MC908QY8CDWE Datasheet, PDF (147/232 Pages) Freescale Semiconductor, Inc – Addendum to MC68HC908QB8, rev. 3

14.6 Exception Control
Normal sequential program execution can be changed in three different ways:
1. Interrupts
a. Maskable hardware CPU interrupts
b. Non-maskable software interrupt instruction (SWI)
2. Reset
3. Break interrupts
Exception Control
14.6.1 Interrupts
An interrupt temporarily changes the sequence of program execution to respond to a particular event.
Figure 14-7 flow charts the handling of system interrupts.
Interrupts are latched, and arbitration is performed in the SIM at the start of interrupt processing. The
arbitration result is a constant that the CPU uses to determine which vector to fetch. Once an interrupt is
latched by the SIM, no other interrupt can take precedence, regardless of priority, until the latched
interrupt is serviced (or the I bit is cleared).
At the beginning of an interrupt, the CPU saves the CPU register contents on the stack and sets the
interrupt mask (I bit) to prevent additional interrupts. At the end of an interrupt, the RTI instruction recovers
the CPU register contents from the stack so that normal processing can resume. Figure 14-8 shows
interrupt entry timing. Figure 14-9 shows interrupt recovery timing.
14.6.1.1 Hardware Interrupts
A hardware interrupt does not stop the current instruction. Processing of a hardware interrupt begins after
completion of the current instruction. When the current instruction is complete, the SIM checks all pending
hardware interrupts. If interrupts are not masked (I bit clear in the condition code register), and if the
corresponding interrupt enable bit is set, the SIM proceeds with interrupt processing; otherwise, the next
instruction is fetched and executed.
If more than one interrupt is pending at the end of an instruction execution, the highest priority interrupt is
serviced first. Figure 14-10 demonstrates what happens when two interrupts are pending. If an interrupt
is pending upon exit from the original interrupt service routine, the pending interrupt is serviced before the
LDA instruction is executed.
The LDA opcode is prefetched by both the INT1 and INT2 return-from-interrupt (RTI) instructions.
However, in the case of the INT1 RTI prefetch, this is a redundant operation.
NOTE
To maintain compatibility with the M6805 Family, the H register is not
pushed on the stack during interrupt entry. If the interrupt service routine
modifies the H register or uses the indexed addressing mode, software
should save the H register and then restore it prior to exiting the routine.
MC68HC908QB8 Data Sheet, Rev. 3
Freescale Semiconductor
145