Comment 26 for bug 1432871

So the regression I discovered was actually just behavior change. Due to the change to /proc/self/mountinfo the filesystem type is now more explicit. In this case that means that nfs mounts are now labeled nfs4.

Additionally I discovered that the test that is now failing was previously being skipped because df was not able to read mount entries on trusty *(which is now fixed in this bug). I'm still working through traces and building tests packages in a ppa to get more output.