From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tejun Heo Subject: Re: libata interface fatal error Date: Mon, 18 Jun 2007 19:56:54 +0900 Message-ID: <467664F6.4060607@gmail.com> References: <46580134.7010107@effenberger.org> <465BEF50.8050207@gmail.com> <466725E5.9030405@effenberger.org> <4667D4F5.1060302@gmail.com> <4673BA28.2020106@effenberger.org> <4675F84D.3060207@gmail.com> <46766208.1080305@effenberger.org> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-15 Content-Transfer-Encoding: 7bit Return-path: Received: from wa-out-1112.google.com ([209.85.146.179]:35074 "EHLO wa-out-1112.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1760429AbXFRK5A (ORCPT ); Mon, 18 Jun 2007 06:57:00 -0400 Received: by wa-out-1112.google.com with SMTP id v27so2124495wah for ; Mon, 18 Jun 2007 03:56:59 -0700 (PDT) In-Reply-To: <46766208.1080305@effenberger.org> Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: Florian Effenberger Cc: linux-ide@vger.kernel.org, jeff@garzik.org Florian Effenberger wrote: > Hi, > >> The controller being ich8, I'm pretty sure it isn't a driver problem. > > I think so, too. The Intel chipsets have shown to be very good in the past. > >> Do the errors occur on all four drives? Also, if things work after >> speed is downgraded to 1.5Gbps, it doesn't really matter. There's no >> noticeable performance difference for single disk anyway. > > Yes, they do occur on all drives, as far as I know. With 1.5Gbps, the > error doesn't occur much as often and not under normal circumstances, > only when doing a real hard stress test. Hmmm... Can you use a separate PSU to power two of the four drives and see what happens? Just power up a PSU as directed in the following webpage and connect two of the harddrives to the PSU. http://modtown.co.uk/mt/article2.php?id=psumod > Would it make sense to downgrade to 1.5 Gbps via a boot option? I don't know. Till now all the problem cases have been isolated to a specific controller / drive combination (sata_promise and newer seagate drives) or hardware configuration problem (most of them being PSU issues), so I don't think we need such option yet. If you have a problematic hardware which pukes on 3.0Gbps, libata should do the right thing after complaining a bit which IMHO isn't too bad. -- tejun