Interrupt Processing

As a program is executing, it is blissfully unaware of events that may demand the immediate attention of the system. Timers and other I/O devices operate independently of whatever program is executing, and there is no facility to coordinate their execution with other software in the computer. Raising an interrupt abruptly halts the execution of the currently executing program and transfers control to the address assigned by the hardware. Since an interrupt from a tinier requires different processing than that from a keyboard interrupt, a means to differentiate between them is also necessary. By assigning a unique address in storage to each device or timer that can generate an interrupt, the resource that requires attention is identified. For example, powering up the AGC or performing a fresh start causes an immediate transfer of control to location 04000. While pressing a key on the DSKY routes control to location 04024. Note that these interrupts are very different from counter interrupts produced by the IMU, radar and optics. Counters do not perform a context switch (which suspends a program and permits executing an interrupt routine) and do not require saving and restoring registers. The types of interrupts discussed here require logic that is more complex than counter increments or decrements, and frequently involve scheduling new work in the system. During the processing of an interrupt, the hardware prevents any other interrupts taking control of the system, and devices or timers that need attention must wait until the current interrupt has completed.

A program running at the time of the interrupt can never know when, and for how long its suspension will last. When an interrupt routine completes, no trace of its work can be visible to the resumed program. From the perspective of the interrupted program, absolutely nothing has occurred.

Questionnaire:

• Explain about interrupt processing.