On Mon 2019-05-20 09:03:50, Ran Wang wrote:
> Hi Pavel,
> 
> On Monday, May 20, 2019 16:57, Pavel Machek wrote:
> > 
> > Hi!
> > 
> > > > > +static int rcpm_pm_prepare(struct device *dev) {
> > > > > +	struct device_node *np = dev->of_node;
> > > > > +	struct wakeup_source *ws;
> > > > > +	struct rcpm *rcpm;
> > > > > +	u32 value[RCPM_WAKEUP_CELL_MAX_SIZE + 1], tmp;
> > > > > +	int i, ret;
> > > > > +
> > > > > +	rcpm = dev_get_drvdata(dev);
> > > > > +	if (!rcpm)
> > > > > +		return -EINVAL;
> > > > > +
> > > > > +	/* Begin with first registered wakeup source */
> > > > > +	ws = wakeup_source_get_next(NULL);
> > > > > +	while (ws) {
> > > >
> > > > while (ws = wakeup_source_get_next(NULL)) ?
> > >
> > > Actually, we only pass NULL to wakeup_source_get_next() at very first
> > > call to get 1st wakeup source. Then in the while loop, we will fetch
> > > next source but not 1st, that's different. I am afraid your suggestion
> > > is not quite correct.
> > 
> > Sorry, I seen your next version before seeing this explanation.
> > 
> > You are right, but the current code is "interesting". What about
> > 
> >     ws = NULL;
> >     while (ws = wakeup_source_get_next(NULL)) ...
> > 
> > then?
> 
> Did you mean:
>      ws = NULL;
>      while (ws = wakeup_source_get_next(ws)) ...
> 
>    Yes, that will be the same to my original logic, do you recommend to change
> to this? :)

Yes please. It will be less confusing to the reader.

Thanks (and sorry for cross-talk),
								Pavel
-- 
DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany