From mboxrd@z Thu Jan 1 00:00:00 1970 From: Graham Knap Subject: Re: Bug#338089: New aic7xxx driver fails spectacularly on 2940UW Date: Sun, 13 Nov 2005 11:47:32 -0500 (EST) Message-ID: <20051113164732.73492.qmail@web88010.mail.re2.yahoo.com> References: <1131855418.4179.7.camel@mulgrave> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7BIT Return-path: Received: from web88010.mail.re2.yahoo.com ([206.190.37.197]:32598 "HELO web88010.mail.re2.yahoo.com") by vger.kernel.org with SMTP id S964933AbVKMQrd (ORCPT ); Sun, 13 Nov 2005 11:47:33 -0500 In-Reply-To: <1131855418.4179.7.camel@mulgrave> Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: James Bottomley Cc: Horms , 338089@bugs.debian.org, linux-scsi@vger.kernel.org James Bottomley wrote: > My best guess would be that the bus is slightly marginal. The > aic7xxx drivers are notoriously sensitive to bus problems. Could you > try lowering the bus speed to 10MHz in the aic7xxx bios and see if > that helps? Sure. I changed the device to 20MByte/s max. Now the driver shows 10MHz, 16-bit wide... and DV still fails, with what looks like the same messages. As for the bus being marginal, I don't see why it would be, but anything is possible. Host adapter termination is set to "Automatic". If I understand things correctly, since there's no cable attached to the internal 50pin connector or the external connector, I could try forcing it to "Low ON / High ON"... right? The cabling looks like this: [HBA] -- [disk] -- [unused connector] -- [terminator] I'm pretty sure the cable is U160 rated (it wasn't cheap). The terminator isn't integrated into the cable -- it's a plug-in Active SE terminator. The drive is LVD capable, but I've never had trouble running it in SE mode. The "Force SE" jumper is not installed, but I suppose I could try that too. http://www.hitachigst.com/tech/techlib.nsf/products/Ultrastar_36LZX BTW -- Another oddity I should mention is that if I warm-reboot the system after DV fails, the drive isn't detected during POST. Power cycling fixes it. (In other cases, warm-rebooting works fine.) Let me know what I should try next. -- graham