amforth-devel

[Amforth-devel] Amforth on ATmega644P

[Robert L. <xu...@xu...>]

Hello-

Has anyone successfully gotten amforth working on an ATmega644P?  I
tried to do this today, and immediately ran afoul of the fact that
avra (AVR assembler, on Linux) doesn't support the ATmega644P.  I
patched avra to recognize the '644P (patch below), and was able to
build amforth.  After uploading it to the device, I got an amforth
prompt as expected, but creating a new word (e.g. : hello ." Hello
world" cr ; ) caused the device to apparently hang.  Still in search
of problems, I took a closer look at core/devices/atmega644.asm, and
noticed the number of interrupts listed there do not line up with the
datasheet.  I patched atmega644.asm (patch below), rebuilt, and
uploaded to the device, with identical results.

Can anyone suggest a course of action?  I'm relatively new to Forth,
so while my normal next step would be to delve into the amforth code,
it would be a steep learning curve; I thought I'd try here first.

My development device is a PDIP-40 ATmega644P, installed in an STK500
board with the crystal oscillator selected and a 10MHz crystal
installed.  My fuse bits are at the bottom of this email.

Transcript:
amforth 3.6 ATmega644P
> : hello ." Hello World " cr ;
(hangs)



AVRA Patch:

--- avra-1.2.3.orig/device.c    2007-06-25 14:43:10.000000000 -0700
+++ avra-1.2.3/device.c    2010-01-04 20:54:02.000000000 -0800
@@ -104,6 +104,7 @@
   {   "ATmega88",        4096,     0x100,     1024,    512,
DF_NO_EICALL|DF_NO_EIJMP|DF_NO_ELPM|DF_NO_ESPM},
   {  "ATmega168",        8192,     0x100,     1024,    512,
DF_NO_EICALL|DF_NO_EIJMP|DF_NO_ELPM|DF_NO_ESPM},
   { "ATmega8515",        8192,      0x60,      512,    512,
DF_NO_EICALL|DF_NO_EIJMP|DF_NO_ELPM|DF_NO_ESPM},
+  { "ATmega644P",       32768,     0x100,     4096,   2048,
DF_NO_EICALL|DF_NO_EIJMP|DF_NO_ESPM},
   {NULL, 0, 0, 0, 0}
 };

atmega644.asm Patch:
--- atmega644.asm	2010-01-04 21:06:12.000000000 -0800
+++ atmega644P.asm	2010-01-04 21:06:54.000000000 -0800
@@ -2,7 +2,7 @@
 ; Built using part description XML file version 69
 ; generated automatically
 .nolist
-	.include "m644def.inc"
+	.include "m644Pdef.inc"
 .list

 .equ ramstart =  $100
@@ -43,7 +43,7 @@
 	rol zh
 .endmacro
 .equ intvecsize = 2 ; please verify; flash size: 65536 bytes
-.equ INTVECTORS = 28
+.equ INTVECTORS = 31
 .org $002
 	 rcall isr ; External Interrupt Request 0
 .org $004
@@ -98,7 +98,13 @@
 	 rcall isr ; 2-wire Serial Interface
 .org $036
 	 rcall isr ; Store Program Memory Read
+.org $038
+	 rcall isr ; USART1, Rx Complete
+.org $03A
+	 rcall isr ; USART1 Data register Empty
+.org $03C
+	 rcall isr ; USART1, Tx Complete
 mcustring:
-	.dw 9
-	.db "ATmega644",0
+	.dw 10
+	.db "ATmega644P"
 .set codestart=pc

Fuses:
# set the fuses according to your MCU
LFUSE=0xFE
HFUSE=0xDF
# some MCU have this one, see write-fuses target below
EFUSE=0xFF

Re: [Amforth-devel] Amforth on ATmega644P

[Erich W. <ew....@na...>]

Hello Robert,

> Has anyone successfully gotten amforth working on an ATmega644P?  I
> tried to do this today, and immediately ran afoul of the fact that
> avra (AVR assembler, on Linux) doesn't support the ATmega644P.

about a year ago, I experimented with atmega-644p controllers, too.
This was with amforth 3.3

I did *not* patch avra. And if I remember correctly, the include files
from AVRStudio are of a newer format, and avra didn't like them. So I went
and installed wine and the assembler from avrstudio. Yapp, not the
best solution, but it worked for the moment.

