I really like the simplicity of the Microchip stack where you can ask if
how many bytes are in a buffer (for us to pick up data on input), or how
much space is available in a buffer (for us to put data in on output). If
we're waiting for data, and there's nothing there, we just go on to the
next task. The next time we're in the task, we ask again. On the output
side, we can either send a part of what we need to send or exit the task
and on the next call see if there's then enough room for us to output. I
like this second approach in that complete messages are always sent in a
single piece.

MANY years ago, I wrote a simple multitasking system for an 80286. A
separate stack was maintained for each task. The stack held local
variables for that task plus the subroutine call stack. A function called
NextTask() would move the stack pointer to point to the stack for the next
task in the loop, then do a return, loading the program counter with where
we left off in that stack.

Getting back to I/O (and interprocess communications), each of my
processes had an input fifo. When a task was called, it would see if there
was anything in its input FIFO. If so, it would take the complete message
(which is why I like the complete message transmission described above),
process it, then put the results (another complete message) in the input
FIFO of another process or I/O device.

I wrote a similar system for the PIC18 several years back (see
http://www.piclist.com/techref/microchip/language/c/multitask.c.htm ). A
difference here was that the stack was copied to a buffer, then copied
back to the stack when the task came around again since it was not really
practical to keep several stacks on the PIC18 and just move the stack
pointer.




--=20
FCC Rules Updated Daily at http://www.hallikainen.com - Advertising
opportunities available!
Not sent from an iPhone.
--=20
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.