Return-Path: Received: from mx2.netapp.com ([216.240.18.37]:15405 "EHLO mx2.netapp.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756349Ab1CWSR5 convert rfc822-to-8bit (ORCPT ); Wed, 23 Mar 2011 14:17:57 -0400 Subject: RE: [PATCH 3/3] NFS: Detect loops in a readdir due to bad cookies From: Trond Myklebust To: peter.staubach@emc.com Cc: bjschuma@netapp.com, linux-nfs@vger.kernel.org In-Reply-To: <5E6794FC7B8FCA41A704019BE3C70E8B0693F02D@MX31A.corp.emc.com> References: <4D8A3060.807@netapp.com> <5E6794FC7B8FCA41A704019BE3C70E8B0693F02D@MX31A.corp.emc.com> Content-Type: text/plain; charset="UTF-8" Date: Wed, 23 Mar 2011 14:17:47 -0400 Message-ID: <1300904267.11677.47.camel@lade.trondhjem.org> Sender: linux-nfs-owner@vger.kernel.org List-ID: MIME-Version: 1.0 On Wed, 2011-03-23 at 14:10 -0400, peter.staubach@emc.com wrote: > Doesn't returning identical cookies for multiple, different entries in the same directory constitute a protocol violation? Yes. Unfortunately quite a few common filesystems seem quite happy to do so. Most of the time they get away with it, but occasionally we do get support requests from people asking 'why is my NFS client looping forever?'. With Bryan's patches, we now have a way to have the client error out instead of looping, and notify the admin that their server is exporting a broken underlying filesystem. -- Trond Myklebust Linux NFS client maintainer NetApp Trond.Myklebust@netapp.com www.netapp.com