[PATCH 5/5] xfs_io: print dedupe errors to stderr, not stdout

[Darrick J. Wong]

On Tue, Feb 02, 2016 at 04:14:10PM +1100, Dave Chinner wrote:
> On Fri, Jan 22, 2016 at 04:35:31PM -0800, Darrick J. Wong wrote:
> > Signed-off-by: Darrick J. Wong <darrick.wong at oracle.com>
> > ---
> >  io/reflink.c |    6 ++++--
> >  1 file changed, 4 insertions(+), 2 deletions(-)
> > 
> > 
> > diff --git a/io/reflink.c b/io/reflink.c
> > index def01be..a09e82d 100644
> > --- a/io/reflink.c
> > +++ b/io/reflink.c
> > @@ -78,11 +78,13 @@ dedupe_ioctl(
> >  			goto done;
> >  		}
> >  		if (info->status < 0) {
> > -			printf("dedupe: %s\n", _(strerror(-info->status)));
> > +			fprintf(stderr, "dedupe: %s\n",
> > +					_(strerror(-info->status)));
> >  			goto done;
> >  		}
> >  		if (info->status == XFS_EXTENT_DATA_DIFFERS) {
> > -			printf(_("Extents did not match.\n"));
> > +			fprintf(stderr, "dedupe: %s\n",
> > +					_("Extents did not match."));
> >  			goto done;
> >  		}
> >  		if (args->length != 0 &&
> 
> I think this patch breaks dedupe detection in xfstests. I see the
> dedupe tests running instead of being detected as not supported, and
> all the diffs from test failures contain:
> 
>     +dedupe: Invalid argument
> 
> This is on 4.5-rc2, so I suspect it's not returning the expected
> "not supported" error to xfstests from the ioctl.

Christoph sent a patch "[PATCH 3/4] xfstests: also treat EINVAL as
reflink/dedup not supported" to take care of that.

--D

> 
> Cheers,
> 
> Dave.
> -- 
> Dave Chinner
> david at fromorbit.com