From: Bob Peterson Date: Wed, 28 Feb 2018 16:16:18 -0500 (EST) Subject: [Cluster-devel] [PATCH] gfs2: Dirty source inode during rename In-Reply-To: <20180227115011.9878-1-agruenba@redhat.com> References: <20180227115011.9878-1-agruenba@redhat.com> Message-ID: <1561112443.6989500.1519852578766.JavaMail.zimbra@redhat.com> List-Id: To: cluster-devel.redhat.com MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit ----- Original Message ----- | Mark the source inode dirty during a rename instead of just updating the | underlying buffer head. Otherwise, fsync may find the inode clean and | will then skip flushing the journal. A subsequent power failure will | cause the rename to be lost. This happens in command sequences like: | | xfs_io -f -c 'pwrite 0 4096' -c 'fsync' foo | mv foo bar | xfs_io -c 'fsync' bar | # power failure | | Signed-off-by: Andreas Gruenbacher | --- | fs/gfs2/dir.c | 13 ++----------- | fs/gfs2/inode.c | 10 +--------- | 2 files changed, 3 insertions(+), 20 deletions(-) Hi, Thanks. This is now pushed to the for-next branch of the linux-gfs2 tree: https://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2.git/commit/fs/gfs2?h=for-next&id=e49f900d6e610ba6c9e6138f132eb37e91bddb37 Regards, Bob Peterson Red Hat File Systems