Fig. 6.

The circuit of Fig. 5 can work with the same instance of $X_i$ if it switches continually between state 1 (forward operation of $X_i$) and state 2 (inverse operation of $X_i$). Buffers are needed for retaining the result of the previous state