Page 172 - Embedded Microprocessor Systems Real World Design

I 1 I
INTERRUPT 1
INTERRUPT 2 I I
I
SHARED INTERRUPT I
-
TO MICROPROCESSOR 7
ISR FOR INTERUPT 1 I
I
ISR FOR INTERRUPT 2 0
I
f
ISR 2 OCCURS WHILE ISR 1 IS STILL HIGH
MICROPROCESSOR SEES ONLY Oh€ EDGE
CAUSED BY ISR 1. AND INTERRUW 2 IS
NEVER SERWCED

Figure 5.6
Shared Edge-Sensitive Interrupts.

Another method to handle this problem is to arbitrate the interrupt input so
that no device can request an interrupt while the line is active. Each interrupting
device must hold the line active by using a resettable flipflop or something similar
until the processor clears it. The software does not clear the interrupt line until it
has serviced the interrupt and is ready to accept another. Most edge-sensitive inter-
rupt circuits permit the internal interrupt to be cleared before another actually is
enabled, which keeps the ISR from being a reentrant.
A single device connected to an edge-sensitive interrupt can have the same
missed interrupt problem as multiple devices if it generates interrupts at a rate
faster than the processor can service them.
Many processors, such as the 80188, have timer inputs that can be used as
interrupts. The timer is programmed to count external edges on the input pin and
generate an interrupt on rollover. The timer is then loaded with a count that is one
less than the rollover value. The first edge causes a rollover and generates an inter-
rupt. The ISR must reload the counter for the next interrupt. If the timer function
is not needed, the input can function as an interrupt. Even if the extra interrupt
is not needed because you ran out of interrupt inputs, using a timer input as an
interrupt can be useful in some applications. If it is essential to know if an inter-
rupt is missed, this can be determined by looking at the timer in the ISR. If no
interrupts are missed, the timer will have rolled over to zero. If one or more inter-
rupts were missed, the timer will keep incrementing them and the value will be
greater than zero.
If you have a use for this technique, remember that some timers can be pro-
grammed either to keep incrementing after a rollover or to roll over and stop. Be
sure you program the timer to keep going after a rollover. Figure 5.7 shows a timing
diagram of the use of a timer as an interrupt source.

Interrupts in Embedded Systems 153