Ken Nakata <kenn@synap.ne.jp> wrote:

>On Fri, 21 Aug 1998 23:51:58 +1000, Paul Ripke wrote:
>> 
>> I'm afraid I'm not particularly willing to wipe my IDE drive, tho...
>> but booting a MacOS based kernel and doing a probe, or the HFS
>> test from NetBSD should be fine.
>
>That's quite all right.  There are some other things we must first
>take care of.  First of all, you should contact Scott Reynolds and Bob
>Nestor to update ROM vector table, intvid_info[], and so on, for the
>PB190.

I didn't realize we were still trying to fill in the tables used by the 
MRG code since the switch to ADBDIRECT code.  There are some limitations 
with the MRG code in that it doesn't always work for NetBSD.  In some 
cases the Gestalt ID isn't enough to distinguish between different 
members of the same family that use different ROM based code.  Also there 
are some Macs that seem to require patches to the ROM routines that we 
don't have in NetBSD. MacOS gets around this by loading INITs at boot 
time I think.

Anyway, if the addresses are still being used in some configurations the 
process of getting them is pretty simple.  Just grab a copy of the "ROM 
Sniffer" program from puma and run it on the Mac system in question. This 
is a MacOS application that will attempt to grab the necessary ROM 
addresses that are used by NetBSD.  Once this has been done all that's 
needed is to post the results to the mailing list so people who have the 
ability to build test kernels can do so.  Once the addresses have been 
verifed under NetBSD with a test kernel we can ask Scott Reynolds to 
include them in the source tree.

I can provide a copy of the ROM Sniffer program for anyone who needs it, 
but I believe the one on puma is more up-to-date than the original I 
have.  Unless I'm mistaken Hauke Fath(?) made some additions to the 
program to get information on floppies to support his floppy driver 
development effort.  Posting the results from his version might help him 
as well.

Oh, don't be alarmed or worried about some of the bogus informaton the 
program might display about the Mac it runs on.  These errors are part of 
the original code I wrote and are not in any way related to the additions 
that Hauke made, and they in no way affect the intregrity of the critical 
ROM addresses that are displayed.

-bob