Re: [PATCH] mm/compaction: count pages and stop correctly during page isolation.

From: Zi Yan <ziy@nvidia.com>
To: Michal Hocko <mhocko@suse.com>
Cc: Andrew Morton <akpm@linux-foundation.org>, <linux-mm@kvack.org>,
	Rik van Riel <riel@surriel.com>, <linux-kernel@vger.kernel.org>,
	Vlastimil Babka <vbabka@suse.cz>
Subject: Re: [PATCH] mm/compaction: count pages and stop correctly during page isolation.
Date: Fri, 30 Oct 2020 10:53:52 -0400	[thread overview]
Message-ID: <3B3B176C-F74A-4B5B-B4B2-D527060217C2@nvidia.com> (raw)
In-Reply-To: <20201030144917.GK1478@dhcp22.suse.cz>

[-- Attachment #1: Type: text/plain, Size: 3199 bytes --]

On 30 Oct 2020, at 10:49, Michal Hocko wrote:

> On Fri 30-10-20 10:35:43, Zi Yan wrote:
>> On 30 Oct 2020, at 9:36, Michal Hocko wrote:
>>
>>> On Fri 30-10-20 08:20:50, Zi Yan wrote:
>>>> On 30 Oct 2020, at 5:43, Michal Hocko wrote:
>>>>
>>>>> [Cc Vlastimil]
>>>>>
>>>>> On Thu 29-10-20 16:04:35, Zi Yan wrote:
>>>>>> From: Zi Yan <ziy@nvidia.com>
>>>>>>
>>>>>> In isolate_migratepages_block, when cc->alloc_contig is true, we are
>>>>>> able to isolate compound pages, nr_migratepages and nr_isolated did not
>>>>>> count compound pages correctly, causing us to isolate more pages than we
>>>>>> thought. Use thp_nr_pages to count pages. Otherwise, we might be trapped
>>>>>> in too_many_isolated while loop, since the actual isolated pages can go
>>>>>> up to COMPACT_CLUSTER_MAX*512=16384, where COMPACT_CLUSTER_MAX is 32,
>>>>>> since we stop isolation after cc->nr_migratepages reaches to
>>>>>> COMPACT_CLUSTER_MAX.
>>>>>>
>>>>>> In addition, after we fix the issue above, cc->nr_migratepages could
>>>>>> never be equal to COMPACT_CLUSTER_MAX if compound pages are isolated,
>>>>>> thus page isolation could not stop as we intended. Change the isolation
>>>>>> stop condition to >=.
>>>>>>
>>>>>> Signed-off-by: Zi Yan <ziy@nvidia.com>
>>>>>> ---
>>>>>>  mm/compaction.c | 8 ++++----
>>>>>>  1 file changed, 4 insertions(+), 4 deletions(-)
>>>>>>
>>>>>> diff --git a/mm/compaction.c b/mm/compaction.c
>>>>>> index ee1f8439369e..0683a4999581 100644
>>>>>> --- a/mm/compaction.c
>>>>>> +++ b/mm/compaction.c
>>>>>> @@ -1012,8 +1012,8 @@ isolate_migratepages_block(struct compact_control *cc, unsigned long low_pfn,
>>>>>>
>>>>>>  isolate_success:
>>>>>>  		list_add(&page->lru, &cc->migratepages);
>>>>>> -		cc->nr_migratepages++;
>>>>>> -		nr_isolated++;
>>>>>> +		cc->nr_migratepages += thp_nr_pages(page);
>>>>>> +		nr_isolated += thp_nr_pages(page);
>>>>>
>>>>> Does thp_nr_pages work for __PageMovable pages?
>>>>
>>>> Yes. It is the same as compound_nr() but compiled
>>>> to 1 when THP is not enabled.
>>>
>>> I am sorry but I do not follow. First of all the implementation of the
>>> two is different and also I was asking about __PageMovable which should
>>> never be THP IIRC. Can they be compound though?
>>
>> __PageMovable, non-lru movable pages, can be compound and thp_nr_page cannot
>> be used for it, since when THP is off, thp_nr_page will return the wrong number.
>> I got confused by its name, sorry.
>
> OK, this matches my understanding. Good we are on the same page.
>
>> But __PageMovable is irrelevant to this patch, since we are using
>> __isolate_lru_page to isolate pages. non-lru __PageMovable should not appear
>> after isolate_succes. thp_nr_pages can be used here.
>
> But this is still not clear to me. __PageMovable pages are isolated by
> isolate_movable_page and then jump to isolate_succes. Does that somehow
> changes the nature of the page being compound or tat thp_nr_page would
> start working on those pages.

Ah, I missed that part of the code. If __PageMovable can reach the code, we
should use compound_nr instead. Will send v2 to fix it. Thanks.

—
Best Regards,
Yan Zi

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 854 bytes --]