Skip to contentSkip to conversation

mm,compaction: Do not schedule work on other CPUs for compaction

PATCH In review Mel Gorman 4 participants 4 messages May 6, 2010 – May 10, 2010

Subsystems (3)
lore
Share & tools
lore

Mel Gorman

author
#0 lore raw

PATCH 1/1 mm,compaction: Do not schedule work on other CPUs for compaction

Migration normally requires a call to migrate_prep() as a preparation step. This schedules work on all CPUs for pagevecs to be drained. This makes sense for move_pages and memory hot-remove but is unnecessary for memory compaction.

To avoid queueing work on multiple CPUs, this patch introduces migrate_prep_local() which drains just local pagevecs.

This patch can be either merged with mmcompaction-memory-compaction-core.patch or placed immediately after it to clarify why migrate_prep_local() was introduced.

---
 include/linux/migrate.h |    2 ++
 mm/compaction.c         |    2 +-
 mm/migrate.c            |   11 ++++++++++-
 3 files changed, 13 insertions(+), 2 deletions(-)
include/linux/migrate.h +2 -0
index 05d2292..6dec3ef 100644
@@ -19,6 +19,7 @@ extern int fail_migrate_page(struct address_space *,
 			struct page *, struct page *);
 
 extern int migrate_prep(void);
+extern int migrate_prep_local(void);
 extern int migrate_vmas(struct mm_struct *mm,
 		const nodemask_t *from, const nodemask_t *to,
 		unsigned long flags);
@@ -32,6 +33,7 @@ static inline int migrate_pages(struct list_head *l, new_page_t x,
 		unsigned long private, int offlining) { return -ENOSYS; }
 
 static inline int migrate_prep(void) { return -ENOSYS; }
+static inline int migrate_prep_local(void) { return -ENOSYS; }
 
 static inline int migrate_vmas(struct mm_struct *mm,
 		const nodemask_t *from, const nodemask_t *to,
mm/compaction.c +1 -1
index bd13560..94cce51 100644
@@ -383,7 +383,7 @@ static int compact_zone(struct zone *zone, struct compact_control *cc)
 	cc->free_pfn = cc->migrate_pfn + zone->spanned_pages;
 	cc->free_pfn &= ~(pageblock_nr_pages-1);
 
-	migrate_prep();
+	migrate_prep_local();
 
 	while ((ret = compact_finished(zone, cc)) == COMPACT_CONTINUE) {
 		unsigned long nr_migrate, nr_remaining;
mm/migrate.c +10 -1
index 053fd39..d99ec15 100644
@@ -40,7 +40,8 @@
 
 /*
  * migrate_prep() needs to be called before we start compiling a list of pages
- * to be migrated using isolate_lru_page().
+ * to be migrated using isolate_lru_page(). If scheduling work on other CPUs is
+ * undesirable, use migrate_prep_local()
  */
 int migrate_prep(void)
 {
@@ -55,6 +56,14 @@ int migrate_prep(void)
 	return 0;
 }
 
+/* Do the necessary work of migrate_prep but not if it involves other CPUs */
+int migrate_prep_local(void)
+{
+	lru_add_drain();
+
+	return 0;
+}
+
 /*
  * Add isolated pages on the list back to the LRU under page lock
  * to avoid leaking evictable pages back onto unevictable list.
Reviewed-by on v1 1/1
#1 lore raw
Quoted 5 lines · Mel Gorman

Migration normally requires a call to migrate_prep() as a preparation step. This schedules work on all CPUs for pagevecs to be drained. This makes sense for move_pages and memory hot-remove but is unnecessary for memory compaction.

Quoted 8 lines

To avoid queueing work on multiple CPUs, this patch introduces migrate_prep_local() which drains just local pagevecs.

This patch can be either merged with mmcompaction-memory-compaction-core.patch or placed immediately after it to clarify why migrate_prep_local() was introduced.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>

Reviewed-by on v1 1/1
#2 lore raw
In reply to Mel Gorman
Quoted 5 lines · On Thu, 6 May 2010 16:08:09 +0100 Mel Gorman <[email protected]> wrote:

Migration normally requires a call to migrate_prep() as a preparation step. This schedules work on all CPUs for pagevecs to be drained. This makes sense for move_pages and memory hot-remove but is unnecessary for memory compaction.

Quoted 8 lines

To avoid queueing work on multiple CPUs, this patch introduces migrate_prep_local() which drains just local pagevecs.

This patch can be either merged with mmcompaction-memory-compaction-core.patch or placed immediately after it to clarify why migrate_prep_local() was introduced.