Btrfs: use the inode own lock to protect its delalloc_bytes

We need not use a global lock to protect the delalloc_bytes of the
inode, just use its own lock. In this way, we can reduce the lock
contention and ->delalloc_lock will just protect delalloc inode
list.
Signed-off-by: Miao Xie <miaox@cn.fujitsu.com>
Signed-off-by: Josef Bacik <jbacik@fusionio.com>

fs/btrfs/btrfs_inode.h

@@ -40,6 +40,7 @@
 #define BTRFS_INODE_HAS_ASYNC_EXTENT		6
 #define BTRFS_INODE_NEEDS_FULL_SYNC		7
 #define BTRFS_INODE_COPY_EVERYTHING		8
+#define BTRFS_INODE_IN_DELALLOC_LIST		9
 
 /* in memory btrfs inode */
 struct btrfs_inode {

fs/btrfs/disk-io.c

@@ -3690,6 +3690,8 @@ static void btrfs_destroy_delalloc_inodes(struct btrfs_root *root)
 				    delalloc_inodes);
 
 		list_del_init(&btrfs_inode->delalloc_inodes);
+		clear_bit(BTRFS_INODE_IN_DELALLOC_LIST,
+			  &btrfs_inode->runtime_flags);
 
 		btrfs_invalidate_inodes(btrfs_inode->root);
 	}

fs/btrfs/inode.c

@@ -1514,16 +1514,23 @@ static void btrfs_set_bit_hook(struct inode *inode,
 			spin_unlock(&BTRFS_I(inode)->lock);
 		}
 
-		spin_lock(&root->fs_info->delalloc_lock);
-		BTRFS_I(inode)->delalloc_bytes += len;
 		__percpu_counter_add(&root->fs_info->delalloc_bytes, len,
 				     root->fs_info->delalloc_batch);
-		if (do_list && list_empty(&BTRFS_I(inode)->delalloc_inodes)) {
+		spin_lock(&BTRFS_I(inode)->lock);
+		BTRFS_I(inode)->delalloc_bytes += len;
+		if (do_list && !test_bit(BTRFS_INODE_IN_DELALLOC_LIST,
+					 &BTRFS_I(inode)->runtime_flags)) {
+			spin_lock(&root->fs_info->delalloc_lock);
+			if (list_empty(&BTRFS_I(inode)->delalloc_inodes)) {
 				list_add_tail(&BTRFS_I(inode)->delalloc_inodes,
 					      &root->fs_info->delalloc_inodes);
+				set_bit(BTRFS_INODE_IN_DELALLOC_LIST,
+					&BTRFS_I(inode)->runtime_flags);
 			}
 			spin_unlock(&root->fs_info->delalloc_lock);
 		}
+		spin_unlock(&BTRFS_I(inode)->lock);
+	}
 }
 
 /*
@@ -1557,17 +1564,23 @@ static void btrfs_clear_bit_hook(struct inode *inode,
 		    && do_list)
 			btrfs_free_reserved_data_space(inode, len);
 
-		spin_lock(&root->fs_info->delalloc_lock);
 		__percpu_counter_add(&root->fs_info->delalloc_bytes, -len,
 				     root->fs_info->delalloc_batch);
+		spin_lock(&BTRFS_I(inode)->lock);
 		BTRFS_I(inode)->delalloc_bytes -= len;
-
 		if (do_list && BTRFS_I(inode)->delalloc_bytes == 0 &&
-		    !list_empty(&BTRFS_I(inode)->delalloc_inodes)) {
+		    test_bit(BTRFS_INODE_IN_DELALLOC_LIST,
+			     &BTRFS_I(inode)->runtime_flags)) {
+			spin_lock(&root->fs_info->delalloc_lock);
+			if (!list_empty(&BTRFS_I(inode)->delalloc_inodes)) {
 				list_del_init(&BTRFS_I(inode)->delalloc_inodes);
+				clear_bit(BTRFS_INODE_IN_DELALLOC_LIST,
+					  &BTRFS_I(inode)->runtime_flags);
 			}
 			spin_unlock(&root->fs_info->delalloc_lock);
 		}
+		spin_unlock(&BTRFS_I(inode)->lock);
+	}
 }
 
 /*
@@ -7326,14 +7339,19 @@ int btrfs_init_cachep(void)
 static int btrfs_getattr(struct vfsmount *mnt,
 			 struct dentry *dentry, struct kstat *stat)
 {
+	u64 delalloc_bytes;
 	struct inode *inode = dentry->d_inode;
 	u32 blocksize = inode->i_sb->s_blocksize;
 
 	generic_fillattr(inode, stat);
 	stat->dev = BTRFS_I(inode)->root->anon_dev;
 	stat->blksize = PAGE_CACHE_SIZE;
+
+	spin_lock(&BTRFS_I(inode)->lock);
+	delalloc_bytes = BTRFS_I(inode)->delalloc_bytes;
+	spin_unlock(&BTRFS_I(inode)->lock);
 	stat->blocks = (ALIGN(inode_get_bytes(inode), blocksize) +
-		ALIGN(BTRFS_I(inode)->delalloc_bytes, blocksize)) >> 9;
+			ALIGN(delalloc_bytes, blocksize)) >> 9;
 	return 0;
 }
 
@@ -7620,8 +7638,11 @@ int btrfs_start_delalloc_inodes(struct btrfs_root *root, int delay_iput)
 		list_del_init(&binode->delalloc_inodes);
 
 		inode = igrab(&binode->vfs_inode);
-		if (!inode)
+		if (!inode) {
+			clear_bit(BTRFS_INODE_IN_DELALLOC_LIST,
+				  &binode->runtime_flags);
 			continue;
+		}
 
 		list_add_tail(&binode->delalloc_inodes,
 			      &root->fs_info->delalloc_inodes);