Modified: 8 July 1996 by Karl Lunt
Last Modified: 7 April 1999 by Tom Almy
The as12 assembler is a two-pass cross-assembler for the Motorola 68hc12 microcontroller (MCU). as12 was written in C and derived from the source code for the original asm11 68hc11 assembler. So far, as12 has been ported to Sun Sparcstations, HPs, IBM PC-compatibles, and Apple Macintoshes.
as12 was designed to run from a command-line interpreter (CLI), so it works well with DOS batch files and Unix shell scripts. The Mac version of as12 has been integrated with the Macintosh Programmer's Workbench (MPW).
Unless otherwise noted, this document describes the PC version of as12.
Additional changes by Tom Almy corrects documentation, adds a unary complement operator, and provides a Windows console mode compiler which handles long file names.
To start the as12 assembler, enter the following command at the prompt:
as12 file.ext
where file.ext is the path, name, and extension of the file you want to assemble. The as12 assembler will assemble the file, sending the listing output to the console and writing the S19 object output to a file named m.out. To save the listing output in a text file for later review, use the -L option. For example:
as12 foo.asm -Lfoo.lst
will assemble the file foo.asm and write the output listing to a file named foo.lst.
Entering as12 with no arguments will display a short help file describing the available command-line options.
Command-line options allow you to specify input file names, define global symbols, and declare paths for library files. For example:
as12 -dTIMERS -l\mylib\hc12 foo -L
will assemble the file foo.asm, using the library files found in the directory \mylib\hc12. Additionally, the label TIMERS will be defined with a value of 1 and the listing output will be written to file foo.lst.
When specifying the assembler source file, the extension ".asm" is assumed if no extension is explicitly given. If more than one source file is specified, then they will be read in the order listed.
The full set of command-line options includes:
The -d option allows you to define a label on the command line. Labels defined in this manner are treated by the as12 assembler as if they had been defined within your source file and assigned a value of 1. Your source file can then refer to these labels in #ifdef tests. For example:
as12 -dMY_ALGORITHM myprog.asm
causes the as12 assembler to behave as if your source file began with the line:
#define MY_ALGORITHM
This ability to define labels from the command line adds great power to the as12 assembler. You can use this feature to selectively assemble blocks of source code based on arguments you specify in the command line, without first having to edit the source code before each assembly.
Normally, as12 first checks in the current directory for needed source and include files. If as12 cannot find a needed file in the current directory, it then checks the path for a previous source file and searches that directory. If that search also fails, as12 will use the library path specified with the -l option on the command line, if any. For example:
as12 -lc:\mypath myfile.asm
as12 allows you to pass in a special string which will identify the part you are compiling the program for. When combined with the #ifpart conditional assembly directive, can give users a powerful way to compile source code which may depend upon what part is being targete d.
An illustration of it usage...
as12 -pb32
Turns on the internal debug features of as12. Mostly for the developer of as12, but may help you if you are having a problem.
as12 -d source_file
Specifies a listing file. If no filename extension is given, ".lst" is assumed. If no file name is given, then the file name will be that of the first source file, with an extension ".lst".
This help is designed as a reminder to the other command line options.
Example...
as12 -h
The include directive allows other assembly source files to be inserted in the code immediately after the include statement, as if the c ontents of the included file were actually in the file that contained the include statement. Stated differently, the include statement works as you m ight expect. The syntax of the include statement is shown below...
#include /my_dir/myfile.s
It is important to note that the filename expansion will only be as good as the filename expansion as the shell that you are operating in. For ex ample, if you are running shell (/bin/sh) then the tilde username (~user) lookup may not work correctly. It is best to put in relative or absolute fi lepath specifications that are not shell dependent.
include statements may be used within #ifdef statements.
The define statement allows labels to be defined. This statement is simply an alternate form for an equ assembler directive. T he alternate form is provided so that users will be alerted to the opportunities to write more sophisticated code that the #ifeq, and related statements allow. The proper use of the define statement is...
#define MY_LABEL expression
The define statement is as if the user had typed the following...
MY_LABEL EQU expression
Both forms are equally valid, and both forms are implemented internally thee same way. The EQU is probably more portable of the two constructs.
The ifeq command allows for the user to conditionally compile different sections of assembly language based on whether or not a label is equal to a value. Example...
#ifeq MY_SYMBOL expression_to_compare_to ... (this code will be executed if MY_SYMBOL has the same value as the expression_to_compare_to) ... #endif
I show the #endif statement because for every form of if there needs to be a marker so that as12 knows what code is to be conditionally compiled. Restated, for every if there needs to be an endif.
If the expression resolves to the same value as the label in an #ifeq directive, then every line between the #ifeq and the #endif is executed. If the expression resolves to a different value than the label, all of the lines between the #ifeq and the #endif are ignored.
The ifneq command allows for the user to conditionally compile different sections of assembly language based on whether or not a label is not equ al to a value. Example...
#ifneq MY_LABEL expression_to_compare_to ... (this code will be executed if MY_LABEL has a different value as the expression_to_compare_to) ... #endif
I show the #endif statement because for every form of if there needs to be a marker so that as12 knows what code is to be conditionally compiled. Restated, for every if there needs to be an endif.
If the expression resolves to the same value as the label in an #ifeq directive, then every line between the #ifeq and the #endif is ignored. If the expression resolves to a different value than the label, all of the lines between the #ifeq and the #endif are executed.
The ifdef command allows for the user to conditionally compile different sections of assembly language based on whether or not a label is defined (via a #define or an EQU. Example...
#ifdef MY_LABEL ... (this code will be executed if MY_LABEL has been defined) ... #endif
I show the #endif statement because for every form of if there needs to be a marker so that as12 knows what code is to be conditionally compiled. Restated, for every if there needs to be an endif.
If the label in an #ifdef directive is defined, then every line between the #ifeq and the #endif is executed. If the label is not defined, all o f the lines between the #ifdef and the #endif are ignored.
This is the only directive that allows for a string comparison. A special internal variable is the only variable which is a string variable. The only way to set that variable is with the -p command line option. The sole purpose of this directive is to allow for conditional as sembly based upon the value of the string. This seemed natural for handling the different part types. Example...
#ifpart b32 ... (assembly code) (will be executed if the string <b32> is same as string in -p option ... #endif
I show the #endif statement because for every form of if there needs to be a marker so that as12 knows what code is to be conditionally compiled. Restated, for every if there needs to be an endif.
If the string that follows the #ifpart directive matches the string that was passed in via the -p option, then the lines betwee n the #ifpart and the #endif will be executed. If the strings do not match, the lines between the #ifpart and the #endif will be ignored.
This directive must be coupled with any of the if directives. This allows either or compilation and performs just like you expect an else to perf orm. Example...
#ifdef MY_LABEL ... (assembly code) (will be executed if MY_LABEL is defined) #else ... (assembly code) (will be executed if MY_LABEL is NOT defined) #endif
I show the #endif statement because for every form of if there needs to be a marker so that as12 knows what code is to be conditionally compiled. Restated, for every if there needs to be an endif.
If the if statement that goes with the else statement is true, the statements between the if and the else will be assembled, and the statements b etween the else and the endif will be ignored. If the if statement is false, the statements between the if and the else will be ignored and the state ments between the else and the endif will be executed.
There can only be one else for each if statement.
The endif statement tells the assembler when the conditional assembly section of the code is finished. Otherwise the assembler would have no way of knowing when to quit.
For every if statement there needs to be one endif. If there is an if and an else, then there should be one end statement also.
Examples...
#ifpart part_name #else #endif #ifndef MY_LABEL #endif #ifndef MY_LABEL #else #endif
#ifdef EXPANDED_MODE org START_OF_EXTERNAL_RAM_TESTS #else org START_OF_FLASH_RAM #endif
#ifpart b32 RAM_START EQU $800 FEE_START EQU $8000 REG_START EQU $0000 PWM_START EQU $c7 #endif #ifpart a4 RAM_START EQU $600 REG_START EQU $0100 #endif
Notice how easy you could build a library of different parts and make your source code compile accordingly.
This executable file was created with Borland C++ version 5.0, targeting the Win32 Console.
A file with the same name as the first source file but with the extension ".s19" is always produced by the assembler. It cannot be suppressed. It is the s-records that are created by assembling the source file that is given to as12 on the as12 command line.
99 times out of 100 times, this is the file of interest when using the assembler.
For information regarding s-records (like a spec, but not quite) go my little s-record description.
The listing file is useful for debugging. Simply add the command line option "-L" to create the listing file.
Standard ASCII source files. These should be created with the extension ".asm" since that is the default used by the assembler.
Expressions may consist of symbols, constants or the character '*' (denoting the current value of the program counter) joined together by one of the operators: +-*/%&|^. You may nest expressions using parentheses up to 5 levels deep. The operators are the same as in C:
+ add - subtract * multiply / divide % remainder after division & bitwise and | bitwise or ^ bitwise exclusive-or
In addition, the unary minus (-) and complement (~) operators are allowed when preceding a symbol, constant, or character '*' only.
Examples of valid expressions... (5*8) (my_val-10+20*(16-label)/10) 10 $10 * %10010 my_value ~$20
Note: When the asterisk (*) is used in a context where the as12 is expecting a label, the asterisk (*) represents the value of the current program counter.
Symbols consist of one or more characters where the first character is alphabetic and any remaining characters are alphanumeric. Symbol are case sensitive.
Constants are constructed with the same syntax as the Motorola MDOS assembler (oh, now thats a real useful piece of information - hey I just copied this anyway):
' followed by ASCII character $ followed by hexadecimal constant @ followed by octal constant % followed by binary constant digit decimal constant
A symbol starting in the first column is a label and may optionally be ended with a ':'. A label may appear on a line by itself and is then interpreted as:
Label EQU *
Note that labels are case sensitive. "Label" and "label" are different labels.
Here are some notes about comments...
Zeros memory. Please use zmb instead.
Actually, don't use these at all.
db Byte_Definition[,Byte_Definition] db $55,$66,%11000011 db 10 half db 0.5*100
Defines the value of a byte or bytes that will be placed at a given address.
The db directive assigns the value of the expression to the current program counter. Then the program counter is incremented.
Multiple bytes can be defined at a time by comma separating the arguments. Each comma separated argument can be a separate expression that the as 12 will evaluate.
Identical to db. My preference is to use the db and not this one. This is only to help read other peoples software that may get sent to us.
Identical to dw.
ds Number_of_Bytes_To_Advance_Program_Counter
The ds increments the program counter by the value indicated in the Number of Bytes argument.
Identical to ds. I don't care what form that you use, but I imagine that the ds is better than the ds.b. I hope you don't mind clicking one more time to get to the right spot in the manual.
ds.w Number_of_Words_To_Advance_Program_Counter
The ds.w increments the program counter by the value indicated in the Number of Words argument multiplied by two. In other words, if the ds.w expression evaluates to 4 then the program counter is advanced by 8.
dw Word_Definition[,Word_Definition] dw $55aa,$66,%11000011 dw 10 half dw 0.5*65536
Defines the value of a word or words that will be placed at a given address.
The dw directive assigns the value of the expression to the current program counter. Then the program counter is incremented by 2.
Multiple words can be defined at a time by comma separating the arguments. Each comma separated argument can be a separate expression that the as 12 will evaluate.
Identical to ttl.
Label EQU Value_To_Assign_To_The_Label
Directly assigns a value to a label.
Identical to db. The db is the preferred command. This flies in the face of Motorola history, but I believe that externally our files may be more compatible (Although I can't prove it so do what you want).
fcc delim_characterstring_to_encodedelim_character fcc /my_string/ fcc *// string with slashes //* fcc 'best to use single quotes'
FCC allow the encoding of a string.
The first character is the delimiter. By allowing the flexibility of selecting delimiters, you can easily make strings which have slashes and tick marks in them. The only catch is that if you choose a delimiter, it
In the second example, my_string will be encoded as an ASCII string. The /'s simply mark the ending and beginning of the string. This also lets you put spaces in the string.
In the third example, the * (asterisk) is the delimiter and the slashes will be encoded with their ASCII values into the ASCII string.
I like single quotes the best as a delimiter. You could argue that double quotes are even better because it follows 'C' convention.
Identical to dw. I prefer the dw usage.
fill byte_to_fill_memory_with,num_of_bytes_to_fill
FILL allows a user to fill memory with a byte. See my comments in zmb about the value of these pseudo opcodes.
Another Pseudo OP called LOC basically increments and produces an internal counter used in conjunctions with the backwards tick mark (`). By using LOC's and the ` mark you can do code like the following without worrying about thinking up new labels.
LOC ldaa #1 loop` deca bra loop` LOC loop` brset 0,x $55 loop`
This code will work perfectly fine because the second loops label is really loop002 and the first ones is loop001. The assembler really sees this...
LOC ldaa #1 loop001 deca bra loop001 LOC loop002 brset 0,x $55 loop002
You may also seed the LOC with a valid expression or number by putting that expression or number in the operand field. This gives you the ability to over ride the automatic numbering. This is also sometimes handy if you need to keep track of what your local variable is. (you lose track in the source if you aren't careful, because the tick ' mark is the only thing you see).
Ignored
There are five permissible operands for this instruction:
org value_to_set_program_counter_to org $800 ;not my preferred form org MY_PROGRAM_START ;better form org LAST_MEMORY_LOCATION-(LAST_PROGRAM_BYTE-FIRST_PROGRAM_BYTE);best but complex and has reference problems
The org pseudo opcode allows the assembler's program counter to be set to a value. This is useful for locating your software and its elements (tables, ram, constants, etc) in useful (intelligent) locations within the memory space of the microcontroller.
In better multi-pass assemblers (not as12), the org statement is rarely used because the code is located at the link, and not during compilation. Since as12 is a simple two-pass assembler, orgs must be used so that the code is compiled where it is supposed to.
Ignored
Used to redefine first operand (which must be a label) to value of second operand (an expression)
Example:
foo equ 10 ldaa #foo ; Accumulator A gets value 10 redef foo 12 ldab #foo ; Accumulator B gets value 12
Equivalent to ds.b or ds, which is preferred.
Equivalent to ds.w.
Ignored
Ignored
Operand specifies number of bytes to allocate and fill with zero. Use is not recommended.