All of lore.kernel.org
 help / color / mirror / Atom feed
From: Vlastimil Babka <vbabka@suse.cz>
To: Michal Hocko <mhocko@kernel.org>,
	Andrew Morton <akpm@linux-foundation.org>
Cc: linux-mm@kvack.org, linux-kernel@vger.kernel.org,
	Mel Gorman <mgorman@techsingularity.net>,
	Joonsoo Kim <iamjoonsoo.kim@lge.com>,
	David Rientjes <rientjes@google.com>,
	Rik van Riel <riel@redhat.com>, Vlastimil Babka <vbabka@suse.cz>
Subject: [PATCH v2 11/18] mm, compaction: add the ultimate direct compaction priority
Date: Tue, 31 May 2016 15:08:11 +0200	[thread overview]
Message-ID: <20160531130818.28724-12-vbabka@suse.cz> (raw)
In-Reply-To: <20160531130818.28724-1-vbabka@suse.cz>

During reclaim/compaction loop, it's desirable to get a final answer from
unsuccessful compaction so we can either fail the allocation or invoke the OOM
killer. However, heuristics such as deferred compaction or pageblock skip bits
can cause compaction to skip parts or whole zones and lead to premature OOM's,
failures or excessive reclaim/compaction retries.

To remedy this, we introduce a new direct compaction priority called
COMPACT_PRIO_SYNC_FULL, which instructs direct compaction to:

- ignore deferred compaction status for a zone
- ignore pageblock skip hints
- ignore cached scanner positions and scan the whole zone

The new priority should get eventually picked up by should_compact_retry() and
this should improve success rates for costly allocations using __GFP_REPEAT,
such as hugetlbfs allocations, and reduce some corner-case OOM's for non-costly
allocations.

Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Acked-by: Michal Hocko <mhocko@suse.com>
---
 include/linux/compaction.h | 3 ++-
 mm/compaction.c            | 5 ++++-
 2 files changed, 6 insertions(+), 2 deletions(-)

diff --git a/include/linux/compaction.h b/include/linux/compaction.h
index 22a5fb9c509c..29dc7c05bd3b 100644
--- a/include/linux/compaction.h
+++ b/include/linux/compaction.h
@@ -6,8 +6,9 @@
  * Lower value means higher priority, analogically to reclaim priority.
  */
 enum compact_priority {
+	COMPACT_PRIO_SYNC_FULL,
+	MIN_COMPACT_PRIORITY = COMPACT_PRIO_SYNC_FULL,
 	COMPACT_PRIO_SYNC_LIGHT,
-	MIN_COMPACT_PRIORITY = COMPACT_PRIO_SYNC_LIGHT,
 	DEF_COMPACT_PRIORITY = COMPACT_PRIO_SYNC_LIGHT,
 	COMPACT_PRIO_ASYNC,
 	INIT_COMPACT_PRIORITY = COMPACT_PRIO_ASYNC
diff --git a/mm/compaction.c b/mm/compaction.c
index af50f20de369..a399e7ca4630 100644
--- a/mm/compaction.c
+++ b/mm/compaction.c
@@ -1586,6 +1586,8 @@ static enum compact_result compact_zone_order(struct zone *zone, int order,
 		.alloc_flags = alloc_flags,
 		.classzone_idx = classzone_idx,
 		.direct_compaction = true,
+		.whole_zone = (prio == COMPACT_PRIO_SYNC_FULL),
+		.ignore_skip_hint = (prio == COMPACT_PRIO_SYNC_FULL)
 	};
 	INIT_LIST_HEAD(&cc.freepages);
 	INIT_LIST_HEAD(&cc.migratepages);
@@ -1631,7 +1633,8 @@ enum compact_result try_to_compact_pages(gfp_t gfp_mask, unsigned int order,
 								ac->nodemask) {
 		enum compact_result status;
 
-		if (compaction_deferred(zone, order)) {
+		if (prio > COMPACT_PRIO_SYNC_FULL
+					&& compaction_deferred(zone, order)) {
 			rc = max_t(enum compact_result, COMPACT_DEFERRED, rc);
 			continue;
 		}
-- 
2.8.3

WARNING: multiple messages have this Message-ID (diff)
From: Vlastimil Babka <vbabka@suse.cz>
To: Michal Hocko <mhocko@kernel.org>,
	Andrew Morton <akpm@linux-foundation.org>
Cc: linux-mm@kvack.org, linux-kernel@vger.kernel.org,
	Mel Gorman <mgorman@techsingularity.net>,
	Joonsoo Kim <iamjoonsoo.kim@lge.com>,
	David Rientjes <rientjes@google.com>,
	Rik van Riel <riel@redhat.com>, Vlastimil Babka <vbabka@suse.cz>
Subject: [PATCH v2 11/18] mm, compaction: add the ultimate direct compaction priority
Date: Tue, 31 May 2016 15:08:11 +0200	[thread overview]
Message-ID: <20160531130818.28724-12-vbabka@suse.cz> (raw)
In-Reply-To: <20160531130818.28724-1-vbabka@suse.cz>

