Subject: Re: Summer Of Code - Filesystem What do you think ?
To: None <>
From: Matthew Orgass <>
List: tech-kern
Date: 06/09/2005 14:56:42
On 2005-06-09 wrote:
> On 6/9/05, Jonathan Buschmann <> wrote:
> > Hello everyone,
> >
> > Currently i'm searching in which way i can help BSD system and in this
> > case NetBSD.
> > And i would help adding another filesystem like JFS or HFS+.
> >
> > I'm not sure that all the capabilities of these filesystem can be
> > developped during this 2 months.

  :) Even so, getting core functionality of a file system working in that
time can make it much easier for others to work on it.  Support for HFS+
at least has major interoperability benefits (I don't know about JFS), and
even just a core implementation could be very helpful.  There could easily
be non-conflicting projects related to HFS+, such as native APIs and tar
support for "resource forks" (which may be the same as for ACLs, if no one
is working on that project) or a fsck_hfs (it should be possible to work
on this with one of the other implementations pending native support).
fsck for any file system currently supported other than ffs could be used
as a project also (a good fsck_msdos would be quite helpful, IMO).  There
should be projects available for LFS if you are interested in that; it
sounds like LFS may be getting close to being reliable if that appeals to
you (which is highly unlikely to be the case for any new file system for
quite some time).

> > What do you think about that ? Which filesystem is the most interesting
> > to learn ?
> This is an interesting question and it is a shame this thread went
> almost no where with it.

   My understanding of LFS is that it is a very flexable file system, so
that in addition to the core file system there is an opportunity to make
various tradeoffs in usage and cleaner policies.

> I propose the idea of designing and creating a new file system and not
> worrying about all of the other file system implementations.  File
> structures and storage mechanisms are well documented through-out
> computer science books, including several ideas I have yet to see
> fully implemented and exploited.

  That would be good too, and hopefully any new file system would minimize
complexity as much as possible (taking advantage of layering) and could
be easier to implement than most existing ones.

Matthew Orgass