You have spent government funds on a boxcar full of 8-bit stack chips whose interface is shown in figure 12.25. The STK chip has a CLK input and active low $\overline{\mathrm{PUSH}}$ and $\overline{\mathrm{POP}}$ inputs. At most one of $\overline{\mathrm{PUSH}}$ and $\overline{\mathrm{POP}}$ can be asserted (low) at a time. If $\overline{\mathrm{PUSH}}$ is asserted at the time of an active clock edge, data on the eight data lines are pushed onto a stack stored inside the chip. If $\overline{\mathrm{POP}}$ is asserted at the time of an active clock edge, data from the top of the stack are popped off and driven onto the eight data lines.
You propose to use STK modules to implement the microstack of the MAYBE machine. The 8 data bits of the STK module are connected to the MAYBE's data bus, and $\overline{\mathrm{PUSH}}$ and $\overline{\mathrm{POP}}$ control inputs are connected to outputs of the load-select and drive-select decoders, respectively. Thus STK can be used as either a source or a destination for the MAYBE microarchitecture.
A. Give nanocode for the push and pop microinstructions using the new stack device. Use symbolic notation, as in the comment field of figure 12.8; STK may appear as either a source or destination.
B. You have likewise modified the call and $\mathrm{rtn}$ microinstructions to save the return point on the STK module. Since the capacity of the STK modules is quite large, you consider a microcode calling convention in which arguments to microsubroutines are passed on the microstack. Briefly explain why this proposal is awkward.