Threads and Semaphores: 2.1 Two Models of Multitasking: Cooperative Multitasking

A cooperative multitasking system, as the name suggests, is one in which the various programs cooperate. They agree to share the system and its resources. Each program running under a system has complete control while it is actually running. Every so often, when it reaches a convenient place, it calls a special system routine (called a context-switch routine) to see if any other program has work do to. If so, that program takes control of the system until it in turn reaches a convenient stopping point and passes control on to the next waiting program. If several programs are ready to run, they wait in a queue so that each one gets a chance to run before the first program runs again. (It is also possible to implement a cooperative multitasking system where some programs have a higher priority than others. In this case, a more complicated algorithm might be used by the context-switch routine to determine which program gets to run next.)

Smooth operation of a cooperative multitasking system requires that all programs be written to call the context-switch routine frequently. When large calculations are being performed, programmers tend to find this requirement inconvenient. Writing well-behaved programs (i.e., programs that do not keep control of the processor for too long at a stretch) is especially difficult because most cooperative multitasking systems impose restrictions on when a context switch can take place.

GEOS SDK TechDocs |

2 Two Models of Multitasking | Next:

2.2 Preemptive Multitasking