hi,

> Okay, the problem with the partitions seems to be solved: They are
> organized similar to SunOS, partition C: is sd0d and not sd0a, right?
Yep, the partition sceme is:
     partition a: A root-filesystem (if any)
     partition b: Swap space (if any)
     partition c: Whole disk (always :-) )
     partition d: First user/other partition
              ...
              ...
     partition p: Last user/other partition
> 
> But: Even repartitioning the harddisk doesn't fix the problem with the
> "MMU fault".
I just don't get it, why the MMU-fault! I haven't had MMU faults with this
kernel since januari (exept when resizing consoles). Where exactly are
you in the installation-process. As get it, you are stuck at the point
where the first filesystems are created. As you gave an example of
 "newfs /dev/sd1a". _Did_ this work manually? 

As 4 or 5 people before you installed it on a TT by now, there must be
something specific to your configuration. I think the disk is not the 
problem unless it reports read/write or other errors. Maybe you can
include a "disklabel sd1" in your next mail. The next important thing
is the memory configuration. You should try 'loadbsd -tD netbsd' and
check the memory the bootloader _thinks_ you have against the real
configuration. Especially the st/tt ram sizes and the tt-ram start
address. To do this, you shouldn't use 'loadbsd' on the bootfloppy but
the newer one in the 'atari/tools' directory.
One last point, do you have a very old TT? I used to have an old motherboard
with a little daughter-board on it. There was something wrong with DMA
to TT-ram (atari confirmed it). As you have little st-ram, nearly all
DMA goes to TT-ram in your case.

  *** Is  there anybody out there who installed it on a TT with ***
  *** 2 Mb st-ram and XMb tt-ram?????                           ***

I have a define in the SCSI driver to handle it but it isn't activated on
your kernel. As you have run Linux before it should not be the problem
(unless Linux bounces all DMA above 16Mb what I consider unlikely). Maybe
I should put up a very conservative kernel on sun-lamp to outrule this kind
of problems.

> 
> The problem is: I can't read the last messages that are printed on the
> screen because the debugger output scrolls everything up... :-(
> 
> But I think I even read something about "fstab" before the error happened.
> 
I think it might indeed be a nice feature to pause before the stack dump.

> Is there a newer kernel arount? My version is from Mar 30, I think.
Dependent on the outcome, maybe there should be one soon.

Bye, Leo.