Re: [PATCH -next] mm: Mark undo_dev_pagemap as __maybe_unused

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

On 7/2/19 1:54 PM, Andrew Morton wrote:
> On Tue,  2 Jul 2019 06:02:03 -0700 Guenter Roeck <linux@roeck-us.net> wrote:
> 
>> Several mips builds generate the following build warning.
>>
>> mm/gup.c:1788:13: warning: 'undo_dev_pagemap' defined but not used
>>
>> The function is declared unconditionally but only called from behind
>> various ifdefs. Mark it __maybe_unused.
>>
>> ...
>>
>> --- a/mm/gup.c
>> +++ b/mm/gup.c
>> @@ -1785,7 +1785,8 @@ static inline pte_t gup_get_pte(pte_t *ptep)
>>   }
>>   #endif /* CONFIG_GUP_GET_PTE_LOW_HIGH */
>>   
>> -static void undo_dev_pagemap(int *nr, int nr_start, struct page **pages)
>> +static void __maybe_unused undo_dev_pagemap(int *nr, int nr_start,
>> +					    struct page **pages)
>>   {
>>   	while ((*nr) - nr_start) {
>>   		struct page *page = pages[--(*nr)];
> 
> It's not our preferred way of doing it but yes, it would be a bit of a
> mess and a bit of a maintenance burden to get the ifdefs correct.
> 
That is why I did it here. I understand that some maintainers don't like it,
and I noticed that it wasn't used elsewhere in the file, but it seemed to be
to most straightforward solution.

> And really, __maybe_unused isn't a bad way at all - it ensures that the
> function always gets build-tested and the compiler will remove it so we
> don't have to play the chase-the-ifdefs game.
> 
Yes, it does have its advantages. I like it myself, but usually I would not
impose my opinion on others. In this case, anything else would have been
quite awkward and be prone to never-ending adjustments.

Thanks,
Guenter