Re: [PATCH] Revert "mm/page_isolation: unset migratetype directly for non Buddy page"

[Chen Wandun <chenwandun@huawei.com>]

在 2022/1/17 23:32, David Hildenbrand 写道:
> On 17.01.22 15:27, Chen Wandun wrote:
>> This reverts commit 075782149abff45ee22f27315eced44d02b96779.
>>
>> commit 075782149abf ("mm/page_isolation: unset migratetype directly for
>> non Buddy page") will result memory that should in buddy disappear by
>> mistake. move_freepages_block move all pages in pageblock instead of
>> pages indicated by input parameter, so if input pages is not in buddy
>> but other pages in pageblock is in buddy, it will result in page out of
>> control.
>>
> I was wondering about exactly that condition when skimming over the
> patch but didn't have time review in detail.
when the first valid page in pageblock is non-buddy page and the page order
is not equal pagblock_order,  unset_migratetype_isolate will missing 
move free
page in this pageblock.
>
> Acked-by: David Hildenbrand <david@redhat.com>
>
>> Reported-by: "kernelci.org bot" <bot@kernelci.org>
>> Signed-off-by: Chen Wandun <chenwandun@huawei.com>
>> ---
>>   mm/page_isolation.c | 2 +-
>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/mm/page_isolation.c b/mm/page_isolation.c
>> index 6a0ddda6b3c5..f67c4c70f17f 100644
>> --- a/mm/page_isolation.c
>> +++ b/mm/page_isolation.c
>> @@ -115,7 +115,7 @@ static void unset_migratetype_isolate(struct page *page, unsigned migratetype)
>>   	 * onlining - just onlined memory won't immediately be considered for
>>   	 * allocation.
>>   	 */
>> -	if (!isolated_page && PageBuddy(page)) {
>> +	if (!isolated_page) {
>>   		nr_pages = move_freepages_block(zone, page, migratetype, NULL);
>>   		__mod_zone_freepage_state(zone, nr_pages, migratetype);
>>   	}
>