Re: [PATCH v2 09/10] mm: migrate_device: use more folio in migrate_device_unmap()

[Kefeng Wang <wangkefeng.wang@huawei.com>]

On 2024/4/25 17:31, David Hildenbrand wrote:
> On 25.04.24 10:40, Kefeng Wang wrote:
>> The page for migrate_device_unmap() already has a reference, so it is
>> safe to convert the page to folio to save a few calls to compound_head(),
>> which removes the last isolate_lru_page() call.
>>
>> Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com>
>> ---
>>   mm/migrate_device.c | 16 ++++++++--------
>>   1 file changed, 8 insertions(+), 8 deletions(-)
>>
>> diff --git a/mm/migrate_device.c b/mm/migrate_device.c
>> index a68616c1965f..423d71ad736a 100644
>> --- a/mm/migrate_device.c
>> +++ b/mm/migrate_device.c
>> @@ -379,33 +379,33 @@ static unsigned long 
>> migrate_device_unmap(unsigned long *src_pfns,
>>               continue;
>>           }
>> +        folio = page_folio(page);
>>           /* ZONE_DEVICE pages are not on LRU */
>> -        if (!is_zone_device_page(page)) {
>> -            if (!PageLRU(page) && allow_drain) {
>> +        if (!folio_is_zone_device(folio)) {
>> +            if (!folio_test_lru(folio) && allow_drain) {
>>                   /* Drain CPU's lru cache */
>>                   lru_add_drain_all();
>>                   allow_drain = false;
>>               }
>> -            if (!isolate_lru_page(page)) {
>> +            if (!folio_isolate_lru(folio)) {
>>                   src_pfns[i] &= ~MIGRATE_PFN_MIGRATE;
>>                   restore++;
>>                   continue;
>>               }
>>               /* Drop the reference we took in collect */
>> -            put_page(page);
>> +            folio_put(folio);
>>           }
>> -        folio = page_folio(page);
>>           if (folio_mapped(folio))
>>               try_to_migrate(folio, 0);
>>           if (page_mapped(page) ||
> 
> folio_mapped(), just as above

ah, don't know why missing this one, will fix, thanks.

>