Re: [PATCH] mm/migrate: rename PAGE_ migration flags to FOLIO_MF_

[David Hildenbrand (Arm)]

On 3/23/26 15:46, Matthew Wilcox wrote:
> On Mon, Mar 23, 2026 at 02:19:37PM +0000, Shivank Garg wrote:
>> +++ b/mm/migrate.c
>> @@ -1135,26 +1135,26 @@ static int move_to_new_folio(struct folio *dst, struct folio *src,
>>   * This is safe because nobody is using it except us.
>>   */
>>  enum {
>> -	PAGE_WAS_MAPPED = BIT(0),
>> -	PAGE_WAS_MLOCKED = BIT(1),
>> -	PAGE_OLD_STATES = PAGE_WAS_MAPPED | PAGE_WAS_MLOCKED,
>> +	FOLIO_MF_WAS_MAPPED = BIT(0),
>> +	FOLIO_MF_WAS_MLOCKED = BIT(1),
>> +	FOLIO_MF_OLD_STATES = FOLIO_MF_WAS_MAPPED | FOLIO_MF_WAS_MLOCKED,
> 
> what does 'MF' stand for?  i mean, presumably not the pop culture
> meaning of MF ...

Migration FLag

> 
>>  	locked = true;
>>  	if (folio_test_mlocked(src))
>> -		old_page_state |= PAGE_WAS_MLOCKED;
>> +		old_folio_state |= FOLIO_MF_WAS_MLOCKED;
> 
> I mean, this doesn't make sense any more.  PAGE_WAS_MLOCKED is a
> sentence that means something.  FOLIO_MF_WAS_MLOCKED doesn't ... unless
> you imagine Samuel L Jackson saying it ...

:)

PAGE_WAS_MLOCKED was suboptimal given that the names overlap with things
like PAGE_SHARED or even PAGE_IS_PRESENT.

So using some prefix to indicate that these flag belong logically
together makes sense. Even if it's not a beautiful sentence.

FOLIO_MIGRATION_* could be done, but "FOLIO_MIGRATION_WAS_MLOCKED" is
also does not make sense. FOLIO_MIGRATION_FLAG_* is mouthful.

-- 
Cheers,

David