Re: [RFC PATCH v2 16/17] irqchip/riscv-rpmi-sysmsi: Add ACPI support

[Andy Shevchenko]

On Mon, Feb 03, 2025 at 02:19:05PM +0530, Anup Patel wrote:
> From: Sunil V L <sunilvl@xxxxxxxxxxxxxxxx>
> 
> Add ACPI support for the RISC-V RPMI system MSI based irqchip driver.

...

> +	if (!is_of_node(dev->fwnode)) {

Please, use dev_fwnode(),

But why do you need this? Can't the below simply become a no-op without
this check?

> +		rc = riscv_acpi_get_gsi_info(dev->fwnode, &priv->gsi_base, &id,

Ditto.

> +					     &priv->nr_irqs, NULL);
> +		if (rc) {
> +			dev_err(dev, "failed to find GSI mapping\n");
> +			return rc;
> +		}
> +	}

...

>  		 * then we need to set it explicitly before using any platform
>  		 * MSI functions.
>  		 */
> -		if (is_of_node(dev->fwnode))
> +		if (is_of_node(dev->fwnode)) {
>  			of_msi_configure(dev, to_of_node(dev->fwnode));
> +		} else {
> +			struct irq_domain *msi_domain;
> +
> +			msi_domain = irq_find_matching_fwnode(imsic_acpi_get_fwnode(dev),
> +							      DOMAIN_BUS_PLATFORM_MSI);

> +			if (msi_domain)

Hmm... The OF case above assumes this check is not needed. Why is it special
otherwise?

> +				dev_set_msi_domain(dev, msi_domain);
> +		}
>  
>  		if (!dev_get_msi_domain(dev))

Even here you have a check for NULL, so I believe the conditional is simply
redundant.

...

> +#ifdef CONFIG_ACPI

> +	if (!acpi_disabled)

Why?

> +		acpi_dev_clear_dependencies(ACPI_COMPANION(dev));
> +#endif

...

> +#ifdef CONFIG_ACPI

Drop this ugly ifdeffery along with ACPI_PTR(). They are more harmful than
useful.

...

> +static const struct acpi_device_id acpi_rpmi_sysmsi_match[] = {
> +	{ "RSCV0006", 0 },

Drop ', 0' part as it may be converted to a pointer in the future.

> +	{}
> +};
> +MODULE_DEVICE_TABLE(acpi, acpi_rpmi_sysmsi_match);
> +#endif

-- 
With Best Regards,
Andy Shevchenko