[RESEND,v4,2/3] fs: Don't need to put list_lru into its own cacheline

Commit Message

Waiman Long Dec. 14, 2018, 9:53 p.m. UTC

The list_lru structure is essentially just a pointer to a table of
per-node LRU lists. Even if CONFIG_MEMCG_KMEM is defined, the list
field is just used for LRU list registration and shrinker_id is set
at initialization. Those fields won't need to be touched that often.

So there is no point to make the list_lru structures to sit in their
own cachelines.

Signed-off-by: Waiman Long <[e-mail]>
Reviewed-by: Dave Chinner <[e-mail]>
---
 include/linux/fs.h | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

Patch

diff --git a/include/linux/fs.h b/include/linux/fs.h
index c95c080..7707c4a 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -1474,11 +1474,12 @@  struct super_block {
 	struct user_namespace *s_user_ns;
 
 	/*
-	 * Keep the lru lists last in the structure so they always sit on their
-	 * own individual cachelines.
+	 * The list_lru structure is essentially just a pointer to a table
+	 * of per-node lru lists, each of which has its own spinlock.
+	 * There is no need to put them into separate cachelines.
 	 */
-	struct list_lru		s_dentry_lru ____cacheline_aligned_in_smp;
-	struct list_lru		s_inode_lru ____cacheline_aligned_in_smp;
+	struct list_lru		s_dentry_lru;
+	struct list_lru		s_inode_lru;
 	struct rcu_head		rcu;
 	struct work_struct	destroy_work;