fs: Lock the inode LRU list separately

Introduce the inode_lru_lock to protect the inode_lru list. This
lock is nested inside the inode->i_lock to allow the inode to be
added to the LRU list in iput_final without needing to deal with
lock inversions. This keeps iput_final() clean and neat.

Further, where marking the inode I_FREEING and removing it from the
LRU, move the LRU list manipulation within the inode->i_lock to keep
the list manipulation consistent with iput_final. This also means
that most of the open coded LRU list removal + unused inode
accounting can now use the inode_lru_list_del() wrappers which
cleans the code up further.

However, this locking change means what the LRU traversal in
prune_icache() inverts this lock ordering and needs to use trylock
semantics on the inode->i_lock to avoid deadlocking. In these cases,
if we fail to lock the inode we move it to the back of the LRU to
prevent spinning on it.

Signed-off-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>

fs/inode.c

@@ -32,10 +32,13 @@
  *
  * inode->i_lock protects:
  *   inode->i_state, inode->i_hash, __iget()
+ * inode_lru_lock protects:
+ *   inode_lru, inode->i_lru
  *
  * Lock ordering:
  * inode_lock
  *   inode->i_lock
+ *     inode_lru_lock
  */
 
 /*
@@ -85,6 +88,7 @@ static unsigned int i_hash_shift __read_mostly;
  */
 
 static LIST_HEAD(inode_lru);
+static DEFINE_SPINLOCK(inode_lru_lock);
 static struct hlist_head *inode_hashtable __read_mostly;
 
 /*
@@ -356,18 +360,22 @@ EXPORT_SYMBOL(ihold);
 
 static void inode_lru_list_add(struct inode *inode)
 {
+	spin_lock(&inode_lru_lock);
 	if (list_empty(&inode->i_lru)) {
 		list_add(&inode->i_lru, &inode_lru);
 		inodes_stat.nr_unused++;
 	}
+	spin_unlock(&inode_lru_lock);
 }
 
 static void inode_lru_list_del(struct inode *inode)
 {
+	spin_lock(&inode_lru_lock);
 	if (!list_empty(&inode->i_lru)) {
 		list_del_init(&inode->i_lru);
 		inodes_stat.nr_unused--;
 	}
+	spin_unlock(&inode_lru_lock);
 }
 
 static inline void __inode_sb_list_add(struct inode *inode)
@@ -543,10 +551,9 @@ void evict_inodes(struct super_block *sb)
 		}
 
 		inode->i_state |= I_FREEING;
-		if (!(inode->i_state & (I_DIRTY | I_SYNC)))
-			inodes_stat.nr_unused--;
+		inode_lru_list_del(inode);
 		spin_unlock(&inode->i_lock);
-		list_move(&inode->i_lru, &dispose);
+		list_add(&inode->i_lru, &dispose);
 	}
 	spin_unlock(&inode_lock);
 
@@ -596,10 +603,9 @@ int invalidate_inodes(struct super_block *sb, bool kill_dirty)
 		}
 
 		inode->i_state |= I_FREEING;
-		if (!(inode->i_state & (I_DIRTY | I_SYNC)))
-			inodes_stat.nr_unused--;
+		inode_lru_list_del(inode);
 		spin_unlock(&inode->i_lock);
-		list_move(&inode->i_lru, &dispose);
+		list_add(&inode->i_lru, &dispose);
 	}
 	spin_unlock(&inode_lock);
 
@@ -623,7 +629,7 @@ static int can_unuse(struct inode *inode)
 
 /*
  * Scan `goal' inodes on the unused list for freeable ones. They are moved to a
- * temporary list and then are freed outside inode_lock by dispose_list().
+ * temporary list and then are freed outside inode_lru_lock by dispose_list().
  *
  * Any inodes which are pinned purely because of attached pagecache have their
  * pagecache removed.  If the inode has metadata buffers attached to
@@ -645,6 +651,7 @@ static void prune_icache(int nr_to_scan)
 
 	down_read(&iprune_sem);
 	spin_lock(&inode_lock);
+	spin_lock(&inode_lru_lock);
 	for (nr_scanned = 0; nr_scanned < nr_to_scan; nr_scanned++) {
 		struct inode *inode;
 
@@ -653,11 +660,20 @@ static void prune_icache(int nr_to_scan)
 
 		inode = list_entry(inode_lru.prev, struct inode, i_lru);
 
+		/*
+		 * we are inverting the inode_lru_lock/inode->i_lock here,
+		 * so use a trylock. If we fail to get the lock, just move the
+		 * inode to the back of the list so we don't spin on it.
+		 */
+		if (!spin_trylock(&inode->i_lock)) {
+			list_move(&inode->i_lru, &inode_lru);
+			continue;
+		}
+
 		/*
 		 * Referenced or dirty inodes are still in use. Give them
 		 * another pass through the LRU as we canot reclaim them now.
 		 */
-		spin_lock(&inode->i_lock);
 		if (atomic_read(&inode->i_count) ||
 		    (inode->i_state & ~I_REFERENCED)) {
 			spin_unlock(&inode->i_lock);
@@ -676,17 +692,21 @@ static void prune_icache(int nr_to_scan)
 		if (inode_has_buffers(inode) || inode->i_data.nrpages) {
 			__iget(inode);
 			spin_unlock(&inode->i_lock);
+			spin_unlock(&inode_lru_lock);
 			spin_unlock(&inode_lock);
 			if (remove_inode_buffers(inode))
 				reap += invalidate_mapping_pages(&inode->i_data,
 								0, -1);
 			iput(inode);
 			spin_lock(&inode_lock);
+			spin_lock(&inode_lru_lock);
 
 			if (inode != list_entry(inode_lru.next,
 						struct inode, i_lru))
 				continue;	/* wrong inode or list_empty */
-			spin_lock(&inode->i_lock);
+			/* avoid lock inversions with trylock */
+			if (!spin_trylock(&inode->i_lock))
+				continue;
 			if (!can_unuse(inode)) {
 				spin_unlock(&inode->i_lock);
 				continue;
@@ -703,6 +723,7 @@ static void prune_icache(int nr_to_scan)
 		__count_vm_events(KSWAPD_INODESTEAL, reap);
 	else
 		__count_vm_events(PGINODESTEAL, reap);
+	spin_unlock(&inode_lru_lock);
 	spin_unlock(&inode_lock);
 
 	dispose_list(&freeable);