Re: [PATCH v7 21/22] xfs: minimal conversion to errseq_t writeback error reporting

["Darrick J. Wong" <darrick.wong@oracle.com>]

On Mon, Jun 26, 2017 at 01:58:32PM -0400, jlayton@redhat.com wrote:
> On Mon, 2017-06-26 at 08:22 -0700, Darrick J. Wong wrote:
> > On Fri, Jun 16, 2017 at 03:34:26PM -0400, Jeff Layton wrote:
> > > Just check and advance the data errseq_t in struct file before
> > > before returning from fsync on normal files. Internal filemap_*
> > > callers are left as-is.
> > > 
> > > Signed-off-by: Jeff Layton <jlayton@redhat.com>
> > > ---
> > >  fs/xfs/xfs_file.c | 15 +++++++++++----
> > >  1 file changed, 11 insertions(+), 4 deletions(-)
> > > 
> > > diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c
> > > index 5fb5a0958a14..bc3b1575e8db 100644
> > > --- a/fs/xfs/xfs_file.c
> > > +++ b/fs/xfs/xfs_file.c
> > > @@ -134,7 +134,7 @@ xfs_file_fsync(
> > >  	struct inode		*inode = file->f_mapping-
> > > >host;
> > >  	struct xfs_inode	*ip = XFS_I(inode);
> > >  	struct xfs_mount	*mp = ip->i_mount;
> > > -	int			error = 0;
> > > +	int			error = 0, err2;
> > >  	int			log_flushed = 0;
> > >  	xfs_lsn_t		lsn = 0;
> > >  
> > > @@ -142,10 +142,12 @@ xfs_file_fsync(
> > >  
> > >  	error = filemap_write_and_wait_range(inode->i_mapping,
> > > start, end);
> > >  	if (error)
> > > -		return error;
> > > +		goto out;
> > >  
> > > -	if (XFS_FORCED_SHUTDOWN(mp))
> > > -		return -EIO;
> > > +	if (XFS_FORCED_SHUTDOWN(mp)) {
> > > +		error = -EIO;
> > > +		goto out;
> > > +	}
> > >  
> > >  	xfs_iflags_clear(ip, XFS_ITRUNCATED);
> > >  
> > > @@ -197,6 +199,11 @@ xfs_file_fsync(
> > >  	    mp->m_logdev_targp == mp->m_ddev_targp)
> > >  		xfs_blkdev_issue_flush(mp->m_ddev_targp);
> > >  
> > > +out:
> > > +	err2 = filemap_report_wb_err(file);
> > 
> > Could we have a comment here to remind anyone reading the code a year
> > from now that filemap_report_wb_err has side effects?  Pre-coffee me
> > was
> > wondering why we'd bother calling filemap_report_wb_err in the
> > XFS_FORCED_SHUTDOWN case, then remembered that it touches data
> > structures.
> > 
> > The first sentence of the commit message (really, the word 'advance')
> > added as a comment was adequate to remind me of the side effects.
> > 
> > Once that's added,
> > Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
> > 
> > --D
> > 
> 
> Yeah, definitely. I'm working on a respin of the series now to
> incorporate HCH's suggestion too. I'll add that in as well.
> 
> Maybe I should rename that function to file_check_and_advance_wb_err()
> ? It would be good to make it clear that it does advance the errseq_t
> cursor.

Seems like a good idea.

--D

> 
> > > +	if (!error)
> > > +		error = err2;
> > > +
> > >  	return error;
> > >  }
> > >  
> > > -- 
> > > 2.13.0
> > > 
> > > --
> > > To unsubscribe from this list: send the line "unsubscribe linux-
> > > xfs" in
> > > the body of a message to majordomo@vger.kernel.org
> > > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> > 
> > --
> > To unsubscribe from this list: send the line "unsubscribe linux-
> > btrfs" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html