HD6475348R Datasheet, PDF (112/487 Pages) Hitachi Semiconductor – Single-Chip Microcomputer

currently executing one of the five instructions listed above. After executing this instruction the
CPU always proceeds to the next instruction. (And if the next instruction is one of these five, the
CPU also proceeds to the next instruction after that.) The exception-handling sequence starts after
the next instruction that is not one of these five has been executed. The following is an example:
(Example)
.
.
.
.
.
.
LDC.B #H'00,TP
Program flow
← Interrupt controller notifies CPU of interrupt
MOV.W #H'FE80,SP
CPU executes the instruction next to LDC before
starting exception handling
MOV.B #H'00,@WCR
.
.
.
To exception-handling sequence
4.8.2 Disabling of Exceptions Immediately after a Reset
If an interrupt is accepted after a reset and before the stack pointer (SP) is initialized, the program
counter and status register will not be saved correctly, leading to a program crash. To prevent this,
when the chip comes out of the reset state all interrupts, including the NMI, are disabled, so the
first instruction of the reset routine is always executed. As noted earlier, in the minimum mode,
this instruction should initialize the stack pointer (SP). In the maximum mode, the first instruction
should be an LDC instruction that initializes the stack page register (TP); the next instruction
should initialize the stack pointer.
4.8.3 Disabling of Interrupts after a Data Transfer Cycle
If an interrupt starts the data transfer controller and another interrupt is requested during the data
transfer cycle, when the data transfer cycle ends, the CPU always executes the next instruction
before handling the second interrupt.
Even if a nonmaskable interrupt (NMI) occurs during a data transfer cycle, it is not accepted until
the next instruction has been executed. An example of this is shown below.
92