9 files changed, 459 insertions, 168 deletions

diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h
index ea1ea0af8c0e..82ee56bba299 100644
--- a/fs/btrfs/btrfs_inode.h
+++ b/fs/btrfs/btrfs_inode.h
@@ -138,6 +138,7 @@ struct btrfs_inode {
          * of these.
          */
         unsigned ordered_data_close:1;
+        unsigned dummy_inode:1;
 
         struct inode vfs_inode;
 };
diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index 746a9acd2189..6ade48b227e9 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -114,6 +114,10 @@ struct btrfs_ordered_sum;
  */
 #define BTRFS_DEV_ITEMS_OBJECTID 1ULL
 
+#define BTRFS_BTREE_INODE_OBJECTID 1
+
+#define BTRFS_EMPTY_SUBVOL_DIR_OBJECTID 2
+
 /*
  * we can actually store much bigger names, but lets not confuse the rest
  * of linux
@@ -792,6 +796,8 @@ struct btrfs_fs_info {
 
         /* the log root tree is a directory of all the other log roots */
         struct btrfs_root *log_root_tree;
+
+        spinlock_t fs_roots_radix_lock;
         struct radix_tree_root fs_roots_radix;
 
         /* block group cache stuff */
@@ -1011,6 +1017,8 @@ struct btrfs_root {
         u64 highest_objectid;
         int ref_cows;
         int track_dirty;
+        int in_radix;
+
         u64 defrag_trans_start;
         struct btrfs_key defrag_progress;
         struct btrfs_key defrag_max;
@@ -2111,12 +2119,15 @@ int btrfs_drop_subtree(struct btrfs_trans_handle *trans,
                         struct extent_buffer *parent);
 /* root-item.c */
 int btrfs_find_root_ref(struct btrfs_root *tree_root,
-                   struct btrfs_path *path,
-                   u64 root_id, u64 ref_id);
+                        struct btrfs_path *path,
+                        u64 root_id, u64 ref_id);
 int btrfs_add_root_ref(struct btrfs_trans_handle *trans,
                        struct btrfs_root *tree_root,
-                       u64 root_id, u8 type, u64 ref_id,
-                       u64 dirid, u64 sequence,
+                       u64 root_id, u64 ref_id, u64 dirid, u64 sequence,
+                       const char *name, int name_len);
+int btrfs_del_root_ref(struct btrfs_trans_handle *trans,
+                       struct btrfs_root *tree_root,
+                       u64 root_id, u64 ref_id, u64 dirid, u64 *sequence,
                        const char *name, int name_len);
 int btrfs_del_root(struct btrfs_trans_handle *trans, struct btrfs_root *root,
                    struct btrfs_key *key);
@@ -2149,6 +2160,10 @@ btrfs_lookup_dir_index_item(struct btrfs_trans_handle *trans,
                             struct btrfs_path *path, u64 dir,
                             u64 objectid, const char *name, int name_len,
                             int mod);
+struct btrfs_dir_item *
+btrfs_search_dir_index_item(struct btrfs_root *root,
+                            struct btrfs_path *path, u64 dirid,
+                            const char *name, int name_len);
 struct btrfs_dir_item *btrfs_match_dir_item_name(struct btrfs_root *root,
                               struct btrfs_path *path,
                               const char *name, int name_len);
@@ -2171,6 +2186,7 @@ int btrfs_insert_orphan_item(struct btrfs_trans_handle *trans,
                              struct btrfs_root *root, u64 offset);
 int btrfs_del_orphan_item(struct btrfs_trans_handle *trans,
                           struct btrfs_root *root, u64 offset);
+int btrfs_find_orphan_item(struct btrfs_root *root, u64 offset);
 
 /* inode-map.c */
 int btrfs_find_free_objectid(struct btrfs_trans_handle *trans,
@@ -2243,6 +2259,10 @@ int btrfs_unlink_inode(struct btrfs_trans_handle *trans,
 int btrfs_add_link(struct btrfs_trans_handle *trans,
                    struct inode *parent_inode, struct inode *inode,
                    const char *name, int name_len, int add_backref, u64 index);
+int btrfs_unlink_subvol(struct btrfs_trans_handle *trans,
+                        struct btrfs_root *root,
+                        struct inode *dir, u64 objectid,
+                        const char *name, int name_len);
 int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
                                struct btrfs_root *root,
                                struct inode *inode, u64 new_size,
diff --git a/fs/btrfs/dir-item.c b/fs/btrfs/dir-item.c
index 1d70236ba00c..f3a6075519cc 100644
--- a/fs/btrfs/dir-item.c
+++ b/fs/btrfs/dir-item.c
@@ -281,6 +281,53 @@ btrfs_lookup_dir_index_item(struct btrfs_trans_handle *trans,
         return btrfs_match_dir_item_name(root, path, name, name_len);
 }
 
