Re: [PATCH for v5.9] mm/page_alloc: handle a missing case for memalloc_nocma_{save/restore} APIs

[Vlastimil Babka]

On 8/26/20 7:12 AM, Joonsoo Kim wrote:
> 2020년 8월 25일 (화) 오후 6:43, Vlastimil Babka <vbabka@xxxxxxx>님이 작성:
>>
>>
>> On 8/25/20 6:59 AM, js1304@xxxxxxxxx wrote:
>> > From: Joonsoo Kim <iamjoonsoo.kim@xxxxxxx>
>> >
>> > memalloc_nocma_{save/restore} APIs can be used to skip page allocation
>> > on CMA area, but, there is a missing case and the page on CMA area could
>> > be allocated even if APIs are used. This patch handles this case to fix
>> > the potential issue.
>> >
>> > Missing case is an allocation from the pcplist. MIGRATE_MOVABLE pcplist
>> > could have the pages on CMA area so we need to skip it if ALLOC_CMA isn't
>> > specified.
>> >
>> > This patch implements this behaviour by checking allocated page from
>> > the pcplist rather than skipping an allocation from the pcplist entirely.
>> > Skipping the pcplist entirely would result in a mismatch between watermark
>> > check and actual page allocation.
>>
>> Are you sure? I think a mismatch exists already. Pages can be on the pcplist but
>> they are not considered as free in the watermark check. So passing watermark
>> check means there should be also pages on free lists. So skipping pcplists would
>> be safe, no?
> 
> You are right.
> 
>> > And, it requires to break current code
>> > layering that order-0 page is always handled by the pcplist. I'd prefer
>> > to avoid it so this patch uses different way to skip CMA page allocation
>> > from the pcplist.
>>
>> Well it would be much simpler and won't affect most of allocations. Better than
>> flushing pcplists IMHO.
> 
> Hmm...Still, I'd prefer my approach. There are two reasons. First,
> layering problem
> mentioned above. In rmqueue(), there is a code for MIGRATE_HIGHATOMIC.
> As the name shows, it's for high order atomic allocation. But, after
> skipping pcplist
> allocation as you suggested, we could get there with order 0 request.
> We can also
> change this code, but, I'd hope to maintain current layering. Second,
> a performance
> reason. After the flag for nocma is up, a burst of nocma allocation
> could come. After
> flushing the pcplist one times, we can use the free page on the
> pcplist as usual until
> the context is changed.

Both solutions are ugly and we should have CMA in ZONE_MOVABLE or get rid of it
completely. Let's CC Mel what he thinks.

> How about my reasoning?
> 
> Thanks.
>