Issuing Commands to Devices - Introduction

Now that we know how to determine when a device needs service (interrupts or polling), we have a couple more questions to ask. The first one, answered on this page, is "How does the ISR control (issue commands to and otherwise communicate with) the device?"

There are two possibilities:

• memory-mapped control, and
• control using special CPU instructions

In both cases the basic principle is the same. A typical device appears to the O/S as little more than a set of memory locations or registers (though more complicated devices can have a more complex interface than this - but don't worry about that now).

There will be:

• registers to hold data,
• registers to accept commands, and
• registers to reveal the device status.

Controlling a device consists of reading and writing those registers. The hardware manual for a device (needed if you expect to progarm that device) will give a description of each of that device's registers and what they are used for.