Re: [PATCH 2/2] regmap: Allow setting IRQ domain name suffix

[Matti Vaittinen]

On 7/7/24 21:13, Thomas Gleixner wrote:
> On Mon, Jul 01 2024 at 13:59, Matti Vaittinen wrote:
> > +static int regmap_irq_create_domain(struct fwnode_handle *fwnode, int irq_base,
> > +				    const struct regmap_irq_chip *chip,
> > +				    struct regmap_irq_chip_data *d)
> > +{
> > +	struct irq_domain_info info = {
> > +		.fwnode = fwnode,
> > +		.size = irq_base + chip->num_irqs,
> > +		.hwirq_max = irq_base + chip->num_irqs,
>
> This is not correct. irq_base is the linux interrupt number base. The
> first_hwirq argument of irq_domain_create_legacy() is 0.

I tried to pick the logic from the implementation of the 
irq_domain_create_legacy() - but I must've missed something. I will 
re-check this.

> > +		.ops = &regmap_domain_ops,
> > +		.host_data = d,
> > +		.name_suffix = chip->domain_suffix,
> > +	};
> > +
> > +	d->domain = irq_domain_instantiate(&info);
> > +	if (IS_ERR(d->domain)) {
> > +		dev_err(d->map->dev, "Failed to create IRQ domain\n");
> > +		return PTR_ERR(d->domain);
> > +	}
> > +
> > +	if (irq_base)
> > +		irq_domain_associate_many(d->domain, irq_base, 0, chip->num_irqs);
>
> I wonder whether this can be handled at the core. Let me stare at it.

Thanks Thomas! I'll wait for your ideas before re-spinning this series :)

Yours,
	-- Matti


--
Matti Vaittinen
Linux kernel developer at ROHM Semiconductors
Oulu Finland

~~ When things go utterly wrong vim users can always type :help! ~~