mm, compaction: kcompactd should not ignore pageblock skip

Kcompactd is needlessly ignoring pageblock skip information.  It is
doing MIGRATE_SYNC_LIGHT compaction, which is no more powerful than
MIGRATE_SYNC compaction.

If compaction recently failed to isolate memory from a set of
pageblocks, there is nothing to indicate that kcompactd will be able to
do so, or that it is beneficial from attempting to isolate memory.

Use the pageblock skip hint to avoid rescanning pageblocks needlessly
until that information is reset.

Link: http://lkml.kernel.org/r/alpine.DEB.2.10.1708151638550.106658@chino.kir.corp.google.com
Signed-off-by: David Rientjes <rientjes@google.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Mel Gorman <mgorman@techsingularity.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This commit is contained in:
David Rientjes 2017-11-17 15:26:27 -08:00 committed by Linus Torvalds
parent 09af5ccea2
commit a0647dc920

View File

@ -1928,9 +1928,8 @@ static void kcompactd_do_work(pg_data_t *pgdat)
.total_free_scanned = 0, .total_free_scanned = 0,
.classzone_idx = pgdat->kcompactd_classzone_idx, .classzone_idx = pgdat->kcompactd_classzone_idx,
.mode = MIGRATE_SYNC_LIGHT, .mode = MIGRATE_SYNC_LIGHT,
.ignore_skip_hint = true, .ignore_skip_hint = false,
.gfp_mask = GFP_KERNEL, .gfp_mask = GFP_KERNEL,
}; };
trace_mm_compaction_kcompactd_wake(pgdat->node_id, cc.order, trace_mm_compaction_kcompactd_wake(pgdat->node_id, cc.order,
cc.classzone_idx); cc.classzone_idx);