Re: [PATCH] fix usb skeleton driver

[Stefani Seibold <stefani@seibold.net>]

Am Mittwoch, den 06.06.2012, 09:32 +0200 schrieb Oliver Neukum:
> Am Mittwoch, 6. Juni 2012, 09:00:36 schrieb stefani@seibold.net:
> > @@ -126,32 +122,21 @@ exit:
> >  
> >  static int skel_release(struct inode *inode, struct file *file)
> >  {
> > -       struct usb_skel *dev;
> > -
> > -       dev = file->private_data;
> > -       if (dev == NULL)
> > -               return -ENODEV;
> > +       struct usb_skel *dev = file->private_data;
> >  
> >         /* allow the device to be autosuspended */
> > -       mutex_lock(&dev->io_mutex);
> > -       if (dev->interface)
> > -               usb_autopm_put_interface(dev->interface);
> > -       mutex_unlock(&dev->io_mutex);
> > +       usb_autopm_put_interface(dev->interface);
> 
> That is a bug. You must check for disconnect here, because
> after a disconnect the interface may be bound already to another
> driver.
> 
> 	Regards
> 		Oliver

Yes, you are right.

But as i now figured out, the whole usb_skeleton.c drivers is full of
races in the skel_open and skel_release path.

I will send a clean and tested patch set in the next few days.

Greetings,
Stefani