Figure 2.7 illustrates the situation with a dedicated CPU
performing message passing as well as executing the handler. At
no time is the currently running process on the user CPU interrupted!
Without considering the time savings derived from running the handler on a
separate CPU, we get the following result:
Figure 2.7: General Purpose CPU, Handler in Kernel:
An arriving message is read into a buffer by the CPU
dedicated to message transfer. It then executes the
handler. The user process running on the other CPU is
never interrupted!