mm, page_alloc: check once if a zone has isolated pageblocks
authorMel Gorman <mgorman@techsingularity.net>
Fri, 20 May 2016 00:13:58 +0000 (17:13 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Fri, 20 May 2016 02:12:14 +0000 (19:12 -0700)
When bulk freeing pages from the per-cpu lists the zone is checked for
isolated pageblocks on every release.  This patch checks it once per
drain.

[mgorman@techsingularity.net: fix locking radce, per Vlastimil]
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/page_alloc.c

index 48afc1a42bbd442f70d972af372dfe3a24517179..a3b7eb86f9129e0b87d3ede5da2bddddb4f30ac2 100644 (file)
@@ -831,8 +831,10 @@ static void free_pcppages_bulk(struct zone *zone, int count,
        int batch_free = 0;
        int to_free = count;
        unsigned long nr_scanned;
+       bool isolated_pageblocks;
 
        spin_lock(&zone->lock);
+       isolated_pageblocks = has_isolate_pageblock(zone);
        nr_scanned = zone_page_state(zone, NR_PAGES_SCANNED);
        if (nr_scanned)
                __mod_zone_page_state(zone, NR_PAGES_SCANNED, -nr_scanned);
@@ -870,7 +872,7 @@ static void free_pcppages_bulk(struct zone *zone, int count,
                        /* MIGRATE_ISOLATE page should not go to pcplists */
                        VM_BUG_ON_PAGE(is_migrate_isolate(mt), page);
                        /* Pageblock could have been isolated meanwhile */
-                       if (unlikely(has_isolate_pageblock(zone)))
+                       if (unlikely(isolated_pageblocks))
                                mt = get_pageblock_migratetype(page);
 
                        __free_one_page(page, page_to_pfn(page), zone, 0, mt);