Re: 6502 Multitasking OS announce

[cdahello@comp.brad.ac.uk (xian the desk lisard)]

PLEASE don't go over about 75 columns!!!  i have to edit the
paragraphs by hand if they don't fit.  it's what has to be done in
vi...

thus spake Michael Ellis in comp.os.misc...

. 2) Passing parameters via zero page has problems with the fact that
. your code will NOT be reentrant.  If you propose to do ANY multitasking
. whatsoever, you must write reentrant code.  If you don't, the following
. will happen:

i dispute this.  it is possible to multitask with non-re-entrant bits
- that's what shared-memory synchronisation is all about.  but you
have to make sure that no more than one thread is using any one bit of
memory at a time.  which is easy enough if you have fixed desched
points - look at forth's classical tasking model.

.   1. parameters are installed and program jumps to routine

.   2. timer interrupt comes in and invokes the dispatcher which then
.      decides to give time to another task

solution: dispatcher says "ok, i'll do that when this task's got out
of the kernel code".  yes, it's messy, but it would work for that
purpose.  (think of it as a semaphore on the parameter area.)

[]
. The way to get around this is to pass all parameters on the stack,
. and use separate stacks for the separate tasks.

and this option is even less usable on the 6502 because you only get
256 bytes of stack space, unless you use a couple of zp locations to
define your own.
--
xian the desk lisard -- cdahello@comp.brad.ac.uk
    time's taught the killing game herself
  NO        i taught the killing game
                                              first