Re: [PATCH v1 2/2] Input: matrix_keypad - replace header inclusions by forward declarations

[Guenter Roeck <linux@roeck-us.net>]

On 10/29/22 21:33, Dmitry Torokhov wrote:
> On Sat, Oct 29, 2022 at 04:02:56PM -0700, Guenter Roeck wrote:
>> On 10/29/22 15:28, Dmitry Torokhov wrote:
>>> On Sat, Oct 29, 2022 at 07:25:51AM -0700, Guenter Roeck wrote:
>>>> On Fri, Sep 23, 2022 at 09:46:32PM +0300, Andy Shevchenko wrote:
>>>>> When the data structure is only referred by pointer, compiler may not need
>>>>> to see the contents of the data type. Thus, we may replace header inclusions
>>>>> by respective forward declarations.
>>>>>
>>>>> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
>>>>> ---
>>>>>    include/linux/input/matrix_keypad.h | 5 +++--
>>>>>    1 file changed, 3 insertions(+), 2 deletions(-)
>>>>>
>>>>> diff --git a/include/linux/input/matrix_keypad.h b/include/linux/input/matrix_keypad.h
>>>>> index 9476768c3b90..b8d8d69eba29 100644
>>>>> --- a/include/linux/input/matrix_keypad.h
>>>>> +++ b/include/linux/input/matrix_keypad.h
>>>>> @@ -3,8 +3,9 @@
>>>>>    #define _MATRIX_KEYPAD_H
>>>>>    #include <linux/types.h>
>>>>> -#include <linux/input.h>
>>>>
>>>> Possibly, but may other drivers rely on those includes.
>>>> This results in widespread build failures such as
>>>>
>>>> Building arm:allmodconfig ... failed
>>>> --------------
>>>> Error log:
>>>> In file included from include/linux/input/samsung-keypad.h:12,
>>>>                    from arch/arm/mach-s3c/keypad.h:12,
>>>>                    from arch/arm/mach-s3c/mach-crag6410.c:57:
>>>> arch/arm/mach-s3c/mach-crag6410.c:183:19: error: 'KEY_VOLUMEUP' undeclared here
>>>
>>> I fixed this particular instance, hopefully it is one of the very
>>> last of them...
>>>
>>
>> Sorry, I didn't bother listing all of them. There is at least one more.
>>
>> Error log:
>> arch/arm/mach-pxa/spitz.c:410:11: error: 'EV_PWR' undeclared here (not in a function)
>>    410 |   .type = EV_PWR,
>>        |           ^~~~~~
>>
>> with arm:pxa_defconfig.
> 
> Ah, I see. Arnd is trying to delete bunch of PXA code, so I'll pull the
> header change until after he merges his, and hopefully spitz will be
> gone.
> 
> Do you have more examples by chance?
> 

No, but no guarantees - there are various other build failures in -next,
and I may have missed some.

Guenter