From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tejun Heo Subject: Re: libata interface fatal error Date: Thu, 14 Jun 2007 18:43:56 +0900 Message-ID: <46710DDC.1030900@gmail.com> References: <46580134.7010107@effenberger.org> <465BEF50.8050207@gmail.com> <466725E5.9030405@effenberger.org> <4667D4F5.1060302@gmail.com> <466FC901.4060907@effenberger.org> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Return-path: Received: from nz-out-0506.google.com ([64.233.162.239]:10191 "EHLO nz-out-0506.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751316AbXFNJoE (ORCPT ); Thu, 14 Jun 2007 05:44:04 -0400 Received: by nz-out-0506.google.com with SMTP id n1so529763nzf for ; Thu, 14 Jun 2007 02:44:03 -0700 (PDT) In-Reply-To: <466FC901.4060907@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 Tejun, > >> Okay, there was a bug in link speed limit logic. That's probably why >> speed down to 1.5Gbps didn't kick in. The attached patch contains the >> fix and hack to force 1.5Gbps. Please give it a shot. > > thanks a lot for your patch, it seems to work, at least better than > without patch. :-) > > When rsyncing about 12 GB, no trouble occured. When doing heavy stress > tests, I receive errors again, but okay, maybe that's due to a hardware > bug. > > Will your patch go into the vanilla kernel? I'm currently not sure what the root cause is 1. if the controller is at fault, we need to force 1.5Gbps on the controller. 2. if the drive model is broken, we need to blacklist the drives. 3. if your specific configuration is broken (faulty hw, PSU, bad karma), the upstream speed limit fix patch should be enough. Can you post the result of 'hdparm -I /dev/sdX'? -- tejun