From mboxrd@z Thu Jan  1 00:00:00 1970
From: lee.jones@linaro.org (Lee Jones)
Date: Fri, 18 May 2012 13:36:54 +0100
Subject: [PATCH 12/18] ARM: ux500: Disable platform setup of the ab8500
 when DT is enabled
In-Reply-To: <201205181112.17254.arnd@arndb.de>
References: <1337262323-27692-1-git-send-email-lee.jones@linaro.org>
 <201205181038.36949.arnd@arndb.de> <4FB62BE5.7030003@linaro.org>
 <201205181112.17254.arnd@arndb.de>
Message-ID: <4FB64266.6040301@linaro.org>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

On 18/05/12 12:12, Arnd Bergmann wrote:
> On Friday 18 May 2012, Lee Jones wrote:
>>> Well, as far as I can tell, you duplicated a lot of functions or
>>> data structures so you can remove the parts one by one, but when you
>>> duplicated snowball_platform_devs[] you did not also duplicate
>>> mop500_platform_devs[].
>>
>> That's true, but why would I have done? I would think that whoever
>> writes the DT for that platform would do that?
>
> In order to not break them when you add the nodes to their device tree? ;-)
>
>> So I see four ways round this: 1) disable all the nodes in db8500.dtsi
>> and force snowball.dts and mop500.dts et. al to re-enable them as
>> appropriate. 2) duplicate mop500_platform_devs[] for DT booting and
>> strip out the ab8500 entry, essentially only leaving in only
>> mop500_gpio_keys_device, 3) DT gpio-keys for the mop500 and do away with
>> platform_add()ing mop500_platform_devs[] in u8500_init_machine()
>> altogether. Or my personal favorite 4) as there is no DT or compatible
>> entry for st-ericsson,mop500 or st-ericsson,hrefv60+ yet, don't worry
>> about it and let whoever enables DT for those platforms (me?) to deal
>> with at a later date.
>
> Well, you've got a point there that no device tree has been written
> yet for those boards. I think my preference would be 2), since you can
> do that in a way that actually removes code (see below), but it's also
> fine if you want to leave it as is for now.
>
> 	Arnd
>
> diff --git a/arch/arm/mach-ux500/board-mop500.c b/arch/arm/mach-ux500/board-mop500.c
> index 182dc4e..4ca56615 100644
> --- a/arch/arm/mach-ux500/board-mop500.c
> +++ b/arch/arm/mach-ux500/board-mop500.c
> @@ -806,8 +806,7 @@ static void __init u8500_init_machine(void)
>
>   	parent = u8500_init_devices();
>
> -	for (i = 0; i<  ARRAY_SIZE(mop500_platform_devs); i++)
> -		mop500_platform_devs[i]->dev.parent = parent;
> +	mop500_gpio_keys_device.parent = parent;
>   	for (i = 0; i<  ARRAY_SIZE(snowball_platform_devs); i++)
>   		snowball_platform_devs[i]->dev.parent = parent;
>
> @@ -816,9 +815,7 @@ static void __init u8500_init_machine(void)
>
>   	if (of_machine_is_compatible("st-ericsson,mop500")) {
>   		mop500_gpio_keys[0].gpio = GPIO_PROX_SENSOR;
> -
> -		platform_add_devices(mop500_platform_devs,
> -				ARRAY_SIZE(mop500_platform_devs));
> +		platform_device_register(&mop500_gpio_keys_device);
>
>   		mop500_sdi_init(parent);
>
> @@ -839,8 +836,7 @@ static void __init u8500_init_machine(void)
>   		 * instead.
>   		 */
>   		mop500_gpio_keys[0].gpio = HREFV60_PROX_SENSE_GPIO;
> -		platform_add_devices(mop500_platform_devs,
> -				ARRAY_SIZE(mop500_platform_devs));
> +		platform_device_register(&mop500_gpio_keys_device);
>
>   		hrefv60_sdi_init(parent);

This is nice. Can I leave it for a now and submit this with the rest of 
my clean-ups? I have a nice TODO list now that this would fit nicely 
into. The way I see it, as long as this is submitted before we add a 
mop500.dts then it's not going to cause anyone issues.

-- 
Lee Jones
Linaro ST-Ericsson Landing Team Lead
M: +44 77 88 633 515
Linaro.org ? Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog