If we have the following sequence of events
btrfs sub create /a
btrfs sub create /a/b
btrfs sub snap /a /c
mkdir /c/foo
mv /a/b /mnt/test/c/foo
rm -rf /mnt/test/*
We will end up with a transaction abort.
The reason for this is because we create a root ref for b pointing to a.
When we create a snapshot of c we still have b in our tree, but because
the root ref points to a and not c we will make it appear to be empty.
The problem happens when we move b into c. This removes the root ref
for b pointing to a and adds a ref of b pointing to c. When we rmdir c
we'll see that we have a ref to our root and remove the root ref,
despite not actually matching our reference name.
Now btrfs_del_root_ref() allowing this to work is a bug as well, however
we know that this inode does not actually point to a root ref in the
first place, so we shouldn't be calling btrfs_del_root_ref() in the
first place and instead simply look up our dir index for this item and
do the rest of the removal.
Signed-off-by: Josef Bacik <josef@xxxxxxxxxxxxxx>
---
fs/btrfs/inode.c | 27 +++++++++++++++++++--------
1 file changed, 19 insertions(+), 8 deletions(-)
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 3cbdca2749bd..db67e1984c91 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -4279,13 +4279,16 @@ static int btrfs_unlink_subvol(struct btrfs_trans_handle *trans,
}
btrfs_release_path(path);
- ret = btrfs_del_root_ref(trans, objectid, root->root_key.objectid,
- dir_ino, &index, name, name_len);
- if (ret < 0) {
- if (ret != -ENOENT) {
- btrfs_abort_transaction(trans, ret);
- goto out;
- }
+ /*
+ * This is a placeholder inode for a subvolume we didn't have a
+ * reference to at the time of the snapshot creation. In the meantime
+ * we could have renamed the real subvol link into our snapshot, so
+ * depending on btrfs_del_root_ref to return -ENOENT here is incorret.
+ * Instead simply lookup the dir_index_item for this entry so we can
+ * remove it. Otherwise we know we have a ref to the root and we can
+ * call btrfs_del_root_ref, and it _shouldn't_ fail.
+ */
+ if (btrfs_ino(inode) == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID) {
di = btrfs_search_dir_index_item(root, path, dir_ino,
name, name_len);
if (IS_ERR_OR_NULL(di)) {
@@ -4300,8 +4303,16 @@ static int btrfs_unlink_subvol(struct btrfs_trans_handle *trans,
leaf = path->nodes[0];
btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
index = key.offset;
+ btrfs_release_path(path);
+ } else {
+ ret = btrfs_del_root_ref(trans, objectid,
+ root->root_key.objectid, dir_ino,
+ &index, name, name_len);
+ if (ret) {
+ btrfs_abort_transaction(trans, ret);
+ goto out;
+ }
}
- btrfs_release_path(path);
ret = btrfs_delete_delayed_dir_index(trans, BTRFS_I(dir), index);
if (ret) {
--
2.23.0
