Return-Path: linux-nfs-owner@vger.kernel.org Received: from mail-qa0-f50.google.com ([209.85.216.50]:42856 "EHLO mail-qa0-f50.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932506AbaCQLHF (ORCPT ); Mon, 17 Mar 2014 07:07:05 -0400 Received: by mail-qa0-f50.google.com with SMTP id o15so5165958qap.9 for ; Mon, 17 Mar 2014 04:07:05 -0700 (PDT) From: Jeff Layton To: trond.myklebust@primarydata.com Cc: linux-nfs@vger.kernel.org Subject: [PATCH v2 0/5] nfs: consolidate rename code into one set of functions Date: Mon, 17 Mar 2014 07:06:53 -0400 Message-Id: <1395054418-14504-1-git-send-email-jlayton@redhat.com> Sender: linux-nfs-owner@vger.kernel.org List-ID: Summary of changes: v2: remove -ERESTARTSYS error handling in nfs_rename hardcode isdir value in fsnotify_nameremove call for sillyrenames As Trond pointed out recently, it makes little sense to maintain two separate sets of functions for handling RENAMEs. This patchset converts nfs_rename to use the asynchronous RENAME infrastructure that is already in place for sillyrenames. It also does some cleanup to remove some minor layering violations, and adds a patch to make the sillyrename code emit a fsnotify_nameremove when a sillyrename succeeds. I've lightly tested this set and it seems to do the right thing, but it obviously could stand some time in linux-next. Jeff Layton (5): nfs: abstract out code needed to complete a sillyrename nfs: make nfs_async_rename non-static nfs: convert nfs_rename to use async_rename infrastructure nfs: remove synchronous rename code nfs: emit a fsnotify_nameremove call in sillyrename codepath fs/nfs/dir.c | 13 +++++++++++-- fs/nfs/internal.h | 7 +++++++ fs/nfs/nfs3proc.c | 36 ------------------------------------ fs/nfs/nfs4proc.c | 44 -------------------------------------------- fs/nfs/proc.c | 25 ------------------------- fs/nfs/unlink.c | 35 ++++++++++++++++++++++++++++++----- include/linux/nfs_fs.h | 1 - include/linux/nfs_xdr.h | 3 +-- 8 files changed, 49 insertions(+), 115 deletions(-) -- 1.8.5.3