The inode size and layout were changed to allow more directory entries and hard links. This means that 6fs cannot mount file systems with the old layout anymore. You need to recreate your 6fs file systems from scratch. (A conversion of the old format to the new format would be possible, but is not implemented yet because 6fs is still experimental anyway.)
6fs now uses mmap on the six storage files by default, which improves performance a lot.
(However, to get anywhere near decent performance on a larger scale, much better data structures
would be required. Since 6fs is mainly written for learning about FUSE and file systems, a
switch to more sophisticated data structures is unlikely to happen.)
The initial version of 6fs is now in the public git repository.