Work in progress to make simple graphical display for EV's.

EV-display part done by Henry Palonen

Most of the work done by Martin Thomas, original README below

----------------------------------------------
Port of the Atmel AVR DB101 (ATAVRDB101)
example-application code to the "GNU toolchain"

done by Martin Thomas, Kaiserslautern, Germany
http://www.siwawi.arubi.uni-kl.de/avr_projects
<eversmith(at)heizung-thomas(dot)de> or
<mthomas(at)rhrk(dot)uni-kl(dot)de>

GNU/avr-gcc Port Version 20080301

based on the Atmel's code for IAR EWAVR/ICCAVR
Atmel cvs tag: release_A_code_20070917
(download available from atmel.com)

----------------------------------------------



Used tools for this version of the port:
----------------------------------------

GNU toolchain: Binutils 2.17, avr-gcc 4.1.2
and avr-libc 1.4.6 as in WinAVR 20070525

and

AVR-Studio 4.13 SP2

Unix/Linux/OS X users can build with the 
makefile available in the directory 
production_demo_rev_A/make_standalone


General information:
--------------------

I'm not working for Atmel so this gcc-port is
not supported by Atmel. I will try to support
it as good as I can (send an e-mail).

I tried to keep the source compatible for both 
compilers by adding a "Compiler Abstraction Layer"
(CAL). This is a set of preprocessor defines, 
macros and inline-functions in the file cal.h. 

I have never used the IAR compiler for AVR so I 
don't know exactly if the IAR-part of the CAL
implements the "best methods". But I tried to
create the IAR-part of the CAL so the preprocessed
source of the port should be very similar to the 
original one.

The doxygen documentation has been removed from
this port-package to keep the distribution archive
small. You can get the documentation with the 
original source available from atmel.com:
http://www.atmel.com/dyn/products/tools_card.asp?tool_id=4221
I did not change any interfaces so the documentation
is valid for both toolsets.

I marked a few but not all of my modifications 
with ///mt. Diff with the original source to see 
everything that has been changed.


Enjoy the DB101 gcc-port
- Martin Thomas


TODOs:
------

- attribute "used" for dummy reads or other method
  to avoid some warnings (they are currently just 
  disabled)
- check for more "missing" volatiles
- add German umlauts in TERMFONT_fontdata


Version History / Diary:
------------------------

1. March 2008

Just some modification to increase compatiblity
with IAR-EWARM.

- Added prod_demo.ewp with MISRA-checking disable.
  File provided by Colin O'Flynn - Thanks.
  (renamed original to prod_demo.ewp_org)
- timing_lib.c: changed order of && operation
  to avoid a MISRA-warning (Rule 33), tested by 
  Colin O Flynn. Changed marked with 20080301
- The IAR MISRA checker will still complain on code
  in forms_lib.c. I had to use some dirty tricks
  which include casting of function-pointers
  to get the "traits" working with avr-gcc.

25. Feb. 2008

- Modifications for IAR-EWAVR compatiblity
  from Kevin Rosenberg. Thanks.
- IAR build now tested by Kevin Rosenberg and 
  Colin O'Flynn with MISRA-checking disabled.
- timinglib.c: "Rework while() statement to 
  explictly set order of volatile variable access"
  by Kevin Rosenberg
- Replaced absolute with relative pathes 
  in aps-file (again, it seems AVR-Studio
  sometimes overwrites with absolute pathes)
- added  -Wno-unused back to gcc-plugin options
- GNU build tested with AVR-Studio 4.13SP2
  and (still) WinAVR 5/2007
- commonn/cal.h: new macros CAL_MAIN and CAL_MAIN_LAST
  to avaid warning cause by gcc function-attribute 
  noreturn and to keep support for IAR __task
- prod_demo/main.c: use CAL_MAIN and CAL_MAIN_LAST
- Information on current drawn by the DB101 from
  Colin O'Flynn: "[...] In sleep it's around 1mA. 
  Normal operation is like 7mA, backlight is around 
  70mA. But it seems to depend on which DB101 I 
  use[...]"


16. Oct. 2007

- widgets_libc.c/WIDGETS_SelectCtrl_InitCheck():
  added initialisation for OnCheck. This fixes 
  an issue when selecting tunes in the "Music Demo"
  or when using the select widget in own applications.
  (The issue is also present in the preinstalled
  firmware and not caused by the GNU tools.)
- db101_demo.aps: "regenerated" - again
- song_data.h: "Fur Elise" -> "Fuer Elise"
  (It's "Für" but umlauts currently not present
  in the font)
- makefile_standalone/Makefile: added compiler
  options -ffunction-sections -fdata-sections
  and linker-option --gc-sections. This removes
  unused code from the binary and reduces the size
  around 14% (currently 67578->58420). But the 
  disassambly looks strange. You may want to 
  disable these options for debugging. They are 
  not enabled in in the AVR Studio worksspace 
  settings, just in the "standalone" makefile.
  
