Re: [RFC PATCH 14/26] mm: compaction: simplify should_compact_retry()

[Johannes Weiner <hannes@cmpxchg.org>]

On Tue, Apr 25, 2023 at 08:56:47AM +0800, Huang, Ying wrote:
> Johannes Weiner <hannes@cmpxchg.org> writes:
> 
> > The different branches for retry are unnecessarily complicated. There
> > is really only three outcomes: progress, skipped, failed. Also, the
> > retry counter only applies to loops that made progress, move it there.
> >
> > Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
> > ---
> >  mm/page_alloc.c | 60 +++++++++++++++++--------------------------------
> >  1 file changed, 20 insertions(+), 40 deletions(-)
> >
> > diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> > index c3b7dc479936..18fa2bbba44b 100644
> > --- a/mm/page_alloc.c
> > +++ b/mm/page_alloc.c
> > @@ -4608,7 +4608,6 @@ should_compact_retry(struct alloc_context *ac, int order, int alloc_flags,
> >  		     enum compact_priority *compact_priority,
> >  		     int *compaction_retries)
> >  {
> > -	int max_retries = MAX_COMPACT_RETRIES;
> >  	int min_priority;
> >  	bool ret = false;
> >  	int retries = *compaction_retries;
> > @@ -4621,19 +4620,27 @@ should_compact_retry(struct alloc_context *ac, int order, int alloc_flags,
> >  		return false;
> >  
> >  	/*
> > -	 * Compaction managed to coalesce some page blocks, but the
> > -	 * allocation failed presumably due to a race. Retry some.
> > +	 * Compaction coalesced some page blocks, but the allocation
> > +	 * failed, presumably due to a race. Retry a few times.
> >  	 */
> > -	if (compact_result == COMPACT_SUCCESS)
> > -		(*compaction_retries)++;
> > +	if (compact_result == COMPACT_SUCCESS) {
> > +		int max_retries = MAX_COMPACT_RETRIES;
> >  
> > -	/*
> > -	 * All zones were scanned completely and still no result. It
> > -	 * doesn't really make much sense to retry except when the
> > -	 * failure could be caused by insufficient priority
> > -	 */
> > -	if (compact_result == COMPACT_COMPLETE)
> > -		goto check_priority;
> > +		/*
> > +		 * !costly requests are much more important than
> > +		 * __GFP_RETRY_MAYFAIL costly ones because they are de
> > +		 * facto nofail and invoke OOM killer to move on while
> > +		 * costly can fail and users are ready to cope with
> > +		 * that. 1/4 retries is rather arbitrary but we would
> > +		 * need much more detailed feedback from compaction to
> > +		 * make a better decision.
> > +		 */
> > +		if (order > PAGE_ALLOC_COSTLY_ORDER)
> > +			max_retries /= 4;
> > +
> > +		ret = ++(*compaction_retries) <= MAX_COMPACT_RETRIES;
>                                                  ~~~~~~~~~~~~~~~~~~~
> 
> Should be max_retries?

Good catch. max_retries is deleted in a later patch, but this one
should be fixed regardless. Thanks, I will correct it.