Skip to contentSkip to conversation

mm,compaction: serialize waitqueue_active() checks (for real)

PATCH In review Davidlohr Bueso 1 participant 2 messages Sep 14, 2017

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 mm,compaction: serialize waitqueue_active() checks (for real)

Andrea brought to my attention that the L->{L,S} guarantees are completely bogus for this case. I was looking at the diagram, from the offending commit, when that _is_ the race, we had the load reordered already.

What we need is at least S->L semantics, thus simply use wq_has_sleeper() to serialize the call for good.

Fixes 46acef048a6 (mm,compaction: serialize waitqueue_active() checks)
---
 mm/compaction.c | 11 +++++------
 1 file changed, 5 insertions(+), 6 deletions(-)
mm/compaction.c +5 -6
index fb548e4c7bd4..d1ad2cf6c758 100644
@@ -1999,17 +1999,16 @@ void wakeup_kcompactd(pg_data_t *pgdat, int order, int classzone_idx)
 	if (pgdat->kcompactd_max_order < order)
 		pgdat->kcompactd_max_order = order;
 
-	/*
-	 * Pairs with implicit barrier in wait_event_freezable()
-	 * such that wakeups are not missed in the lockless
-	 * waitqueue_active() call.
-	 */
 	smp_acquire__after_ctrl_dep();
 
 	if (pgdat->kcompactd_classzone_idx > classzone_idx)
 		pgdat->kcompactd_classzone_idx = classzone_idx;
 
-	if (!waitqueue_active(&pgdat->kcompactd_wait))
+	/*
+	 * Pairs with implicit barrier in wait_event_freezable()
+	 * such that wakeups are not missed.
+	 */
+	if (!wq_has_sleeper(&pgdat->kcompactd_wait))
 		return;
 
 	if (!kcompactd_node_suitable(pgdat))
Signature · 1 line
2.12.0
#1 lore raw
Quoted 6 lines · Davidlohr Bueso
(quoted diff) +0 -5
-	/*
-	 * Pairs with implicit barrier in wait_event_freezable()
-	 * such that wakeups are not missed in the lockless
-	 * waitqueue_active() call.
-	 */
	smp_acquire__after_ctrl_dep();

sorry, forgot to delete this. Here's v2.

-----8<----------------------------------------------------- [PATCH v2] mm,compaction: serialize waitqueue_active() checks (for real)

Andrea brought to my attention that the L->{L,S} guarantees are completely bogus for this case. I was looking at the diagram, from the offending commit, when that _is_ the race, we had the load reordered already.

What we need is at least S->L semantics, thus simply use wq_has_sleeper() to serialize the call for good.

Fixes 46acef048a6 (mm,compaction: serialize waitqueue_active() checks)

---
mm/compaction.c | 13 +++++--------

1 file changed, 5 insertions(+), 8 deletions(-)
mm/compaction.c +0 -0
index fb548e4c7bd4..03d31a875341 100644
@@ -1999,17 +1999,14 @@ void wakeup_kcompactd(pg_data_t *pgdat, int order, int classzone_idx)
	if (pgdat->kcompactd_max_order < order)
		pgdat->kcompactd_max_order = order;

- /*
- * Pairs with implicit barrier in wait_event_freezable()
- * such that wakeups are not missed in the lockless
- * waitqueue_active() call.
- */
- smp_acquire__after_ctrl_dep();
-

	if (pgdat->kcompactd_classzone_idx > classzone_idx)
		pgdat->kcompactd_classzone_idx = classzone_idx;

- if (!waitqueue_active(&pgdat->kcompactd_wait))
+ /*
+ * Pairs with implicit barrier in wait_event_freezable()
+ * such that wakeups are not missed.
+ */
+ if (!wq_has_sleeper(&pgdat->kcompactd_wait))

		return;

	if (!kcompactd_node_suitable(pgdat))
Signature · 1 line
2.12.0