This results in the following call to produce the hex files:

wine ~/wine/AvrAssembler2/avrasm2.exe -I ~/Forth/amforth/releases/3.3/core \
  -I . -fI  -l main.lst -m main.map -e main.eep.hex main.asm

So for the heck of it, I just grabbed a 644P and powered it up.
The hex file loads and the controller does talk with me:
> ver
amforth 3.3 ATmega644P ok
> .res
free FLASH cells      25207
free RAM cells        3795
used EEPROM cells     42
used data stack cells 0
used return stack     10
free return stack     70
 ok
> : hi ." hello, world!" cr ;
 ok
> hi
hello, world!
 ok
>

So far so good. How did I make it work??? Hm.
. wine + avrassembler2 I mentioned above.
. Fuse bits:
  # external crystal at 11.0592 MHz
  LFUSE=0xee
  HFUSE=0x89
. there are 31 interrupts now, you found that.

It is certainly possible to dig out the differences to the old
version 3.3, however, it would take some time.

I could obviously send you my files, if you are interested
in "code archeaology" :-)  I have not used the 644p any more yet.

Cheers,
Erich

Re: [Amforth-devel] Amforth on ATmega644P

[Robert L. <xu...@xu...>]

Erich-

Thank you for your response.  My patch to AVRA was somewhat slapdash and as
such, I tend to suspect that AVRA is doing something wrong.  I'll try
assembling with AVRStudio's assembler, see if that gets me anywhere.  It's
also possible (even likely) there was a change between 3.3 and 3.6 that is
causing this.  Would you mind sending me the template.asm and so forth you
used to build 3.3?  It would be educational, if nothing else.

Thanks again.


On Tue, Jan 5, 2010 at 9:03 AM, Erich Waelde <ew....@na...> wrote:

> Hello Robert,
>
> > Has anyone successfully gotten amforth working on an ATmega644P?  I
> > tried to do this today, and immediately ran afoul of the fact that
> > avra (AVR assembler, on Linux) doesn't support the ATmega644P.
>
> about a year ago, I experimented with atmega-644p controllers, too.
> This was with amforth 3.3
>
> I did *not* patch avra. And if I remember correctly, the include files
> from AVRStudio are of a newer format, and avra didn't like them. So I went
> and installed wine and the assembler from avrstudio. Yapp, not the
> best solution, but it worked for the moment.
>
> This results in the following call to produce the hex files:
>
> wine ~/wine/AvrAssembler2/avrasm2.exe -I ~/Forth/amforth/releases/3.3/core
> \
>  -I . -fI  -l main.lst -m main.map -e main.eep.hex main.asm
>
> So for the heck of it, I just grabbed a 644P and powered it up.
> The hex file loads and the controller does talk with me:
> > ver
> amforth 3.3 ATmega644P ok
> > .res
> free FLASH cells      25207
> free RAM cells        3795
> used EEPROM cells     42
> used data stack cells 0
> used return stack     10
> free return stack     70
>  ok
> > : hi ." hello, world!" cr ;
>  ok
> > hi
> hello, world!
>  ok
> >
>
> So far so good. How did I make it work??? Hm.
> . wine + avrassembler2 I mentioned above.
> . Fuse bits:
>  # external crystal at 11.0592 MHz
>  LFUSE=0xee
>  HFUSE=0x89
> . there are 31 interrupts now, you found that.
>
> It is certainly possible to dig out the differences to the old
> version 3.3, however, it would take some time.
>
> I could obviously send you my files, if you are interested
> in "code archeaology" :-)  I have not used the 644p any more yet.
>
> Cheers,
> Erich
>
>
> ------------------------------------------------------------------------------
> This SF.Net email is sponsored by the Verizon Developer Community
> Take advantage of Verizon's best-in-class app development support
> A streamlined, 14 day to market process makes app distribution fast and
> easy
> Join now and get one step closer to millions of Verizon customers
> http://p.sf.net/sfu/verizon-dev2dev
> _______________________________________________
> Amforth-devel mailing list
> Amf...@li...
> https://lists.sourceforge.net/lists/listinfo/amforth-devel
>