On Thu, Mar 17, 2016 at 07:00:43AM +1100, Dave Chinner wrote: > On Wed, Mar 16, 2016 at 10:54:18PM +0800, Eryu Guan wrote: > > On Wed, Mar 16, 2016 at 05:04:16AM -0700, Christoph Hellwig wrote: > > > On Wed, Mar 16, 2016 at 07:53:24PM +0800, Eryu Guan wrote: > > > > When testing xfs/073 with MKFS_OPTIONS="-m crc=1,finobt=1" set, it fails > > > > due to extra warning about disabling finobt feature: > > > > > > > > +warning: finobt not supported without CRC support, disabled. > > > > > > > > Because xfs/073 disables crc unconditionally and finobt can not be > > > > enabled either. > > > > > > How about explicitly disabling finobt as well? > > > > I think that's a cleaner way to fix it. Just that some extra work is > > needed to let the test pass with old mkfs.xfs without metadata suport. > > This is a downhill slope. e.g. if I add "-i sparse=1" to the mkfs > line, it will warning differently and a new filter is needed. I > would suggest just ignoring the mkfs output altogether (maybe just > dumping it all to $seqres.full) as it's not necessary to determine > whether the test ran correctly or not. Yes, other tests that call _scratch_mkfs_xfs with explicit "-m crc=0" all redirect output to null or $seqres.full, so I only see the warning in xfs/073. I'd like to do the same for it. I think the update to _scratch_mkfs_xfs_opts() in v2 patch is still worth doing, it belongs to a seperate patch though. Thanks, Eryu _______________________________________________ xfs mailing list xfs@xxxxxxxxxxx http://oss.sgi.com/mailman/listinfo/xfs