During reclaim/compaction loop, it's desirable to get a final answer from
unsuccessful compaction so we can either fail the allocation or invoke the OOM
killer. However, heuristics such as deferred compaction or pageblock skip bits
can cause compaction to skip parts or whole zones and lead to premature OOM's,
failures or excessive reclaim/compaction retries.

To remedy this, we introduce a new direct compaction priority called
COMPACT_PRIO_SYNC_FULL, which instructs direct compaction to:

- ignore deferred compaction status for a zone
- ignore pageblock skip hints
- ignore cached scanner positions and scan the whole zone

The new priority should get eventually picked up by should_compact_retry() and
this should improve success rates for costly allocations using __GFP_REPEAT,
such as hugetlbfs allocations, and reduce some corner-case OOM's for non-costly
allocations.

Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Acked-by: Michal Hocko <mhocko@suse.com>
---
 include/linux/compaction.h | 3 ++-
 mm/compaction.c            | 5 ++++-
 2 files changed, 6 insertions(+), 2 deletions(-)

diff --git a/include/linux/compaction.h b/include/linux/compaction.h
index 22a5fb9c509c..29dc7c05bd3b 100644
--- a/include/linux/compaction.h
+++ b/include/linux/compaction.h
@@ -6,8 +6,9 @@
  * Lower value means higher priority, analogically to reclaim priority.
  */
 enum compact_priority {
+	COMPACT_PRIO_SYNC_FULL,
+	MIN_COMPACT_PRIORITY = COMPACT_PRIO_SYNC_FULL,
 	COMPACT_PRIO_SYNC_LIGHT,
-	MIN_COMPACT_PRIORITY = COMPACT_PRIO_SYNC_LIGHT,
 	DEF_COMPACT_PRIORITY = COMPACT_PRIO_SYNC_LIGHT,
 	COMPACT_PRIO_ASYNC,
 	INIT_COMPACT_PRIORITY = COMPACT_PRIO_ASYNC
diff --git a/mm/compaction.c b/mm/compaction.c
index af50f20de369..a399e7ca4630 100644
--- a/mm/compaction.c
+++ b/mm/compaction.c
@@ -1586,6 +1586,8 @@ static enum compact_result compact_zone_order(struct zone *zone, int order,
 		.alloc_flags = alloc_flags,
 		.classzone_idx = classzone_idx,
 		.direct_compaction = true,
+		.whole_zone = (prio == COMPACT_PRIO_SYNC_FULL),
+		.ignore_skip_hint = (prio == COMPACT_PRIO_SYNC_FULL)
 	};
 	INIT_LIST_HEAD(&cc.freepages);
 	INIT_LIST_HEAD(&cc.migratepages);
@@ -1631,7 +1633,8 @@ enum compact_result try_to_compact_pages(gfp_t gfp_mask, unsigned int order,
 								ac->nodemask) {
 		enum compact_result status;
 
-		if (compaction_deferred(zone, order)) {
+		if (prio > COMPACT_PRIO_SYNC_FULL
+					&& compaction_deferred(zone, order)) {
 			rc = max_t(enum compact_result, COMPACT_DEFERRED, rc);
 			continue;
 		}
-- 
2.8.3

--
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: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

  parent reply	other threads:[~2016-05-31 13:10 UTC|newest]

