> > I suppose it is more suitable, but since at the moment there is only one(?) > > such feature and there is an open question whether it should or should not > > be configurable, I myself would have taken the easy path, but Miklos > > often has a different perspective on these sort of things... > > What exactly are we testing? > What are testing against silent regressions. If some change breaks whiteout share we won't know about it without a test. > Hard linked whiteouts are an optimization, not something to be relied > on in any case. The test should succeed even if overlayfs decides for > some reason not to share the inode. > The best practice would be to ask overlay if feature is supported in the kernel AND make sure that overlay mount did not disable it on a specific instance (because of upper fs capabilities). That is what _check_overlay_feature does for "features" that have both module param AND a mount option. If overlay says that the feature is expected to be enabled on the instance, we check correctness of the feature. Otherwise, we report that "feature" is not supported on this instance. Thanks, Amir.
