In <URL:news:local.netbsd> on Mon 01 May, Ben Harris wrote:
> Quick one this.  In the usual place on chiark, there's a new kernel which
> might have support for "ea" devices (ether3 etc).  I've no idea if the
> driver works -- I just hacked at the arm32 driver until it compiled.  
> I've done the same thing with the "asc" (Acorn SCSI) driver, and it seems
> to work a little bit, but hangs probing non-existent targets.  Since
> someone over in arm32-land is apparently rewriting the "asc" driver, I'll
> postpone further work on it for now.  
First of all, thanks for your continuing work on the port.

I use a ACORN SCSI card and have less luck with it then you have because it
fails to be probed. My dmesg/out file folows:

[ preserving 95636 bytes of netbsd ELF symbol table ]
Copyright (c) 1996, 1997, 1998, 1999, 2000
    The NetBSD Foundation, Inc.  All rights reserved.
Copyright (c) 1982, 1986, 1989, 1991, 1993
    The Regents of the University of California.  All rights reserved.

NetBSD 1.4X (GENERIC) #420: Mon May  1 21:44:08 BST 2000
    bjh21@cromarty:/usr/src/sys/arch/arm26/compile/GENERIC
total memory = 4096 KB
avail memory = 2080 KB
using 16 buffers containing 288 KB of memory
cpu0 (root): ARM3 (rev. 0), cache enabled
iobus0 (root)
ioc0 at iobus0 base 0x200000
iic0 at ioc0
rtc0 at iic0 addr 0xa0: PCF8583, 32.768 kHz clock
arckbd0 at ioc0 bank 0 offset 0x04
arckbd0: interrupting at IRQ 15 (rx) and IRQ 14 (tx)
arcwskbd0 at arckbd0
wskbd0 at arcwskbd0: console keyboard
arcwsmouse0 at arckbd0
wsmouse0 at arcwsmouse0
podulebus0 at ioc0 bank 4
podule <0000:0002> at podulebus0 slot 1 not configured

At this point the kernel stops.

With the first kernel I get the following dmesg/out file

 [ no symbols available ]
Copyright (c) 1996, 1997, 1998, 1999, 2000
    The NetBSD Foundation, Inc.  All rights reserved.
Copyright (c) 1982, 1986, 1989, 1991, 1993
    The Regents of the University of California.  All rights reserved.

NetBSD 1.4X (GENERIC) #351: Thu Apr 13 23:55:49 BST 2000
    bjh21@trafalgar:/export/share/opt/local/src/netbsd/sys/arch/arm26/compile/GENERIC
    
SKIP

podulebus0 at ioc0 bank 4
CDFS & SCSI Expansion Card <0000:0002> at podulebus0 slot 1 not configured
ANT Ethernet (00:02:07:01:1a:40) Acorn Access / AUN <0053:00a4> at podulebus0 slot 3 not configured
arcvideo0 (root): VSYNC interrupts at IRQ 3

SKIP

init: copying out path `/sbin/init' 11
FIXME: pmap_clear_modify not implemented

This kernel did recognise the card (or did it just read the copyright
message from the card 8-). 
The problem may also be the probing of the ethernet card. And indeed that it
is. After removing the ethernet card the kernel booted till it asked for
a root device. The following dmesg/out file was generated:

START

[ preserving 95636 bytes of netbsd ELF symbol table ]
Copyright (c) 1996, 1997, 1998, 1999, 2000
    The NetBSD Foundation, Inc.  All rights reserved.
Copyright (c) 1982, 1986, 1989, 1991, 1993
    The Regents of the University of California.  All rights reserved.

NetBSD 1.4X (GENERIC) #420: Mon May  1 21:44:08 BST 2000
    bjh21@cromarty:/usr/src/sys/arch/arm26/compile/GENERIC
total memory = 4096 KB
avail memory = 2080 KB
using 16 buffers containing 288 KB of memory
cpu0 (root): ARM3 (rev. 0), cache enabled
iobus0 (root)
ioc0 at iobus0 base 0x200000
iic0 at ioc0
rtc0 at iic0 addr 0xa0: PCF8583, 32.768 kHz clock
arckbd0 at ioc0 bank 0 offset 0x04
arckbd0: interrupting at IRQ 15 (rx) and IRQ 14 (tx)
arcwskbd0 at arckbd0
wskbd0 at arcwskbd0: console keyboard
arcwsmouse0 at arckbd0
wsmouse0 at arcwsmouse0
podulebus0 at ioc0 bank 4
podule <0000:0002> at podulebus0 slot 1 not configured
arcvideo0 (root): VSYNC interrupts at IRQ 3
wsdisplay0 at arcvideo0: console (arccons, vt100 emulation), using wskbd0
ioc0: 100 Hz clock interrupting at IRQ 5
ioc0: 100 Hz statclock interrupting at IRQ 6
arckbd0: layout 1
cpu0: 10000 loops in 1671 microseconds, delay factor = 1
root device: 
use one of: halt
root device: 

END

I do not know if the SCSI driver actualy works but the probing of the
ethernet card failes.

> We apologise for the lack of
> documentation with this snapshot -- I can't be bothered just now.
No problem
> 
> Oh, and the Ether1 driver's now called "ei" to avoid clashing with an
> atari framebuffer driver.
> 
> On a different front, my desk seems to have sprouted an R140 running
> RISC iX, which makes an interesting comparison.  Our console drivers, at
> least, need some serious work to compete.
> 
I hope this gives some info that helps you creating a stable kernel.

Leo

-- 
Leo Smiers
System specialist FB/SP                    bv Nederland Haarlem
!PDF 0.73a | !Flash 0.49b                  http://people.a2000.nl/lsmiers