Re: [PATCH v4 5/8] mfd: lpc_ich: Add support for pinctrl in non-ACPI system

[Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>]

On Tue, Feb 15, 2022 at 05:29:51PM +0000, Lee Jones wrote:
> On Tue, 15 Feb 2022, Andy Shevchenko wrote:
> > On Tue, Feb 15, 2022 at 04:54:00PM +0000, Lee Jones wrote:
> > > On Mon, 31 Jan 2022, Andy Shevchenko wrote:

> > Thank you for the review, my answers below.

...

> > > > +static struct resource apl_gpio_resources[APL_GPIO_NR_DEVICES][2] = {
> > > > +	[APL_GPIO_NORTH] = {
> > > > +		DEFINE_RES_MEM(APL_GPIO_NORTH_OFFSET, 0x1000),
> > > 
> > > Are these 0x1000's being over-written in lpc_ich_init_pinctrl()?
> > > 
> > > If so, why pre-initialise?
> > 
> > You mean to pre-initialize the offsets, but leave the length to be added
> > in the function? It can be done, but it feels inconsistent, since we would
> > have offsets and lengths in different places for the same thingy. That said,
> > I prefer current way for the sake of consistency.
> 
> Don't you over-write this entry entirely?
> 
>   for (i = 0; i < ARRAY_SIZE(apl_gpio_devices); i++) {
>         struct resource *mem = &apl_gpio_resources[i][0];
> 
>         /* Fill MEM resource */
>         mem->start += base.start;
>         mem->end += base.start;
>         mem->flags = base.flags;
>   }
> 
> Oh wait, you're just adding the base value to the offsets.
> 
> In which case that comment is also confusing!

I have realised that in current form it has a bug (*), so I re-do a bit the way
that comment stays and the actual actions will be to *fill* the resource.

*) unbinding and binding back will bring us to the completely wrong resources.

-- 
With Best Regards,
Andy Shevchenko