Not being able to use variables on pic18fxx50

[Lope Vega]

Hi list!

I'm new to microcontrolers, and I'm tying to program a pic18f4550,
under gnu/linux and/or freebsd.

The main problem I'm finding is with variables. I can't make them work.

gpasm would say "Warning [231] No memory has been reserved by this instruction." when it sees:

myvar res 1

I tried with:

myvar res 2

with no succesful results, I've tried something like this:

        ;; setup processor and output format                                    
        list p=p18f4550,f=inhx32
        include <p18f4550.inc>

        ;; disable watchdog timer                                              
        config wdt = off
        ;; chip will use internal oscillator (4 mhz)                            
        config fosc = intoscio_ec                                    

        org 0
        goto main

        ;; let `count1' and `count2' live after the previous
        ;; `goto' instruction on code-memory, where shouldn't
        ;; clash with program's instructions
count1 equ 0x6
count2 equ 0x8

        ;; offset our program by 32 bytes, so we have space for the previous and other variables I might want to use
        org 0x20
main:
        .....

After I started "decoding" hex-files' dumps, doesn't work either, so I'm a bit lost, have tried even more different approaches but I can't get the pic to "respect" and work with any variables other than constants.

This is an example program I wanted to start with:

        ;; setup processor and output format                                                
        list p=pic18f4550,f=inhx32
        include <p18f4550.inc>

        ;; disable watchdog timer                                                          
        config wdt = off
        ;; chip will use internal oscillator (4 mhz)                                        
        config fosc = intoscio_ec

        org 0

counter res     2

main:
        ;; portd will be all outputs
        clrf    PORTD
        clrf    TRISD

restart:
        ;; give the memory address location `0x0' the value `5'
        movlw   5
        movwf   0x0

loop:
        ;; decrement this value till it's 0, what will cause the

        decfsz  0x0, f
        goto    loop

        btg     PORTD, RD1
        goto    restart
        end



__________________________________________________
Correo Yahoo!
Espacio para todos tus mensajes, antivirus y antispam ¡gratis!
Regístrate ya - http://correo.yahoo.es


---------------------------------------------------------------------
To unsubscribe, e-mail: gnupic-unsubscribe@...
For additional commands, e-mail: gnupic-help@...

[David Barnett-3]

It looks like you're reserving memory in ROM and then accessing addresses in
RAM. The RES directive allocates memory in the section you're defining, so
look at UDATA and similar directives in the manual.

David

On Mon, Aug 18, 2008 at 12:20 PM, Lope Vega <vega_lope@...> wrote:

> ...but I can't get the pic to "respect" and work with any variables other
> than constants.
>
> This is an example program I wanted to start with:
>
> ...

       org 0

>
> counter res     2
>
> main:
>        ;; portd will be all outputs
>        clrf    PORTD
>        clrf    TRISD
>
> restart:
>        ;; give the memory address location `0x0' the value `5'
>        movlw   5
>        movwf   0x0
>
> loop:
>        ;; decrement this value till it's 0, what will cause the
>
>        decfsz  0x0, f
>        goto    loop
>
>        btg     PORTD, RD1
>        goto    restart
>        end
>

[Lope Vega]

Hello again,

Thanks a lot, but it didn't do the trick either.

By looking at this a bit more indeep, while in relocatable mode I've noticed I can't use `udata_shr' (it throws an error about a not existing .udata_shr on the linker script, even when I gave that section a name to avoid it making use of the default value).

I still haven't been able to use variables reliably even with the most simple of the programs.

I'm pasting today, the program, and dumps from the chip's code and eeprom memories, and if somebody could tell me something about what's going on here, that would be great.

Here is the code:

        ;; setup processor and output format                                                
        list p=p18f4550,f=inhx32
        include <p18f4550.inc>

        config plldiv = 1
        config fosc = intoscio_ec
        config fcmen = off
        config ieso = off
        config pwrt = off
        config bor = off
        config borv = 0
        config vregen = off
        config wdt = off
        config mclre = on
        config lpt1osc = off
        config pbaden = off
        config ccp2mx = on
        config stvren = off
        config lvp = on
        config icprt = off
        config xinst = on
        config debug = off
        config cp0 = off
        config cp1 = off
        config cp2 = off
        config cp3 = off
        config cpb = off
        config cpd = off
        config wrt0 = off
        config wrt1 = off
        config wrt2 = off
        config wrt3 = off
        config wrtb = off
        config wrtc = off
        config wrtd = off
        config ebtr0 = off
        config ebtr1 = off
        config ebtr2 = off
        config ebtr3 = off
        config ebtrb = off

        udata
counter         res     1

        code    0
        pagesel main
        goto main

        code
main
        ;; PORTB is all inputs, it just got an active-high switch                          
        ;; on RB4
        clrf    PORTB
        movlw   0xff
        movwf   TRISB

