Re: Linux doing "cooperative multitasking"?
Steven N. Hirsch (shirsch@ibm.net)
Wed, 27 Nov 1996 18:10:44 -0500
Wayne Schlitt wrote:
>
Cooperative multitasking is only a severely broken method when it is
> used for the wrong purposes. The same can be said about must about
> any other method.
>
> I recently finished a project where I used cooperative multitasking
> for the first time and I found it to be a big plus, because:
>
> * I couldn't afford the 16k of memory that the preemptive
> multitasking kernel took up. (This was an embedded system that
> had 64k of memory.)
>
> * cooperative multitasking simplifies a lot of the interprocess
> communication problems. You don't have to worry about semaphores
> as much since you know that you won't lose control until you say
> so.
Very true. Embedded Systems mag ran a great article by Do-while Jones
on the virtues of cooperative multitasking for implementing real-time
control systems. Changed my mind on the subject at any rate! Does
anyone else read ES?
Steve