From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mga14.intel.com ([192.55.52.115]:12136 "EHLO mga14.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751044AbcDRJEb (ORCPT ); Mon, 18 Apr 2016 05:04:31 -0400 Message-ID: <1460970266.7369.3.camel@gmail.com> Subject: Re: Reconsidering exportable UBIFS From: Artem Bityutskiy Reply-To: dedekind1@gmail.com To: Richard Weinberger , linux-fsdevel Cc: "linux-mtd@lists.infradead.org" , Christoph Hellwig Date: Mon, 18 Apr 2016 12:04:26 +0300 In-Reply-To: <5702E7F5.1050807@nod.at> References: <5702E7F5.1050807@nod.at> Content-Type: text/plain; charset="UTF-8" Mime-Version: 1.0 Content-Transfer-Encoding: 8bit Sender: linux-fsdevel-owner@vger.kernel.org List-ID: On Tue, 2016-04-05 at 00:17 +0200, Richard Weinberger wrote: > The biggest problem I see is that UBIFS does not really support > telldir() > and seekdir(). > Directory offsets in UBIFS are plain hash values, so > telldir()/seekdir() won't > correctly work if UBIFS faces hash collisions. > Currently UBIFS implements a hack which stores the UBIFS dent object > into > file->private_data such that consecutive readdir()s are guaranteed to > work. > A comment on UBIFS's readdir states: >  * This means that UBIFS cannot support NFS which requires full >  * 'seekdir()'/'telldir()' support. Richard, I do not remember much about NFS support. I can only say that I never really put efforts into this, unfortunately. Here is an old discussion which could give you some more hints: http://marc.info/?l=linux-nfsv4&m=126480082613293&w=2 Artem.