Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 4 Oct 1999 22:58:19 +0000 (GMT)
From:      Terry Lambert <tlambert@primenet.com>
To:        luoqi@watermarkgroup.com (Luoqi Chen)
Cc:        Tor.Egge@fast.no, woju@bbs.ee.ntu.edu.tw, freebsd-smp@FreeBSD.ORG
Subject:   Re: SMP on 4 Pentium III(450NX) failed
Message-ID:  <199910042258.PAA19179@usr05.primenet.com>
In-Reply-To: <199910041814.OAA04534@lor.watermarkgroup.com> from "Luoqi Chen" at Oct 4, 99 02:14:58 pm

next in thread | previous in thread | raw e-mail | index | archive | help
> > > Does anybody have successful experience for running FreeBSD 3.X-STABLE
> > > SMP on 4 Pentium III 550 with Intel 450NX chipset?
> > > 
> > > When I run FreeBSD 3.3-STABLE SMP on AcerAltos 21000, Pentium III 550 x 4,
> > > the booting process halt at "Lauching CPU #1".
> > > 
> > > With 4 CPU installed, it halted at "SMP: AP CPU #3(or #2) Launched!".
> > 
> > [....]
> > 
> > >                INT     active-hi       level        0   7:D          2   19
> > 			 ^^^^^^^^^				       ^^^^^^
> > 
> > >                INT     active-lo       level        0  11:A          2   19
> > 			 ^^^^^^^^^				       ^^^^^^
> > 			 Same pin, opposite polarity.
> > 			 This might cause an infinite interrupt loop.
> > 
> > >                INT     active-lo       level        1   9:A          2   23
> > >                INT     active-lo       level        1   9:B          2   23
> > 
> > The MP table seems broken.  You'll probably have to edit mptable_pass1
> > and mptable_pass2 to ignore the entry for pci0.7 (the active-hi level
> > one).
> > 
> > - Tor Egge
>
> I'm wondering how NT handles this case?

Most likely by either keeping it's own table of the pins that it fills
in based on the pin number (which would result in the second entry
overwriting the first, which is probably correct behaviour, since it
would allow a soft patch of a ROM bug via a BIOS update), by ignoring
"active-hi" pins entirely, or by prioritizing which entry to believe
(via "7:D"/"11:A" comparison).

It would be interesting to set up an email alias at FreeBSD.org, and
attach a script to it that knows how to characterize the "brokenness"
of "broken" MP tables, and then have everyone with a "broken" MP table
send it to the alias to better characterize the classes of "broken" so
that the software could resolve the problem itself, without needing a
list of "known rogues" or user kernel modification/(re)configuration.


					Terry Lambert
					terry@lambert.org
---
Any opinions in this posting are my own and not those of my present
or previous employers.


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-smp" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199910042258.PAA19179>