On Tue, Jun 22, 2021 at 10:02:24PM -0400, Theodore Ts'o wrote: > On Tue, Jun 22, 2021 at 11:18:45AM -0700, Leah Rumancik wrote: > > > diff --git a/tests/ext4/048 b/tests/ext4/048 > > > index 51189618..35e6aa7f 100755 > > > --- a/tests/ext4/048 > > > +++ b/tests/ext4/048 > > > @@ -93,6 +93,14 @@ _scratch_mkfs_sized $((128 * 1024 * 1024)) >> $seqres.full 2>&1 > > > # create scratch dir for testing > > > # create some files with no name a substr of another name so we can grep later > > > _scratch_mount >> $seqres.full 2>&1 > > > + > > > +# Use the presence of the journal checkpoint ioctl as a proxy of filename > > > +# wipe being supported > > > +if test -x $here/src/checkpoint_journal && \ > > > + ! $here/src/checkpoint_journal $SCRATCH_MNT --dry-run ; then > > > + _notrun "filename wipe not supported" > > > +fi > > > > What if checkpoint_journal is not there? Should the test be > > skipped in that case as well? > > I went back and forth on that one. In actual practice > checkpoint_journal should always be built in a valid xfstests > installation, so the case of it not existing should rarely if ever > arise. We don't actually _need_ checkpoint_journal to run the test; > we're just using it as a hint as to whether the filename wipe feature > is present. So I decided to let the test run if we couldn't find it, > on the theory that in the long run, all future kernels will have the > feature. But the case could be made in other direction.... > > - Ted Works for me. -Leah