+struct btrfs_dir_item *
+btrfs_search_dir_index_item(struct btrfs_root *root,
+                            struct btrfs_path *path, u64 dirid,
+                            const char *name, int name_len)
+{
+        struct extent_buffer *leaf;
+        struct btrfs_dir_item *di;
+        struct btrfs_key key;
+        u32 nritems;
+        int ret;
+
+        key.objectid = dirid;
+        key.type = BTRFS_DIR_INDEX_KEY;
+        key.offset = 0;
+
+        ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+        if (ret < 0)
+                return ERR_PTR(ret);
+
+        leaf = path->nodes[0];
+        nritems = btrfs_header_nritems(leaf);
+
+        while (1) {
+                if (path->slots[0] >= nritems) {
+                        ret = btrfs_next_leaf(root, path);
+                        if (ret < 0)
+                                return ERR_PTR(ret);
+                        if (ret > 0)
+                                break;
+                        leaf = path->nodes[0];
+                        nritems = btrfs_header_nritems(leaf);
+                        continue;
+                }
+
+                btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
+                if (key.objectid != dirid || key.type != BTRFS_DIR_INDEX_KEY)
+                        break;
+
+                di = btrfs_match_dir_item_name(root, path, name, name_len);
+                if (di)
+                        return di;
+
+                path->slots[0]++;
+        }
+        return NULL;
+}
+
 struct btrfs_dir_item *btrfs_lookup_xattr(struct btrfs_trans_handle *trans,
                                           struct btrfs_root *root,
                                           struct btrfs_path *path, u64 dir,
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 790f4b61a3d7..a4f531047c4a 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -41,6 +41,7 @@
 
 static struct extent_io_ops btree_extent_io_ops;
 static void end_workqueue_fn(struct btrfs_work *work);
+static void free_fs_root(struct btrfs_root *root);
 
 static atomic_t btrfs_bdi_num = ATOMIC_INIT(0);
 
@@ -951,14 +952,16 @@ static int find_and_setup_root(struct btrfs_root *tree_root,
                      root, fs_info, objectid);
         ret = btrfs_find_last_root(tree_root, objectid,
                                    &root->root_item, &root->root_key);
+        if (ret > 0)
+                return -ENOENT;
         BUG_ON(ret);
 
         generation = btrfs_root_generation(&root->root_item);
         blocksize = btrfs_level_size(root, btrfs_root_level(&root->root_item));
         root->node = read_tree_block(root, btrfs_root_bytenr(&root->root_item),
                                      blocksize, generation);
-        root->commit_root = btrfs_root_node(root);
         BUG_ON(!root->node);
+        root->commit_root = btrfs_root_node(root);
         return 0;
 }
 
@@ -1176,39 +1179,66 @@ struct btrfs_root *btrfs_read_fs_root_no_name(struct btrfs_fs_info *fs_info,
                 return fs_info->dev_root;
         if (location->objectid == BTRFS_CSUM_TREE_OBJECTID)
                 return fs_info->csum_root;
-
+again:
+        spin_lock(&fs_info->fs_roots_radix_lock);
         root = radix_tree_lookup(&fs_info->fs_roots_radix,
                                  (unsigned long)location->objectid);
+        spin_unlock(&fs_info->fs_roots_radix_lock);
         if (root)
                 return root;
 
+        ret = btrfs_find_orphan_item(fs_info->tree_root, location->objectid);
+        if (ret == 0)
+                ret = -ENOENT;
+        if (ret < 0)
+                return ERR_PTR(ret);
+
         root = btrfs_read_fs_root_no_radix(fs_info->tree_root, location);
         if (IS_ERR(root))
                 return root;
 
+        WARN_ON(btrfs_root_refs(&root->root_item) == 0);
         set_anon_super(&root->anon_super, NULL);
 
+        ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM);
+        if (ret)
+                goto fail;
+
+        spin_lock(&fs_info->fs_roots_radix_lock);
         ret = radix_tree_insert(&fs_info->fs_roots_radix,
                                 (unsigned long)root->root_key.objectid,
                                 root);
+        if (ret == 0)
+                root->in_radix = 1;
+        spin_unlock(&fs_info->fs_roots_radix_lock);
+        radix_tree_preload_end();
         if (ret) {
-                free_extent_buffer(root->node);
-                kfree(root);
-                return ERR_PTR(ret);
+                if (ret == -EEXIST) {
+                        free_fs_root(root);
+                        goto again;
+                }
+                goto fail;
         }
-        if (!(fs_info->sb->s_flags & MS_RDONLY)) {
-                ret = btrfs_find_dead_roots(fs_info->tree_root,
-                                            root->root_key.objectid);
-                BUG_ON(ret);
+
+        ret = btrfs_find_dead_roots(fs_info->tree_root,
+                                    root->root_key.objectid);
+        WARN_ON(ret);
+
+        if (!(fs_info->sb->s_flags & MS_RDONLY))
                 btrfs_orphan_cleanup(root);
-        }
+
         return root;
+fail:
+        free_fs_root(root);
+        return ERR_PTR(ret);
 }
 
 struct btrfs_root *btrfs_read_fs_root(struct btrfs_fs_info *fs_info,
                                       struct btrfs_key *location,
                                       const char *name, int namelen)
 {
+        return btrfs_read_fs_root_no_name(fs_info, location);
+#if 0
         struct btrfs_root *root;
         int ret;
 
