From: Bernd Schubert Subject: RFC: expensive lookup for unlink Date: Fri, 20 Apr 2012 21:47:31 +0200 Message-ID: <4F91BD53.2060602@itwm.fraunhofer.de> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit To: "linux-fsdevel@vger.kernel.org" , Al Viro Return-path: Received: from mailgw1.uni-kl.de ([131.246.120.220]:57426 "EHLO mailgw1.uni-kl.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757495Ab2DTTrf (ORCPT ); Fri, 20 Apr 2012 15:47:35 -0400 Received: from itwm2.itwm.fhg.de (itwm2.itwm.fhg.de [131.246.191.3]) by mailgw1.uni-kl.de (8.14.3/8.14.3/Debian-5+lenny1) with ESMTP id q3KJlXlW031517 (version=TLSv1/SSLv3 cipher=EDH-RSA-DES-CBC3-SHA bits=168 verify=NOT) for ; Fri, 20 Apr 2012 21:47:33 +0200 Sender: linux-fsdevel-owner@vger.kernel.org List-ID: Hello, for a network filesystem one should try to avoid all remote operations, if possible. So when I just want to unlink a file, it does not make too much sense to first send a lookup and once the dentry is fully established to unlink the file. I think the simplest way to solve that is just to add a another nameidata flag (LOOKUP_UNLINK) and the file system may then just ignore the lookup request and fill in fake values. Or are there any other suggestions to avoid a rather useless network-lookup + network+unlink, so so combine it into one step? Thanks, Bernd