On Fri, Mar 6, 2020 at 1:18 PM Marc Zyngier <maz@xxxxxxxxxx> wrote:
> On 2020-03-06 12:12, Linus Walleij wrote:
> > +static void msm_gpio_irq_eoi(struct irq_data *d)
> > +{
> > + if (d->parent_data)
> > + irq_chip_eoi_parent(d);
> > +}
> > +
(...)
> > - pctrl->irq_chip.irq_eoi = irq_chip_eoi_parent;
> > + pctrl->irq_chip.irq_eoi = msm_gpio_irq_eoi;
> > pctrl->irq_chip.irq_set_type = msm_gpio_irq_set_type;
> > pctrl->irq_chip.irq_set_wake = msm_gpio_irq_set_wake;
> > pctrl->irq_chip.irq_request_resources = msm_gpio_irq_reqres;
>
> Long term, it may me better to offer a different set of callbacks
> for interrupts that are terminated at the pinctrl level.
Yeah, it's gpiolib here actually. Pin control is pretty much innocent,
it's just that the driver doubles as a GPIO driver with an irqchip
inside it.
> In the meantime, and as a fix:
>
> Acked-by: Marc Zyngier <maz@xxxxxxxxxx>
>
> I assume you'll take this via the pinctrl tree?
Yeps, thanks Marc!
Yours,
Linus Walleij
