From mboxrd@z Thu Jan 1 00:00:00 1970 From: Breno Leitao Subject: Re: Is VFS behavior fine? Date: Thu, 03 Jul 2008 11:29:27 -0300 Message-ID: <486CE247.5020904@linux.vnet.ibm.com> References: <486B419F.7050407@nokia.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Cc: Linux Kernel Mailing List , Christoph Hellwig , linux-fsdevel@vger.kernel.org, Hunter Adrian To: Artem.Bityutskiy@nokia.com Return-path: Received: from igw2.br.ibm.com ([32.104.18.25]:42954 "EHLO igw2.br.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751344AbYGCO6f (ORCPT ); Thu, 3 Jul 2008 10:58:35 -0400 In-Reply-To: <486B419F.7050407@nokia.com> Sender: linux-fsdevel-owner@vger.kernel.org List-ID: Artem Bityutskiy wrote: > The result of it was that the '->delete_inode()' operation for > the 'xxx' directory inode is not called. It is not called even > after 'cd /'. However, if we do not do the 'touch tmp' command > (which actually fails), '->delete_inode()' _is_ called for 'xxx'. > > So 'touch tmp' has a side effect. You may notice this by watching > 'cat /proc/slabinfo | grep dentry' while running the following > script: I tried to reproduce this on my system using an EXT3 filesystem, leaving the script running the entire night. When I wake up I found that the system disk was full. So, besides the memory issue, this problem also consume some bytes on your filesystem.