@@ -1225,7 +1255,7 @@ struct btrfs_root *btrfs_read_fs_root(struct btrfs_fs_info *fs_info,
                 kfree(root);
                 return ERR_PTR(ret);
         }
-#if 0
+
         ret = btrfs_sysfs_add_root(root);
         if (ret) {
                 free_extent_buffer(root->node);
@@ -1233,9 +1263,9 @@ struct btrfs_root *btrfs_read_fs_root(struct btrfs_fs_info *fs_info,
                 kfree(root);
                 return ERR_PTR(ret);
         }
-#endif
         root->in_sysfs = 1;
         return root;
+#endif
 }
 
 static int btrfs_congested_fn(void *congested_data, int bdi_bits)
@@ -2229,20 +2259,25 @@ int write_ctree_super(struct btrfs_trans_handle *trans,
 
 int btrfs_free_fs_root(struct btrfs_fs_info *fs_info, struct btrfs_root *root)
 {
-        WARN_ON(!RB_EMPTY_ROOT(&root->inode_tree));
+        spin_lock(&fs_info->fs_roots_radix_lock);
         radix_tree_delete(&fs_info->fs_roots_radix,
                           (unsigned long)root->root_key.objectid);
+        spin_unlock(&fs_info->fs_roots_radix_lock);
+        free_fs_root(root);
+        return 0;
+}
+
+static void free_fs_root(struct btrfs_root *root)
+{
+        WARN_ON(!RB_EMPTY_ROOT(&root->inode_tree));
         if (root->anon_super.s_dev) {
                 down_write(&root->anon_super.s_umount);
                 kill_anon_super(&root->anon_super);
         }
-        if (root->node)
-                free_extent_buffer(root->node);
-        if (root->commit_root)
-                free_extent_buffer(root->commit_root);
+        free_extent_buffer(root->node);
+        free_extent_buffer(root->commit_root);
         kfree(root->name);
         kfree(root);
-        return 0;
 }
 
 static int del_fs_roots(struct btrfs_fs_info *fs_info)
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 9e81f3184f24..6036b36789cc 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -2373,6 +2373,69 @@ static int btrfs_unlink(struct inode *dir, struct dentry *dentry)
         return ret;
 }
 
+int btrfs_unlink_subvol(struct btrfs_trans_handle *trans,
+                        struct btrfs_root *root,
+                        struct inode *dir, u64 objectid,
+                        const char *name, int name_len)
+{
+        struct btrfs_path *path;
+        struct extent_buffer *leaf;
+        struct btrfs_dir_item *di;
+        struct btrfs_key key;
+        u64 index;
+        int ret;
+
+        path = btrfs_alloc_path();
+        if (!path)
+                return -ENOMEM;
+
+        di = btrfs_lookup_dir_item(trans, root, path, dir->i_ino,
+                                   name, name_len, -1);
+        BUG_ON(!di || IS_ERR(di));
+
+        leaf = path->nodes[0];
+        btrfs_dir_item_key_to_cpu(leaf, di, &key);
+        WARN_ON(key.type != BTRFS_ROOT_ITEM_KEY || key.objectid != objectid);
+        ret = btrfs_delete_one_dir_name(trans, root, path, di);
+        BUG_ON(ret);
+        btrfs_release_path(root, path);
+
+        ret = btrfs_del_root_ref(trans, root->fs_info->tree_root,
+                                 objectid, root->root_key.objectid,
+                                 dir->i_ino, &index, name, name_len);
+        if (ret < 0) {
+                BUG_ON(ret != -ENOENT);
+                di = btrfs_search_dir_index_item(root, path, dir->i_ino,
+                                                 name, name_len);
+                BUG_ON(!di || IS_ERR(di));
+
+                leaf = path->nodes[0];
+                btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
+                btrfs_release_path(root, path);
+                index = key.offset;
+        }
+
+        di = btrfs_lookup_dir_index_item(trans, root, path, dir->i_ino,
+                                         index, name, name_len, -1);
+        BUG_ON(!di || IS_ERR(di));
+
+        leaf = path->nodes[0];
+        btrfs_dir_item_key_to_cpu(leaf, di, &key);
+        WARN_ON(key.type != BTRFS_ROOT_ITEM_KEY || key.objectid != objectid);
+        ret = btrfs_delete_one_dir_name(trans, root, path, di);
+        BUG_ON(ret);
+        btrfs_release_path(root, path);
+
+        btrfs_i_size_write(dir, dir->i_size - name_len * 2);
+        dir->i_mtime = dir->i_ctime = CURRENT_TIME;
+        ret = btrfs_update_inode(trans, root, dir);
+        BUG_ON(ret);
+        dir->i_sb->s_dirt = 1;
+
+        btrfs_free_path(path);
+        return 0;
+}
+
 static int btrfs_rmdir(struct inode *dir, struct dentry *dentry)
 {
         struct inode *inode = dentry->d_inode;
@@ -2382,29 +2445,31 @@ static int btrfs_rmdir(struct inode *dir, struct dentry *dentry)
         struct btrfs_trans_handle *trans;
         unsigned long nr = 0;
 
