Re: [PATCH RFC 3/4] mm/page_alloc: always move pages to the tail of the freelist in unset_migratetype_isolate()

[David Hildenbrand]

On 24.09.20 13:13, Vlastimil Babka wrote:
> On 9/16/20 8:34 PM, David Hildenbrand wrote:
>> Page isolation doesn't actually touch the pages, it simply isolates
>> pageblocks and moves all free pages to the MIGRATE_ISOLATE freelist.
>>
>> We already place pages to the tail of the freelists when undoing
>> isolation via __putback_isolated_page(), let's do it in any case
>> (e.g., if order == pageblock_order) and document the behavior.
>>
>> This change results in all pages getting onlined via online_pages() to
>> be placed to the tail of the freelist.
>>
>> Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
>> Cc: Alexander Duyck <alexander.h.duyck@xxxxxxxxxxxxxxx>
>> Cc: Mel Gorman <mgorman@xxxxxxxxxxxxxxxxxxx>
>> Cc: Michal Hocko <mhocko@xxxxxxxxxx>
>> Cc: Dave Hansen <dave.hansen@xxxxxxxxx>
>> Cc: Vlastimil Babka <vbabka@xxxxxxx>
>> Cc: Wei Yang <richard.weiyang@xxxxxxxxxxxxxxxxx>
>> Cc: Oscar Salvador <osalvador@xxxxxxx>
>> Cc: Mike Rapoport <rppt@xxxxxxxxxx>
>> Cc: Scott Cheloha <cheloha@xxxxxxxxxxxxx>
>> Cc: Michael Ellerman <mpe@xxxxxxxxxxxxxx>
>> Signed-off-by: David Hildenbrand <david@xxxxxxxxxx>
>> ---
>>  include/linux/page-isolation.h |  2 ++
>>  mm/page_alloc.c                | 36 +++++++++++++++++++++++++++++-----
>>  mm/page_isolation.c            |  8 ++++++--
>>  3 files changed, 39 insertions(+), 7 deletions(-)
>>
>> diff --git a/include/linux/page-isolation.h b/include/linux/page-isolation.h
>> index 572458016331..a36be2cf4dbb 100644
>> --- a/include/linux/page-isolation.h
>> +++ b/include/linux/page-isolation.h
>> @@ -38,6 +38,8 @@ struct page *has_unmovable_pages(struct zone *zone, struct page *page,
>>  void set_pageblock_migratetype(struct page *page, int migratetype);
>>  int move_freepages_block(struct zone *zone, struct page *page,
>>  				int migratetype, int *num_movable);
>> +int move_freepages_block_tail(struct zone *zone, struct page *page,
>> +			      int migratetype);
>>  
>>  /*
>>   * Changes migrate type in [start_pfn, end_pfn) to be MIGRATE_ISOLATE.
>> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
>> index bba9a0f60c70..75b0f49b4022 100644
>> --- a/mm/page_alloc.c
>> +++ b/mm/page_alloc.c
>> @@ -899,6 +899,15 @@ static inline void move_to_free_list(struct page *page, struct zone *zone,
>>  	list_move(&page->lru, &area->free_list[migratetype]);
>>  }
>>  
>> +/* Used for pages which are on another list */
>> +static inline void move_to_free_list_tail(struct page *page, struct zone *zone,
>> +					  unsigned int order, int migratetype)
>> +{
>> +	struct free_area *area = &zone->free_area[order];
>> +
>> +	list_move_tail(&page->lru, &area->free_list[migratetype]);
>> +}
> 
> There are just 3 callers of move_to_free_list() before this patch, I would just
> add the to_tail parameter there instead of new wrapper. For callers with
> constant parameter, the inline will eliminate it anyway.
> 

So, I'll leave this as is for now, it nicely pairs with
add_to_free_list()/add_to_free_list_tail() and ...

[...]

>> +int move_freepages_block_tail(struct zone *zone, struct page *page,
>> +			      int migratetype)
>> +{
>> +	return __move_freepages_block(zone, page, migratetype, NULL, true);
>>  }
> 

Drop these wrappers, converting callers of move_freepages_block().

Thanks!


-- 
Thanks,

David / dhildenb