        ;; PORTD is all outputs                                                              
        clrf    TRISD

        banksel counter

restart
        movlw   0x05
        movwf   counter

loop
        ;; has the switch been pressed?                                                    
        btfsc   PORTB, RB4
        ;; no, continue polling                                                            
        goto    loop

        ;; it's been, turn on the led @ RD0, just to have some feedback                    
        bsf     PORTD, RD0

debounce
        ;; has the switch been released?                                                    
        btfss   PORTB, RB4
        ;; no, poll some more                                                              
        goto    debounce

        ;; yes, turn the led @ RD0 off                                                      
        bcf     PORTD, RD0

decrement
        ;; decrement `counter', store result in that variable itself, and                  
        ;; and skip the instruction after this one if `counter == 0'                        
        decfsz  counter, f
        goto    loop

        ;; toggle led @ RD1                              
        btg     PORTD, RD1

        ;; goto start, where the counter gets set to 5 again, in theory                    
        goto    restart

        end

This is a dump of the code's memory once burned:

ADDRESS   DATA
--------  -----------------------------------------------
0x000000  81 6A FF 0E 93 6E 95 6A 00 01 05 0E 60 6F 81 B8
0x000010  07 EF 00 F0 83 80 81 A8 0B EF 00 F0 83 90 60 2F
0x000020  07 EF 00 F0 83 72 05 EF 00 F0


And this a dump of the epprom as well:

ADDRESS   DATA
--------  -----------------------------------------------
0x000000  9F FF

If somebody could compile the above program and provide me with similar device's dumps that would be same as helpful as any explanation probably. The problem could very be due to the only program I could use for burning the chips on my freebsd: Broccoli18.

Many thanks.


--- El mar, 19/8/08, David Barnett <daviebdawg@...> escribió:

> De: David Barnett <daviebdawg@...>
> Asunto: Re: [gnupic] Not being able to use variables on pic18fxx50
> Para: gnupic@...
> Fecha: martes, 19 agosto, 2008 12:37
> It looks like you're reserving memory in ROM and then
> accessing addresses in
> RAM. The RES directive allocates memory in the section
> you're defining, so
> look at UDATA and similar directives in the manual.
>
> David
>
> On Mon, Aug 18, 2008 at 12:20 PM, Lope Vega
> <vega_lope@...> wrote:
>
> > ...but I can't get the pic to "respect"
> and work with any variables other
> > than constants.
> >
> > This is an example program I wanted to start with:
> >
> > ...
>
>        org 0
> >
> > counter res     2
> >
> > main:
> >        ;; portd will be all outputs
> >        clrf    PORTD
> >        clrf    TRISD
> >
> > restart:
> >        ;; give the memory address location `0x0'
> the value `5'
> >        movlw   5
> >        movwf   0x0
> >
> > loop:
> >        ;; decrement this value till it's 0, what
> will cause the
> >
> >        decfsz  0x0, f
> >        goto    loop
> >
> >        btg     PORTD, RD1
> >        goto    restart
> >        end
> >

__________________________________________________
Correo Yahoo!
Espacio para todos tus mensajes, antivirus y antispam ¡gratis!
Regístrate ya - http://correo.yahoo.es


---------------------------------------------------------------------
To unsubscribe, e-mail: gnupic-unsubscribe@...
For additional commands, e-mail: gnupic-help@...

[Peter Stuge-4]

On Thu, Aug 21, 2008 at 03:13:19PM +0000, Lope Vega wrote:
>         udata
> counter         res     1

Good so far.


>         code    0
>         pagesel main
>         goto main
>
>         code

These lines should read:

        code
        org 0
        pagesel main
        goto main


A small note on the code:

> debounce
>         ;; has the switch been released?

This debounce is not really debouncing but the program will fall
through the code on the first bouncy switch press, making the LED
flash for only a few instructions which will not be noticeable to
the eye. :) You could capture the edge with an oscilloscope.


//Peter

---------------------------------------------------------------------
To unsubscribe, e-mail: gnupic-unsubscribe@...
For additional commands, e-mail: gnupic-help@...

[Robert Pearce-3]

On Fri, 22 Aug 2008, Peter Stuge <peter@...> wrote :

>
>>         code    0
>>         pagesel main
>>         goto main
>>
>>         code
>
>These lines should read:
>
>       code
>       org 0
>       pagesel main
>       goto main
>

Actually the former is acceptable - the 'code' directive accepts an
argument which acts as an "org" hint for the linker. In fact Peter's
suggestion is wrong, because the "org 0" takes the assembler out of
relocatable mode and back to absolute, so it makes no sense to put one
immediately after a "code" directive.

However, use of the "pagesel" macro is redundant on 18F series devices,
since the "goto" instruction on them has a 24-bit range.

