From owner-freebsd-mobile@FreeBSD.ORG Thu May 29 07:17:41 2003 Return-Path: Delivered-To: freebsd-mobile@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id A7EFC37B401 for ; Thu, 29 May 2003 07:17:41 -0700 (PDT) Received: from mail.distalzou.net (203.141.139.231.user.ad.il24.net [203.141.139.231]) by mx1.FreeBSD.org (Postfix) with ESMTP id C65B743F75 for ; Thu, 29 May 2003 07:17:40 -0700 (PDT) (envelope-from devin@spamcop.net) Received: from borosilicate.pun-pun.prv ([192.168.7.29]) by mail.distalzou.net with esmtp (TLSv1:AES256-SHA:256) (Exim 4.20) id 19LODq-0001Ez-Sy; Thu, 29 May 2003 23:17:38 +0900 Date: Thu, 29 May 2003 23:17:38 +0900 (JST) From: Tod McQuillin X-X-Sender: devin@borosilicate.pun-pun.prv To: "M. Warner Losh" In-Reply-To: <20030529.062946.90825563.imp@bsdimp.com> Message-ID: <20030529230721.M20777@borosilicate.pun-pun.prv> References: <20030527020616.J48601@borosilicate.pun-pun.prv> <20030527120556.M48601@borosilicate.pun-pun.prv> <20030529.062946.90825563.imp@bsdimp.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII cc: freebsd-mobile@freebsd.org Subject: Re: Orinoco Gold trouble with SMP server X-BeenThere: freebsd-mobile@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Mobile computing with FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 29 May 2003 14:17:41 -0000 On Thu, 29 May 2003, M. Warner Losh wrote: > OK. I've been reviewing my notes. We have a SMP box at work and > haven't been able to get ISA interrupts to work with OLDCARD on that > box. :-( Thanks Warner! This is actually the best news I've heard all day because it means I can get down to business hacking on FreeBSD without worrying that I'm overlooking something obvious. I found another ISA card (old adaptec AVA-1502 that I think is supported by the aic driver) which can be configured for IRQ 9 or 10. I will begin by verifying that I can in fact use both of those IRQs with another card, and then put the pcic back in and start tracing the code path for setting up an inserted card. Could it have something to do with missing some step for programming the APIC? I notice this line from the verbose dmesg: SMP: enabled INTs: 1, 3, 4, 5, 6, 7, 9, 14, 15, 16, 17, 18, 19, apic_imen: 0x00f03d05 I note that IRQ 10 (the one pccardd is trying to allocate for wi0) is not in this list. This makes sense since wi0 has not been probed yet at this point, but I wonder if there is something the pcic code is overlooking? Also I wonder if the machine where you observed similar symptoms is also a compaq afflicted with the dreaded "broken mptable" glitch. There is a workaround for this which prints this message: APIC_IO: MP table broken: 8259->APIC entry missing! but perhaps it's related. If I have questions about the kernel code for allocating IRQs should I ask here, on -hackers, on -smp, or what? Thanks again, -- Tod McQuillin