Re: [PATCH v6 06/11] mm, compaction: more reliably increase direct compaction priority

[Vlastimil Babka]

On 08/16/2016 08:07 AM, Joonsoo Kim wrote:
> > Signed-off-by: Vlastimil Babka <vbabka@xxxxxxx>
> > ---
> >  mm/page_alloc.c | 18 +++++++++++-------
> >  1 file changed, 11 insertions(+), 7 deletions(-)
> >
> > diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> > index fb975cec3518..b28517b918b0 100644
> > --- a/mm/page_alloc.c
> > +++ b/mm/page_alloc.c
> > @@ -3155,13 +3155,8 @@ should_compact_retry(struct alloc_context *ac, int order, int alloc_flags,
> >  	 * so it doesn't really make much sense to retry except when the
> >  	 * failure could be caused by insufficient priority
> >  	 */
> > -	if (compaction_failed(compact_result)) {
> > -		if (*compact_priority > MIN_COMPACT_PRIORITY) {
> > -			(*compact_priority)--;
> > -			return true;
> > -		}
> > -		return false;
> > -	}
> > +	if (compaction_failed(compact_result))
> > +		goto check_priority;
> >
> >  	/*
> >  	 * make sure the compaction wasn't deferred or didn't bail out early
> > @@ -3185,6 +3180,15 @@ should_compact_retry(struct alloc_context *ac, int order, int alloc_flags,
> >  	if (compaction_retries <= max_retries)
> >  		return true;
> >
> > +	/*
> > +	 * Make sure there is at least one attempt at the highest priority
> > +	 * if we exhausted all retries at the lower priorities
> > +	 */
> > +check_priority:
> > +	if (*compact_priority > MIN_COMPACT_PRIORITY) {
> > +		(*compact_priority)--;
> > +		return true;
> > +	}
> >  	return false;
>
> The only difference that this patch makes is increasing priority when
> COMPACT_PARTIAL(COMPACTION_SUCCESS) returns. In that case, we can

Hm it's true that I adjusted this patch from the previous version, 
before realizing that PARTIAL is now SUCCESS.

> usually allocate high-order freepage so we would not enter here. Am I
> missing something? Is it really needed behaviour change?

It will likely be rare when this triggers, when compaction success 
doesn't lead to allocation success due to parallel allocation activity.

> Thanks.