This section is about the internals of the msp430.asm
module. It may be interesting for developers that work on this module or who are interested in using the functions the module provides in their own code.
The file format of .o4
files is a bit unusual. It actually contains something that could be labeled as (specialized) Forth code. So the linker is some sort of Forth interpreter. This has the advantage that the object files can be debugged without any special tools, just a text editor. It also makes the format quite universal; it could produce binaries for all sorts of CPUs (single special case: the directive JMP
is MSP430 specific).
A list of supported words can be found in the following document:
linker_words
For more details also take a look at the sources of ld.py
.
MCU memory definitions can be provided in a file with Forth like syntax.
A list of supported words can be found in the following document:
mcdudef_words
For more details also take a look at the sources of mcu_definition_parser.py
.
msp430.asm
msp430.asm.as
This module implements the MSP430(X) assembler. When the module is executed (e.g. using python -m msp430.asm.as
), it acts as a command line tool.
__init__(msp430x=False, debug=False)
- param msp430x
Set to true to enable MSP430X instruction set.
- param debug
When set to true dump some internal data so sys.stderr while compiling.
Create an instance of the assembler.
assemble(f, filename=None, output=sys.stdout)
- param f
A file like object that supports iterating over lines.
- param filename
An optional string that is used in error messages.
- param output
File like object used to write the object code to.
This method takes assembler source and transforms it to object code that can be forwarded to the linker.
AssemblerError
This instances of this class are raised by the MSP430Assembler
in case of errors in the source. It may be annotated with the source filename and line number where the error occurred.
filename
line
msp430.asm.ld
This module implements the linker. When the module is executed (e.g. using python -m msp430.asm.ld
), it acts as a command line tool.
__init__(name, start_address=None, end_address=None, align=True, programmable=False, little_endian=True, parent=None, mirror_of=None)
__getitem__(segment_name)
- param segment_name
name of an sub segment.
- raises KeyError
when no segment with given name is found
Easy access to subsegment by name.
sort_subsegments(by_address=False)
- param by_address
Sort by address if true, otherwise sort by name.
Sort list of subsegments either by order of definition or by order of start address.
clear()
Clear data. Recursively with all subsegments. Segments are not removed, just their data.
__len__()
Get the number of data bytes contained in the segment.
__cmp__(other)
Compare function that allows to sort segments by their start_address.
__lt__(other)
Compare function that allows to sort segments by their start_address.
print_tree(output, indent='', hide_empty=False)
- param output
a file like object (supporting
write
)- param indent
a prefix put before each line.
- param hide_empty
when set to true omit empty segments (no data) in output.
Output segment and subsegments.
shrink_to_fit(address=None)
Modify start- and end_address of segment to fit the data that it contains. Recursively applied to the tree of segments. Typically called with address=None
.
write_8bit(value)
- param value
an integer (8 significant bits)
Write one byte.
write_16bit(value)
- param value
an integer (16 significant bits)
Write two bytes. Order in memory depends on endianness of segment.
write_32bit(value)
- param value
an integer (32 significant bits)
Write four bytes. Order in memory depends on endianness of segment.
__init__(instructions)
- param instructions
list of directives for the linker
Initialize a linker instance. The given instructions are essentially what is read from a .o4
file as sequence of words.
segments_from_definition(segment_definitions)
- param segment_definitions
dictionary describing the memory map
This sets the memory map used for linking. See mcu_definition_parser
for a way to load this description.
update_mirrored_segments()
Called before writing the final output. In case the memory map contains segments that mirror the contents of other segments, they are updated. This is typically used for .data_init
which contains the initial values that are copied by startup code to the .data
segment in RAM.
pass_one()
Run the linkers 1st pass. It iterates through the instructions and places the data into segments.
pass_two()
Run the linkers 2nd pass. It iterates through the instructions and finds all the labels and saves their position.
pass_three()
Run the linkers 3rd pass. It iterates through the instructions and creates the final binary with all known labels set to their target address.
LinkError
Exception object raised when errors during linking occur. May be annotated with the location of the line within the original source file causing the error.
filename
lineno
column
msp430.asm.cpp
This module implements the preprocessor. When the module is executed (e.g. using python -m msp430.asm.cpp
), it acts as a command line tool.
line_joiner(next_line)
Given an iterator for lines, yield lines. It joins consecutive lines with the continuation marker (\\
) to a single line.
This object is used by the preprocessor to write out the preprocessed text. It adds notes in the form #line <line> "<filename>"
. These notes are used by the assembler to know where a source line originally came from (as preprocessed text may contain additional lines etc.)
__init__(output, filename)
- param output
file like object to write to
- param filename
the filename used in the notes
write(lineno, text)
- param linno
line number being written
- param text
the actual contents of the line
preprocess(infile, outfile, filename)
- param infile
file like object to read from
- param outfile
file like object to write to
- param filename
original file name of the input (infile)
This runs the preprocessor over the given input.
PreprocessorError
Exception object raised when errors during preprocessing occur.
msp430.asm.disassemble
This module implements the disassembler. When the module is executed (e.g. using python -m msp430.asm.disassemble
), it acts as a command line tool.
