From mboxrd@z Thu Jan  1 00:00:00 1970
Date: Sun,  1 Oct 1995 02:30:12 -0400
From: peter@plan9.com peter@plan9.com
Subject: Exception #4 with net booting a NeXT Cube
Topicbox-Message-UUID: 29ecf7a8-eac8-11e9-9e20-41e7f4b1d025
Message-ID: <19951001063012.8Hz6DRP4KuWkRw3okAddYyLqKYCZjcMgGpwBkzLjv4Q@z>

> I was trying to boot my NeXT Cube (ROM 2.4) into a Plan 9 terminal
> with u9fs running on a PC running NeXTSTEP. While loading
> 9nextstation, I got Exception #4. What does this mean and how can I
> fix it?

The problem is the 2l program that produced the 9nextstation kernel  
does not setup for the entry point properly.  This problem will not  
show itself if you use an early verson of the NeXT boot program.

A quick fix if you don't have an early version boot program is to  
start the kernel manually with the following commands to the ROM  
monitor, after the exception occurs:

> r pc
? 4000000
> c

This changes the program counter to the correct entry point 4000000  
and then execution is begun at that point.

Geoff Collyer of Bell Labs gave me considerable help while I was  
tracking this problem down.  Be sure you have his bug fix installed  
ih the kernel.  If you don't it won't get very far.  I repeat it  
here.  These diffs are relative to the cd rom.

diff ./io.h /sys/src/9/next/io.h
19a20
> #define VIDEOOUTLIMIT	IO(0x02004184)
diff ./main.c /sys/src/9/next/main.c
66a67,76
>
> 	/*
> 	 * the boot rom is supposed to set all the video out  
registers sensibly,
> 	 * and indeed the v2.2 boot rom does so.  The v2.1 boot rom  
does not,
> 	 * and the video limit register contains a very large  
garbage number;
> 	 * thus we need to set the limit register here in order to  
get `clock'
> 	 * interrupts near each frame end to run the scheduler.   
The value 0xea
> 	 * is what the v2.2 boot rom sets the limit register to.
> 	 */
> 	*(ulong*)VIDEOOUTLIMIT = 0xea;

Regards,

Peter G. Weiner