Subject: Re: Kernel startup and entry vectors
To: Dan Winship <danw@ximian.com>
From: Louis Gerbarg <gerbal@cs.rpi.edu>
List: port-macppc
Date: 04/23/2001 19:21:08
On 24 Apr 2001, Dan Winship wrote:

> > 1. Where exactly is your equivalent of head.s/start.s
> > 2. What is your kernel's current entry point
> 
> __start in sys/arch/macppc/macppc/locore.S
> 
> You might also find sys/arch/macppc/stand/ofwboot/ interesting (that's
> the current booter).

Thanks. Both of those should be very useful. I actually kind of meant,
where is the entry vector in memory, but I suppose I can read that out of
the ELF header or something... Since that could be done with linker tricks
or some such I figured I would ask.
 
> > 3. Are there multiple entry points
> 
> Nope

Okay. Are you adverse to having a second entry point if necessary?

> > 4. Does the kernel expect to be in real-mode on entry
> 
> Blah. Haven't looked at low-level ppc stuff in too long. Does "real
> mode" mean no memory mapping? If so, then yes.

Yeah, that was what I wanted to know.

> > 5. Does the kernel expect OF to quietsce before entry
> 
> No.

This one is kind of tricky. The OS X kernel expects the system to be
quietsce before entry. I might be able to convinve them to move that from
the end of the booter to the beginning of start.s (their equivalent of
locore.S), but I am not optimistic. What exactly do you guys use OF for
after ofwboot, actual interpretted commands, or just building the device
tree.

> > 6. What exactly does your FFS look like (4.2BSD? Apple's is a derivative
> > of 4.4BSD)
> 
> 4.4ish. (I don't know more than that...)

It looks to me, upon cursory inspection, that ofwboot can't read UFS at
the moment, so I suppose that can take a back burner. Since Apple's is
also 4.4ish it should just take someone staring at it for a few minutes to
find the differences (famous last words).

Thanks.

Louis