Currently tests always return success even on failures. For example if fio is not installed, exit code of all tests run is 0.
Also, it's hard to determine if some particular tests are failed, for example generic/001 output is not different if fio is installed or not, but perhaps it didn't work correctly without fio.
It would be good to have
- Tests return failure if they fail
- Define small subset tests that not taking much time for smoke testing after build.