lpt[n] = mklp(m, m0, LP_FREE, m1), n < # lpt, ∀ n. n < # lpt ∧ lpt[n].flags = LP_FREE → lpt[n].ref-size = 0 ⊦ m = 0, n ∈ nx