From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Tue, 21 May 2002 13:33:38 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Tue, 21 May 2002 13:33:37 -0400 Received: from parcelfarce.linux.theplanet.co.uk ([195.92.249.252]:22025 "EHLO www.linux.org.uk") by vger.kernel.org with ESMTP id ; Tue, 21 May 2002 13:33:36 -0400 Message-ID: <3CEA85DC.8623C6B8@zip.com.au> Date: Tue, 21 May 2002 10:37:32 -0700 From: Andrew Morton X-Mailer: Mozilla 4.79 [en] (X11; U; Linux 2.4.19-pre8 i686) X-Accept-Language: en MIME-Version: 1.0 To: Antti Salmela CC: linux-kernel@vger.kernel.org, ext2-devel@lists.sourceforge.net Subject: Re: ext3 assertion failure and oops, 2.4.18 In-Reply-To: <20020521114244.GA29043@otitsun.oulu.fi> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Antti Salmela wrote: > > I can reliably reproduce an assertion failure and oops in ext3 by simply > restarting cyrus21, if directories used by cyrus have +j flag set with > chattr. Filesystem was mounted with default journalling mode data=orderded, > kernels tested were 2.4.18 and 2.4.19-pre3-ac4. Recent -pre or -ac kernels > wouldn't compile with my .config. > > Assertion failure in journal_revoke() at revoke.c:330: > "!(__builtin_constant_p(BH_Revoked) ? constant_test_bit((BH_Revoked),( > &bh->b_state)) : variable_test_bit((BH_Revoked),( &bh->b_state)))" > > I'll capture whole oops if requested. > > I found two similar cases from lkml archives, but they were left > unresponded (atleast lkml wasn't cc'ed). I couldn't judge from changelogs if > this problem has been already fixed. > > http://groups.google.com/groups?selm=2446DD7E.7F1AEC90.00A5E169%40netscape.net&output=gplain > http://groups.google.com/groups?as_umsgid=%3C2446DD7E.7F1AEC90.00A5E169%40netscape.net%3E&lr=&hl=en > It seems that mixing journalling modes in this manner doesn't work. Please turn off +j for the while - we'll aim to fix this in 2.4.20. -