-        /*
-         * the FIRST_FREE_OBJECTID check makes sure we don't try to rmdir
-         * the root of a subvolume or snapshot
-         */
         if (inode->i_size > BTRFS_EMPTY_DIR_SIZE ||
-            inode->i_ino == BTRFS_FIRST_FREE_OBJECTID) {
+            inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)
                 return -ENOTEMPTY;
-        }
 
         trans = btrfs_start_transaction(root, 1);
         btrfs_set_trans_block_group(trans, dir);
 
+        if (unlikely(inode->i_ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)) {
+                err = btrfs_unlink_subvol(trans, root, dir,
+                                          BTRFS_I(inode)->location.objectid,
+                                          dentry->d_name.name,
+                                          dentry->d_name.len);
+                goto out;
+        }
+
         err = btrfs_orphan_add(trans, inode);
         if (err)
-                goto fail_trans;
+                goto out;
 
         /* now the directory is empty */
         err = btrfs_unlink_inode(trans, root, dir, dentry->d_inode,
                                  dentry->d_name.name, dentry->d_name.len);
         if (!err)
                 btrfs_i_size_write(inode, 0);
-
-fail_trans:
+out:
         nr = trans->blocks_used;
         ret = btrfs_end_transaction_throttle(trans, root);
         btrfs_btree_balance_dirty(root, nr);
@@ -3091,29 +3156,67 @@ out_err:
  * is kind of like crossing a mount point.
  */
 static int fixup_tree_root_location(struct btrfs_root *root,
-                             struct btrfs_key *location,
-                             struct btrfs_root **sub_root,
-                             struct dentry *dentry)
+                                    struct inode *dir,
+                                    struct dentry *dentry,
+                                    struct btrfs_key *location,
+                                    struct btrfs_root **sub_root)
 {
-        struct btrfs_root_item *ri;
+        struct btrfs_path *path;
+        struct btrfs_root *new_root;
+        struct btrfs_root_ref *ref;
+        struct extent_buffer *leaf;
+        int ret;
+        int err = 0;
 
-        if (btrfs_key_type(location) != BTRFS_ROOT_ITEM_KEY)
-                return 0;
-        if (location->objectid == BTRFS_ROOT_TREE_OBJECTID)
-                return 0;
+        path = btrfs_alloc_path();
+        if (!path) {
+                err = -ENOMEM;
+                goto out;
+        }
 
-        *sub_root = btrfs_read_fs_root(root->fs_info, location,
-                                        dentry->d_name.name,
-                                        dentry->d_name.len);
-        if (IS_ERR(*sub_root))
-                return PTR_ERR(*sub_root);
+        err = -ENOENT;
+        ret = btrfs_find_root_ref(root->fs_info->tree_root, path,
+                                  BTRFS_I(dir)->root->root_key.objectid,
+                                  location->objectid);
+        if (ret) {
+                if (ret < 0)
+                        err = ret;
+                goto out;
+        }
 
-        ri = &(*sub_root)->root_item;
-        location->objectid = btrfs_root_dirid(ri);
-        btrfs_set_key_type(location, BTRFS_INODE_ITEM_KEY);
-        location->offset = 0;
+        leaf = path->nodes[0];
+        ref = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_root_ref);
+        if (btrfs_root_ref_dirid(leaf, ref) != dir->i_ino ||
+            btrfs_root_ref_name_len(leaf, ref) != dentry->d_name.len)
+                goto out;
 
-        return 0;
+        ret = memcmp_extent_buffer(leaf, dentry->d_name.name,
+                                   (unsigned long)(ref + 1),
+                                   dentry->d_name.len);
+        if (ret)
+                goto out;
+
+        btrfs_release_path(root->fs_info->tree_root, path);
+
+        new_root = btrfs_read_fs_root_no_name(root->fs_info, location);
+        if (IS_ERR(new_root)) {
+                err = PTR_ERR(new_root);
+                goto out;
+        }
+
+        if (btrfs_root_refs(&new_root->root_item) == 0) {
+                err = -ENOENT;
+                goto out;
+        }
+
+        *sub_root = new_root;
+        location->objectid = btrfs_root_dirid(&new_root->root_item);
+        location->type = BTRFS_INODE_ITEM_KEY;
+        location->offset = 0;
+        err = 0;
+out:
+        btrfs_free_path(path);
+        return err;
 }
 
 static void inode_tree_add(struct inode *inode)
@@ -3246,11 +3349,34 @@ struct inode *btrfs_iget(struct super_block *s, struct btrfs_key *location,
         return inode;
 }
 
