From: Richard Kojedzinszky Subject: Re: minix/ext2 + rd problem Date: Wed, 15 Oct 2008 10:19:44 +0200 (CEST) Message-ID: References: <20081015041644.GA24613@wotan.suse.de> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed Cc: linux-ext4@vger.kernel.org To: Nick Piggin Return-path: Received: from krichy.tvnetwork.hu ([80.95.68.194]:53924 "HELO krichy.tvnetwork.hu" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with SMTP id S1751454AbYJOITr (ORCPT ); Wed, 15 Oct 2008 04:19:47 -0400 In-Reply-To: <20081015041644.GA24613@wotan.suse.de> Sender: linux-ext4-owner@vger.kernel.org List-ID: dear nick, i have tried a sync after the remount, but that did not help. what helped is dropping the cache by echoing 3 to /proc/sys/vm/drop_caches, but this still didnt solve the problem in 100%, only in 95% of the cases. But when i read the device with # dd if=/dev/ram0 iflag=direct ... then it worked. I think this bypassed some caches, and thus read the actual data. But a sad result is that I experienced with it, and only with ramdisk does it work as expected. for example with a logical volume it behaves in the wrong way. thanks in advance, Kojedzinszky Richard TvNetWork Nyrt. E-mail: krichy (at) tvnetwork [dot] hu PGP: 0x24E79141 Fingerprint = 6847 ECFF EF58 0C09 18A5 16CF 270F 0C6F 24E7 9141 On Wed, 15 Oct 2008, Nick Piggin wrote: > Date: Wed, 15 Oct 2008 06:16:44 +0200 > From: Nick Piggin > To: Richard Kojedzinszky > Cc: linux-ext4@vger.kernel.org > Subject: Re: minix/ext2 + rd problem > > Hi, > Thanks for reporting this. > > On Tue, Oct 14, 2008 at 04:43:55PM +0200, Richard Kojedzinszky wrote: >> dear all, >> >> I got an embedded system, where I use ramdisk, minix on it as a filesystem >> for /etc. With kernels 2.4 and with 2.6.19 kernels also, the following >> code did exactly what i wanted, creating an image of the /etc without >> unmounting it: >> >> # mount -o remount,ro /etc >> # cat /dev/ram0 > /tmp/image >> # mount -o remount,rw /etc >> >> And then I had a consistent image from /etc in /tmp/image. >> >> This worked still with kernel version 2.6.23.14, but nowadays i upgraded >> to 2.6.26.2, and noticed that the little code didnt work anymore. I wrote >> a simple test script for checking, and reproducing the issue, which is >> also attached. Unfortunately this does not work with ext2 also. >> >> my linux version is: >> Linux version 2.6.26 (root@krichy.tvnetwork.hu) (gcc version 4.2.4 (Debian >> 4.2.4-3)) #27 SMP PREEMPT Tue Oct 14 15:19:30 CEST 2008 >> >> i use debian lenny/sid. >> >> Were there any intended change that made this behaviour change or was it >> by an accident? > > That shouldn't have been changed on purpose, unless it is doing something > funny that worked by accident before. > > /dev/ram will behave much more like any other block device now (with brd) > wheras previously it was probably more coherent between filesystem and > block device node. Hmm, does all the filesystems cache get written back > before remount ro return? > > Could you try sticking a sync after remount,ro? > > Thanks, > Nick >