interrupt

programming 1. An asynchronous event that suspends normal

processing and temporarily diverts the flow of control

through an "interrupt handler" routine.

Interrupts may be caused by both hardware (I/O, timer,

machine check) and software (supervisor, system call or

trap instruction).

In general the computer responds to an interrupt by storing

the information about the current state of the running

program; storing information to identify the source of the

interrupt; and invoking a first-level interrupt handler.

This is usually a kernel level privileged process that can

discover the precise cause of the interrupt (e.g. if several

devices share one interrupt) and what must be done to keep

operating system tables (such as the process table) updated.

This first-level handler may then call another handler,

e.g. one associated with the particular device which generated

the interrupt.

2. Under MS-DOS, nearly synonymous with "system call"

because the OS and BIOS routines are both called using the

INT instruction (see interrupt list) and because programmers

so often have to bypass the operating system (going directly

to a BIOS interrupt) to get reasonable performance.

[Jargon File]

(1995-02-07)