[PATCH] mm, compaction: raise compaction priority after it withdrawns

[Vlastimil Babka]

Mike Kravetz reports that "hugetlb allocations could stall for minutes or hours
when should_compact_retry() would return true more often then it should.
Specifically, this was in the case where compact_result was COMPACT_DEFERRED
and COMPACT_PARTIAL_SKIPPED and no progress was being made."

The problem is that the compaction_withdrawn() test in should_compact_retry()
includes compaction outcomes that are only possible on low compaction priority,
and results in a retry without increasing the priority. This may result in
furter reclaim, and more incomplete compaction attempts.

With this patch, compaction priority is raised when possible, or
should_compact_retry() returns false.

The COMPACT_SKIPPED result doesn't really fit together with the other outcomes
in compaction_withdrawn(), as that's a result caused by insufficient order-0
pages, not due to low compaction priority. With this patch, it is moved to
a new compaction_needs_reclaim() function, and for that outcome we keep the
current logic of retrying if it looks like reclaim will be able to help.

Reported-by: Mike Kravetz <mike.kravetz@xxxxxxxxxx>
Signed-off-by: Vlastimil Babka <vbabka@xxxxxxx>
---
 include/linux/compaction.h | 22 +++++++++++++++++-----
 mm/page_alloc.c            | 16 ++++++++++++----
 2 files changed, 29 insertions(+), 9 deletions(-)

diff --git a/include/linux/compaction.h b/include/linux/compaction.h
index 9569e7c786d3..4b898cdbdf05 100644
--- a/include/linux/compaction.h
+++ b/include/linux/compaction.h
@@ -129,11 +129,8 @@ static inline bool compaction_failed(enum compact_result result)
 	return false;
 }
 
-/*
- * Compaction  has backed off for some reason. It might be throttling or
- * lock contention. Retrying is still worthwhile.
- */
-static inline bool compaction_withdrawn(enum compact_result result)
+/* Compaction needs reclaim to be performed first, so it can continue. */
+static inline bool compaction_needs_reclaim(enum compact_result result)
 {
 	/*
 	 * Compaction backed off due to watermark checks for order-0
@@ -142,6 +139,16 @@ static inline bool compaction_withdrawn(enum compact_result result)
 	if (result == COMPACT_SKIPPED)
 		return true;
 
+	return false;
+}
+
+/*
+ * Compaction has backed off for some reason after doing some work or none
+ * at all. It might be throttling or lock contention. Retrying might be still
+ * worthwhile, but with a higher priority if allowed.
+ */
+static inline bool compaction_withdrawn(enum compact_result result)
+{
 	/*
 	 * If compaction is deferred for high-order allocations, it is
 	 * because sync compaction recently failed. If this is the case
@@ -207,6 +214,11 @@ static inline bool compaction_failed(enum compact_result result)
 	return false;
 }
 
+static inline bool compaction_needs_reclaim(enum compact_result result)
+{
+	return false;
+}
+
 static inline bool compaction_withdrawn(enum compact_result result)
 {
 	return true;
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 272c6de1bf4e..bd0f00f8cfa3 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -3965,15 +3965,23 @@ should_compact_retry(struct alloc_context *ac, int order, int alloc_flags,
 	if (compaction_failed(compact_result))
 		goto check_priority;
 
+	/*
+	 * compaction was skipped because there are not enough order-0 pages
+	 * to work with, so we retry only if it looks like reclaim can help.
+	 */
+	if (compaction_needs_reclaim(compact_result)) {
+		ret = compaction_zonelist_suitable(ac, order, alloc_flags);
+		goto out;
+	}
+
 	/*
 	 * make sure the compaction wasn't deferred or didn't bail out early
 	 * due to locks contention before we declare that we should give up.
-	 * But do not retry if the given zonelist is not suitable for
-	 * compaction.
+	 * But the next retry should use a higher priority if allowed, so
+	 * we don't just keep bailing out endlessly.
 	 */
 	if (compaction_withdrawn(compact_result)) {
-		ret = compaction_zonelist_suitable(ac, order, alloc_flags);
-		goto out;
+		goto check_priority;
 	}
 
 	/*
-- 
2.22.0