  Note: size of the "raw binary" memory-dump when 
  reading the flash memory back with avrdude:
    preinstalled application probaly created
      with the IAR tools: 59606 bytes 
    application created with the ported sources
      and avr-gcc 4.2.1 with unused code removal options
      and -os: 58420 bytes --> ca. 2% smaller


13. Oct. 2007

cal.h: modified license as proposed by Eric Weddington
updated AVR-Studio workspace
just replaced package on web-server (still "1st release")


12. Oct. 2007

no longer TODO:
- test with real hardware when available

This says it all. The DB101 is here and I have spent
a few hours debugging the gnu-port (after a lot
more hours fighting with commercial ARM and
AVR projects this has been fun). Thanks again
to Atmel for sending the board (Upd. 2/2008:
While the board is a gift from Atmel I have 
received an invoice from the German costums
in the meantime (around 40EUR).)

The port should be more or less done. At least I 
could not find a difference between the preinstalled 
application and my gcc-port but I have not 
tested that much. The only unneeded modification 
is a different "Introduction" message which informs 
that a gcc-port of the Atmel code is running.

The implementation of the traits ("message 
handlers") has been a little difficult to 
transfer to avr-libc "progmem access-style". 
The modifications for this do not look that nice 
but hey, it works. JTAGICE MK2 has not been 
that useful, my old workbench-PC seems to be too
slow for such a rather large application. But
the AVR-Studio simulator helped a few times.

It plays "Für Elise" and this is of cause most 
important...

==> fist public release, timestamp 20071012

Feedback welcome - Have fun


10. Oct. 2007

Used _delay_cycles from delay_x_gcc.h for the GCC-part
of the CAL. So less modifications in the original source 
needed. delay_x_gcc.h is the copy of the file delay_x.h 
which has been found to the avrfreaks gcc forum. It has
been created by Hans-Juergen Heinrichs. Thanks for
providing it.


9. Oct. 2007

removed redundant CAL_pgm (has been alias for CAL_PGM)


8. Oct. 2007

minor modifications - still untested with hardware
no longer TODOs:
- get rid of the remaining warnings 
  (maybe just disable or add attribute used)
  -> fixed origin of all warnings but the
     warning about "unused variables" (dummy reads)
     -> just disabled by -wno-unused for now
- create standalone makefile
  -> WinARM 5/07 template not realy usable (sources 
     from different directories)
     -> using AVR-Studio generated makefile as template 
        and modified a little bit see:
        production_demo_rev_A/make_standalone/Makefile


7. Oct. 20007

received e-mail from Atmel (Thanks Jacob, Thanks Eric 
for the arranging the contact): I will get a DB101 to 
test the gcc-port. Thanks to Atmel for providing it. 
It seems the Atmel's own gcc-port is not "high priority". 
So maybe my porting-work will not be just useful for me. 
Additonaly I can test the full port with the "real" 
hardware once the DB101 is here and not just parts of 
it with my own ATmega644/EA DOGM128 prototype.


6. Oct. 2007 

Initial porting work. Most of the dirty work
should be done, no DB101 board here for tests.
Why am I doing this port?
Eric Weddington mentioned in the avrfreaks forum
that Atmel plans to do an own gcc-port. But in a
current project some parts of the Atmel code 
(esp. the LCD-handling and "Widgets") might be
very useful and hopefully save me some time and work.
So I tried to create an initial port of the
DB101-code which will be modified/reduced to match 
the hardware I have here (hopefully the 
ATmega644P can handle it...). I will try to backport 
modifications from my project's code into the 
DB101 code-port if possible.






Scratchpad - ignore everything below

# mega1281, 7,3MHz, -wno-unused, short-enums *not* enabled

# List C source files here. (C dependencies are automatically generated.)
# (a) Demo sources:
SRC = main.c
SRC += walkabout.c configuart.c flashpics.c gameoflife.c lcdcontrast.c
SRC += memory.c slideshow.c smokeydemo.c snake.c sounddemo.c terminal.c clock.c
# (b) Driver sources:
SRC += ../gfx/s6b1713_driver.c ../gfx/lcd_lib.c ../gfx/popup_lib.c ../gfx_lib.c
SRC += ../joystick_driver/joystick_driver.c
SRC += ../power_driver/power_driver.c
SRC += ../backlight_driver/backlight_driver.c
SRC += ../fifo_lib/fifo_lib.c
SRC += ../memblock_lib/memblock_lib.c
SRC += ../Picture_lib/picture_lib.c
SRC += ../forms_lib/widgets_lib.c
SRC += ../forms_lib/forms_lib.c
SRC += ../forms_lib/dialog_lib.c
SRC += ../uart_driver/uart_driver.c
SRC += ../rtc_driver/rtc_driver.c
SRC += ../timing_lib/timing_lib.c
SRC += ../termfont_lib/termfont_lib.c
SRC += ../Sound/sound_driver.c ../Sound/song_lib.c
SRC += ../terminal_lib/terminal_lib.c

EXTRAINCDIRS  = ../common ../Picture_lib ../power_driver ../rtc_driver ../Sound
EXTRAINCDIRS += ../termfont_lib ../terminal_lib ../timing_lib ../uart_driver
EXTRAINCDIRS += ../backlight_driver ../fifo_lib ../forms_lib ../gfx ../img 
EXTRAINCDIRS += ../joystick_driver ../memblock_lib