From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753512Ab0LJCF2 (ORCPT ); Thu, 9 Dec 2010 21:05:28 -0500 Received: from mail-bw0-f45.google.com ([209.85.214.45]:56868 "EHLO mail-bw0-f45.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751423Ab0LJCF1 convert rfc822-to-8bit (ORCPT ); Thu, 9 Dec 2010 21:05:27 -0500 MIME-Version: 1.0 In-Reply-To: <1291946360-sup-154@think> References: <20101207142145.GA27861@think> <20101207182243.GB21112@redhat.com> <20101207193514.GA2921@thunk.org> <20101209180111.GF2921@thunk.org> <20101209201359.GG2921@thunk.org> <20101209231616.GA12515@basil.fritz.box> <1291945065-sup-1838@think> <1291946360-sup-154@think> From: Jon Nelson Date: Thu, 9 Dec 2010 20:05:05 -0600 Message-ID: Subject: Re: hunt for 2.6.37 dm-crypt+ext4 corruption? (was: Re: dm-crypt barrier support is effective) To: Chris Mason Cc: Mike Fedyk , Andi Kleen , "Ted Ts'o" , Mike Snitzer , Matt , Milan Broz , linux-btrfs , dm-devel , Linux Kernel , htd , htejun , linux-ext4 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8BIT Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Thu, Dec 9, 2010 at 8:00 PM, Chris Mason wrote: > Excerpts from Mike Fedyk's message of 2010-12-09 20:58:40 -0500: >> On Thu, Dec 9, 2010 at 5:38 PM, Chris Mason wrote: >> > Excerpts from Andi Kleen's message of 2010-12-09 18:16:16 -0500: >> >> > 512MB. >> >> > >> >> > 'free' reports 75MB, 419MB free. >> >> > >> >> > I originally noticed the problem on really real hardware (thinkpad >> >> > T61p), however. >> >> >> >> If you can easily reproduce it could you try a git bisect? >> > >> > Do we have a known good kernel?  I looked back through the thread and >> > didn't see any reports where the postgres test on ext4 passed in this >> > config. >> > >> >> 2.6.34.something.  -- Any chance a newer kernel can be tested to be found good? > > But he is triggering the ext4 corruption without dm-crypt.  I think > dm-crypt was still used somewhere on the system during the test, just > not on the partitions that actually hit the corruption. I now have my doubts about being able to trigger it without dm-crypt. I stand by my report, but I'm also unable to reproduce it after after 50 test iterations. I did run 2.6.36 for a few weeks without any issue that I recall. -- Jon