Re: [PATCH 12/16] mm/migration: fix potential page refcounts leak in migrate_pages

[Huang, Ying]

Baolin Wang <baolin.wang@xxxxxxxxxxxxxxxxx> writes:

> Hi Miaohe,
>
> On 3/4/2022 5:34 PM, Miaohe Lin wrote:
>> In -ENOMEM case, there might be some subpages of fail-to-migrate THPs
>> left in thp_split_pages list. We should move them back to migration
>> list so that they could be put back to the right list by the caller
>> otherwise the page refcnt will be leaked here. Also adjust nr_failed
>> and nr_thp_failed accordingly to make vm events account more accurate.
>> Fixes: b5bade978e9b ("mm: migrate: fix the return value of
>> migrate_pages()")
>> Signed-off-by: Miaohe Lin <linmiaohe@xxxxxxxxxx>
>> ---
>>   mm/migrate.c | 9 +++++++++
>>   1 file changed, 9 insertions(+)
>> diff --git a/mm/migrate.c b/mm/migrate.c
>> index e0db06927f02..6c2dfed2ddb8 100644
>> --- a/mm/migrate.c
>> +++ b/mm/migrate.c
>> @@ -1422,6 +1422,15 @@ int migrate_pages(struct list_head *from, new_page_t get_new_page,
>>   				}
>>     				nr_failed_pages += nr_subpages;
>> +				/*
>> +				 * There might be some subpages of fail-to-migrate THPs
>> +				 * left in thp_split_pages list. Move them back to migration
>> +				 * list so that they could be put back to the right list by
>> +				 * the caller otherwise the page refcnt will be leaked.
>> +				 */
>> +				list_splice_init(&thp_split_pages, from);
>> +				nr_failed += retry;
>> +				nr_thp_failed += thp_retry;
>
> Yes, I think we missed this case before, and your patch looks
> right. But we should also update the 'rc' to return the correct number
> of pages that were not migrated, right?

Per my understanding, -ENOMEM should be returned to indicate an fatal
error.

Best Regards,
Huang, Ying