From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from aserp1040.oracle.com ([141.146.126.69]:24971 "EHLO aserp1040.oracle.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753119AbcLHPTK (ORCPT ); Thu, 8 Dec 2016 10:19:10 -0500 Date: Thu, 8 Dec 2016 07:19:00 -0800 From: Liu Bo To: dsterba@suse.cz Cc: linux-btrfs@vger.kernel.org, Chris Mason , Jan Kara Subject: Re: [PATCH 2/6] Btrfs: set single device limit for dax usecase Message-ID: <20161208151900.GA20111@localhost.localdomain> Reply-To: bo.li.liu@oracle.com References: <1481147110-20048-1-git-send-email-bo.li.liu@oracle.com> <1481147110-20048-3-git-send-email-bo.li.liu@oracle.com> <20161208133559.GE12522@suse.cz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: <20161208133559.GE12522@suse.cz> Sender: linux-btrfs-owner@vger.kernel.org List-ID: On Thu, Dec 08, 2016 at 02:35:59PM +0100, David Sterba wrote: > On Wed, Dec 07, 2016 at 01:45:06PM -0800, Liu Bo wrote: > > Dax on btrfs is not ready for multiple device. > > How about DUP? Technically it's not multi-device but still stores > multiple copies, so I don't know if the implementation is ok with that. Good question. For this patch set, meta DUP is OK while data DUP is not, the main obstacle is that the actual copy-to-device operation happens in iomap code (iomap_dax_actor), not inside filesystem itself, and it only takes one pair to copy to device. Thanks, -liubo