From owner-freebsd-smp Wed Jul 28 1:25:38 1999 Delivered-To: freebsd-smp@freebsd.org Received: from mercury.gfit.net (ns.gfit.net [209.41.124.90]) by hub.freebsd.org (Postfix) with ESMTP id 8C2B814F46 for ; Wed, 28 Jul 1999 01:25:33 -0700 (PDT) (envelope-from tom@embt.com) Received: from paranor.embt.net (timembt.iinc.com [206.67.169.229]) by mercury.gfit.net (8.8.8/8.8.8) with SMTP id DAA15103 for ; Wed, 28 Jul 1999 03:26:15 -0500 (CDT) (envelope-from tom@embt.com) Message-Id: <3.0.3.32.19990728042442.007617f0@mail.embt.com> X-Sender: tembt@mail.embt.com X-Mailer: QUALCOMM Windows Eudora Pro Version 3.0.3 (32) Date: Wed, 28 Jul 1999 04:24:42 -0400 To: freebsd-smp@freebsd.org From: Tom Embt Subject: Re: SMP kernel on dual board with single proc? Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org >If you are using a slot-architecture board, then you need a >terminator in the empty slot. > >In general, "BSP" means "Boot System Processor" and "AP" means >"Auxillary Processor" -- in other words, it thinks it can't >find the BIOS designated boot processor. Ahh, that's what it means. :) >You may find that swapping slots fixes the problem (if, as Mike >Smith said, you are running a -current that is no more than two >weeks old). Already tried that. Not eager to keep swapping, as I don't want to make a mess of the thermal goop between the heatsink & CPU (it's not the drying kind). >For standard slot processers (e.g., my dual P90), the BIOS has >always given the right information to the BP (including the APIC >ID), so it has been rare that it would not run with one processor >removed (the last time was the first time somone kicked it up >from 2 to 4 processors). > > >> Shall I assume it is not possible to run an SMP kernel on this system w/o >> another CPU? > >1) How old is the build? cvsup'ed Tuesday 7/27 3.2-STABLE just to be sure, still panics >2) Socket 370 = needs terminator, IIRC... Hmm, I've never heard of a 370 terminator before.. I suspect that the mobo might be auto-terminating the unpopulated socket and disabling some of the APIC hardware. When using 1 CPU on a MP board, is it normal that the board require a particular socket/slot to be populated first? This board does not, so I am wondering if perhaps the logic required to enable this "feature" is throwing off some of the SMP functionality (which of course would normally not be needed in a UP situation anyway). Keep in mind this is naive speculation be someone who has no experience with SMP anything whatsoever. > > Terry Lambert > terry@lambert.org Oh bother, I should probably just get another processor while they're still available. Then it _better_ work! ;) Tom Embt tom@embt.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message