[PATCH] gpio: Device tree support for LPC32xx

[stigge@antcom.de (Roland Stigge)]

Hi Arnd,

thanks for your further suggestions. I'm integrating them, comments below:

On 04/03/2012 10:29 AM, Arnd Bergmann wrote:
>> +- gpio-lines: Number of GPIOs in that subnode/GPIO group
> 
> While I suggested the gpio-lines property, I'm not sure if it's worth
> including it when you don't actually use it.

Right. I'm actually removing it now since for enabling individual GPIO 
lines selectively, it's not flexible enough, anyway: only the first n 
lines of a GPIO "group" could be chosen.

However, I will keep (and document) the possible status="disabled" 
property to disable whole GPIO groups of LPC32xx, if really necessary.

Looks like all LPC32xx variants have and will have the same GPIO layout.

To fix disruptiveness of this patch, I'll make it support both DT and 
non-DT so it doesn't depend on LPC32xx being switched together 
atomically. (Posting below).

Roland



        if (pdev->dev.of_node) {
                for_each_child_of_node(pdev->dev.of_node, node) {
                        if (of_device_is_available(node)) {
                                u32 index;
                                struct gpio_chip *chip;
                                if (of_property_read_u32(node, "reg", &index) <  0)
                                        continue;
                                if (index >= ARRAY_SIZE(lpc32xx_gpiochip))
                                        continue;
                                chip = &lpc32xx_gpiochip[index].chip;
                                chip->of_node = of_node_get(node);
                                gpiochip_add(chip);
                        }
                }
        } else {
                for (i = 0; i < ARRAY_SIZE(lpc32xx_gpiochip); i++)
                	gpiochip_add(&lpc32xx_gpiochip[i].chip);
        }