Return-Path: Received: from mx142.netapp.com ([216.240.21.19]:35223 "EHLO mx142.netapp.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752291AbbHYQB7 (ORCPT ); Tue, 25 Aug 2015 12:01:59 -0400 Subject: Re: [PATCH RFC 00/11] NFS/NFSD: add NFSv42 COPY operation support To: Peng Tao , Linux NFS Mailing List References: <1440516829-116041-1-git-send-email-tao.peng@primarydata.com> CC: Trond Myklebust , Christoph Hellwig , Zach Brown , Darren Hart , Bruce Fields , Jeff Layton From: Anna Schumaker Message-ID: <55DC9162.4080303@Netapp.com> Date: Tue, 25 Aug 2015 12:01:38 -0400 MIME-Version: 1.0 In-Reply-To: Content-Type: text/plain; charset="utf-8" Sender: linux-nfs-owner@vger.kernel.org List-ID: Hey Tao, On 08/25/2015 11:42 AM, Peng Tao wrote: > On Tue, Aug 25, 2015 at 11:33 PM, Peng Tao wrote: >> Hi all, >> >> This patchset adds NFSv42 COPY support to nfs and nfsd. As suggested by Christoph, >> it pulls btrfs BTRFS_IOC_CLONE/BTRFS_IOC_CLONE_RANGE ioctls to generic layer and >> adds a new file operation to call down to file systems. Then the CLONE ioctl >> is implemented for NFS and NFSD. >> >> I took a slightly different approach from Anna's NFS42 COPY implementation >> (http://www.spinics.net/lists/linux-nfs/msg53009.html) mainly because CLONE >> has different requirements than COPY. And the main differences between the two are: >> 1. well, CLONE is supported rather than COPY. >> 2. vfs_file_clone_range() does not expect file systems to do data copy, and thus >> no rw_verify_area() required. >> 3. clone does not expect partial success. >> 4. clone does not fall back to data copy if underlying storage does not support it. >> 5. no new system call required. We reuse existing BTRFS_IOC_CLONE/BTRFS_IOC_CLONE_RANGE >> and make it generic to all file systems. > I should have mentioned that this overlaps with Anna's COPY work. > We'll deal with the conflicts after figuring out what to do with both. Since we're both working on this at the same time, can you change the title of this patch series to "NFSv42 CLONE"? I got confused when I saw these patches :) Thanks, Anna > > Cheers, > Tao > >> >> Cheers, >> Tao >> >> Anna Schumaker (2): >> nfsd: Pass filehandle to nfs4_preprocess_stateid_op() >> NFSD: Implement the CLONE call >> >> Peng Tao (9): >> vfs: pull btrfs clone API to vfs layer >> vfs/btrfs: add .clone_range file operation >> nfs42: decode_layoutstats does not need res parameter >> nfs42: remove unused declaration >> nfs42: add CLONE xdr functions >> nfs42: add CLONE proc functions >> nfs42: add .copy_range file operation >> nfs: get clone_blksize when probing fsinfo >> nfs42: respect clone_blksize >> >> fs/btrfs/ctree.h | 2 + >> fs/btrfs/file.c | 1 + >> fs/btrfs/ioctl.c | 68 ++++-------------------------- >> fs/ioctl.c | 30 +++++++++++++ >> fs/nfs/client.c | 1 + >> fs/nfs/nfs42.h | 3 +- >> fs/nfs/nfs42proc.c | 71 +++++++++++++++++++++++++++++++ >> fs/nfs/nfs42xdr.c | 102 +++++++++++++++++++++++++++++++++++++++++++-- >> fs/nfs/nfs4file.c | 58 ++++++++++++++++++++++++++ >> fs/nfs/nfs4proc.c | 4 +- >> fs/nfs/nfs4xdr.c | 26 ++++++++++++ >> fs/nfsd/nfs4proc.c | 91 ++++++++++++++++++++++++++++++++++++---- >> fs/nfsd/nfs4state.c | 5 +-- >> fs/nfsd/nfs4xdr.c | 21 ++++++++++ >> fs/nfsd/state.h | 4 +- >> fs/nfsd/vfs.c | 7 ++++ >> fs/nfsd/vfs.h | 1 + >> fs/nfsd/xdr4.h | 10 +++++ >> fs/read_write.c | 45 ++++++++++++++++++++ >> include/linux/fs.h | 4 ++ >> include/linux/nfs4.h | 7 +++- >> include/linux/nfs_fs_sb.h | 2 + >> include/linux/nfs_xdr.h | 20 +++++++++ >> include/uapi/linux/btrfs.h | 10 ----- >> include/uapi/linux/fs.h | 9 ++++ >> 25 files changed, 510 insertions(+), 92 deletions(-) >> >> -- >> 1.8.3.1 >>