+static struct inode *new_simple_dir(struct super_block *s,
+                                    struct btrfs_key *key,
+                                    struct btrfs_root *root)
+{
+        struct inode *inode = new_inode(s);
+
+        if (!inode)
+                return ERR_PTR(-ENOMEM);
+
+        init_btrfs_i(inode);
+
+        BTRFS_I(inode)->root = root;
+        memcpy(&BTRFS_I(inode)->location, key, sizeof(*key));
+        BTRFS_I(inode)->dummy_inode = 1;
+
+        inode->i_ino = BTRFS_EMPTY_SUBVOL_DIR_OBJECTID;
+        inode->i_op = &simple_dir_inode_operations;
+        inode->i_fop = &simple_dir_operations;
+        inode->i_mode = S_IFDIR | S_IRUGO | S_IWUSR | S_IXUGO;
+        inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
+
+        return inode;
+}
+
 struct inode *btrfs_lookup_dentry(struct inode *dir, struct dentry *dentry)
 {
         struct inode *inode;
-        struct btrfs_inode *bi = BTRFS_I(dir);
-        struct btrfs_root *root = bi->root;
+        struct btrfs_root *root = BTRFS_I(dir)->root;
         struct btrfs_root *sub_root = root;
         struct btrfs_key location;
         int ret;
@@ -3263,17 +3389,25 @@ struct inode *btrfs_lookup_dentry(struct inode *dir, struct dentry *dentry)
         if (ret < 0)
                 return ERR_PTR(ret);
 
-        inode = NULL;
-        if (location.objectid) {
-                ret = fixup_tree_root_location(root, &location, &sub_root,
-                                                dentry);
-                if (ret < 0)
-                        return ERR_PTR(ret);
-                if (ret > 0)
-                        return ERR_PTR(-ENOENT);
+        if (location.objectid == 0)
+                return NULL;
+
+        if (location.type == BTRFS_INODE_ITEM_KEY) {
+                inode = btrfs_iget(dir->i_sb, &location, root);
+                return inode;
+        }
+
+        BUG_ON(location.type != BTRFS_ROOT_ITEM_KEY);
+
+        ret = fixup_tree_root_location(root, dir, dentry,
+                                       &location, &sub_root);
+        if (ret < 0) {
+                if (ret != -ENOENT)
+                        inode = ERR_PTR(ret);
+                else
+                        inode = new_simple_dir(dir->i_sb, &location, sub_root);
+        } else {
                 inode = btrfs_iget(dir->i_sb, &location, sub_root);
-                if (IS_ERR(inode))
-                        return ERR_CAST(inode);
         }
         return inode;
 }
@@ -3283,9 +3417,6 @@ static struct dentry *btrfs_lookup(struct inode *dir, struct dentry *dentry,
 {
         struct inode *inode;
 
-        if (dentry->d_name.len > BTRFS_NAME_LEN)
-                return ERR_PTR(-ENAMETOOLONG);
-
         inode = btrfs_lookup_dentry(dir, dentry);
         if (IS_ERR(inode))
                 return ERR_CAST(inode);
@@ -3691,26 +3822,35 @@ int btrfs_add_link(struct btrfs_trans_handle *trans,
                    struct inode *parent_inode, struct inode *inode,
                    const char *name, int name_len, int add_backref, u64 index)
 {
-        int ret;
+        int ret = 0;
         struct btrfs_key key;
         struct btrfs_root *root = BTRFS_I(parent_inode)->root;
 
-        key.objectid = inode->i_ino;
-        btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY);
-        key.offset = 0;
+        if (unlikely(inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)) {
+                memcpy(&key, &BTRFS_I(inode)->root->root_key, sizeof(key));
+        } else {
+                key.objectid = inode->i_ino;
+                btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY);
+                key.offset = 0;
+        }
+
+        if (unlikely(inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)) {
+                ret = btrfs_add_root_ref(trans, root->fs_info->tree_root,
+                                         key.objectid, root->root_key.objectid,
+                                         parent_inode->i_ino,
+                                         index, name, name_len);
+        } else if (add_backref) {
+                ret = btrfs_insert_inode_ref(trans, root,
+                                             name, name_len, inode->i_ino,
+                                             parent_inode->i_ino, index);
+        }
 
