CPU Definition

A CPU definition defines the syntax of the assembly instructions and how they are encoded.

Directives

.addressing_mode

.addressing_mode name {
    parameters
}

notation

This parameter is optional. If it is missing, the notation is just the mnemonic of the instruction.

arguments

This parameter is optional. If it is missing, the addressing mode has no arguments.

encoding

.opcode

.pc

.argument_type

.enum

.argument_type name map {
    name: encoded-value
    ...
}

The argument must be one of the specified name and is encoded as the corresponding encoded-value. All names must be keywords, all values must be integers.

.fits

  .argument_type name fits integer_encoding

The argument must fit in the specified integer encoding.

For example:

.argument_type byte fits 1
.argument_type signed_byte fits -1
.argument_type word fits 2

.map

.argument_type name map {
    source-value: encoded-value
    ...
}

The argument must be one of the specified source-value and is encoded as the corresponding encoded-value. All values must be integers.

.range

Warning: This argument type is currently broken. Don't use it.

.byte_order

Specify the native byte order of the CPU.

  .byte_order byte_order

Each digit in byte_order corresponds to one byte, with 1 representing the least significant and 8 the most significant byte. Bytes are encode in memory in the order they are listed. For example, little endian is 12345678, big endian is 87654321.

.instruction

.instruction mnemonic {
    addressing-mode: opcode
    ...
}

.syntax

  .syntax [keywords|punctuation] "symbol" ...

This specifies punctuation and keywords used the assembler syntax. Keywords that are valid identifiers (begin with a letter or underscore and contain only letters, underscore, and digits) need not be specified.