From mboxrd@z Thu Jan 1 00:00:00 1970 From: "J. Bruce Fields" Subject: [PATCH 00/11] mainly d_splice_alias fixes Date: Wed, 4 Jun 2014 17:20:52 -0400 Message-ID: <1401916863-7916-1-git-send-email-bfields@redhat.com> Cc: Christoph Hellwig , linux-fsdevel@vger.kernel.org, "J. Bruce Fields" To: Al Viro Return-path: Received: from fieldses.org ([174.143.236.118]:48768 "EHLO fieldses.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751634AbaFDVVa (ORCPT ); Wed, 4 Jun 2014 17:21:30 -0400 Sender: linux-fsdevel-owner@vger.kernel.org List-ID: From: "J. Bruce Fields" The following are fixes mainly to d_splice_alias and related code. I think they'd be appropriate for 3.16, though some of them could probably use another pair of eyes. For future work: - d_splice_alias and d_materialise_unique are now doing nearly the same thing and should probably be combined. - __d_unalias's dependence on trylock's can cause spurious lookup failures on NFS. I've been meaning to look into whether it's reasonable (and good enough) to just retry in those cases. --b. J. Bruce Fields (11): namei: trivial fix to vfs_rename_dir comment dcache: move d_splice_alias dcache: close d_move race in d_splice_alias dcache: d_splice_alias mustn't create directory aliases dcache: d_splice_alias should ignore DCACHE_DISCONNECTED dcache: d_obtain_alias callers don't all want DISCONNECTED dcache: remove unused d_find_alias parameter dcache: d_find_alias needn't recheck IS_ROOT && DCACHE_DISCONNECTED exportfs: update Exporting documentation dcache: d_materialise_unique isn't GPL-only dcache: d_splice_alias should detect loops Documentation/filesystems/nfs/Exporting | 38 +++--- fs/btrfs/super.c | 9 +- fs/ceph/super.c | 2 +- fs/dcache.c | 198 +++++++++++++++++++------------ fs/namei.c | 2 +- fs/nfs/getroot.c | 2 +- fs/nilfs2/super.c | 2 +- include/linux/dcache.h | 1 + 8 files changed, 148 insertions(+), 106 deletions(-) -- 1.7.9.5