From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tejun Heo Subject: Re: sata_promise error handling Date: Tue, 03 Jul 2007 15:45:07 +0900 Message-ID: <4689F073.4070405@gmail.com> References: <200706191859.07050.tbaumgartner@swissonline.ch> 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 ([64.233.162.232]:19221 "EHLO nz-out-0506.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753683AbXGCGpS (ORCPT ); Tue, 3 Jul 2007 02:45:18 -0400 Received: by nz-out-0506.google.com with SMTP id s18so1127762nze for ; Mon, 02 Jul 2007 23:45:17 -0700 (PDT) In-Reply-To: Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: Mark Cc: linux-ide@vger.kernel.org, Mikael Pettersson Mark wrote: >> Can I ignore them (EH handled them) or do i have to worry that the dd images > are corrupted >> (don't wanna make an md5sum of a 250gb disk and image)? Yeah, data should be okay. All failed commands are retried. > I get the same errors with my Promise SATA300 TX4 with 3 Samsung SATA II > drives. I've been playing with it for a few days and I couldn't get to work > reliably in anything but 2.6.22-rc5. > > Now that I use 2.6.22-rc5, I get those errors for a few minutes then the driver > hard resets and locks at 1.5 Gb/s. It seems reliable after that but I do get > those errors every now and then under heavy I/O. That part of code got improved during 2.6.22 devel phase, so that's the expected behavior. It seems that the promise controller has some problems when connected to certain 3Gbps drives (IIRC newer Seagate ones). It's most likely a hardware problem and we might need to default to 1.5 Gbps on such combinations, but even without such modifications, libata EH should do the right thing. -- tejun