Very strange behaviour if M2 multitasking enabled incorrectly. Not detected by editor

[PhilHornby]

You could argue this is a firmware bug - but it's probably most easily fixed in the editor. If you stop people like me typing rubbish in, you don't get rubbish out


The following program loops indefinitely, incrementing and printing the value of W0. It pauses 1000mS between loops. The value of OSCCON is printed, to determine if multi-tasking is actually in use.

#picaxe 08M2
#terminal 4800

;Demonstration of doubled-up commands when multi-tasking enabled (incorrectly).
;Change "start" to "start1:" to show the issue:-
start:
      peeksfr $39,b0                      ;Doesn't work in simulator.
      sertxd (cr,lf,"OSCCON=",#b0)        ;Show clock freq (104=4MHz,120=16MHz)
     
      W0 = 0
     
      do
            sertxd(cr,lf,"Msg #",#W0,",TIME:",#TIME)
            inc W0
            pause 1000
      loop 

;start1: pause 65535 : goto start1

If the initial label is set to "start", you get the expected output:-

OSCCON=104 (4MHz.)​

​

​

Msg #0,TIME:0​

Msg #1,TIME:1​

Msg #2,TIME:2​

Msg #3,TIME:3​

Msg #4,TIME:4​

Msg #5,TIME:5​

Changing the label to "start0" doesn't change anything.

OSCCON=104 (4MHz)​

Msg #0,TIME:0​

Msg #1,TIME:1​

Msg #2,TIME:2​

Msg #3,TIME:3​

Msg #4,TIME:4​

Msg #5,TIME:5​

Changing the label to "start1" is very weird... I would suggest this condition should be detected and prohibited...it's as though it runs Task #1 twice, because there's no Task #0

OSCCON=120 (16MHz)​

Msg #0,TIME:0​

Msg #0,TIME:0​

Msg #2,TIME:1​

Msg #2,TIME:1​

Msg #4,TIME:2​

Msg #4,TIME:2​

Msg #6,TIME:3​

Msg #6,TIME:3​

Changing the label back to "start0" and adding a 2nd task: "Start1: pause 65535 : goto Start1" :-

OSCCON=120 (16MHz)​

Msg #0,TIME:0​

Msg #1,TIME:1​

Msg #2,TIME:2​

Msg #3,TIME:3​

Msg #4,TIME:4​

Msg #5,TIME:5​

What I find equally strange, is that this is reproducible in the simulator too (though not the OSCCON check).

 

[hippy]

Changing the label to "start1" is very weird... I would suggest this condition should be detected and prohibited...it's as though it runs Task #1 twice, because there's no Task #0

Thats' pretty much it. The implied Start0 at the start of the program falls into what also happens to be Start1 so you end up with two programs running in lock-step with each other.