Return-Path: Received: from mail-yw1-f68.google.com ([209.85.161.68]:43610 "EHLO mail-yw1-f68.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1727738AbeJTCFh (ORCPT ); Fri, 19 Oct 2018 22:05:37 -0400 MIME-Version: 1.0 References: <20181019153018.32507-1-olga.kornievskaia@gmail.com> <20181019153018.32507-2-olga.kornievskaia@gmail.com> <20181019174446.GA28891@bombadil.infradead.org> In-Reply-To: <20181019174446.GA28891@bombadil.infradead.org> From: Amir Goldstein Date: Fri, 19 Oct 2018 20:58:17 +0300 Message-ID: Subject: Re: [PATCH v1 02/11] VFS permit cross device vfs_copy_file_range To: Matthew Wilcox Cc: olga.kornievskaia@gmail.com, linux-fsdevel , Linux NFS Mailing List , fweimer@redhat.com, Steve French , linux-api@vger.kernel.org, Christoph Hellwig Content-Type: text/plain; charset="UTF-8" Sender: linux-nfs-owner@vger.kernel.org List-ID: On Fri, Oct 19, 2018 at 8:44 PM Matthew Wilcox wrote: > > On Fri, Oct 19, 2018 at 07:14:23PM +0300, Amir Goldstein wrote: > > > On Fri, Oct 19, 2018 at 6:30 PM Olga Kornievskaia > > > wrote: > > > > Allow copy_file_range to copy between different superblocks but only > > > > of the same file system types. This feature was of interest to CIFS > > > > as well as NFS. > > > > Sorry, you meant to allow cross sb copy from client, so you do want > > to change user facing behavior. > > For that you need to declare a flag to opt in for the new behavior > > and document it. > > Please keep linux-api CC-ed when posting the flag. > > Referencing a draft for man page is recommended. > > Huh? We're allowed to make things work that used to fail. There's no > need to add a flag that the user must set. OK, you are right. Maybe this case doesn't call for an opt-in flag. I'll let other people weight in. I should note that the behavior of copy_file_range() has already changed silently to try to clone_file_range() first and that went quietly... Thanks, Amir.