Re: [PATCH] gpio: pl061: Support implementations without GPIOINTR line

[Linus Walleij <linus.walleij@xxxxxxxxxx>]

Hi Alexander,

I think I answered some stuff around this patch in my previous
mail but just reiterating so it's clear:

On Wed, Mar 17, 2021 at 4:59 PM Alexander A Sverdlin
<alexander.sverdlin@xxxxxxxxx> wrote:

> @@ -426,6 +426,7 @@ config GPIO_PL061
>         depends on ARM_AMBA
>         select IRQ_DOMAIN
>         select GPIOLIB_IRQCHIP
> +       select IRQ_DOMAIN_HIERARCHY

I think this needs to be optional otherwise you activate hierarchical
IRQs on a lot of systems that don't have it.

select IRQ_DOMAIN_HIERARCHY if ARCH_OMAP_...

This leads to having to use some if IS_ENABLED and
maybe even ifdef to make it compile without hierarchies.

> +       if (!adev->irq[PL061_GPIO_NR - 1]) {
> +               WARN_ON(adev->irq[1]);
> +
> +               girq->parent_handler = pl061_irq_handler;
> +               girq->num_parents = 1;
> +               girq->parents = devm_kcalloc(dev, 1, sizeof(*girq->parents),
> +                                            GFP_KERNEL);
> +               if (!girq->parents)
> +                       return -ENOMEM;
> +               girq->parents[0] = irq;
> +       } else {
> +               girq->fwnode = dev->fwnode;
> +               girq->parent_domain =
> +                       irq_get_irq_data(adev->irq[PL061_GPIO_NR - 1])->domain;
> +               girq->child_to_parent_hwirq = pl061_child_to_parent_hwirq;
> +       }

This is starting to look right :)

But use the top-level board DT compatible to determine that
hiearchy is needed, and implement a per-soc child_to_parent_hwirq()
and do not attempt to get the IRQs from the device tree.

Yours,
Linus Walleij