Qing Li - Real-Time Concepts for Embedded Systems

Exceptions raised by internal events, such as events generated by the execution of processor instructions, are called synchronous exceptions. Examples of synchronous exceptions include the following:

· On some processor architectures, the read and the write operations must start at an even memory address for certain data sizes. Read or write operations that begin at an odd memory address cause a memory access error event and raise an exception (called an alignment exception ).

· An arithmetic operation that results in a division by zero raises an exception.

Exceptions raised by external events, which are events that do not relate to the execution of processor instructions, are called asynchronous exception s. In general, these external events are associated with hardware signals. The sources of these hardware signals are typically external hardware devices. Examples of asynchronous exceptions include the following:

· Pushing the reset button on the embedded board triggers an asynchronous exception (called the system reset exception) .

· The communications processor module that has become an integral part of many embedded designs is another example of an external device that can raise asynchronous exceptions when it receives data packets.

An interrupt , sometimes called an external interrupt , is an asynchronous exception triggered by an event that an external hardware device generates. Interrupts are one class of exception. What differentiates interrupts from other types of exceptions, or more precisely what differentiates synchronous exceptions from asynchronous exceptions, is the source of the event. The event source for a synchronous exception is internally generated from the processor due to the execution of some instruction. On the other hand, the event source for an asynchronous exception is an external hardware device.

Because the term interrupt has been used extensively in other texts, therefore, the text that follows uses exceptions to mean synchronous exceptions and interrupts to mean asynchronous exceptions. The book uses general exceptions to mean both. The term interrupts and external interrupts are used interchangeably throughout the text.

Exceptions and interrupts are the necessary evils that exist in the majority of embedded systems. This facility, specific to the processor architecture, if misused, can become the source of troubled designs. While exceptions and interrupts introduce challenging design complications and impose strict coding requirements, they are nearly indispensable in embedded applications. The following sections describe the most common and important uses of these mechanisms.

From an application's perspective, exceptions and external interrupts provide a facility for embedded hardware (either internal or external to the processor) to gain the attention of application code. Interrupts are a means of communicating between the hardware and an application currently running on an embedded processor.

In general, exceptions and interrupts help the embedded engineer in three areas:

· internal errors and special conditions management,

· hardware concurrency, and

· service requests management.

Handling and appropriately recovering from a wide range of errors without coming to a halt is often necessary in the application areas in which embedded systems are typically employed.

Exceptions are either error conditions or special conditions that the processor detects while executing instructions. Error conditions can occur for a variety of reasons. The embedded system might be implementing an algorithm, for example, to calculate heat exchange or velocity for a cruise control. If some unanticipated condition occurs that causes a division by zero, over-flow, or other math error, the application must be warned. In this case, the execution of the task performing the calculation halts, and a special exception service routine begins. This process gives the application an opportunity to evaluate and appropriately handle the error. Other types of errors include memory read or write failures (a common symptom of a stray pointer), or attempts to access floating-point hardware when not installed.

Many processor architectures have two modes of execution: normal and privileged. Some instructions, called privileged instructions , are allowed to execute only when the processor is in the privileged execution mode. An exception is raised when a privileged instruction is issued while the processor is in normal execution mode.

Special conditions are exceptions that are generated by special instructions, such as the TRAP instruction on the Motorola 68K processor family. These instructions allow a program to force the processor to move into privileged execution mode, consequently gaining access to a privileged instruction set. For example, the instruction used to disable external interrupts must be issued in privileged mode.

Another example of a special condition is the trace exception generated by the break point feature available on many processor architectures. The debugger agent, a special software program running on the embedded device, handles this exception, which makes using a host debugger to perform software break point and code stepping possible.

Although not all microcontrollers or embedded processors define the same types of exceptions or handle them in the same way, an exception facility is available and can assist the embedded systems engineer design a controlled response to these internal errors and special conditions.

The ability to perform different types of work simultaneously is important in embedded systems. Many external hardware devices can perform device-specific operations in parallel to the core processor. These devices require minimum intervention from the core processor. The key to concurrency is knowing when the device has completed the work previously issued so that additional jobs can be given. External interrupts are used to achieve this goal.

For example, an embedded application running on a core processor issues work commands to a device. The embedded application continues execution, performing other functions while the device tries to complete the work issued. After the work is complete, the device triggers an external interrupt to the core processor, which indicates that the device is now ready to accept more commands. This method of hardware concurrency and use of external interrupts is common in embedded design.

Another use of external interrupts is to provide a communication mechanism to signal or alert an embedded processor that an external hardware device is requesting service. For example, an initialized programmable interval timer chip communicates with the embedded processor through an interrupt when a preprogrammed time interval has expired. (Chapter 11 discusses programmable interval timers in detail.) Similarly, the network interface device uses an interrupt to indicate the arrival of packets after the received packets have been stored into memory.

The capabilities of exceptions and their close cousins, external interrupts, empower embedded designs. Applying the general exception facility to an embedded design, however, requires properly handling general exceptions according to the source and associated cause of each particular general exception in question. The following section provides the needed background knowledge.