Interrupt problem.

Hello,

It's the first time I try to use ON INTERRUPT.
If I compile my program (PBP2.45a), I get an error.
Error test.mac 46: [235] Opcode Expected Instead of '1'
If I disable ON INTERRUPT with a " ' " , compiling is ok.

What's wrong?

@ DEVICE PIC16F819, INTRC_OSC_NOCLKOUT, MCLR_OFF
OPTION_REG = %10000000 ' Turn off pull-up resistors
DEFINE ONINT_USED 1
OSCCON=$60 ' osc. controller, 70=8MHz, 60=4MHz
adcon1=7 ' comperator off
TRISA=%00011111
TRISB=%11110000
led VAR PORTB.0

ON INTERRUPT GoTo myint ' Define interrupt handler
INTCON = %10001000 ' Enable RBIE interrupt
loop:
led=1 ' Turn LED on
GoTo loop ' Do it forever
myint:
Disable ' No interrupts past this point
led=0 ' If we get here, turn LED off
Pause 500 ' Wait .5 seconds
INTCON.0 = 0 ' Clear interrupt flag
Resume
Enable

Stop

Thanks for the help,
Peter

myint:
Disable ' No interrupts past this point


is this work:

DISABLE ‘ Disable interrupts in handler
myint:

Resume ‘ Return to main program
Enable ‘ Enable interrupts after handler

Hello Kima,

I tried what you wrote.
It still gives the same error.
What else could it be?

Peter

(1) Ensure that the 'myint' in the ON INTERRUPT statement is the same spelling as the Label in the Interrupt Service Routine.

(2) Ensure that the myint label in the ISR has a COLON and not a SEMI-COLON following it.

(3) Ensure that the myint label in the ISR is against the left-hand margin of your source code, and is not prefixed by a Space or Tab or other character (this forum by default strips out formatting characters so we can't see if this is the case).

(4) Delete out both occurances of myint and retype them in.

See PM.TXT in your PBP directory.

Hello,

After deleting DEFINE ONINT_USED 1, compiling wasn't a problem anymore.
I have to test it in real.
Thanks for the help so far.

Peter

Peter, to my knowlege there is no command "DEFINE ONINT_USED 1, at least in the PBP 2.45 manual listings.

Dave Purola,

On versions of the PicBasic Pro Compiler earlier than 2.33, Define ONINT_USED 1 was necessary when using the MeLabs boot-loader to keep library code from being inserted in the first 4 locations.

Define LOADER_USED 1 is used for newer versions. You'll receive the error message Opcode Expected Instead of '1' when using the older define with a newer version of PBP.

Bruce, it's true.
The first version of PBP for me was earlier than 2.33.
At the top of my programs I always put the PIC-chip pinning and the "standard" initalisation of that chip.
For a new program I copy that top of the program and thus also "old" commands......
That happend before with the command 'ansel' for the 12F675.
Thanks everybody for the reactions.

Hello,
Next program works fine.
I found that you are not allowed to place any command after
resume and enable !!
So, place this always at the end of the program.
Is this a bug in PBP?

@ DEVICE PIC16F819, INTRC_OSC_NOCLKOUT, WDT_OFF, MCLR_OFF
OPTION_REG=0 ' Enable PORTB pullups RB4,5 EN 6, %00000000
DEFINE LOADER_USED 1 ' Boot loader is being used
OSCCON=$60 ' osc. controller, 70=8MHz, 60=4MHz

init:
Output PORTA.7 ' Output LED
option_reg = %0
ON INTERRUPT GoTo int_routine ' Define interrupt handler
INTCON = $88 ' GIE=1, enable INTE interrupt
main:
PORTA.7=0 ' LED off
GoTo main
End
int_routine:
Disable
PORTA.7=1 ' LED on
Pause 300
intcon.0=0
Resume
Enable
' No commands allowed after here!!

Peter

It is best to pretend that your Interrupt Service Routine is like a subroutine, and the RESUME statement is the RETURN that exits you back out from that routine. Naturally, unless you explictly place code AFTER a RETURN statement, and somehow then jump to it, it'll never be executed will it? Likewise with your ISR.

Also, you have one fatal error in your program, remove that END statement and place it at the very end of your code NOT in the middle. The END statement is there to tell the compiler that there is no more code, so technically your ISR shouldn't even compile in. This too is perhaps why further code is ignored.

Hi Peter,

It's not a bug in PBP.

RESUME inserts RETFIE (return from interrupt) after your interrupt handler is finished.

If you insert code after the RETFIE (RESUME), there's no way for it to ever execute since program flow is returned to where it was before the interrupt.

I don't understand the logic behind having END mid-stream in code like this, but the END command won't prevent any code following it from being compiled. It generates code to enter a continuous loop executing the SLEEP command.

If you don't include a GOTO before END, and program execution lands on the END statement, then you're trapped in a continuous loop similar to the one shown below.

@@0000 sleep
clrf PCLATH
goto @@0000

This (shown below) should work as-is since program execution should never land on END, but I still don't see the logic in placing it mid-stream like this!

GoTo main
End
int_routine:

I have developed a PBP code to be used on an embedded sms sender...

My requirement is when my PIC (SMS) is interrupted by a centralized controlling PIC it should execute

the necessary routines and send sms via a T10 phone.... As far as sending sms is concerned the code

works ok...But the interrupt mechanism is not working properly.

