Subject: Re: DUOTEST kernel
To: Daishi Kato <daishi@axlight.com>
From: Bob Nestor <rnestor@augustmail.com>
List: port-mac68k
Date: 11/08/1999 07:06:48
Daishi Kato  (daishi@axlight.com) wrote:

>At 0:44 PM +0100 1999.11.8, svinto@ita.chalmers.se wrote:
>> usually it hangs at "ncrscsi0 at obio0", sometimes it hangs at "adb0
>> at obio0".
>
>Do I need to make SBC version? I don't know.
>maybe later...
>
I don't believe the system would hang here if it was just a matter of the 
ncrscsi vs. sbc.

>> with normal console i usually get this:
>>  mrg: 'Mac II class ROMs' ROM glue, tracing off, debug off, silent
>> traps
>
>Really?
>In my case, it is
>  mrg: '2nd Powerbook class ROMs' ROM glue, ......
>
>There seems an entry of "Duo 270C ROMs" which does not used.
>We need somebody to help us about this.
>
The entry which is used is based on the machine ID obtained under MacOS 
using the Gestalt API system call.  Unfortunately there are a number of 
machines that return the same machine ID but use a different set of ROM 
code, so whatever MacOS is using to uniquely identify these systems, it's 
more than just the Machine ID.  That's one of the reasons the MRG 
approach was dropped in favor of the direct ADB code.  The other problem 
with the MRG approach is that MacOS, in some cases, patches the ROM 
routines using system INITs.  Without these patches the ROM code doesn't 
work on some of the newer 68k machines.

As a test you can try changing the Machine ID of the other entries in the 
table so there's only one match for your system.  Or change the Machine 
ID of the "Duo 270c" entry to a unique value and force that value to be 
used in the Booter. I think there's an option to do this in newer 
versions of the Booter.

Ultimately you'll want to get the system using the ADB direct code that's 
now in GENERIC builds.

Hope this helps,
-bob