Subject: Re: Strange dmesg-Ouput on NetBSD2.0_RC4 SMP-Kernel
To: =?ISO-8859-1?Q?R=FCdiger?= Engel <granada@a500.de>
From: MINOURA Makoto <makoto@hauN.ORG>
List: tech-smp
Date: 11/15/2004 23:42:00
|> In <41967EA3.2010109@a500.de>
|> R=FCdiger Engel <granada@a500.de> wrote:
> Yepp! Already did so as I learnt that NetBSD's scheduler still does not=20
> support Hyperthreading. Now the whole ACPI-crap is commented out and I=20
> have two CPUs runnig fine and with sufficient performance. But I'd love=20
> to have a system with 4 CPUs, even if it is "only" a dual Xeon with=20
> Hyperthreading enabled. ;-)
The problem is not due to our HTT support.
MPACPI first probes the processors from ACPI MADT (detects
cpu0 and cpu1). After that it gives to configure busses
from the ACPI DSDT since it finds your ACPI BIOS listed on
the acpi_quirks table. The kernel has to fall back to
MPBIOS but your kernel does not have it enabled (here the
kernel had to report it and should panic IMHO). The kernel
misunderstands that it does not yet configured the
processors and falls back to uniprocessor mode (detects cpu2).
--=20
Minoura Makoto <makoto@hauN.org>