__init__(memory, msp430x=False, named_symbols=None)
- param memory
A msp43.memory.Memory instance containing the binary.
- param msp430x
Set to true to enable MSP430X instruction set.
- param named_symbols
An (optional) instance of
NamedSymbols
which is used to label peripherals and bits.
Initialize the disassembler with data.
disassemble(output, source_only=False)
- param output
A file like object used for the resulting text.
- param source_only
When set to true, the address and data columns are omitted from the output.
Run the disassembler, result is written to output.
msp430.asm.rpn
This module implements the an RPN calculator. The calculator can be tested by executing the module (e.g. using python -m msp430.asm.rpn
).
This class is used to wrap words so that their source location can be tracked. This is useful for error messages.
__new__(cls, word, filename, lineno, text)
- param cls
Class for __new__
- param word
The word (unicode)
- param filename
Filename where the word was read from.
- param lineno
Line number within the file.
- param text
The complete line (or context).
- type filename
unicode or None
- type lineno
int or None
- type text
unicode or None
Create new instance with a word that was read from given location.
An RPN calculator. It provides a data stack and implements a number of basic operations (arithmetical and stack)
interpret(next_word)
- param next_word
A function return the next word from input when called.
Interpret a sequence of words given by the iterator next_word.
annotated_words(sequence, filename=None, lineno=None, offset=None, text=None)
Create an generator for Word
, all annotated with the given information.
words_in_string(data, name='<string>')
- param data
String with (lines) of text.
- param name
Optional name, used in error messages.
Create a generator for annotated Word
in string (splitlines()
is used).
words_in_file(filename)
- param filename
Name of a file to read from.
Create a generator for annotated Word
read from file given by name.
rpn_function(code)
- param code
A string in RPN notation
- return
A Python function.
Return a wrapper - a function that evaluates the given RPN code when called. This can be used to insert functions implemented as RPN into the name space.
word(name)
Function decorator used to tag methods that will be visible in the RPN built-in name space.
val(words, stack=[], namespace={})
- param words
Sequence of words.
- param stack
Optional initial stack.
- param namespace
Optional namespace.
- return
The top element from the stack
Evaluate sequence of words.
python_function(code, namespace={})
- param code
RPN code to execute.
- param namespace
Optional namespace.
- return
A python function that executes
code
when called.
Create a Python function that will execute given code when called. All parameters given to the Python function will be placed on the stack and the top of stack will be returned.
interpreter_loop(namespace={}, debug=False)
Run an interactive loop. Can be used as calculator.
RPNError
Exception type used for errors when parsing or executing RPN code. It may be annotated with the source position where the word causing the error came from.
filename
lineno
offset
text
msp430.asm.peripherals
This module implements a parser for a file format describing the peripherals and their bits of a MCU. The module can be executed (e.g. using python -m msp430.asm.peripherals
) to test definition files.
This class implements the parser and keeps the result. It inherits from RPN
.
load_symbols(filename)
- param filename
Load symbols from a file named like this.
- return
instance of
SymbolDefinitions
Load definitions from a file of given name.
load_internal(name)
- param name
Name of an internal file.
- return
instance of
SymbolDefinitions
This tries to load internal data (using pkgutil
).
SymbolError
Exception object used for errors in the definition file.
msp430.asm.mcu_definition_parser
This module implements the a parser for files describing the memory map of a CPU. The module can be executed (e.g. using python -m msp430.asm.mcu_definition_parser
) to test definition files.
This class implements the parser and keeps the result. It inherits from msp430.asm.rpn.RPN
. Loaded definitions may contain the memory maps of many MCUs and also partial maps (that may depend on each other).
load_from_file(filename)
- param filename
Load definitions from file of given name.
- return
instance of
MCUDefintitions
load_internal()
- return
instance of
MCUDefintitions
Load internal list. The default list is included in msp430/asm/definitions/msp430-mcu-list.txt
expand_definition(memory_maps, name)
- param memory_maps
Memory map descriptions.
- param name
Name of an MCU that should be extracted
- type memory_maps
MCUDefintitions
- return
Dictionary with recursively expanded memory map.
Return the memory map of a specific MCU. If the definition depends on others, it is expanded so that a single, complete description is returned.
msp430.asm.infix2postfix
This module implements a converter that can translate infix (arithmetical) notation to postfix notation (RPN). It is used by the preprocessor and assembler when evaluating expressions.
infix2postfix(expression, variable_prefix='', scanner=Scanner, precedence=default_precedence)
- param expression
Input string in infix notation.
- param variable_prefix
A string that is prepended to symbols found in the expression.
- param scanner
The class that is used to parse the expression.
- param precedence
A dictionary returning the priority given an operator as key.
- return
A string with the expression in postfix notation.
convert_precedence_list(precedence_list)
- param precedence_list
A list of lists that defines operator priorities.
- return
A dictionary mapping operators to priorities.
Input will look like this:
default_precedence_list = [
# lowest precedence
['or'],
['and'],
['not'],
['<', '<=', '>', '>=', '==', '!='],
['|', '^', '&'],
['<<', '>>'],
['+', '-'],
['*', '/', '%'],
['~', 'neg', '0 +'],
['(', ')'],
# highest precedence
]