Re: [PATCH v3 2/2] gpio: loongson: add firmware offset parse support

[Yinbo Zhu]

在 2023/8/10 下午4:27, Linus Walleij 写道:
> Hi Yinbo,
>
> thanks for your patch!
>
> On Mon, Aug 7, 2023 at 9:41 AM Yinbo Zhu <zhuyinbo@xxxxxxxxxxx> wrote:
>
> > Loongson GPIO controllers come in multiple variants that are compatible
> > except for certain register offset values.  Add support for device
> > properties allowing to specify them in ACPI or DT.
> >
> > Signed-off-by: Yinbo Zhu <zhuyinbo@xxxxxxxxxxx>
>
> (...)
> > @@ -26,6 +26,7 @@ struct loongson_gpio_chip_data {
> >          unsigned int            conf_offset;
> >          unsigned int            out_offset;
> >          unsigned int            in_offset;
> > +       unsigned int            inten_offset;
>
> Consider just changing all of these from unsigned int to u32.


okay, I got it.

> (...)
> > +       if (device_property_read_u32(dev, "loongson,gpio-conf-offset", (u32 *)&d->conf_offset)
> > +           || device_property_read_u32(dev, "loongson,gpio-in-offset", (u32 *)&d->in_offset)
> > +           || device_property_read_u32(dev, "loongson,gpio-out-offset", (u32 *)&d->out_offset)
> > +           || device_property_read_u32(dev, "loongson,gpio-ctrl-mode", (u32 *)&d->mode))
>
> Because then you can get rid of this annoying forest of cast.


Change offset to u32 and here still need use a (u32 *) cast, because the
chip_data is const type so &chip_data->offset will be (const u32 *) type
and need a (u32 *) cast.

> I'm fine with doing this change in this patch without a need for a separate
> refactoring, as it's just a contained driver and clearly just about typing.


okay, I got it.

Thanks,
Yinbo.