
> > > +++ linux-2.6.26-sierra/drivers/input/touchscreen/usbtouchscreen.c    2008-06-26
15:23:38.000000000 +0200
> > > @@ -973,8 +973,8 @@ static void usbtouch_disconnect(struct u
> > >  
> > >       dbg("%s - usbtouch is initialized, cleaning up", __FUNCTION__);
> > >       usb_set_intfdata(intf, NULL);
> > > -     usb_kill_urb(usbtouch->irq);
> > >       input_unregister_device(usbtouch->input);
> > > +     usb_kill_urb(usbtouch->irq);
> > >       usb_free_urb(usbtouch->irq);
> > >       usbtouch_free_buffers(interface_to_usbdev(intf), usbtouch);
> > >       kfree(usbtouch);
> > 
> > I think you can just remove that usb_kill_urb() call since
> > usbtouch_close() has already called it by that time.
> 
> Are you sure?

Yes, calling dev->close for opened devices as part of unregister is done
by the input core.

--

-- 
Dmitry
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in

(Continue reading)

headers

Oliver Neukum | 26 Jun 16:35

Re: [patch]fix race between disconnect and open in usbtouch

Am Donnerstag 26 Juni 2008 16:30:11 schrieb Dmitry Torokhov:
> > > I think you can just remove that usb_kill_urb() call since
> > > usbtouch_close() has already called it by that time.
> > 
> > Are you sure?
> 
> Yes, calling dev->close for opened devices as part of unregister is done
> by the input core.

OK, thanks.

	Regards
		Oliver

--
To unsubscribe from this list: send the line "unsubscribe linux-input" in
the body of a message to majordomo <at> vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

headers

Alan Stern | 26 Jun 16:46

Re: [patch]fix race between disconnect and open in usbtouch

On Thu, 26 Jun 2008, Ville Syrjälä wrote:

> > --- linux-2.6.26-sierra/drivers/input/touchscreen/usbtouchscreen.c.alt	2008-06-26
15:23:34.000000000 +0200
> > +++ linux-2.6.26-sierra/drivers/input/touchscreen/usbtouchscreen.c	2008-06-26
15:23:38.000000000 +0200
> > @@ -973,8 +973,8 @@ static void usbtouch_disconnect(struct u
> >  
> >  	dbg("%s - usbtouch is initialized, cleaning up", __FUNCTION__);
> >  	usb_set_intfdata(intf, NULL);
> > -	usb_kill_urb(usbtouch->irq);
> >  	input_unregister_device(usbtouch->input);
> > +	usb_kill_urb(usbtouch->irq);
> >  	usb_free_urb(usbtouch->irq);
> >  	usbtouch_free_buffers(interface_to_usbdev(intf), usbtouch);
> >  	kfree(usbtouch);
> 
> I think you can just remove that usb_kill_urb() call since
> usbtouch_close() has already called it by that time.

How do you know that usbtouch_close has been called at this point?

Alan Stern

--
To unsubscribe from this list: send the line "unsubscribe linux-input" in
the body of a message to majordomo <at> vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

(Continue reading)

headers

Dmitry Torokhov | 26 Jun 19:41

Re: [patch]fix race between disconnect and open in usbtouch

On Thu, Jun 26, 2008 at 03:32:05PM +0200, Oliver Neukum wrote:
> Hi,
> 
> if you kill the URB before you unregister, open can come in between
> and resubmit the urb. This patch reverses the order.
> 

Hm, we just talked about it - it should be simply removed, not moved
around.

--

-- 
Dmitry
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@...
More majordomo info at  http://vger.kernel.org/majordomo-info.html

headers

Dmitry Torokhov | 26 Jun 19:44

Re: [patch]fix race between disconnect and open in usbtouch

On Thu, Jun 26, 2008 at 01:41:31PM -0400, Dmitry Torokhov wrote:
> On Thu, Jun 26, 2008 at 03:32:05PM +0200, Oliver Neukum wrote:
> > Hi,
> > 
> > if you kill the URB before you unregister, open can come in between
> > and resubmit the urb. This patch reverses the order.
> > 
> 
> Hm, we just talked about it - it should be simply removed, not moved
> around.
> 

Oops, I ma sorry - it looks lke the original mail just now arrived into
my mailbox ;)

--

-- 
Dmitry
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@...
More majordomo info at  http://vger.kernel.org/majordomo-info.html