Re: [PATCH v1 4/6] mfd: intel_quark_i2c_gpio: Convert to use software nodes

[Andy Shevchenko]

On Wed, Jun 10, 2020 at 01:38:54PM +0200, Linus Walleij wrote:
> On Mon, Jun 8, 2020 at 3:43 PM Andy Shevchenko
> <andriy.shevchenko@xxxxxxxxxxxxxxx> wrote:
> 
> > -/* The base GPIO number under GPIOLIB framework */
> > -#define INTEL_QUARK_MFD_GPIO_BASE      8
> 
> OK I see this was around before, sigh.
> So it's not your fault. It was introduced in commit
> 60ae5b9f5cdd8 which I was not involved in reviewing,
> for the record I would have said "no".
> 
> It is exploiting commit 3d2613c4289ff where I allowed
> pdata to set the base so it is anyway my fault for not
> noticing :(
> 
> But me complaining about this doesn't make things better.
> 
> Can we simply DELETE this assignment and just set base to
> -1 in a separate patch before this patch and see what happens? It's
> really unsafe to hardcode base like this.

I don't see easy way to do so, because as I explained in previous mail it will
affect relation (from firmware) to the fact that the numbering of this
controller is static for that platform.

> +       PROPERTY_ENTRY_U32("snps,nr-gpios", 8),
> 
> This is however fine in principle but just use the existing generic
> property "ngpios" and save this custom property.

This is established (not by ACPI!) property. Completely not my fault :-)

-- 
With Best Regards,
Andy Shevchenko