From owner-freebsd-stable Thu May 13 7:56:37 1999 Delivered-To: freebsd-stable@freebsd.org Received: from mate.kjsl.com (Mate.KJSL.COM [204.87.183.252]) by hub.freebsd.org (Postfix) with ESMTP id F2FE814DD0 for ; Thu, 13 May 1999 07:56:35 -0700 (PDT) (envelope-from fmc@reanimators.org) Received: (from uucp@localhost) by mate.kjsl.com (8.9.2/8.9.2) with UUCP id HAA18233; Thu, 13 May 1999 07:56:32 -0700 (PDT) Received: (from fmc@localhost) by daemonweed.reanimators.org (8.9.3/8.9.3) id HAA00961; Thu, 13 May 1999 07:43:42 -0700 (PDT) (envelope-from fmc) Message-Id: <199905131443.HAA00961@daemonweed.reanimators.org> To: Warner Losh Cc: freebsd-stable@FreeBSD.ORG Subject: Re: Is aha broken? References: <199905121429.HAA01982@daemonweed.reanimators.org> <199905130418.WAA26127@harmony.village.org> From: Frank McConnell Date: 13 May 1999 07:43:42 -0700 In-Reply-To: Warner Losh's message of "Wed, 12 May 1999 22:18:21 -0600" Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG Warner Losh wrote: > I've never used aha on the SMP machine. Usually this error is an > indication that one of two things. First one is a termination error. > The second is a IRQ/DRQ mismatch. Usually a termination. I thought about termination before writing, but the cabling is the same cabling that it had when it was in the 2.2.1-STABLE box, I thought it was terminated properly (but does the 1542CP do auto-termination on its internal connector if there's nothing connected there? -- can't find TFM), I've been over it just in case I wiggled a connector loose during the box swap, and I've replaced a 6' cable with a 3' just in case. No joy doing that. But this morning I changed the card's idea of its IRQ to 9 and it all works with no "create_intr" complaint. I'm guessing that IRQ 11 was falling foul of (and being changed to -1 by) the "#ifdef APIC_IO" code just before the call to register_intr() in config_isadev_c(), but as I'm not sure what is going on in that code I think I'm just going to rejoice and do a backup now that I can get to my tape drive. > The driver doesn't use the hardware status bits that can tell if the > cable configuration is illegal, nor does it tell you if the > termination is bad. Plug and play is poorly supported. It seems that > works sometimes. However, I've had reports of problems which I've > been unable to recreate here. OK, here's a couple of additional data points: plug and play is disabled on this card, and this is all now plugged into a Tyan S1832DL "Tiger 100" motherboard with two CPUs. -Frank McConnell To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message