Subject: Re: Where/how is rootvp set in kernel initialisation?
To: Chris G. Demetriou <>
From: Bill Studenmund <>
List: tech-kern
Date: 09/25/2000 11:54:50
On 21 Sep 2000, Chris G. Demetriou wrote:

> Colin Wood <> writes:
> > certainly, but there still remains one rather tiny problem before sysinst
> > is truly useful on mac68k:  native HFS support.
> that reminds me: is there some great technical hurdle standing in the
> way of that?

Time is the main one.

> IIRC, HFS (and MFS) were adequately described in the old Inside
> Macintosh volumes (which i happen to own copies of -- before i found
> this whole UNIX thing, i used to be a mac hacker 8-).

It'd be nice if whatever we did would do HFS+ too. But that adds a
big-enough can of worms to delay the project. :-)

It'd also be nice to be able to write.

Given that, using the Apple Darwin HFS/HFS+ code would make the most sense
- the we're bug-compatible with Apple. The problem is that they introduced
the concept of Complex vnodes, to reflect the different file forks of an
HFS file. I'm not sure how best to do the locking on those, and I got
quite confused the one time I tried to port Apple's code. Though
admittedly I didn't understand Complex vnodes at that point.

Take care,