On Tue, May 25, 2021 at 11:53:21PM +0100, Daniel Scally wrote:
> On 21/05/2021 13:57, Andy Shevchenko wrote:
...
> >> +static const struct regulator_ops int3472_gpio_regulator_ops;
> > Hmm... Can you use 'reg-fixed-voltage' platform device instead?
> >
> > One example, although gone from upstream, but available in the tree, I can
> > point to is this:
> >
> > git log -p -- arch/x86/platform/intel-mid/device_libs/platform_bcm43xx.c
> >
> > It uses constant structures, but I think you may dynamically generate the
> > necessary ones.
> >
>
> I can experiment with this, though one thing is we have no actual idea
> what voltages these are supplying...it doesn't look like that matters
> from drivers/regulator/fixed.c, but I'd have to try it to be sure.
I believe it is likely 1.8v. But I can check if I have the schematics of a
reference design for something like this.
...
> >> + if (int3472->clock.ena_gpio) {
> >> + ret = skl_int3472_register_clock(int3472);
> >> + if (ret)
> >> + goto out_free_res_list;
> >> + } else {
> > Hmm... Have I got it correctly that we can't have ena_gpio && led_gpio together?
>
>
> No, just that we can only have led_gpio if we also have ena_gpio (at
> least that's the intention...)
Okay, perhaps then a comment above?
--
With Best Regards,
Andy Shevchenko
