Last week there was some discussion regarding copying the PIC18 stack. I
started the thread to look at multitasking. Well, I've got some beginning
code in C18 that seems to work. I'm sure it can be improved (for example,
I'm restoring stack position 0, even though it does not exist). But, it
seems to work. I'd appreciate comments. It's at
http://www.piclist.org/techref/microchip/language/c/MultiTask.c.htm .

Next project is intertask communications. In the past (on MC6800 and
80286), I've had a fifo for each task that was on the task's "input." Any
task could deposit a message on another task's input fifo. When that
task's turn came up, it would read any messages off the fifo, process them
(generally creating messages for other task input fifos). Messages were
always "complete" before they were sent so we would not get a partial
message from one task and a partial from another. This, however, resulted
in a lot of ram being wasted since most of the time all these fifos were
empty. I'm wondering about maybe having a single fifo that is the input to
all tasks. Any task would place messages in this fifo. When a task
resumes, it would see if the first message in the fifo was for it. If not,
it would exit and allow the next task to run, which would also check the
first message in the fifo. Messages would be pulled off the fifo in order
so we don't end up having to keep track of empty spaces in the fifo. This
may result in the system locking up, but it SEEMS like as long as each
task pulls a message off the fifo every time the task runs, it should
work. The only time a task would exit would be when the first message in
the fifo is NOT addressed to this task. So... that's my current thinking.

Comments?

Harold



-- 
FCC Rules Updated Daily at http://www.hallikainen.com - Advertising
opportunities available!
-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist