sh-dis/generate_impl.py

import sys
import os

from lexer import Lexer
from parser import Parser
from ast_to_c_source import generate
from ast_transformers import transform_statements

from instruction_table import untabulate_instructions_sh4
from instruction_table import untabulate_instructions_sh2
from instruction_table import unparse_instruction_code
from instruction_function_name import instruction_function_name
from instruction_file_name import instruction_file_name
from instruction_properties import has_delay_slot

from disabled_instructions import disabled_instructions

def is_instruction_descriptor(s):
    return all(c in {'0', '1', 'n', 'm', 'i', 'd'} for c in s)

def parse_header(buf):
    lines = buf.split('\n')
    code, available, *rest = lines
    assert len(code) == 16 and is_instruction_descriptor(lines[0]), lines[0]
    if not available.startswith('Available only when'):
        return code, None, '\n'.join([available, *rest])
    else:
        return code, available, '\n'.join(rest)

def read_ins_source(ins):
    file_name = instruction_file_name(ins)
    src_path = os.path.join("sh4", file_name)
    with open(src_path) as f:
        return f.read()

def parse_tokens(buf):
    lexer = Lexer(buf)

    while True:
        try:
            token = lexer.next_token()
        except IndexError:
            break
        yield token

def generate_function_declaration(instruction_name, function_name, variables):
    args = ", ".join([
        "struct architectural_state * state",
        "struct memory_map * map",
        *[
            f"const uint32_t {x}"
            for x in variables
        ]
    ])

    yield f"/* {instruction_name} */"
    yield f"void {function_name}({args})"

def generate_file(instruction_name, function_name, ins, delay_slot):
    buf = read_ins_source(ins)
    code, _available, source = parse_header(buf)
    assert code == unparse_instruction_code(ins), (code, unparse_instruction_code(ins))

    tokens = list(parse_tokens(source))
    parser = Parser(tokens)
    statements = []
    while parser.tokens[parser.pos:]:
        statements.append(parser.statement())

    statements = transform_statements(statements)

    yield from generate_function_declaration(instruction_name, function_name, ins.variables)
    yield "{"
    output = []
    for statement in statements:
        src = "".join(generate(statement))
        output.append(src)

    for line in "".join(output).rstrip().split('\n'):
        yield f"  {line}"
    yield ""
    yield f"  state->is_delay_slot = {delay_slot};"
    yield "}"
    yield ""
    yield ""

def main():
    output = [
        '#include "impl.h"',
        '#include "operations.h"',
        '#include "exception.h"',
        '#include "state_helpers.h"',
        '#include "fpu.h"',
        '',
    ]

    header_output = [
        '#pragma once',
        '',
        '#include "state.h"',
        '#include "memory_map.h"',
        '',
    ]

    for ins in untabulate_instructions_sh4():
        if ins.instruction in disabled_instructions:
            continue

        _name = [ins.instruction, ins.operands] if ins.operands else [ins.instruction]
        instruction_name = " ".join(_name)
        function_name = instruction_function_name(ins)
        delay_slot = "true" if has_delay_slot(ins) else "false"
        gen = generate_file(instruction_name, function_name, ins, delay_slot)
        output.extend(gen)

        gen = generate_function_declaration(instruction_name, function_name, ins.variables)
        lgen = list(gen)
        lgen[-1] += ';'
        header_output.extend(lgen)

    with open(sys.argv[1], "w") as f:
        f.write("\n".join(output))

    with open(sys.argv[2], "w") as f:
        f.write("\n".join(header_output))

if __name__ == '__main__':
    main()