Re: [PATCH 2/5] writeback: dirty position control

[Wu Fengguang <fengguang.wu@intel.com>]

On Wed, Aug 17, 2011 at 08:03:56PM +0800, Jan Kara wrote:
> On Wed 17-08-11 02:40:19, David Horner wrote:
> >  I noticed a significant typo below (another of those thousand eyes,
> > thanks to Jan Kara's post that started ne looking) :
> > 
> >  > +static unsigned long bdi_position_ratio(struct backing_dev_info *bdi,
> >  > + unsigned long thresh,
> > ...
> >  > + * For JBOD case, bdi_thresh (not bdi_dirty!) could fluctuate up to its
> >  > + * own size, so move the slope over accordingly.
> >  > + */
> >  > + if (unlikely(bdi_thresh > thresh))
> >  > + bdi_thresh = thresh;
> >  > + /*
> >  > + * scale global setpoint to bdi's:  setpoint *= bdi_thresh / thresh
> >  > + */
> >  > + x = div_u64((u64)bdi_thresh << 16, thresh | 1);
> > 
> >                   ^
> >  I believe should be
> > 
> >     x = div_u64((u64)bdi_thresh << 16, thresh + 1);
>   I've noticed this as well but it's mostly a consistency issue. 'thresh'
> is going to be large in practice so there's not much difference between
> thresh + 1 and thresh | 1.

Right :) Anyway I'll change it to thresh + 1.

Thanks,
Fengguang