next build: 304 warnings 0 failures (next/next-20141013)

[Marek Szyprowski]

Hello,

On 2014-10-13 11:12, Arnd Bergmann wrote:
> On Monday 13 October 2014 09:46:18 Russell King - ARM Linux wrote:
>> On Sun, Oct 12, 2014 at 10:32:02PM -0700, Olof's autobuilder wrote:
>>>        1 arch/arm/mach-cns3xxx/pcie.c:313:1: warning: the frame size of 1072 bytes is larger than 1024 bytes [-Wframe-larger-than=]
>> This should be fixed.
> This is an ancient bug, both Mark Brown and I have suggested fixes,
> we just need to apply one of them.
>
>>>        1 arch/arm/mach-omap1/board-htcherald.c:296:2: warning: initialization makes pointer from integer without a cast [enabled by default]
>>>        1 arch/arm/mach-omap1/board-htcherald.c:296:2: warning: (near initialization for 'gpio_leds[0].gpiod') [enabled by default]
>>>        1 arch/arm/mach-omap1/board-htcherald.c:297:2: warning: initialization makes pointer from integer without a cast [enabled by default]
>>>        1 arch/arm/mach-omap1/board-htcherald.c:297:2: warning: (near initialization for 'gpio_leds[1].gpiod') [enabled by default]
>> ...
>>
>> All of these need fixing.
> Hadn't seen this one before, I guess it's because of
> https://lkml.org/lkml/2014/9/16/222.
>
> adding Mika and Linus to Cc, should be easy to fix using
> named initializers like all other 120+ platforms do,
> but to be on the safe side, we could also move the new
> 'gpiod' member to the end of 'struct gpio_led'.
>
>>>        2 drivers/base/dma-contiguous.c:244:2: warning: initialization from incompatible pointer type
>>>        2 drivers/base/dma-contiguous.c:244:2: warning: (near initialization for 'rmem_cma_ops.device_init')
>> This does too.
>>
>>>        3 drivers/base/dma-coherent.c:303:2: warning: initialization from incompatible pointer type
>>>        3 drivers/base/dma-coherent.c:303:2: warning: (near initialization for 'rmem_dma_ops.device_init')
>>>        8 drivers/base/dma-contiguous.c:244:2: warning: initialization from incompatible pointer type [enabled by default]
>>>        8 drivers/base/dma-contiguous.c:244:2: warning: (near initialization for 'rmem_cma_ops.device_init') [enabled by default]
>>>       50 drivers/base/dma-coherent.c:303:2: warning: initialization from incompatible pointer type [enabled by default]
>>>       50 drivers/base/dma-coherent.c:303:2: warning: (near initialization for 'rmem_dma_ops.device_init') [enabled by default]
>> And these.
>>
>> None of these warnings should hit mainline; they should all be fixed now,
>> before any of the changes which caused them are merged.
> Adding Marek as well. No idea how this slipped in, apparently the device_init
> callback declaration has always used 'void' as the return type, while all
> functions assigned to it have always returned 'int' and the caller
> in __reserved_mem_init_node() actually evaluates the return code.

Patches adding of_reservedmem support for dma-contiguous and 
dma-coherent were posted together
with a patch adding a return value. However, for some reasons the latter 
was dropped. My fault
that I've posted that patch in a hurry and missed adding return value 
check in some places.

I've posted a fix and I hope it can get merged as a 'fix':
https://lkml.org/lkml/2014/10/9/178


> This seems to be the obvious fix:
>
> diff --git a/include/linux/of_reserved_mem.h b/include/linux/of_reserved_mem.h
> index 5b5efae09135..3c2cf6fc5e78 100644
> --- a/include/linux/of_reserved_mem.h
> +++ b/include/linux/of_reserved_mem.h
> @@ -16,7 +16,7 @@ struct reserved_mem {
>   };
>   
>   struct reserved_mem_ops {
> -	void	(*device_init)(struct reserved_mem *rmem,
> +	int	(*device_init)(struct reserved_mem *rmem,
>   			       struct device *dev);
>   	void	(*device_release)(struct reserved_mem *rmem,
>   				  struct device *dev);
>
>
> 	Arnd
>

Best regards
-- 
Marek Szyprowski, PhD
Samsung R&D Institute Poland