Subject: Re: More netbsd-4 woes
To: NetBSD Current Users <current-users@netbsd.org>
From: Peter Eisch <peter@boku.net>
List: current-users
Date: 10/10/2007 18:30:35
[Keeping threads to a minimum at the cost of readability, replies to many
follow...]
On 10/10/07 3:33 PM, "Andrew Doran" <ad@netbsd.org> wrote:
>> acpi0: X/RSDT: OemId <COMPAQ,YORKTOWN,00000002>, AslId < M-R^D,0000162e>
>
> This looks like a DL580 or similar? I have one and it exhibits the same
> problem.
>
Yes, this is a circa 2001 Compaq DL5?0. It's 7U tall, noisy and heavy, but
otherwise quite handy for chugging on code.
On 10/10/07 3:28 PM, "Brian Buhrow" <buhrow@lothlorien.nfbcal.org> wrote:
> Hello. NetBSD-3 used ACPI subsystem version 20040217, or something
> like that. NetBSD-4 uses one from 2006. I wonder if it's possible that
> NetBSD-4 is trying to use features in your BIOS which aren't particularly
> well implemented? In other words, do you have the latest BIOS on your
> machine? Many years ago, I used to work on Compaq machines a lot, and it
> was common to run into BIOS bugs with various OS's from time to time.
> Another thing you might look at, assuming you have the latest BIOS
> available for your machine, are your BIOS OS settings. Is it set to
> "Windows" or "Unixware" or "other?"
I'm pretty sure I selected Solaris back when I selected it. I can check
this though.
> I'd try booting with this setting on various choices to see what breaks
> differently, or, gasp, in hopes that one of them works.
It's a pretty easy test. Standing and waiting through the reboots is the
hard part.
On 10/10/07 5:09 PM, "Chris Ross" <cross+netbsd@distal.com> wrote:
> I have a couple of Compaq i386 machinesd, that have run various
> versions
> of NetBSD. I've actually found that "Solaris" works best for me for the
> "which type of OS are you running?" for an MP i386 NetBSD. I'm running
> a NetBSD 4.0_RC2 (unless it's RC1) on a DL1850, so much older than
> what you have, but it seems to be working alright for me. Does crash
> periodically, but I fear that was because I was brave enough to make
> some lfs filesystems. :-)
I'm not brave enough to try LFS on it, you win on that count. <grin>
I had a bugger of a time getting this system to netbsd-4 back in August
(with beta 2) but I thought it might have been because of cd drive and its
compatibility with the speed CD I was using.
To get this system upgraded I had to copy the contents of the CD onto /root
and boot with the INSTALL kernel. It was a bit of a circus, but with the
BETA_2 kernel it mostly worked. The CPU speeds would appear goofy but I had
all the CPUs and they all seemed to be at full-speed.
Back on the broader topic, is it fair to try to build a non-ACPI MP kernel
from the netbsd-4 source? More specifically, am I just missing some other
options in the config if I'm compiling and getting:
cc1: warnings being treated as errors
/builds/netbsd-4/src/sys/arch/i386/i386/cpu.c: In function 'cpu_init_first':
/builds/netbsd-4/src/sys/arch/i386/i386/cpu.c:187: warning: implicit
declaration of function 'lapic_cpu_number'
/builds/netbsd-4/src/sys/arch/i386/i386/cpu.c: In function 'cpu_hatch':
/builds/netbsd-4/src/sys/arch/i386/i386/cpu.c:602: warning: implicit
declaration of function 'lapic_enable'
/builds/netbsd-4/src/sys/arch/i386/i386/cpu.c:603: warning: implicit
declaration of function 'lapic_initclocks'
/builds/netbsd-4/src/sys/arch/i386/i386/cpu.c:614: warning: implicit
declaration of function 'lapic_set_lvt'
/builds/netbsd-4/src/sys/arch/i386/i386/cpu.c:623: error: 'lapic_tpr'
undeclared (first use in this function)
/builds/netbsd-4/src/sys/arch/i386/i386/cpu.c:623: error: (Each undeclared
identifier is reported only once
/builds/netbsd-4/src/sys/arch/i386/i386/cpu.c:623: error: for each function
it appears in.)
I just use:
include "arch/i386/conf/GENERIC.MP"
no acpi0 at mainbus0
no ioapic* at mainbus?
peter