ext4: Fix possible corruption when moving a directory

[ Upstream commit 0813299 ]

When we are renaming a directory to a different directory, we need to
update '..' entry in the moved directory. However nothing prevents moved
directory from being modified and even converted from the inline format
to the normal format. When such race happens the rename code gets
confused and we crash. Fix the problem by locking the moved directory.

CC: stable@vger.kernel.org
Fixes: 32f7f22 ("ext4: let ext4_rename handle inline dir")
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20230126112221.11866-1-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Sasha Levin <sashal@kernel.org>

fs/ext4/namei.c

@@ -3873,9 +3873,16 @@ static int ext4_rename(struct user_namespace *mnt_userns, struct inode *old_dir,
 			if (new.dir != old.dir && EXT4_DIR_LINK_MAX(new.dir))
 				goto end_rename;
 		}
+		/*
+		 * We need to protect against old.inode directory getting
+		 * converted from inline directory format into a normal one.
+		 */
+		inode_lock_nested(old.inode, I_MUTEX_NONDIR2);
 		retval = ext4_rename_dir_prepare(handle, &old);
-		if (retval)
+		if (retval) {
+			inode_unlock(old.inode);
 			goto end_rename;
+		}
 	}
 	/*
 	 * If we're renaming a file within an inline_data dir and adding or
@@ -4007,6 +4014,8 @@ static int ext4_rename(struct user_namespace *mnt_userns, struct inode *old_dir,
 	} else {
 		ext4_journal_stop(handle);
 	}
+	if (old.dir_bh)
+		inode_unlock(old.inode);
 release_bh:
 	brelse(old.dir_bh);
 	brelse(old.bh);