This mitigates the problem whereby a packet sent in one test is not dropped
by a kernel and affects the subsequent test. The increased timeout works
around this problem. However, at some point there may be a need to
implement more proper solution (test should capture the packet sent and
drop it on its own, then start another test).