I have some subroutines to attend while inside the ISR..I simulated the code and found that upto the

first GOSUB command it works properly and then it reacts as though a new interrupt has occured and it

executes ISR from the begining...

Please be kind enough to help me

I have herewith attached my code

thanks
aruran

INCLUDE "modedefs.bas"

DEFINE OSC 10 '10MHz crystal

' Set receive register to receiver enabled
DEFINE HSER_RCSTA 90h
' Set transmit register to transmitter enabled
DEFINE HSER_TXSTA 20h
' Set baud rate
DEFINE HSER_BAUD 9600
' Set SPBRG directly (normally set by HSER_BAUD)
DEFINE HSER_SPBRG 15

Clear

customer var byte
string var byte [32]'Text Data
pdu VAR BYTE [8] 'pdu data
tmp VAR BYTE [8] 'temporary array
b1 VAR BYTE [4]
x var byte [2]
finalpdu var byte [54]
recvnum var byte [12]
telnum var byte [12]
string1 var byte [8]
ch var byte

OPTION_REG = 0
TRISB = $0F

on Interrupt Goto myint ' Define interrupt handler
INTCON = $90 ' Enable INTE interrupt

loop:
Goto loop ' Infinite Loop


'''''''''''''' Interrupt handler'''''''''''''''''''''''


myint:

Disable ' No interrupts past this point

for x=0 to 9
LookUp X,["0011000B81"],b1
finalpdu(x)=b1
next x

; Taking in Recepient Number

telnum[1]=9
telnum[0]=4
telnum[3]=7
telnum[2]=7
telnum[5]=6
telnum[4]=0
telnum[7]=6
telnum[6]=9
telnum[9]=3
telnum[8]=1
telnum[11]=7
telnum[10]=$F

For X=0 TO 11
LookUp telnum(x),["0123456789"],b1
recvnum(x)=b1
Next X
recvnum(10)="F"

for x=10 to 21
finalpdu(x)=recvnum[x-10]
next x

for x=22 to 29
LookUp (X-22),["0000AA20"],b1
finalpdu(x)=b1
next x

;;;;;;;;;;;;;;;;;;;;;;Message;;;;;;;;;;;;;;;;;;;;; ;

;Taking in Customer

;;;;;;;;;;; Should be deleted ;;;;;;;;;;;;;;;;;;;;;

for x=0 to 16
lookup x,["aaaaaaaaaaaaaaaaa"],b1
customer(x)=b1
next x

;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; ;;

for x=0 to 16
lookup2 x,[customer],b1
string(x)=b1
next x

;Taking in String2
For X=0 TO 14
LookUp2 X,[",Reported Fault"],b1
string(x+17)=b1
Next X

;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; ;;

;HSerout [str string\32,13,10]
HSerout ["AT+CMGF=0",13,10] 'Set to PDU Mode
HSerout ["AT+CMGS=42",13,10] 'Set to PDU length
hserout [str finalpdu\30]

for x=0 to 7
string1(x)=string(x)
next x
Gosub txt2pdu

for x=0 to 7
string1(x)=string(x+8)
next x
;GoSub txt2pdu

for x=0 to 7
string1(x)=string(x+16)
next x
GoSub txt2pdu

for x=0 to 7
string1(x)=string(x+24)
next x
GoSub txt2pdu
HSerout [26]

INTCON.1 = 0 ' Clear interrupt flag
Resume ' Return to main program
Enable

;-----CONVERSION ROUTINES-------
;---TXT2PDU--(sms encoding)-----


TXT2PDU:
for x=1 to 7
tmp(x) = string1(x) << (8-x)
next x
ch(0)=string1(0)
pdu(0)=tmp(1) + ch(0)

for x=1 to 6
ch(x)=string1(x) >> x
pdu(x)=tmp(x+1) + ch(x)
next x

for x=0 to 6
hserout [hex2 pdu(x)]
next x

RETURN

ENABLE causes PBP to continue inserting interrupt test code. If you jump out
of the ISR routine with GOSUB TXT2PDU, and this sub routine has interrupt
test code in it, then it's going to jump directly back to your ISR sub-routine.

Example. In your ISR interrupt routine you have ENABLE at the end which is
causing PBP to insert interrupt test code in your lower sub-routine TXT2PDU.

Once you land on the first GOSUB TXT2PDU statement, you land right on
code inside your TXT2PDU routine that's re-directing you right back to the
beginning of ISR.

' == last part of ISR routine ==
INTCON.1 = 0 ' Clear interrupt flag
Resume ' Return to main program "with RETFIE"
' Enable '<=== CONTINUES PLACING INTERRUPT TEST CODE AFTER HERE

Remove the ENABLE from the last line of routine ISR (like shown above), and
drop it at the bottom of your TXT2PDU sub-routine like this;

TXT2PDU:
for x=1 to 7
tmp(x) = string1(x) << (8-x)
next x
ch(0)=string1(0)
pdu(0)=tmp(1) + ch(0)

for x=1 to 6
ch(x)=string1(x) >> x
pdu(x)=tmp(x+1) + ch(x)
next x

for x=0 to 6
hserout [hex2 pdu(x)]
next x

RETURN ' return from each GOSUB
Enable ' Re-enable auto inserrtion of interrupt checking code beyond here