From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1750819AbVIWI6J (ORCPT ); Fri, 23 Sep 2005 04:58:09 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1750820AbVIWI6J (ORCPT ); Fri, 23 Sep 2005 04:58:09 -0400 Received: from smtp.osdl.org ([65.172.181.4]:64932 "EHLO smtp.osdl.org") by vger.kernel.org with ESMTP id S1750819AbVIWI6H (ORCPT ); Fri, 23 Sep 2005 04:58:07 -0400 Date: Fri, 23 Sep 2005 01:57:19 -0700 From: Andrew Morton To: Chris Sykes Cc: linux-kernel@vger.kernel.org, ext2-devel@lists.sourceforge.net Subject: Re: Hang during rm on ext2 mounted sync (2.6.14-rc2+) Message-Id: <20050923015719.5eb765a4.akpm@osdl.org> In-Reply-To: <20050922163708.GF5898@sigsegv.plus.com> References: <20050922163708.GF5898@sigsegv.plus.com> X-Mailer: Sylpheed version 1.0.4 (GTK+ 1.2.10; i386-redhat-linux-gnu) Mime-Version: 1.0 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 Chris Sykes wrote: > > Kernels: 2.6.14-rc1 -> 2.6.14-rc2 + up to hg changeset 5c9ff0e17a61 > > I'm experiencing processes getting stuck in the 'D' state whilst > rm'ing files on an ext2 fs mounted with the 'sync' option. What I've > tested so far: > > * Ext2 mounted with sync: rm hangs > * Ext2 mounted without sync: OK > * Ext3 mounted with sync: OK > * Ext3 mounted without sync: OK > > I first noticed this on my /boot partition, and wanted to know whether > it was repeatable so I've created a few test ext2 filesystem images > and mounted them via loopback. Odd. Seems OK here. How hard is it to make it occur? I'd be suspecting a lost I/O completion from the device driver. Are you really sure that ext3 cannot be made to do the same thing? Suggest you generate the `dmesg -s 1000000' output for both good and bad kernels, do a `diff -u' on them and look for IDE complaints (or SCSI, if you're on SCSI).