From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-pd0-f169.google.com (mail-pd0-f169.google.com [209.85.192.169]) by kanga.kvack.org (Postfix) with ESMTP id 3CD016B0038 for ; Tue, 16 Jun 2015 01:39:08 -0400 (EDT) Received: by pdbnf5 with SMTP id nf5so6384654pdb.2 for ; Mon, 15 Jun 2015 22:39:08 -0700 (PDT) Received: from lgeamrelo02.lge.com (lgeamrelo02.lge.com. [156.147.1.126]) by mx.google.com with ESMTP id ij4si21013548pbc.142.2015.06.15.22.39.06 for ; Mon, 15 Jun 2015 22:39:07 -0700 (PDT) Date: Tue, 16 Jun 2015 14:41:15 +0900 From: Joonsoo Kim Subject: Re: [PATCH 3/6] mm, compaction: encapsulate resetting cached scanner positions Message-ID: <20150616054115.GC12641@js1304-P5Q-DELUXE> References: <1433928754-966-1-git-send-email-vbabka@suse.cz> <1433928754-966-4-git-send-email-vbabka@suse.cz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1433928754-966-4-git-send-email-vbabka@suse.cz> Sender: owner-linux-mm@kvack.org List-ID: To: Vlastimil Babka Cc: Andrew Morton , linux-mm@kvack.org, linux-kernel@vger.kernel.org, Minchan Kim , Mel Gorman , Michal Nazarewicz , Naoya Horiguchi , Christoph Lameter , Rik van Riel , David Rientjes On Wed, Jun 10, 2015 at 11:32:31AM +0200, Vlastimil Babka wrote: > Resetting the cached compaction scanner positions is now done implicitly in > __reset_isolation_suitable() and compact_finished(). Encapsulate the > functionality in a new function reset_cached_positions() and call it explicitly > where needed. > > Signed-off-by: Vlastimil Babka > Cc: Minchan Kim > Cc: Mel Gorman > Cc: Joonsoo Kim > Cc: Michal Nazarewicz > Cc: Naoya Horiguchi > Cc: Christoph Lameter > Cc: Rik van Riel > Cc: David Rientjes > --- > mm/compaction.c | 22 ++++++++++++++-------- > 1 file changed, 14 insertions(+), 8 deletions(-) > > diff --git a/mm/compaction.c b/mm/compaction.c > index 7e0a814..d334bb3 100644 > --- a/mm/compaction.c > +++ b/mm/compaction.c > @@ -207,6 +207,13 @@ static inline bool isolation_suitable(struct compact_control *cc, > return !get_pageblock_skip(page); > } > > +static void reset_cached_positions(struct zone *zone) > +{ > + zone->compact_cached_migrate_pfn[0] = zone->zone_start_pfn; > + zone->compact_cached_migrate_pfn[1] = zone->zone_start_pfn; > + zone->compact_cached_free_pfn = zone_end_pfn(zone); > +} > + > /* > * This function is called to clear all cached information on pageblocks that > * should be skipped for page isolation when the migrate and free page scanner > @@ -218,9 +225,6 @@ static void __reset_isolation_suitable(struct zone *zone) > unsigned long end_pfn = zone_end_pfn(zone); > unsigned long pfn; > > - zone->compact_cached_migrate_pfn[0] = start_pfn; > - zone->compact_cached_migrate_pfn[1] = start_pfn; > - zone->compact_cached_free_pfn = end_pfn; > zone->compact_blockskip_flush = false; Is there a valid reason not to call reset_cached_positions() in __reset_isolation_suitable? You missed one callsite in __compact_pgdat(). if (cc->order == -1) __reset_isolation_suitable(zone); This also needs reset_cached_positions(). Thanks. -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: email@kvack.org