Hi David! On Tue 06-12-22 00:35:42, David Disseldorp wrote: > On Mon, 5 Dec 2022 13:41:44 +0100, Jan Kara wrote: > > There are some mount options that are incompatible with ext3 filesystem > > type. If they are used, this test fails because it tries to remount the > > filesystem as ext3. The test makes sense even without remounting as ext3 > > so just make the test silently skip the remount. > > > > Signed-off-by: Jan Kara <jack@xxxxxxx> > > --- > > tests/ext4/044 | 7 ++++--- > > 1 file changed, 4 insertions(+), 3 deletions(-) > > > > diff --git a/tests/ext4/044 b/tests/ext4/044 > > index 50de5a40bdc7..96fa70cc0d1e 100755 > > --- a/tests/ext4/044 > > +++ b/tests/ext4/044 > > @@ -53,9 +53,10 @@ _within_tolerance "sec_ctime" $sec_ctime $sec 1 -v > > > > _scratch_unmount >> $seqres.full 2>&1 > > > > -# Test mount to ext3 then mount back to ext4 and check timestamp again > > -_mount -t ext3 `_scratch_mount_options $*` || _fail "ext3 mount failed" > > -_scratch_unmount >> $seqres.full 2>&1 > > +# Test mount to ext3 then mount back to ext4 and check timestamp again. We > > +# ignore if ext3 failed to mount. It can happen because some mount options are > > +# incompatible with ext3. Still the test makes sense. > > +_mount -t ext3 `_scratch_mount_options $*` >> $seqres.full 2>&1 && _scratch_unmount >> $seqres.full 2>&1 > > _scratch_mount > > I suppose this makes sense compared to filtering or dropping all mount > options. Yeah, filtering would need constant updating and dropping all mount options could have some unexpected sideeffects. > $* should always be empty here, right? Yeah, likely yes. I've just copied that without thinking too much about it :). > Looks fine otherwise... > Reviewed-by: David Disseldorp <ddiss@xxxxxxx> Thanks for review! Honza -- Jan Kara <jack@xxxxxxxx> SUSE Labs, CR
