Return-Path: Received: from bombadil.infradead.org ([198.137.202.9]:50182 "EHLO bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751337AbbGMHlj (ORCPT ); Mon, 13 Jul 2015 03:41:39 -0400 Date: Mon, 13 Jul 2015 00:41:39 -0700 From: Christoph Hellwig To: Peng Tao Cc: Christoph Hellwig , Linux NFS Mailing List , Trond Myklebust , Anna Schumaker , Zach Brown , "Darrick J. Wong" Subject: Re: [PATCH 0/7] nfs: minor cleanups and NFSv42 CLONE support Message-ID: <20150713074139.GB23338@infradead.org> References: <1436766655-10696-1-git-send-email-tao.peng@primarydata.com> <20150713065851.GD26245@infradead.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: Sender: linux-nfs-owner@vger.kernel.org List-ID: On Mon, Jul 13, 2015 at 03:36:22PM +0800, Peng Tao wrote: > So the main change is .copy_file_range syscall. Was it rejected before > or can we just push it forward as it is? It wasn't rejected, but we kept bikeshedding over the semantics insted of moving it forward. > If there are objections against a new syscall, we can drop it and move > on with vfs helper and the new copy_file_range file operation, in > order to just make nfsd clone work. I think everyone wants the syscall, but people don't agree on the semantics. For now maybe we should just add the method and move the btrfs/xfs ioctl to common code.