Here is the answer of one house allocation problem of chapter 3. It is problem 3.6.
I have comments related to point 2:
Since agent 3 left the game and gives his house to agent 2 the walrasian equilibrium would be (1,C),(2,A), (4,B). Recall that a walrasian equilibrium is a price allocation such that each agent maximizes his utility and markets clear. Note that in round 1, agent 1 and agent 2 are involved in a trading cycle. Agent 1 gets house C (which belong to agent 2 by hypothesis) and agent 2 gets house A, i.e., according to their preferences, they get their preferable house. About agent 4´s choice, he would take house B since it is his second best choice for him (his first best choice has been already taken in round 1). Given this, each agent would be maximizing his utility according to Top Trading Cycles.
In this context, i would define market clearing not as consumption of goods equals the total endowment (houses), but as all houses available have a single owner (i´m not pretty sure in this point).
But if we take mkt clearing in this sense, then the identified allocation would be the walrasian equilibrium.
Note that agent 4 ends up being owner of house B (the one where he is living in) and house D (the remaining house)
In conclusion, this allocation its also Pareto Efficient, since there´s no way for one agent to be better off, without hurting other agent´s utility (FWT).