[RFC PATCH 07/15] dt: uartlite: merge platform and of_platform driver bindings

Grant Likely grant.likely at secretlab.ca
Thu Feb 24 05:10:42 EST 2011


On Wed, Feb 23, 2011 at 09:58:01AM +0100, Peter Korsgaard wrote:
> >>>>> "Grant" == Grant Likely <grant.likely at secretlab.ca> writes:
> 
>  Grant> of_platform_driver is getting removed, and a single platform_driver
>  Grant> can now support both devicetree and non-devicetree use cases.  This
>  Grant> patch merges the two driver registrations.
> 
> Wee!
> 
>  Grant>  static int __devinit ulite_probe(struct platform_device *pdev)
>  Grant>  {
>  Grant>  	struct resource *res, *res2;
>  Grant> +	const __be32 *prop;
>  Grant> +	int id = pdev->id;
>  
>  Grant>  	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
>  Grant>  	if (!res)
>  Grant> @@ -583,7 +586,13 @@ static int __devinit ulite_probe(struct platform_device *pdev)
>  Grant>  	if (!res2)
>  Grant>  		return -ENODEV;
>  
>  Grant> -	return ulite_assign(&pdev->dev, pdev->id, res->start, res2->start);
>  Grant> +#ifdef CONFIG_OF
>  Grant> +	prop = of_get_property(pdev->dev.of_node, "port-number", NULL);
>  Grant> +	if (prop)
>  Grant> +		id = be32_to_cpup(prop);
>  Grant> +#endif
>  Grant> +
>  Grant> +	return ulite_assign(&pdev->dev, id, res->start, res2->start);
>  Grant>  }
> 
> Isn't this going to give a build warning about prop being unused on
> !CONFIG_OF builds? Perhaps move it down into the CONFIG_OF conditional?

I'd actually like to be rid of 'port-number' entirely; but that's
another day's debate.  I'll fix this up.

g.



More information about the Linuxppc-dev mailing list