What Lope has not told us, which does matter given the errors he
reports, is what his linker file contains. Also, the dump from the chip
is of far less use than the hex/lst/map files built by the linker.
--
Rob Pearce                       http://www.bdt-home.demon.co.uk

The contents of this | Windows NT crashed.
message are purely   | I am the Blue Screen of Death.
my opinion. Don't    | No one hears your screams.
believe a word.      |

---------------------------------------------------------------------
To unsubscribe, e-mail: gnupic-unsubscribe@...
For additional commands, e-mail: gnupic-help@...

[Lope Vega]

--- El jue, 21/8/08, Peter Stuge <peter@...> escribió:

> De: Peter Stuge <peter@...>
> Asunto: Re: [gnupic] Not being able to use variables on pic18fxx50
> Para: gnupic@...
> Fecha: jueves, 21 agosto, 2008 11:40
> On Thu, Aug 21, 2008 at 03:13:19PM +0000, Lope Vega wrote:
> >         udata
> > counter         res     1
>
> Good so far.
>
>
> >         code    0
> >         pagesel main
> >         goto main
> >
> >         code
>
> These lines should read:
>
> code
> org 0
> pagesel main
> goto main
>
>
> A small note on the code:
>
> > debounce
> >         ;; has the switch been released?
>
> This debounce is not really debouncing but the program will
> fall
> through the code on the first bouncy switch press, making
> the LED
> flash for only a few instructions which will not be
> noticeable to
> the eye. :) You could capture the edge with an
> oscilloscope.
>

Hello, and thanks a lot.

Yes, I know the "debouncing" part is not reliable, but I was more interested in having some code working before I could get on that, which already happened, see my next message, and many thanks for replying.

>
> //Peter
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: gnupic-unsubscribe@...
> For additional commands, e-mail:
> gnupic-help@...

__________________________________________________
Correo Yahoo!
Espacio para todos tus mensajes, antivirus y antispam ¡gratis!
Regístrate ya - http://correo.yahoo.es


---------------------------------------------------------------------
To unsubscribe, e-mail: gnupic-unsubscribe@...
For additional commands, e-mail: gnupic-help@...

[Lope Vega]

--- El vie, 22/8/08, Robert Pearce <rob@...> escribió:

> De: Robert Pearce <rob@...>
> Asunto: Re: [gnupic] Not being able to use variables on pic18fxx50
> Para: gnupic@...
> Fecha: viernes, 22 agosto, 2008 7:55
> On Fri, 22 Aug 2008, Peter Stuge <peter@...>
> wrote :
> >
> >>         code    0
> >>         pagesel main
> >>         goto main
> >>
> >>         code
> >
> >These lines should read:
> >
> >       code
> >       org 0
> >       pagesel main
> >       goto main
> >
> Actually the former is acceptable - the 'code'
> directive accepts an
> argument which acts as an "org" hint for the
> linker. In fact Peter's
> suggestion is wrong, because the "org 0" takes
> the assembler out of
> relocatable mode and back to absolute, so it makes no sense
> to put one
> immediately after a "code" directive.
>
> However, use of the "pagesel" macro is redundant
> on 18F series devices,
> since the "goto" instruction on them has a 24-bit
> range.

Ops, I didn't know it where redundant on 18f's, still I'm new to this whole thing, thanks.

>
> What Lope has not told us, which does matter given the
> errors he
> reports, is what his linker file contains. Also, the dump
> from the chip
> is of far less use than the hex/lst/map files built by the
> linker.

I didn't think of that either, sorry, your completely right though.

The linker script I've been using is the default one.

Now, I solved it yesterdeay at least!.

At some point, I took the PORTB part out, and got back to a "count  from 5 to 0 and turn the led on/off" program-model, and started checking with the multimeter pin by pin.

Then I noticed that when I had the + on RB3, and the - in ground, the pic started behaving completely normal, such like if a reset would have taken place.

I then recompiled the program initializing properly the ports I wasn't using and all the problems solved!. Now I can use variables as reliably as I was meant to.

While I can compile and make relocatable code work, still I haven't looked at why .udata_shr doesn't work (probably because the original defaul script needs some addendums), and in absolute mode, still I wouldn't know how to setup a variable of 1 byte, since `res' throws that message, but by now I'll be using relocatable code anyways.

So thanks so much for your help and your attention guys, I might be able to start exploring the pic and the loads of things it comes with (timers, adc, etc).

Regards.

> --
> Rob Pearce                      
> http://www.bdt-home.demon.co.uk
>
> The contents of this | Windows NT crashed.
> message are purely   | I am the Blue Screen of Death.
> my opinion. Don't    | No one hears your screams.
> believe a word.      |
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: gnupic-unsubscribe@...
> For additional commands, e-mail:
> gnupic-help@...

__________________________________________________
Correo Yahoo!
Espacio para todos tus mensajes, antivirus y antispam ¡gratis!
Regístrate ya - http://correo.yahoo.es


---------------------------------------------------------------------
To unsubscribe, e-mail: gnupic-unsubscribe@...
For additional commands, e-mail: gnupic-help@...