Re: Making an HFS+ Filesystem

[Jeff V. Merkey]

You should probably look at the existing HFS that's already there.  Word
of caution from someone who's fighting this now ..... 2.0 and 2.2/2.3
are different, and it looks like the VFS will continue to evolve more
and more towards and "EXT-ish" flavor of doing stuff (i.e. all the
dcache stuff, etc.).  Most of it's very good work, and is comparable to
the level of capability in Windows NT/2000, and it scales very well (2.2
and forward that is).  2.3 is a moving target, and you may expose lots
of bugs you think are yours but aren't if you try to do first pass FS
development on this one.  You may want to start with 2.0/2.2 first since
these are more stable (this is what we are doing) get your stuff feature
complete and stable on 2.0/2.2 then tackle 2.3 (and hope it's not
changing underneath you).    Just a suggestion.  

Good Luck,

Your friend,

Jeff

Brian Kidder wrote:
> 
> I've been unable to locate any current effort to make Apple's HFS+
> format available to Linux, so I'm considering making a stab at it.
> 
> Before I get to code, though, I need to learn how the kernel and VFS
> system talk with the file system modules.  I'd like to learn from an
> existing module that is (relatively) intelligently written, and
> up-to-date with where the kernel and VFS are going.  Using the old HFS
> module is sounding more and more like a no, no, given the changes that
> have happened since then.  I am familiar with, and have detailed specs
> on, the HFS+ format; its the linux side of things that are weak.
> 
> What single-user-centric file system would make a good example?  I'm
> assuming that ext2 and ext3 would make good multi-user-centric file
> system examples.
> 
> Cheers,
> -Brian Kidder
> 
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [e-mail]
> Please read the FAQ at http://www.tux.org/lkml/