-        ret = btrfs_insert_dir_item(trans, root, name, name_len,
-                                    parent_inode->i_ino,
-                                    &key, btrfs_inode_type(inode),
-                                    index);
         if (ret == 0) {
-                if (add_backref) {
-                        ret = btrfs_insert_inode_ref(trans, root,
-                                                     name, name_len,
-                                                     inode->i_ino,
-                                                     parent_inode->i_ino,
-                                                     index);
-                }
+                ret = btrfs_insert_dir_item(trans, root, name, name_len,
+                                            parent_inode->i_ino, &key,
+                                            btrfs_inode_type(inode), index);
+                BUG_ON(ret);
+
                 btrfs_i_size_write(parent_inode, parent_inode->i_size +
                                    name_len * 2);
                 parent_inode->i_mtime = parent_inode->i_ctime = CURRENT_TIME;
@@ -4800,31 +4940,29 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 {
         struct btrfs_trans_handle *trans;
         struct btrfs_root *root = BTRFS_I(old_dir)->root;
+        struct btrfs_root *dest = BTRFS_I(new_dir)->root;
         struct inode *new_inode = new_dentry->d_inode;
         struct inode *old_inode = old_dentry->d_inode;
         struct timespec ctime = CURRENT_TIME;
         u64 index = 0;
+        u64 root_objectid;
         int ret;
 
-        /* we're not allowed to rename between subvolumes */
-        if (BTRFS_I(old_inode)->root->root_key.objectid !=
-            BTRFS_I(new_dir)->root->root_key.objectid)
+        /* we only allow rename subvolume link between subvolumes */
+        if (old_inode->i_ino != BTRFS_FIRST_FREE_OBJECTID && root != dest)
                 return -EXDEV;
 
-        if (S_ISDIR(old_inode->i_mode) && new_inode &&
-            new_inode->i_size > BTRFS_EMPTY_DIR_SIZE) {
+        if (old_inode->i_ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID ||
+            (new_inode && new_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID))
                 return -ENOTEMPTY;
-        }
 
-        /* to rename a snapshot or subvolume, we need to juggle the
-         * backrefs.  This isn't coded yet
-         */
-        if (old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)
-                return -EXDEV;
+        if (S_ISDIR(old_inode->i_mode) && new_inode &&
+            new_inode->i_size > BTRFS_EMPTY_DIR_SIZE)
+                return -ENOTEMPTY;
 
         ret = btrfs_check_metadata_free_space(root);
         if (ret)
-                goto out_unlock;
+                return ret;
 
         /*
          * we're using rename to replace one file with another.
@@ -4837,6 +4975,9 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 
         trans = btrfs_start_transaction(root, 1);
 
+        if (dest != root)
+                btrfs_record_root_in_trans(trans, dest);
+
         /*
          * make sure the inode gets flushed if it is replacing
          * something.
@@ -4846,18 +4987,22 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry,
                 btrfs_add_ordered_operation(trans, root, old_inode);
         }
 
-        /*
-         * this is an ugly little race, but the rename is required to make
-         * sure that if we crash, the inode is either at the old name
-         * or the new one.  pinning the log transaction lets us make sure
-         * we don't allow a log commit to come in after we unlink the
-         * name but before we add the new name back in.
-         */
-        btrfs_pin_log_trans(root);
+        if (old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID) {
+                /* force full log commit if subvolume involved. */
+                root->fs_info->last_trans_log_full_commit = trans->transid;
+        } else {
+                /*
+                 * this is an ugly little race, but the rename is required
+                 * to make sure that if we crash, the inode is either at the
+                 * old name or the new one.  pinning the log transaction lets
+                 * us make sure we don't allow a log commit to come in after
+                 * we unlink the name but before we add the new name back in.
+                 */
+                btrfs_pin_log_trans(root);
+        }
 
         btrfs_set_trans_block_group(trans, new_dir);
 
-        btrfs_inc_nlink(old_dentry->d_inode);
         old_dir->i_ctime = old_dir->i_mtime = ctime;
         new_dir->i_ctime = new_dir->i_mtime = ctime;
         old_inode->i_ctime = ctime;
@@ -4865,47 +5010,58 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry,
         if (old_dentry->d_parent != new_dentry->d_parent)
                 btrfs_record_unlink_dir(trans, old_dir, old_inode, 1);
 
-        ret = btrfs_unlink_inode(trans, root, old_dir, old_dentry->d_inode,
-                                 old_dentry->d_name.name,
-                                 old_dentry->d_name.len);
-        if (ret)
-                goto out_fail;
+        if (unlikely(old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)) {
+                root_objectid = BTRFS_I(old_inode)->root->root_key.objectid;
+                ret = btrfs_unlink_subvol(trans, root, old_dir, root_objectid,
+                                        old_dentry->d_name.name,
+                                        old_dentry->d_name.len);
+        } else {
+                btrfs_inc_nlink(old_dentry->d_inode);
+                ret = btrfs_unlink_inode(trans, root, old_dir,
+                                         old_dentry->d_inode,
+                                         old_dentry->d_name.name,
+                                         old_dentry->d_name.len);
+        }
+        BUG_ON(ret);
 
