In femto/macro cellular networks, the stability and fairness problems caused by the unplanned and random characteristic of femtocells must be solved. By applying queueing theory in IP based femto/macro cellular networks, we found the stability condition, and described two kinds of cell section policies of users. As a main contribution, we provided the adaptive channel distribution algorithm which minimizes the average packet sojourn time at transmitting systems and keeps the whole systems stable and fair among cells. Through experiments in various environments, we analyzed the influence of channel reuse factor, cell selection policies, and the number of femtocells on system performance.