From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tejun Heo Subject: Re: Still having problems with sata_promise on all late kernels - shutdown too - BISECTED Date: Thu, 16 Aug 2007 20:10:26 +0900 Message-ID: <46C430A2.7040001@gmail.com> References: <46C11BBE.6010802@wasp.net.au> <46C1316F.4010906@wasp.net.au> <46C1DB92.7080806@wasp.net.au> <46C2B046.7060708@gmail.com> <46C3273F.1070306@wasp.net.au> <46C3E3F7.2010803@gmail.com> <46C402AE.6010605@wasp.net.au> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Return-path: Received: from rv-out-0910.google.com ([209.85.198.186]:36561 "EHLO rv-out-0910.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752948AbXHPLKn (ORCPT ); Thu, 16 Aug 2007 07:10:43 -0400 Received: by rv-out-0910.google.com with SMTP id k20so179529rvb for ; Thu, 16 Aug 2007 04:10:43 -0700 (PDT) In-Reply-To: <46C402AE.6010605@wasp.net.au> Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: Brad Campbell Cc: linux-ide@vger.kernel.org Brad Campbell wrote: > Tejun Heo wrote: >> Brad Campbell wrote: >>> Tejun Heo wrote: >>>> Brad Campbell wrote: >>>>> Brad Campbell wrote: >>>>> >>>>>> http://www.fnarfbargle.com/CIMG1029.JPG (I have posted this screen >>>>>> shot previously. It's still doing exactly the same thing on both >>>>>> boxes). >>>>>> >>>>> git-bisect shows up this commit as the guilty party. I can't just >>>>> revert >>>>> it on a recent tree though. >>>>> Will investigate further. >>>> Weird. Does 'hdparm -y /dev/sdX' work? >>> Very interesting indeed.. >>> >>> for i in /dev/sd* ; do echo $i ; hdparm -y $i ; done >>> >>> sda-sdh worked fine.. sdi locked the box. >>> >>> So it's dying on the 3rd or 4th controller. >> >> That's weird. You seem to have the same controllers and drives. What >> happens if you spin down from hdi upwards? >> > > Any attempt to spin down sdi-sdo in any order locks hard. Any attempt to > spin down sda-sdh works in any combination/permutation. Hmmm... I'm out of ideas. It definitely sounds like some sort of hardware problem to me. I think the best way to diagnose the problem is playing the tedious but often rewarding add-remove-swap game with the hardware and see to which piece or combination of pieces the problem is attached. Thanks. -- tejun