         if (new_inode) {
                 new_inode->i_ctime = CURRENT_TIME;
-                ret = btrfs_unlink_inode(trans, root, new_dir,
-                                         new_dentry->d_inode,
-                                         new_dentry->d_name.name,
-                                         new_dentry->d_name.len);
-                if (ret)
-                        goto out_fail;
+                if (unlikely(new_inode->i_ino ==
+                             BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)) {
+                        root_objectid = BTRFS_I(new_inode)->location.objectid;
+                        ret = btrfs_unlink_subvol(trans, dest, new_dir,
+                                                root_objectid,
+                                                new_dentry->d_name.name,
+                                                new_dentry->d_name.len);
+                        BUG_ON(new_inode->i_nlink == 0);
+                } else {
+                        ret = btrfs_unlink_inode(trans, dest, new_dir,
+                                                 new_dentry->d_inode,
+                                                 new_dentry->d_name.name,
+                                                 new_dentry->d_name.len);
+                }
+                BUG_ON(ret);
                 if (new_inode->i_nlink == 0) {
                         ret = btrfs_orphan_add(trans, new_dentry->d_inode);
-                        if (ret)
-                                goto out_fail;
+                        BUG_ON(ret);
                 }
-
         }
         ret = btrfs_set_inode_index(new_dir, &index);
-        if (ret)
-                goto out_fail;
+        BUG_ON(ret);
 
-        ret = btrfs_add_link(trans, new_dentry->d_parent->d_inode,
-                             old_inode, new_dentry->d_name.name,
+        ret = btrfs_add_link(trans, new_dir, old_inode,
+                             new_dentry->d_name.name,
                              new_dentry->d_name.len, 1, index);
-        if (ret)
-                goto out_fail;
+        BUG_ON(ret);
 
-        btrfs_log_new_name(trans, old_inode, old_dir,
-                                       new_dentry->d_parent);
-out_fail:
+        if (old_inode->i_ino != BTRFS_FIRST_FREE_OBJECTID) {
+                btrfs_log_new_name(trans, old_inode, old_dir,
+                                   new_dentry->d_parent);
+                btrfs_end_log_trans(root);
+        }
 
-        /* this btrfs_end_log_trans just allows the current
-         * log-sub transaction to complete
-         */
-        btrfs_end_log_trans(root);
         btrfs_end_transaction_throttle(trans, root);
-out_unlock:
+
         return ret;
 }
 
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index ef0188fb3cc4..9b3a88755e51 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -322,20 +322,9 @@ static noinline int create_subvol(struct btrfs_root *root,
         ret = btrfs_update_inode(trans, root, dir);
         BUG_ON(ret);
 
-        /* add the backref first */
         ret = btrfs_add_root_ref(trans, root->fs_info->tree_root,
-                                 objectid, BTRFS_ROOT_BACKREF_KEY,
-                                 root->root_key.objectid,
+                                 objectid, root->root_key.objectid,
                                  dir->i_ino, index, name, namelen);
-
-        BUG_ON(ret);
-
-        /* now add the forward ref */
-        ret = btrfs_add_root_ref(trans, root->fs_info->tree_root,
-                                 root->root_key.objectid, BTRFS_ROOT_REF_KEY,
-                                 objectid,
-                                 dir->i_ino, index, name, namelen);
-
         BUG_ON(ret);
 
         ret = btrfs_commit_transaction(trans, root);
diff --git a/fs/btrfs/orphan.c b/fs/btrfs/orphan.c
index 3c0d52af4f80..79cba5fbc28e 100644
--- a/fs/btrfs/orphan.c
+++ b/fs/btrfs/orphan.c
@@ -65,3 +65,23 @@ out:
         btrfs_free_path(path);
         return ret;
 }
+
+int btrfs_find_orphan_item(struct btrfs_root *root, u64 offset)
+{
+        struct btrfs_path *path;
+        struct btrfs_key key;
+        int ret;
+
+        key.objectid = BTRFS_ORPHAN_OBJECTID;
+        key.type = BTRFS_ORPHAN_ITEM_KEY;
+        key.offset = offset;
+
+        path = btrfs_alloc_path();
+        if (!path)
+                return -ENOMEM;
+
+        ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+
+        btrfs_free_path(path);
+        return ret;
+}
diff --git a/fs/btrfs/root-tree.c b/fs/btrfs/root-tree.c
index 0ddc6d61c55a..5ef72599a581 100644
--- a/fs/btrfs/root-tree.c
+++ b/fs/btrfs/root-tree.c
@@ -278,31 +278,57 @@ out:
         return ret;
 }
 
