On Mon, 25 Apr 2022 10:37:23 +0200 Peter Zijlstra <peterz@xxxxxxxxxxxxx> wrote:
> On Fri, Apr 22, 2022 at 01:54:13PM -0700, Andrew Morton wrote:
>
> > From: Chen Wandun <chenwandun@xxxxxxxxxx>
> > Subject: mm: rework calculation of bdi_min_ratio in bdi_set_min_ratio
> >
> > In function bdi_set_min_ratio, min_ratio is unsigned int, it will
> > result underflow when setting min_ratio below bdi->min_ratio, it
> > is confusing. Rework it, no functional change.
> >
>
> ...
>
> >
> > --- a/mm/page-writeback.c~mm-rework-calculation-of-bdi_min_ratio-in-bdi_set_min_ratio
> > +++ a/mm/page-writeback.c
> > @@ -650,18 +650,25 @@ static unsigned int bdi_min_ratio;
> >
> > int bdi_set_min_ratio(struct backing_dev_info *bdi, unsigned int min_ratio)
> > {
> > + unsigned int delta;
> > int ret = 0;
> >
> > spin_lock_bh(&bdi_lock);
> > if (min_ratio > bdi->max_ratio) {
> > ret = -EINVAL;
> > } else {
>
> > - min_ratio -= bdi->min_ratio;
> > - if (bdi_min_ratio + min_ratio < 100) {
> > - bdi_min_ratio += min_ratio;
> > - bdi->min_ratio += min_ratio;
> > } else {
> > - ret = -EINVAL;
>
> I really don't see the why we're doing this, code gets to be worse, for
> no actual gain.
Oh. I found the resulting code to be considerably more understandable.
I mean, how does the old code work? afaict it treats `unsigned
min_ratio' as a signed value(!) and relies upon an underflowed unsigned
to be a large number (> 100) when avoiding underflowing global
bdi_min_ratio. Or something like that. Also, modifying an incoming
arg's value is kinda rude.
And there's no change in the object file's .text size.
> > + if (min_ratio < bdi->min_ratio) {
> > + delta = bdi->min_ratio - min_ratio;
> > + bdi_min_ratio -= delta;
> > + bdi->min_ratio = min_ratio;
> > } else {
> > + delta = min_ratio - bdi->min_ratio;
> > + if (bdi_min_ratio + delta < 100) {
> > + bdi_min_ratio += delta;
> > + bdi->min_ratio = min_ratio;
> > + } else {
> > + ret = -EINVAL;
> > + }
> > }
>
> > }
