Subject: Re: NetBSD/mvme (port to the MVME147) available
To: Ignatios Souvatzis <email@example.com>
From: Theo de Raadt <firstname.lastname@example.org>
Date: 06/29/1995 16:36:52
> > NetBSD/mvme - a port of NetBSD 1.0 Release to the MVME 147 processor.
> > This port borrowed from the amiga and the sun3/sparc ports for drivers
> > and part vm system.
> Is this in any way related to Theo de Raadts effort? (besides using the
> same hardware)
not really. but let me tell you a bit of the history.
chuck cranor got the mvme147 up and running (i merely convinced
someone to mail chuck a unit :-). my code is based on chuck's code.
(a lot bunch of things have changed since chuck last touched it!)
dale's stuff is different: quite a bit different actually. some of his
stuff is better, some of our stuff is better. our code is based on the
hp300 tree with device drivers hacked from the sparc and da30 trees.
for the moment i suspect there will be just cross pollination -- i am
definately going to pull in the good things from dale's port -- i do
not know what dale's plans are.
i am under contract with a company to make netbsd run on the
68040-based mvme162 card. i don't have time to switch to dale's
codebase, and i think some of the things i have working now are better
... but, he has scsi working -- our stuff is nfs diskless only at the
moment. but some of those critical things are a 68040-aware pmap
module, a zs driver that can deal with zs-chip memory arrangement
differences, a driver for the idiotic i82596 chip, etc. i've got the
mvme162 somewhat going, while the same kernel runs fine on the
mvme147. it is faster/easier for me to look at what is better in
dale's port and merge it into my codebase.
dale has an mvme167, which is slightly different from the mvme162
(most significant difference is that it uses a cd1400 serial chip).
by the way, there is an mvme177 board which uses the 68060. apparently
it is quite like the mvme167. there is also a 68060-based mvme172
coming out in about two months, but i haven't been able to find out
what chips are on it. motorola definately overdesigns their boards:
the vmechip on the 162 has 31 interrupt sources (there are many other
sources on the board). i have 6 32-bit 1MHz timers at my disposal..
also, you now know that i'm the mysterious 3rd person interested in
Copyright (c) 1982, 1986, 1989, 1991, 1993
The Regents of the University of California. All rights reserved.
NetBSD 1.0A (VME147) #126: Thu Jun 29 15:44:58 MDT 1995
Motorola MVME162-223: 25MHz MC68040 CPU+MMU+FPU, 4k on-chip physical I/D caches)
real mem = 4194304
avail mem = 2187264
using 51 buffers containing 208896 bytes of memory
mc0 at mainbus0 addr 0xfff00000: id 0x84 rev 0
clock0 at mc0 offset 0xc1ff8 ipl 5
zs0 at mc0 offset 0x45000 ipl 4
zs1 at mc0 offset 0x45801 ipl 4
ie0 at mc0 offset 0x46000 ipl 1 address 08:00:3e:23:7a:a0
nfs_boot: using network interface 'ie0'
root on zeus:/zeus.userd/export/root/m162
WARNING: clock lost 9310 days -- CHECK AND RESET THE DATE!
swap on zeus:/zeus.userd/export/swap/m162
vm_fault(98000, fffff000, 1, 0) -> 1
type 8, code [mmu,,ssw]: 2c00525
trap type 8, code = 2c00525, v = ffffffff