-#if 0 /* this will get used when snapshot deletion is implemented */
 int btrfs_del_root_ref(struct btrfs_trans_handle *trans,
                        struct btrfs_root *tree_root,
-                       u64 root_id, u8 type, u64 ref_id)
+                       u64 root_id, u64 ref_id, u64 dirid, u64 *sequence,
+                       const char *name, int name_len)
+
 {
+        struct btrfs_path *path;
+        struct btrfs_root_ref *ref;
+        struct extent_buffer *leaf;
         struct btrfs_key key;
+        unsigned long ptr;
+        int err = 0;
         int ret;
-        struct btrfs_path *path;
 
         path = btrfs_alloc_path();
+        if (!path)
+                return -ENOMEM;
 
         key.objectid = root_id;
-        key.type = type;
+        key.type = BTRFS_ROOT_BACKREF_KEY;
         key.offset = ref_id;
-
+again:
         ret = btrfs_search_slot(trans, tree_root, &key, path, -1, 1);
-        BUG_ON(ret);
-
-        ret = btrfs_del_item(trans, tree_root, path);
-        BUG_ON(ret);
+        BUG_ON(ret < 0);
+        if (ret == 0) {
+                leaf = path->nodes[0];
+                ref = btrfs_item_ptr(leaf, path->slots[0],
+                                     struct btrfs_root_ref);
+
+                WARN_ON(btrfs_root_ref_dirid(leaf, ref) != dirid);
+                WARN_ON(btrfs_root_ref_name_len(leaf, ref) != name_len);
+                ptr = (unsigned long)(ref + 1);
+                WARN_ON(memcmp_extent_buffer(leaf, name, ptr, name_len));
+                *sequence = btrfs_root_ref_sequence(leaf, ref);
+
+                ret = btrfs_del_item(trans, tree_root, path);
+                BUG_ON(ret);
+        } else
+                err = -ENOENT;
+
+        if (key.type == BTRFS_ROOT_BACKREF_KEY) {
+                btrfs_release_path(tree_root, path);
+                key.objectid = ref_id;
+                key.type = BTRFS_ROOT_REF_KEY;
+                key.offset = root_id;
+                goto again;
+        }
 
         btrfs_free_path(path);
-        return ret;
+        return err;
 }
-#endif
 
 int btrfs_find_root_ref(struct btrfs_root *tree_root,
                    struct btrfs_path *path,
@@ -319,7 +345,6 @@ int btrfs_find_root_ref(struct btrfs_root *tree_root,
         return ret;
 }
 
-
 /*
  * add a btrfs_root_ref item.  type is either BTRFS_ROOT_REF_KEY
  * or BTRFS_ROOT_BACKREF_KEY.
@@ -335,8 +360,7 @@ int btrfs_find_root_ref(struct btrfs_root *tree_root,
  */
 int btrfs_add_root_ref(struct btrfs_trans_handle *trans,
                        struct btrfs_root *tree_root,
-                       u64 root_id, u8 type, u64 ref_id,
-                       u64 dirid, u64 sequence,
+                       u64 root_id, u64 ref_id, u64 dirid, u64 sequence,
                        const char *name, int name_len)
 {
         struct btrfs_key key;
@@ -346,13 +370,14 @@ int btrfs_add_root_ref(struct btrfs_trans_handle *trans,
         struct extent_buffer *leaf;
         unsigned long ptr;
 
-
         path = btrfs_alloc_path();
+        if (!path)
+                return -ENOMEM;
 
         key.objectid = root_id;
-        key.type = type;
+        key.type = BTRFS_ROOT_BACKREF_KEY;
         key.offset = ref_id;
-
+again:
         ret = btrfs_insert_empty_item(trans, tree_root, path, &key,
                                       sizeof(*ref) + name_len);
         BUG_ON(ret);
@@ -366,6 +391,14 @@ int btrfs_add_root_ref(struct btrfs_trans_handle *trans,
         write_extent_buffer(leaf, name, ptr, name_len);
         btrfs_mark_buffer_dirty(leaf);
 
+        if (key.type == BTRFS_ROOT_BACKREF_KEY) {
+                btrfs_release_path(tree_root, path);
+                key.objectid = ref_id;
+                key.type = BTRFS_ROOT_REF_KEY;
+                key.offset = root_id;
+                goto again;
+        }
+
         btrfs_free_path(path);
-        return ret;
+        return 0;
 }
diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c
index 94f816cb6e35..915077725fea 100644
--- a/fs/btrfs/transaction.c
+++ b/fs/btrfs/transaction.c
@@ -779,24 +779,14 @@ static noinline int finish_pending_snapshot(struct btrfs_fs_info *fs_info,
         ret = btrfs_update_inode(trans, parent_root, parent_inode);
         BUG_ON(ret);
 
-        /* add the backref first */
         ret = btrfs_add_root_ref(trans, parent_root->fs_info->tree_root,
                                  pending->root_key.objectid,
-                                 BTRFS_ROOT_BACKREF_KEY,
                                  parent_root->root_key.objectid,
                                  parent_inode->i_ino, index, pending->name,
                                  namelen);
 
         BUG_ON(ret);
 
-        /* now add the forward ref */
-        ret = btrfs_add_root_ref(trans, parent_root->fs_info->tree_root,
-                                 parent_root->root_key.objectid,
-                                 BTRFS_ROOT_REF_KEY,
-                                 pending->root_key.objectid,
-                                 parent_inode->i_ino, index, pending->name,
-                                 namelen);
-
         inode = btrfs_lookup_dentry(parent_inode, pending->dentry);
         d_instantiate(pending->dentry, inode);
 fail: