|
|
25 |
|
|
|
26 |
|
|
|
27 |
|
|
|
28 |
|
|
|
29 |
|
|
|
30 |
|
|
|
31 |
|
|
|
32 |
|
|
|
33 |
|
|
|
|
34 |
|
|
|
35 |
|
|
|
36 |
README
|
|
|
37 |
|
|
|
38 |
|
|
|
39 |
|
|
|
40 |
|
|
|
41 |
AVRA v1.3.0 - Assember for the Atmel AVR microcontroller family
|
|
|
42 |
|
|
|
43 |
This program is free software; you can redistribute it and/or modify
|
|
|
44 |
it under the terms of the GNU General Public License as published by |
|
|
45 |
the Free Software Foundation; either version 2 of the License, or |
|
|
46 |
(at your option) any later version. Please read below for for information. |
|
|
47 |
Disclaimer
|
|
|
48 |
This program is distributed in the hope that it will be useful, but
|
|
|
49 |
WITHOUT ANY WARRANTY; without even the implied warranty of |
|
|
50 |
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
|
51 |
GNU General Public License for more details. |
|
|
52 |
GNU General Public License
|
|
|
53 |
You should have received a copy of the GNU General Public License
|
|
|
54 |
along with this program; see the file "COPYING". If not, visit |
|
|
55 |
http://www.gnu.org or write to the Free Software Foundation, Inc., |
|
|
56 |
59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. You can |
|
|
57 |
also contact the authors of AVRA to receive a copy of the COPYING file. |
|
|
58 |
Trademarks and copyright
|
|
|
59 |
Atmel, AVR, AVR Studio, Intel, Windows are registered enterprises, brands
|
|
|
60 |
and registered trademarks. The mentioned companies have no relation to |
|
|
61 |
AVRA and are therefore not responslible for any problems that occur when |
|
|
62 |
using AVRA. Many thanks for your products, support and efforts. |
|
|
63 |
|
|
|
64 |
Introducion
|
|
|
65 |
|
|
|
66 |
AVRA is an assembler for Atmel AVR microcontrollers, and it is almost
|
|
|
67 |
compatible with Atmel’s own assembler AVRASM32. The programming |
|
|
68 |
principles and conceptions are based on the ANSI programming language "C". |
|
|
69 |
The initial version of AVRA was written by John Anders Haugum. He released
|
|
|
70 |
all versions until v0.7. All later versions were released by Tobias Weber. |
|
|
71 |
Differences between AVRA and AVRASM32
|
|
|
72 |
There are some differences between the original Atmel assembler AVRASM32 and AVRA. Basically AVRA is designed to replace AVRASM32 without special changes in your current Atmel AVR Studio enviroment.
|
|
|
73 |
Command line options have been adapted as far as it was possible until now. Jumping to fault containing line directly by double-clicking on the error message in the output window does work as with AVRASM32. |
|
|
74 |
The differences in detail
|
|
|
75 |
|
|
|
76 |
|
|
|
77 |
|
|
|
78 |
Support for some extra preprocessor directives. |
|
|
79 |
|
|
|
80 |
|
|
|
81 |
|
|
|
82 |
.define, .undef, .ifdef, .ifndef, .if, .else, .endif, .elif, .elseif, .warning
|
|
|
83 |
|
|
|
84 |
|
|
|
85 |
|
|
|
86 |
|
|
|
87 |
Not all command line options are supported. |
|
|
88 |
Specifying an eeprom file (-e) is not supported. All eeprom data is |
|
|
89 |
put out into a file called program.eep.hex and always Intel hex |
|
|
90 |
format. Other hex file formats than Intel are currently not supported. |
|
|
91 |
|
|
|
92 |
|
|
|
93 |
|
|
|
94 |
|
|
|
95 |
Forward references not supported for .ifdef and .ifndef directives. |
|
|
96 |
This makes sure, that directives like .ifdef and .undef are working |
|
|
97 |
properly. If you are familiar with the C programming language, you |
|
|
98 |
should get easily into AVRA. See chapter "Programming techniques" for |
|
|
99 |
more information about how to write proper code. |
|
|
100 |
|
|
|
101 |
|
|
|
102 |
|
|
|
103 |
|
|
|
104 |
Enhanced macro support |
|
|
105 |
AVRA has some new features for writing flexible macros. This should |
|
|
106 |
increase the ability to reuse code e.g. build your own library. |
|
|
107 |
|
|
|
108 |
|
|
|
109 |
|
|
|
110 |
|
|
|
111 |
Debugging support |
|
|
112 |
AVRA creates a coff file everytime the assembly was sucessful. This |
|
|
113 |
file allows AVR Studio or any coff compatible debugger to simulate |
|
|
114 |
or emulate the program. |
|
|
115 |
|
|
|
116 |
|
|
|
117 |
|
|
|
118 |
|
|
|
119 |
Meta tags for assembly time |
|
|
120 |
This helps you tracking versions of your software and can also be |
|
|
121 |
used to generate customer specific serial numbers. |
|
|
122 |
|
|
|
123 |
|
|
|
124 |
|
|
|
125 |
Compatibility
|
|
|
126 |
Since AVRA is written in ANSI C, it should be possible to compile it on
|
|
|
127 |
most system platforms. If you have problems compiling AVRA, please leave |
|
|
128 |
a message on the sourceforge message board or send a mail to the |
|
|
129 |
authors of AVRA. |
|
|
130 |
|
|
|
131 |
Installation
|
|
|
132 |
|
|
|
133 |
To install avra you should copy the avra-executable to an apropriate
|
|
|
134 |
location. To compile you should rename the appropriate makefile, and |
|
|
135 |
perform a make (use smake for Amiga SAS/C, and nmake for Mickeysoft |
|
|
136 |
visual c++). |
|
|
137 |
Linux
|
|
|
138 |
To compile avra you need gcc and the automake utilities. These will create
|
|
|
139 |
a ./configure script that evaluates your system enviroment. To get the |
|
|
140 |
AVRA executable, you have to issue the following commands: |
|
|
141 |
|
|
|
142 |
autoconf |
|
|
143 |
automake -a |
|
|
144 |
./configure |
|
|
145 |
make && make install |
|
|
146 |
AmigaOS
|
|
|
147 |
avra can be copied any apropriate directory. If you are using the source
|
|
|
148 |
distribution a make install will do the same. |
|
|
149 |
Microsoft Windows
|
|
|
150 |
If you received the Windows binary package, look into the \bin
|
|
|
151 |
directory where you can find avra.exe. This should be copied to any |
|
|
152 |
apropriate location. You can also overwrite AVRASM32.EXE in your |
|
|
153 |
Atmel AVR Studio. If you want to compile it yourself you could download then |
|
|
154 |
OpenWatcom C/C++ Toolchain for windows and create a new project and add the C |
|
|
155 |
and H files to it and compile. |
|
|
156 |
Apple OS X
|
|
|
157 |
If you recieved the Apple OS X binary package, look into the bin directory this
|
|
|
158 |
file is compiled universal and should run on intel 32 and 64 bit and powerpc. If |
|
|
159 |
you want to compile it yourself go to the src directory and invoke make -f |
|
|
160 |
makefiles/Makefiles.osx and then the executable should be created. |
|
|
161 |
|
|
|
162 |
Synopsis
|
|
|
163 |
|
|
|
164 |
Command line usage
|
|
|
165 |
|
|
|
166 |
|
|
|
167 |
usage: AVRA [-f][O|M|I|G] output file type
|
|
|
168 |
[-o <filename>] output file name |
|
|
169 |
[-l <filename>] generate list file |
|
|
170 |
[-m <mapfile>] generate map file |
|
|
171 |
[--define <symbol>[=<value>]] [--includedir <dir>] [--listmac] |
|
|
172 |
[--max_errors <number>] [--devices] [--version] |
|
|
173 |
[-h] [--help] general help |
|
|
174 |
[-W NoRegDef] supress register redefinition warnings |
|
|
175 |
<file to assemble> |
|
|
176 |
|
|
|
177 |
Parameter list
|
|
|
178 |
|
|
|
179 |
|
|
|
180 |
--listfile -l : Create list file
|
|
|
181 |
--mapfile -m : Create map file |
|
|
182 |
--define -D : Define symbol. |
|
|
183 |
--includedir -I : Additional include dirs. |
|
|
184 |
--listmac : List macro expansion in listfile. |
|
|
185 |
--max_errors : Maximum number of errors before exit |
|
|
186 |
(default: 10) |
|
|
187 |
--devices : List out supported devices. |
|
|
188 |
--version : Version information. |
|
|
189 |
--help, -h : This help text. |
|
|
190 |
|
|
|
191 |
Warning supression
|
|
|
192 |
Since avra 1.1 there is a possibility to supress certain warnings.
|
|
|
193 |
Currently only register reassignment warnings can be supressed. |
|
|
194 |
Example: avra -W NoRegDef
|
|
|
195 |
|
|
|
196 |
Programming techniques
|
|
|
197 |
|
|
|
198 |
Using directives
|
|
|
199 |
AVRA offers a number of directives that are not part of Atmel’s
|
|
|
200 |
assembler. These directives should help you creating versatile code that |
|
|
201 |
can be designed more modular. |
|
|
202 |
Directive .define
|
|
|
203 |
To define a constant, use ".define". This does the same thing as ".equ",
|
|
|
204 |
it is just a little more C style. Keep in mind that AVRA is not case |
|
|
205 |
sensitive. Do not mix ".def" and ".define", because ".def" is used to |
|
|
206 |
assign registers only. This is due to backward compatibility to Atmel’s |
|
|
207 |
AVRASM32. Here is an example on how .define can be used. |
|
|
208 |
|
|
|
209 |
|
|
|
210 |
.define network 1
|
|
|
211 |
|
|
|
212 |
Now "network" is set to the value 1. You may want to assemble a specific
|
|
|
213 |
part of your code depeding on a define or switch setting. You can test |
|
|
214 |
your defined word on existence (.ifdef and .ifndef) as well as on the |
|
|
215 |
value it represents. The following code shows a way to prevent error |
|
|
216 |
messages due to testing undefined constants. Conditional directives must |
|
|
217 |
always end with an .endif directive. |
|
|
218 |
|
|
|
219 |
|
|
|
220 |
.ifndef network
|
|
|
221 |
.define network 0 |
|
|
222 |
.endif |
|
|
223 |
|
|
|
224 |
Directive .if and .else
|
|
|
225 |
The three lines in the last example set the default value of "network".
|
|
|
226 |
In the next example, you see how we can use default values. If a constant |
|
|
227 |
has not defined previously, it is set to zero. Now you can test wether |
|
|
228 |
e.g. network support is included into the assemby process. |
|
|
229 |
|
|
|
230 |
|
|
|
231 |
.if network = 1
|
|
|
232 |
.include "include\tcpip.asm" |
|
|
233 |
.else |
|
|
234 |
.include "include\dummynet.asm" |
|
|
235 |
.endif |
|
|
236 |
|
|
|
237 |
In the second part of the above listing you see the use of .else, which
|
|
|
238 |
defines the part of the condition that is being executed if the equation |
|
|
239 |
of the preceding .if statement is not equal. You can also use the else |
|
|
240 |
statement to test another equasion. For that purpose use .elif, which |
|
|
241 |
means "else if". Always close this conditional part with ".endif" |
|
|
242 |
Directive .error
|
|
|
243 |
This directive can be used to throw errors if a part in the code has reached
|
|
|
244 |
that should not be reached. The following example shows how we can stop |
|
|
245 |
the assembly process if a particular value has not been previously set. |
|
|
246 |
|
|
|
247 |
|
|
|
248 |
.ifndef network
|
|
|
249 |
.error "network is not configured!" ;the assembler stops here |
|
|
250 |
|
|
|
251 |
Directive .nolist and .list
|
|
|
252 |
The ouput to the list file can be paused by this two directives. After
|
|
|
253 |
avra discovers a .nolist while assembling, it stops output to the list file. |
|
|
254 |
After a .list directive is detected, it continues the normal list file output. |
|
|
255 |
Directive .includepath
|
|
|
256 |
By default, any file that is included from within the source file must
|
|
|
257 |
either be a single filename or a complete absolute path. With the directive |
|
|
258 |
.includepath you can set an additional include path . Furthermore you can |
|
|
259 |
set as many include paths as you want. Be sure not no use same filename |
|
|
260 |
in separate includes, because then it is no longer clear which one avra |
|
|
261 |
should take. |
|
|
262 |
|
|
|
263 |
Using include files
|
|
|
264 |
|
|
|
265 |
To avoid multiple inclusions of include files, you may use some pre-
|
|
|
266 |
processor directives. See example file stack.asm that is being included |
|
|
267 |
into the main programm file as well as in other include files. |
|
|
268 |
|
|
|
269 |
|
|
|
270 |
.ifndef _STACK_ASM_
|
|
|
271 |
.define _STACK_ASM_ |
|
|
272 |
|
|
|
273 |
|
|
|
274 |
|
|
|
275 |
.include "include/config.inc"
|
|
|
276 |
|
|
|
277 |
|
|
|
278 |
|
|
|
279 |
; *** stack macro ***
|
|
|
280 |
|
|
|
281 |
|
|
|
282 |
|
|
|
283 |
.dseg
|
|
|
284 |
m_stack: .byte __stack_size__ |
|
|
285 |
.cseg |
|
|
286 |
|
|
|
287 |
|
|
|
288 |
|
|
|
289 |
.macro stack_setup
|
|
|
290 |
load [v:w,m_stack + __stack_size__] |
|
|
291 |
outp [SPREG,v:w] |
|
|
292 |
.endm |
|
|
293 |
|
|
|
294 |
|
|
|
295 |
|
|
|
296 |
.endif ; avoid multiple inclusion of stack.asm
|
|
|
297 |
|
|
|
298 |
|
|
|
299 |
If you like to implement compiler build time and date into your
|
|
|
300 |
program, you can make use of some sepcial tags that avra supports. |
|
|
301 |
|
|
|
302 |
|
|
|
303 |
%MINUTE% is being replaced by the current minute (00-59)
|
|
|
304 |
%HOUR% is being replaced by the current hour (00-23) |
|
|
305 |
%DAY% is being replaced by the current day of month (01-31) |
|
|
306 |
%MONTH% is being replaced by the current month (01-12) |
|
|
307 |
%YEAR% is being replaced by the current year (2004-9999) |
|
|
308 |
|
|
|
309 |
|
|
|
310 |
|
|
|
311 |
buildtime: .db "Release date %DAY%.%MONTH%.%YEAR% %HOUR%:%MINUTE%"
|
|
|
312 |
|
|
|
313 |
This line will then assembled by avra into:
|
|
|
314 |
|
|
|
315 |
|
|
|
316 |
buildtime: .db "Release date 10.05.2004 19:54"
|
|
|
317 |
|
|
|
318 |
You may also create a self defined serial number with meta tags:
|
|
|
319 |
|
|
|
320 |
|
|
|
321 |
.define serialnumber %DAY% + %MONTH%*31 + (%YEAR% - 2000) *31*12
|
|
|
322 |
|
|
|
323 |
The %TAG% is translated before any other parsing happens. The real
|
|
|
324 |
output can be found in the list file. |
|
|
325 |
|
|
|
326 |
Macro features
|
|
|
327 |
|
|
|
328 |
Sometimes you have to work with 16 bit or greater variables stored
|
|
|
329 |
in 8 bit registers. The enhanced macro support allows you to write short |
|
|
330 |
and flexible macros that simplify access to big variables. The extended |
|
|
331 |
mode is active, as soon as you use parenthesis like this "[ ]" to wrap |
|
|
332 |
macro parameters. |
|
|
333 |
Auto type conversion for macros
|
|
|
334 |
Values representing more than 8 Bits are usualy kept in a set of byte
|
|
|
335 |
wide registers. To simplify 16 Bit or greater operations, I added a new |
|
|
336 |
language definitions. Words can be written as r16:r17, whereas register |
|
|
337 |
r16 contains the higher part and register r17 the lower part of this |
|
|
338 |
16 Bit value. |
|
|
339 |
Macro data types
|
|
|
340 |
There are 3 data types that can be used. They will be added as character
|
|
|
341 |
separated by one underline character. |
|
|
342 |
|
|
|
343 |
|
|
|
344 |
immediate values _i
|
|
|
345 |
registers _8,_16,_24,_32,_40,_48,_56,_64 |
|
|
346 |
void parameter _v |
|
|
347 |
|
|
|
348 |
16 Bit Source and Destionation registers dst and src
|
|
|
349 |
|
|
|
350 |
|
|
|
351 |
src = YH:YL
|
|
|
352 |
dst = ZH:ZL |
|
|
353 |
|
|
|
354 |
Within the parenthesis, the two words src and dst are interpreted as YH:YL
|
|
|
355 |
and ZH:ZL. Normal code outside of the macro parameter parenthesis can |
|
|
356 |
still make use of these special key words "src" and "dst". |
|
|
357 |
Examples for automatic type conversion
|
|
|
358 |
To simplify the parameters in the demonstration below, we need to
|
|
|
359 |
redefine some registers. |
|
|
360 |
|
|
|
361 |
|
|
|
362 |
.def a = r16 ; general purpose registers
|
|
|
363 |
.def b = r17 |
|
|
364 |
.def c = r18 |
|
|
365 |
.def d = r19 |
|
|
366 |
|
|
|
367 |
|
|
|
368 |
|
|
|
369 |
.def w = r20 ; working register
|
|
|
370 |
.def v = r21 ; working register |
|
|
371 |
|
|
|
372 |
If we substract 16 Bit values stored in a, higher byte and b, lower byte
|
|
|
373 |
with that in c:d, we usually have to use the following command sequence: |
|
|
374 |
|
|
|
375 |
|
|
|
376 |
sub b,d
|
|
|
377 |
sbc a,c |
|
|
378 |
|
|
|
379 |
Now we can do the following steps to simplify 16 or more Bit manipulations
|
|
|
380 |
|
|
|
381 |
|
|
|
382 |
.macro subs
|
|
|
383 |
.message "no parameters specified" |
|
|
384 |
.endm |
|
|
385 |
|
|
|
386 |
|
|
|
387 |
|
|
|
388 |
.macro subs_16_16
|
|
|
389 |
sub @1,@3 |
|
|
390 |
sbc @0,@2 |
|
|
391 |
.endm |
|
|
392 |
|
|
|
393 |
|
|
|
394 |
|
|
|
395 |
.macro subs_16_8
|
|
|
396 |
sub @1,@2 |
|
|
397 |
sbci @0,0 |
|
|
398 |
.endm |
|
|
399 |
|
|
|
400 |
|
|
|
401 |
|
|
|
402 |
;now we can write a 16 Bit subraction as:
|
|
|
403 |
|
|
|
404 |
|
|
|
405 |
|
|
|
406 |
subs [a:b,c:d]
|
|
|
407 |
|
|
|
408 |
|
|
|
409 |
|
|
|
410 |
;or for calculating 16 minus 8 Bit
|
|
|
411 |
|
|
|
412 |
|
|
|
413 |
|
|
|
414 |
subs [a:b,c]
|
|
|
415 |
|
|
|
416 |
Overloading macros
|
|
|
417 |
Like in you are used to C functions, you can write macros for different
|
|
|
418 |
parameter lists. If you would like to have a versatile macro, you can |
|
|
419 |
specify a unique macro for each parameter situation. See the next sample. |
|
|
420 |
|
|
|
421 |
|
|
|
422 |
.macro load
|
|
|
423 |
|
|
|
424 |
|
|
|
425 |
|
|
|
426 |
; this message is shown if you use the macro within your code
|
|
|
427 |
; specifying no parameters. If your macro allows the case where |
|
|
428 |
; no parameters are given, exchange .message with your code. |
|
|
429 |
|
|
|
430 |
|
|
|
431 |
|
|
|
432 |
.message "no parameters specified"
|
|
|
433 |
.endm |
|
|
434 |
|
|
|
435 |
|
|
|
436 |
|
|
|
437 |
; Here we define the macro "load" for the case it is being used
|
|
|
438 |
; with two registers as first parameter and a immediate (constant) |
|
|
439 |
; value as second parameter. |
|
|
440 |
|
|
|
441 |
|
|
|
442 |
|
|
|
443 |
.macro load_16_i
|
|
|
444 |
ldi @0,high(@2) |
|
|
445 |
ldi @1,low(@2) |
|
|
446 |
.endm |
|
|
447 |
|
|
|
448 |
|
|
|
449 |
|
|
|
450 |
; the same case, but now with a 32 bit register value as first
|
|
|
451 |
; parameter |
|
|
452 |
|
|
|
453 |
|
|
|
454 |
|
|
|
455 |
.macro load_32_i
|
|
|
456 |
ldi @0,BYTE4(@4) |
|
|
457 |
ldi @1,BYTE3(@4) |
|
|
458 |
ldi @2,high(@4) |
|
|
459 |
ldi @3,low(@4) |
|
|
460 |
.endm |
|
|
461 |
|
|
|
462 |
|
|
|
463 |
|
|
|
464 |
; Now let's see how these macros are being used in the code
|
|
|
465 |
|
|
|
466 |
|
|
|
467 |
|
|
|
468 |
load [a:b,15] ;uses macro load_16_i to load immediate
|
|
|
469 |
|
|
|
470 |
|
|
|
471 |
|
|
|
472 |
load [a:b:c:d,15] ;uses macro load_32_i to load immediate
|
|
|
473 |
|
|
|
474 |
More examples
|
|
|
475 |
|
|
|
476 |
|
|
|
477 |
.dseg
|
|
|
478 |
counter .byte 2 |
|
|
479 |
.cseg |
|
|
480 |
|
|
|
481 |
|
|
|
482 |
|
|
|
483 |
.macro poke
|
|
|
484 |
.message "no parameters" |
|
|
485 |
.endm |
|
|
486 |
|
|
|
487 |
|
|
|
488 |
|
|
|
489 |
.macro poke_i_16_i
|
|
|
490 |
ldi @1,high(@3) |
|
|
491 |
sts @0+0,@1 |
|
|
492 |
ldi @2,low(@3) |
|
|
493 |
sts @0+1,@2 |
|
|
494 |
.endm |
|
|
495 |
|
|
|
496 |
|
|
|
497 |
|
|
|
498 |
.macro poke_i_i
|
|
|
499 |
ldi w,@1 |
|
|
500 |
sts @0+0,w |
|
|
501 |
.endm |
|
|
502 |
|
|
|
503 |
|
|
|
504 |
|
|
|
505 |
.macro poke_i_v_i
|
|
|
506 |
ldi w,high(@3) |
|
|
507 |
sts @0+0,w |
|
|
508 |
ldi w,low(@3) |
|
|
509 |
sts @0+1,w |
|
|
510 |
.endm |
|
|
511 |
|
|
|
512 |
|
|
|
513 |
|
|
|
514 |
.macro poke_i_v_v_v_i
|
|
|
515 |
ldi w,high(@3) |
|
|
516 |
sts @0+0,w |
|
|
517 |
ldi w,low(@3) |
|
|
518 |
sts @0+1,w |
|
|
519 |
ldi w,BYTE3(@3) |
|
|
520 |
sts @0+2,w |
|
|
521 |
ldi w,BYTE4(@3) |
|
|
522 |
sts @0+3,w |
|
|
523 |
.endm |
|
|
524 |
|
|
|
525 |
|
|
|
526 |
|
|
|
527 |
; this writes '9999' into the memory at 'counter'
|
|
|
528 |
; uses only the working register for transfering the values. |
|
|
529 |
|
|
|
530 |
|
|
|
531 |
|
|
|
532 |
poke [counter,w:w,9999]
|
|
|
533 |
|
|
|
534 |
|
|
|
535 |
|
|
|
536 |
; works same as above, but the transferred value '9999' is also
|
|
|
537 |
; kept in the pair of register a:b |
|
|
538 |
|
|
|
539 |
|
|
|
540 |
|
|
|
541 |
poke [counter,a:b,9999]
|
|
|
542 |
|
|
|
543 |
|
|
|
544 |
|
|
|
545 |
; in my design 'w' is always working reg. which implies that
|
|
|
546 |
; it cannot be used for normal variables. The following example |
|
|
547 |
; uses poke_i_i because the parameter contains two immediate values. |
|
|
548 |
|
|
|
549 |
|
|
|
550 |
|
|
|
551 |
poke [counter,9999] ;uses poke_i_i
|
|
|
552 |
|
|
|
553 |
|
|
|
554 |
|
|
|
555 |
; to be able to choose between a 8,16 or 32 Bit operation, you just
|
|
|
556 |
; add a void parameter. |
|
|
557 |
|
|
|
558 |
|
|
|
559 |
|
|
|
560 |
poke [counter,,9999] ;uses poke_i_v_i
|
|
|
561 |
|
|
|
562 |
|
|
|
563 |
|
|
|
564 |
; and the same for 32 Bit pokes
|
|
|
565 |
|
|
|
566 |
|
|
|
567 |
|
|
|
568 |
poke [counter,,,,9999] ;uses poke_i_v_v_v_i
|
|
|
569 |
|
|
|
570 |
Loops within macros
|
|
|
571 |
One problem you may have experienced, is that labels defined within macros
|
|
|
572 |
are defined twice if you call the macro for example two times. Now you can |
|
|
573 |
use labels for macro loops. Loops within macros must end with _%. the |
|
|
574 |
"%" symbol is replaced by a running number. |
|
|
575 |
Loop example
|
|
|
576 |
|
|
|
577 |
|
|
|
578 |
; Definition of the macro
|
|
|
579 |
|
|
|
580 |
|
|
|
581 |
|
|
|
582 |
.macro write_8_8
|
|
|
583 |
write_%: |
|
|
584 |
st Z+,@0 |
|
|
585 |
dec @1 |
|
|
586 |
brne write_% |
|
|
587 |
.endm |
|
|
588 |
|
|
|
589 |
|
|
|
590 |
|
|
|
591 |
; Use in user code
|
|
|
592 |
|
|
|
593 |
|
|
|
594 |
|
|
|
595 |
write [a,b]
|
|
|
596 |
write [c,d] |
|
|
597 |
|
|
|
598 |
|
|
|
599 |
|
|
|
600 |
; After assembling this code, the result looks like this
|
|
|
601 |
|
|
|
602 |
|
|
|
603 |
|
|
|
604 |
write_1:
|
|
|
605 |
st Z+,a |
|
|
606 |
dec b |
|
|
607 |
brne write_1 |
|
|
608 |
write_2: |
|
|
609 |
st Z+,c |
|
|
610 |
dec d |
|
|
611 |
brne write_2 |
|
|
612 |
|
|
|
613 |
|
|
|
614 |
Warnings and Errors
|
|
|
615 |
|
|
|
616 |
Some errors and warnings may confuse you a little bit so we will try to
|
|
|
617 |
clear some frequently asked questions about such cases. |
|
|
618 |
Constant out of range
|
|
|
619 |
This warning occurs if a value exceeds the byte or word value of a assignment.
|
|
|
620 |
Read the comment posted by Jim Galbraith: |
|
|
621 |
The expression (~0x80) is a Bitwise Not operation. This
|
|
|
622 |
operator returns the input expression with all its bits |
|
|
623 |
inverted. If 0x80 represents -128, then 0x7f, or +127 |
|
|
624 |
should be ok. If this is considered as a 32-bit expression |
|
|
625 |
(AVRA internal representation), then it appears to be more |
|
|
626 |
like oxffffffff-0x80 or 0xffffffff0x80. The result would then |
|
|
627 |
be 0xffffff7f. The assembler would then have to be told or it |
|
|
628 |
would have to decide, based on context, how much |
|
|
629 |
significance to assign to the higher bits. I have also |
|
|
630 |
encountered such conditions with various assemblers, |
|
|
631 |
including AVRA. To make sure the assembler does what I |
|
|
632 |
really want, I use a construct like 0xff-0x80 or 0xff0x80. |
|
|
633 |
This way the bit significance cannot extend beyond bit-7 and |
|
|
634 |
there cannot be any misunderstanding. |
|
|
635 |
Can’t use .DB directive in data segment
|
|
|
636 |
DB and .DW is only used to assign constant data in eeprom or code space. The reason why using it within data segment is forbidden is, that you
|
|
|
637 |
cannot set ram content at assembly time. The values must be programmed into |
|
|
638 |
ROM area and at boot read from ROM into RAM. This is up to the user code. |
|
|
639 |
You can only allocate memory for your variables using labels and the .byte |
|
|
640 |
directive. |
|
|
641 |
|
|
|
642 |
|
|
|
643 |
.dseg
|
|
|
644 |
my_string: .byte 15 |
|
|
645 |
|
|
|
646 |
BYTE directive
|
|
|
647 |
BYTE directive can only be used in data segment (.DSEG) This directive cannot be used in code or eeprom region because this only
|
|
|
648 |
allocates memory without assgning distinct values to it. Please use .db |
|
|
649 |
or .dw instead. |
|
|
650 |
Internal assembler error
|
|
|
651 |
If you get an "Internal assembler error" please contact the project maintainer
|
|
|
652 |
by sending him a code example and a description of your working enviroment. |
|
|
653 |
|
|
|
654 |
AVRA internals
|
|
|
655 |
|
|
|
656 |
This section provides thoughts of the avra internal design. I have to admit
|
|
|
657 |
that the code of avra is anything else than clean and optimized. To increase |
|
|
658 |
the code readability I will try to give you some standards that should improve |
|
|
659 |
quality. The following standards are similar to what GNU proposes. |
|
|
660 |
Coding standards
|
|
|
661 |
Tab space is always 2 spaces. The Tab character (ascii 9) is not used.
|
|
|
662 |
if,while,for are always opened on the same line but closed on the next line. |
|
|
663 |
The closing bracket is in the same column as the first letter of the loop |
|
|
664 |
directive. |
|
|
665 |
|
|
|
666 |
|
|
|
667 |
Example:
|
|
|
668 |
|
|
|
669 |
|
|
|
670 |
|
|
|
671 |
while(i > 0) {
|
|
|
672 |
do_something(); |
|
|
673 |
} |
|
|
674 |
|
|
|
675 |
|
|
|
676 |
Credits
|
|
|
677 |
|
|
|
678 |
We would like to thank the following people for giving contributions,
|
|
|
679 |
patches and bug reports, as well as suggestions and new ideas. |
|
|
680 |
|
|
|
681 |
|
|
|
682 |
Jon Anders Haugum (project founder)
|
|
|
683 |
Burkhard Arenfeld (release 1.2.0) |
|
|
684 |
Tobias Weber (old maintainer) |
|
|
685 |
Jerry Jacobs (release 1.3.0) |
|
|
686 |
Bernt Hembre |
|
|
687 |
Nils Strøm |
|
|
688 |
Roberto Biancardi |
|
|
689 |
Qwerty Jones |
|
|
690 |
Ben Hitchcock (Maker of the mac port) |
|
|
691 |
Daniel Drotos |
|
|
692 |
Laurence Boyd II |
|
|
693 |
Varuzhan Danielyan |
|
|
694 |
Laurence Turner |
|
|
695 |
Eugene R. O'Bryan |
|
|
696 |
Dmitry Dicky |
|
|
697 |
Bob Harris (Maker of coff support) |
|
|
698 |
Tobias Weber (enhanced macro support) |
|
|
699 |
Lesha Bogdanow |
|
|
700 |
Jim Galbraith |
|
|
701 |
Mark Brinicombe |
|
|
702 |
Igor Nikolayenko |
|
|
703 |
Peter Hettkamp |
|
|
704 |
Herb Poppe |
|
|
705 |
David Burke |
|
|
706 |
Alexey Pavluchenko |
|
|
707 |
Alan Probandt |
|
|
708 |
Mariusz Matuszek |
|
|
709 |
Arne Rossius |
|
|
710 |
Marti Tichacek |
|
|
711 |
Patrick Parity |
|
|
712 |
Johannes Overmann |
|
|
713 |
Roland Riegel |
|
|
714 |
Peter Katzmann |
|
|
715 |
Donald D. Davis |
|
|
716 |
|
|
|
717 |
And all the anonymous people who submitted patches!
|
|
|
718 |
Thank you for your work and support.
|
|
|
719 |
|
|
|
720 |
References
|
|
|
721 |
|
|
|
722 |
|
|
|
723 |
|
|
|
724 |
http://www.suprafluid.com/avra
|
|
|
725 |
http://www.avrfreaks.de |
|
|
726 |
http://www.atmel.com |
|
|
727 |
|
|
|
728 |
|
|
|
729 |
|
|
|
730 |
|
|
|
731 |
|
|
|
732 |
|
|
|
733 |
Version 1.3.0 |