Thread overview: 64+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-05-31 13:08 [PATCH v2 00/18] make direct compaction more deterministic Vlastimil Babka
2016-05-31 13:08 ` Vlastimil Babka
2016-05-31 13:08 ` [PATCH v2 01/18] mm, compaction: don't isolate PageWriteback pages in MIGRATE_SYNC_LIGHT mode Vlastimil Babka
2016-05-31 13:08   ` Vlastimil Babka
2016-05-31 13:08 ` [PATCH v2 02/18] mm, page_alloc: set alloc_flags only once in slowpath Vlastimil Babka
2016-05-31 13:08   ` Vlastimil Babka
2016-05-31 13:08 ` [PATCH v2 03/18] mm, page_alloc: don't retry initial attempt " Vlastimil Babka
2016-05-31 13:08   ` Vlastimil Babka
2016-06-01 13:26   ` Michal Hocko
2016-06-01 13:26     ` Michal Hocko
2016-06-01 14:58     ` Vlastimil Babka
2016-06-01 14:58       ` Vlastimil Babka
2016-05-31 13:08 ` [PATCH v2 04/18] mm, page_alloc: restructure direct compaction handling " Vlastimil Babka
2016-05-31 13:08   ` Vlastimil Babka
2016-05-31 13:08 ` [PATCH v2 05/18] mm, page_alloc: make THP-specific decisions more generic Vlastimil Babka
2016-05-31 13:08   ` Vlastimil Babka
2016-05-31 13:08 ` [PATCH v2 06/18] mm, thp: remove __GFP_NORETRY from khugepaged and madvised allocations Vlastimil Babka
2016-05-31 13:08   ` Vlastimil Babka
2016-06-01 13:33   ` Michal Hocko
2016-06-01 13:33     ` Michal Hocko
2016-05-31 13:08 ` [PATCH v2 07/18] mm, compaction: introduce direct compaction priority Vlastimil Babka
2016-05-31 13:08   ` Vlastimil Babka
2016-05-31 13:08 ` [PATCH v2 08/18] mm, compaction: simplify contended compaction handling Vlastimil Babka
2016-05-31 13:08   ` Vlastimil Babka
2016-05-31 13:08 ` [PATCH v2 09/18] mm, compaction: make whole_zone flag ignore cached scanner positions Vlastimil Babka
2016-05-31 13:08   ` Vlastimil Babka
2016-05-31 13:08 ` [PATCH v2 10/18] mm, compaction: cleanup unused functions Vlastimil Babka
2016-05-31 13:08   ` Vlastimil Babka
2016-06-01 13:45   ` Michal Hocko
2016-06-01 13:45     ` Michal Hocko
2016-05-31 13:08 ` Vlastimil Babka [this message]
2016-05-31 13:08   ` [PATCH v2 11/18] mm, compaction: add the ultimate direct compaction priority Vlastimil Babka
2016-05-31 13:08 ` [PATCH v2 12/18] mm, compaction: more reliably increase " Vlastimil Babka
2016-05-31 13:08   ` Vlastimil Babka
2016-06-01 13:51   ` Michal Hocko
2016-06-01 13:51     ` Michal Hocko
2016-06-23 14:41     ` Vlastimil Babka
2016-06-23 14:41       ` Vlastimil Babka
2016-05-31 13:08 ` [PATCH v2 13/18] mm, compaction: use correct watermark when checking allocation success Vlastimil Babka
2016-05-31 13:08   ` Vlastimil Babka
2016-06-01 13:59   ` Michal Hocko
2016-06-01 13:59     ` Michal Hocko
2016-05-31 13:08 ` [PATCH v2 14/18] mm, compaction: create compact_gap wrapper Vlastimil Babka
2016-05-31 13:08   ` Vlastimil Babka
2016-06-01 14:02   ` Michal Hocko
2016-06-01 14:02     ` Michal Hocko
2016-05-31 13:08 ` [PATCH v2 15/18] mm, compaction: use proper alloc_flags in __compaction_suitable() Vlastimil Babka
2016-05-31 13:08   ` Vlastimil Babka
2016-05-31 13:08 ` [PATCH v2 16/18] mm, compaction: require only min watermarks for non-costly orders Vlastimil Babka
2016-05-31 13:08   ` Vlastimil Babka
2016-06-01 14:08   ` Michal Hocko
2016-06-01 14:08     ` Michal Hocko
2016-05-31 13:08 ` [PATCH v2 17/18] mm, vmscan: make compaction_ready() more accurate and readable Vlastimil Babka
2016-05-31 13:08   ` Vlastimil Babka
2016-06-01 14:14   ` Michal Hocko
2016-06-01 14:14     ` Michal Hocko
2016-05-31 13:08 ` [PATCH v2 18/18] mm, vmscan: use proper classzone_idx in should_continue_reclaim() Vlastimil Babka
2016-05-31 13:08   ` Vlastimil Babka
2016-06-01 14:21   ` Michal Hocko
2016-06-01 14:21     ` Michal Hocko
2016-06-01 15:19     ` Vlastimil Babka
2016-06-01 15:19       ` Vlastimil Babka
2016-06-01 15:45       ` Michal Hocko
2016-06-01 15:45         ` Michal Hocko

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20160531130818.28724-12-vbabka@suse.cz \
    --to=vbabka@suse.cz \
    --cc=akpm@linux-foundation.org \
    --cc=iamjoonsoo.kim@lge.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=mgorman@techsingularity.net \
    --cc=mhocko@kernel.org \
    --cc=riel@redhat.com \
    --cc=rientjes@google.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.