Re: [PATCH v2 1/1] [PATCH] mm/gup: avoid an unnecessary allocation call for FOLL_LONGTERM cases

[John Hubbard <jhubbard@nvidia.com>]

On 11/6/24 1:23 AM, Oscar Salvador wrote:
> On Mon, Nov 04, 2024 at 07:29:44PM -0800, John Hubbard wrote:
...
> Hi John, thanks for doing this.
> 
> Reviewed-by: Oscar Salvador <osalvador@suse.de>

Thanks for the review!

> 
> Nit below:
> 
...
>> @@ -2363,8 +2399,8 @@ static int migrate_longterm_unpinnable_folios(
>>   		 * calling folio_isolate_lru() which takes a reference so the
>>   		 * folio won't be freed if it's migrating.
>>   		 */
>> -		unpin_folio(folios[i]);
>> -		folios[i] = NULL;
>> +		unpin_folio(pofs_get_folio(pofs, i));
> 
> We already retrieved the folio before, cannot we just bypass
> pofs_get_folio() here?
> 

Yes, you are right. Andrew, can we please add this fixup on top of the commits
in today's mm-hotfixes-unstable (or, let me know if you'd prefer a v3 instead):


diff --git a/mm/gup.c b/mm/gup.c
index 0a22f7def83c..ad0c8922dac3 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -2399,7 +2399,7 @@ migrate_longterm_unpinnable_folios(struct list_head *movable_folio_list,
  		 * calling folio_isolate_lru() which takes a reference so the
  		 * folio won't be freed if it's migrating.
  		 */
-		unpin_folio(pofs_get_folio(pofs, i));
+		unpin_folio(folio);
  		pofs_clear_entry(pofs, i);
  	}
  


thanks,
-- 
John Hubbard