On 14/11/2020 17:08:04+0000, Jonathan Cameron wrote:
> > - ret = at91_adc_probe_dt(idev, pdev);
> > - if (ret)
> > - return ret;
> > + st->caps = (struct at91_adc_caps *)
> > + of_match_device(at91_adc_dt_ids, &pdev->dev)->data;
>
> of_device_get match_data - obviously an unrelated change but trivial enough
> I'd just slip it in this patch (unless you have it a later one!)
>
> That returns a void * so no need for the cast.
>
I guess I will change the next patch to use of_device_get_match_data as
Andy doesn't like it when the fo_ and device_ APIs are mixed.
> > + st->registers = &st->caps->registers;
> > + st->num_channels = st->caps->num_channels;
> > +
> > + /* Check if touchscreen is supported. */
> > + if (st->caps->has_ts) {
> > + ret = at91_adc_probe_dt_ts(node, st, &idev->dev);
> > + if (ret)
> > + return ret;
> > + }
> >
> > platform_set_drvdata(pdev, idev);
> >
> > @@ -1091,7 +1072,6 @@ static int at91_adc_probe(struct platform_device
> > *pdev)
> > if (IS_ERR(st->reg_base))
> > return PTR_ERR(st->reg_base);
> >
> > -
> Stray change that shouldn't be in this patch ideally but not that important.
>
I know I sneaked this one in but I didn't want to have that as a separate
patch ;). I'll drop it.
