ck_check_used2() renamed to ck_check_used() as the previous
  function of this name was of no use.
Added ck_check_used_twice() which checks if some position was used
  twice druing a rehash - only after using a position twice can we
  assume that an infinite loop occured
  (can we or is it only a heuristic?)
Rehashing probably working in a basic setup and with no locks.
  (Need to add: item locks, using both hashing functions in
     lookup function, switching the generation after half the table
     is rehashed, etc.)