On Sun, 28 Nov 2021 20:40:33 +0100
Lars-Peter Clausen <lars@xxxxxxxxxx> wrote:
> On 11/28/21 8:26 PM, Jonathan Cameron wrote:
> > static int ad5755_probe(struct spi_device *spi)
> > {
> > @@ -751,8 +736,8 @@ static int ad5755_probe(struct spi_device *spi)
> >
> > mutex_init(&st->lock);
> >
> > - if (spi->dev.of_node)
> > - pdata = ad5755_parse_dt(&spi->dev);
> > + if (dev_fwnode(&spi->dev))
> > + pdata = ad5755_parse_fw(&spi->dev);
> > else
> > pdata = spi->dev.platform_data;
> Considering that it is possible to supply fw_node data through platform
> files we as well just completely drop platform_data support. And make
> the platform data struct local to this file.
>
Good point. I'm in two minds about whether there is any advantage in
leaving the enum definitions in the platform_data header.
They aren't used in the dt binding (which includes an unrelated header - I'll
fix that as well..), so I'm not sure what the logic of exposing them for other
firmware bindings would be...
I'll squash them directly into the c file for now and we can always put them
back in a header if it turns out someone needs them later.
Jonathan
