From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mx141.netapp.com ([216.240.21.12]:21213 "EHLO mx141.netapp.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753100AbdCBQIf (ORCPT ); Thu, 2 Mar 2017 11:08:35 -0500 From: Olga Kornievskaia To: CC: , Subject: [PATCH v1 0/3] VFS changes for NFSv4.2 "inter" server-to-server COPY op Date: Thu, 2 Mar 2017 11:02:08 -0500 Message-ID: <20170302160211.30451-1-kolga@netapp.com> MIME-Version: 1.0 Content-Type: text/plain Sender: linux-fsdevel-owner@vger.kernel.org List-ID: One patch adds a check for validity of arguments. In another patch is crucial to the existance of NFSv4.2 inter SSC COPY. In order to implement the "inter" SSC COPY that involves two files from different superblock, it requires removal of the cross-devices check in VFS. Given that in the future the "cp" command will be written to use copy_file_range(), it would be important for "inter" SSC COPY to use the same interface. This patch has first been met with an objection from Christoph saying that "no other VFS operation is supported between mountpoints" and another objection that since been addressed that "no xfstest coverage". We argue that no other VFS operation needed such support and now there exists one. An alternative approach might be to define a new file system function (eg., bool allow_cross_device_op() that a filesystem can define and NFS will define it and return true. And based on the existence of that function and its result the check could be relaxed)? If that check were to be removed, it would be beneficial to check the the superblocks for the clone operation so that it's not called unnecessarily. Anna Schumaker (1): fs: Don't copy beyond the end of the file Andy Adamson (1): VFS permit cross device vfs_copy_file_range Olga Kornievskaia (1): VFS don't try clone if superblocks are different fs/read_write.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) -- 1.8.3.1