[patch 09/14] fs: inode atomic last_ino, iunique lock

[npiggin]

Make last_ino atomic in preperation for removing inode_lock.
Make a new lock for iunique counter, for removing inode_lock.

Signed-off-by: Nick Piggin <npiggin@xxxxxxxxx>

---
 fs/inode.c |   44 +++++++++++++++++++++++++++++++++-----------
 1 file changed, 33 insertions(+), 11 deletions(-)

Index: linux-2.6/fs/inode.c
===================================================================
--- linux-2.6.orig/fs/inode.c	2010-10-21 23:50:27.000000000 +1100
+++ linux-2.6/fs/inode.c	2010-10-21 23:50:43.000000000 +1100
@@ -817,7 +817,7 @@ struct inode *new_inode(struct super_blo
 	 * error if st_ino won't fit in target struct field. Use 32bit counter
 	 * here to attempt to avoid that.
 	 */
-	static unsigned int last_ino;
+	static atomic_t last_ino = ATOMIC_INIT(0);
 	struct inode *inode;
 
 	spin_lock_prefetch(&inode_lock);
@@ -826,7 +826,7 @@ struct inode *new_inode(struct super_blo
 	if (inode) {
 		spin_lock(&inode_lock);
 		spin_lock(&inode->i_lock);
-		inode->i_ino = ++last_ino;
+		inode->i_ino = (unsigned int)atomic_inc_return(&last_ino);
 		inode->i_state = 0;
 		__inode_add_to_lists(sb, NULL, inode);
 		spin_unlock(&inode->i_lock);
@@ -982,6 +982,29 @@ static struct inode *get_new_inode_fast(
 	return inode;
 }
 
+/* Is the ino for this sb hashed right now? */
+static int is_ino_hashed(struct super_block *sb, unsigned long ino)
+{
+	struct hlist_node *node;
+	struct inode *inode = NULL;
+	struct hlist_head *head = inode_hashtable + hash(sb, ino);
+
+	spin_lock(&inode_hash_lock);
+	hlist_for_each_entry(inode, node, head, i_hash) {
+		if (inode->i_ino == ino && inode->i_sb == sb) {
+			spin_unlock(&inode_hash_lock);
+			return 0;
+		}
+		/*
+		 * Don't bother checking for I_FREEING etc., because
+		 * we don't want iunique to wait on freeing inodes. Just
+		 * skip it and get the next one.
+		 */
+	}
+	spin_unlock(&inode_hash_lock);
+	return 1;
+}
+
 /**
  *	iunique - get a unique inode number
  *	@sb: superblock
@@ -992,34 +1015,33 @@ static struct inode *get_new_inode_fast(
  *	permanent inode numbering system. An inode number is returned that
  *	is higher than the reserved limit but unique.
  *
+ *	Callers must serialise calls to iunique, because this function drops
+ *	all locks after it returns, so if concurrency on the same sb is
+ *	allowed, the value is racy by the time it returns.
+ *
  *	BUGS:
  *	With a large number of inodes live on the file system this function
  *	currently becomes quite slow.
  */
 ino_t iunique(struct super_block *sb, ino_t max_reserved)
 {
+	static DEFINE_SPINLOCK(unique_lock);
 	/*
 	 * On a 32bit, non LFS stat() call, glibc will generate an EOVERFLOW
 	 * error if st_ino won't fit in target struct field. Use 32bit counter
 	 * here to attempt to avoid that.
 	 */
 	static unsigned int counter;
-	struct inode *inode;
-	struct hlist_head *head;
 	ino_t res;
 
 	spin_lock(&inode_lock);
-	spin_lock(&inode_hash_lock);
+	spin_lock(&unique_lock);
 	do {
 		if (counter <= max_reserved)
 			counter = max_reserved + 1;
 		res = counter++;
-		head = inode_hashtable + hash(sb, res);
-		inode = find_inode_fast(sb, head, res);
-	} while (inode != NULL);
-	spin_unlock(&inode_hash_lock);
-	if (inode)
-		spin_unlock(&inode->i_lock);
+	} while (!is_ino_hashed(sb, res));
+	spin_unlock(&unique_lock);
 	spin_unlock(&inode_lock);
 
 	return res;


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/