On Fri, 9 Oct 2026 at 01:00, John Groves <[email protected]> wrote: > Fewer fields is nice, but it's on the honor system that the server allocated > large enough strips - since strips carry the chunk_size rather than their > actual size when CYCLIC is set. IMO it would be an improvement to add the > chunk_size to the header; then the strip sizes could be validated.
Cyclic maps could easily be generalized to not require a fixed chunk size. Giving a different meaning to extent length (strip length) would break that, it would no longer be a CYCLIC map. It's an un-generalization that has dubious value, since the strip length is implicit in either the file size or the parent extent length. In other words: make the server validate the strip size, and all will be good. The kernel really doesn't need this information. > So that's a bug. I've fixed it locally and will share that in the next day > or two. I also needed to refactor your patch the famfs user space because > it disabled ABI compatibility with some older versions of the famfs kernel > that I'm still supporting users with. If you can wait to patch that further > until I share an update, it will be cleaner on my end. Yeah, thanks. > Having said all that, this topic seems to recur and be contentious - we are > working on tests that can demonstrate bottlenecks - but my current ask is > that if I can offer an alternative approach to rbtree that is better on > paper, can we please use it? If it's not adding complexity, then sure. Thanks, Miklos
