68k homebuilts

John Floren wrote: In 1985 I was fresh out of school, moved to Sunnyvale CA to work for BTI Computer Systems, and in within two weeks of starting work had been through the first round of layoffs. They were kind on the college kids and didn't lay us off, but told us to start interviewing. As a result, I had some time on my hands. One thing I did was to design and build a 68000 computer. I used an S-100 backplane and boards, although I didn't adhere to any standard. The design had 256KB of DRAM, a uart, and a timer chip. I debugged it by writing a "JMP *" program and looking at the waves with an oscope until it looked like it was working. Then I wrote a short program that printed some fixed string to the UART and didn't use any DRAM. Then I wrote a simple program to read from the uart and echo it back, again not using DRAM. Once I had I/O capabilities, it was easy to write short memory test programs until I had that debugged. Next I found the source for a 68K monitor program, call VuBug. As I recall the 68K assembler was in the BTI dialect of FORTRAN. I believe it was a version of RM Fortran (Ryan McFarland). The 68K debugger was primitive, so I enhanced it. Luckily, I managed to keep the 5.25" floppies that contained the source, and you are welcomed to use it if you want. Here is the comment block from the debugger: **************************************************************************** * * * DBug is a 68000 debugger. It allows downloading, examining/changing * * memory, disassembling, single-stepping, tracing, breakpointing, * * and examining/changing registers. It was originally called VuBug. * * * * This is its history: * * * * 1) Vubug was written by Edward M. Carter at Vanderbilt University. * * See the copyright notice below. * * * * 2) Mark Isfeld and Dave Conroy modified VuBug for work on the SCC * * project. The modifications they made were: * * -- * * * * 3) Jim Battle took VuBug and modified it in these ways: * * -- made syntax checking stronger and error reporting better * * -- added the Lookup subroutine to allow word commands instead of * * single letter commands * * -- added text substitute to substitute memory (SM) command * * -- changed copy to allow bigger than 64k blocks * * -- rewrote register (R) command * * -- added ability to evaluate expressions and changed all commands to * * allow expressions for parameters * * -- added disassembly capability * * -- added symbol managment and changed the load command to read * * symbols. * * -- added ascii, asm, bsr, clear, clearsym, dis, fill, help, set, * * setsym, and = (calc) commands * * -- changed I/O dependent routines (especially the TERM command) * * * **************************************************************************** ******************************************* * * ******** VUBUG ********* * * * Copyright (C) 1983 * * Vanderbilt Univ. * * Comp. Sci. Dept. * * PO Box 1679 * * Station B * * Nashville, Tenn. * * 37235 * * * * Author: * * Edward M. Carter * * * ******************************************* * * * Although the information contained here-* * in, as well as any information provided * * relative thereto, has been carefully re-* * viewed and is believed correct, Vander- * * bilt University assumes no liability * * arising out of its application or use, * * neither does it convey any license under* * its patent rights nor the rights of * * others. * * * ******************************************* DBUG.68K: 4227 lines (a few authors) ASSEM.68K: 2447 lines (entirely written by me) DISASSEM.68K: 2308 lines (entirely written by me) It was the first 68K code I ever wrote so it is pretty brute-force, which means it is easy to understand: * output in #<quick8>,Dx format. <quick8> is [7:0] and Dx is [11:9] Dis_MOVEQ equ * move.b #'.',(a1)+ send '.L' as a reminder move.b #'L',(a1)+ movea.l a4,a1 point to operand buffer move.b #'#',(a1)+ move.w d0,d5 bsr SymOr_slzsb signed, leading-zero-suppressed format move.b #',',(a1)+ move.w d5,d0 recall opcode rol.w #7,d0 move bits [11:9] to [2:0] bsr Make_Dx rts Once I had it all working, I simply moved on to something else. I had no need to actually use it for anything, since I had quite capable computers all around me anyway. That next project was a modula compiler. I only completed a subset of that before I got my next job, and then there was no time to develop it further. It will need to wait for my retirement in 20 years. :-)