From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tejun Heo Subject: Re: NCQ change on 2007-12-07 Date: Thu, 13 Mar 2008 12:28:10 +0900 Message-ID: <47D89F4A.7090008@gmail.com> References: <47D76B7D.8010808@gmail.com> <47D89A58.6010301@shaw.ca> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Return-path: Received: from wa-out-1112.google.com ([209.85.146.178]:9749 "EHLO wa-out-1112.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751143AbYCMD2P (ORCPT ); Wed, 12 Mar 2008 23:28:15 -0400 Received: by wa-out-1112.google.com with SMTP id v27so3762378wah.23 for ; Wed, 12 Mar 2008 20:28:15 -0700 (PDT) In-Reply-To: <47D89A58.6010301@shaw.ca> Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: Robert Hancock Cc: "Lanstein, Alex C" , linux-ide@vger.kernel.org Hello, Robert Hancock wrote: > I don't think the driver is to blame for this one. CPB resp_flags of 2 > indicates the controller says that tag is "released", i.e. drive has > been given the command but not completed. If it's timing out in that > state, either the NCQ implementation is bad enough that it allows > certain queued commands to starve for over 30 seconds, or the drive is > just forgetting about them entirely. > > Unfortunately, if the firmware revision does indeed appear the same on > both sets of drives, the only automated way we could likely address the > problem would be to blacklist NCQ on all of those drives of that > model/revision.. Hmm... it's weird that this problem hasn't been reported on other controllers. Lanstein, can you be persuaded into trying out another NCQ capable controller? -- tejun