From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tejun Heo Subject: Re: [KERNEL 2.6.26-rc4] bugreport : pata_pcmcia with Sandisk Extreme III 8GB Date: Wed, 09 Jul 2008 13:10:02 +0900 Message-ID: <48743A1A.2010304@gmail.com> References: <20080528225049.9a134e0d.Kristoffer.Ericson@Gmail.com> <20080528222740.GB20893@isilmar.linta.de> <20080530135728.be02d3dc.Kristoffer.Ericson@Gmail.com> <20080607223716.7d142072.komurojun-mbn@nifty.com> <20080611072029.GB28692@comet.dominikbrodowski.net> <20080614120055.ad8fbee0.komurojun-mbn@nifty.com> <48537149.4040102@garzik.org> <20080614213612.96873c25.komurojun-mbn@nifty.com> <4855D8E2.10409@gmail.com> <20080616221145.e3b069b9.komurojun-mbn@nifty.com> <485669A0.6000404@gmail.com> <20080617212345.c452d8c5.komurojun-mbn@nifty.com> <20080706103752.13935ffd.komurojun-mbn@nifty.com> <4870D46C.7000604@gmail.com> <20080707214253.2406f81a.komurojun-mbn@nifty.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Return-path: Received: from wf-out-1314.google.com ([209.85.200.168]:9557 "EHLO wf-out-1314.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750738AbYGIEKH (ORCPT ); Wed, 9 Jul 2008 00:10:07 -0400 Received: by wf-out-1314.google.com with SMTP id 27so2775671wfd.4 for ; Tue, 08 Jul 2008 21:10:07 -0700 (PDT) In-Reply-To: <20080707214253.2406f81a.komurojun-mbn@nifty.com> Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: Komuro Cc: Jeff Garzik , linux-pcmcia , linux-kernel@vger.kernel.org, Dominik Brodowski , linux-ide@vger.kernel.org, Kristoffer Ericson , jgarzik@redhat.com, rjw@sisk.pl Hello, Komuro wrote: >> How come ap->lock can get set to bogus address? Can you please printk >> &host->lock and ap->lock after ata_host_alloc() without the patch? > > Here is the output of printk. > Strangely, without the patch, the pata_pcmcia problem happens again. Ah... super strange. Can you please try to swap memory sticks around and see whether anything changes to rule out bad memory? Another possibility is DMA corrupting the memory and printk makes the timing different and dodges the bullet. Does putting msleep(100) in the pata_pcmcia init_one make any difference? Thanks. -- tejun