[PATCH v2 2/4] ARM: shmobile: r8a7779: add VIN support

[Sergei Shtylyov]

Hello.

On 22-04-2013 8:57, Simon Horman wrote:

>>> From: Vladimir Barinov <vladimir.barinov at cogentembedded.com>

>>> Add VIN clocks and platform devices for R8A7779 SoC; add function to register
>>> the VIN platform devices.

>>> Signed-off-by: Vladimir Barinov <vladimir.barinov at cogentembedded.com>
>>> [Sergei: added 'id' parameter check to r8a7779_add_vin_device(), renamed some
>>> variables.]
>>> Signed-off-by: Sergei Shtylyov <sergei.shtylyov at cogentembedded.com>

>> [...]

>>> Index: renesas/arch/arm/mach-shmobile/setup-r8a7779.c
>>> ===================================================================
>>> --- renesas.orig/arch/arm/mach-shmobile/setup-r8a7779.c
>>> +++ renesas/arch/arm/mach-shmobile/setup-r8a7779.c
>>> @@ -559,6 +559,33 @@ static struct resource ether_resources[]
>>>   	},
>>>   };
>>> +#define R8A7779_VIN(idx) \
>>> +static struct resource vin##idx##_resources[] = {		\
>>> +	DEFINE_RES_MEM(0xffc50000 + 0x1000 * (idx), 0x1000),	\
>>> +	DEFINE_RES_IRQ(gic_iid(0x5f + (idx))),			\
>>> +};								\
>>> +								\
>>> +static struct platform_device_info vin##idx##_info = {		\

>>     Hm, probably should have marked this as '__initdata'... maybe
>> the resources too.

> That doesn't seem to be the case for other devices in
> that or other shmobile files. Am I missing something
> or should numerous other devices be updated?

    If the device is registered using platform_device_register_*(), it seems 
worth marking the resources, the platfrom data and 'struct 
platform_device_info' as '__initdata' as they're copied to the memory 
allocated from heap anyway and hence not needed past the init phase...

WBR, Sergei