Re: [PATCH 10/10] bkl: Fix-up compile problems as a result of the bkl-pushdown.

[Frederic Weisbecker <fweisbec@gmail.com>]

On Wed, Apr 28, 2010 at 07:55:02AM -0700, Linus Torvalds wrote:
> 
> 
> On Tue, 27 Apr 2010, Arnd Bergmann wrote:
> > +static long v4l2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
> >  {
> >  	struct video_device *vdev = video_devdata(filp);
> > +	int ret;
> >  
> >  	/* Allow ioctl to continue even if the device was unregistered.
> >  	   Things like dequeueing buffers might still be useful. */
> > +	if (vdev->fops->unlocked_ioctl) {
> > +		ret = vdev->fops->unlocked_ioctl(filp, cmd, arg);
> > +	} else if (vdev->fops->ioctl) {
> > +		/* TODO: convert all drivers to unlocked_ioctl */
> > +		lock_kernel();
> > +		ret = vdev->fops->ioctl(filp, cmd, arg);
> > +		unlock_kernel();
> > +	} else 
> > +		ret = -ENOTTY;
> >  
> > +	return ret;
> 
> [ Removed the '-' lines so you can see what the end result ends up being ]
> 
> Please, if you do this for the V4L2 layer, then DO NOT make the same 
> mistake we did with the vasic VFS layer.
> 
> In other words, DO NOT keep the "bkl" version named just "ioctl". It was a 
> horrible horrible mistake, and it has resulted in problems years 
> afterwards.
> 
> I realize that it's so easy to just add a new ".unlocked_ioctl" member, 
> and then as people start using it, they get rid of the BKL. But it's a 
> mistake. It was a mistake for the VFS layer, it would be a mistake for the 
> V4L2 layer.
> 
> Instead, spend the 15 minutes just renaming every current 'ioctl' user in 
> the V4L2 layer. It's not that much work, the scripts I documented in my 
> renaming patch do 95% of the work (you just need to change 
> "file_operations" to "v4l2_file_operations"). It's not that painful. And 
> then you don't just push the BKL down, you actually annotate the remaining 
> users so that they can be grepped for.
> 
> So please please please, don't make the same mistake we did long ago. 
> 
> 		Linus


Hmm, there are 92 struct v4l2_file_operations::ioctl but actually a lot
of duplicates ioctl. In fact there are just 26 ioctl functions.

It's probably worth the whole pushdown instead of the rename.

I'm going to do this.