You cannot select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
212 lines
6.1 KiB
Python
212 lines
6.1 KiB
Python
#
|
|
import os
|
|
import re
|
|
import pin
|
|
import assembly as ASM
|
|
from controller import get_hex_str, save_hex_str
|
|
|
|
# 允许操作的寄存器
|
|
ALLOW_REGISTER_MAP = {
|
|
# "a": pin.A, ALU经常使用这俩寄存器, 还是禁止使用算了
|
|
# "b": pin.B,
|
|
"c": pin.C,
|
|
"d": pin.D,
|
|
"ss": pin.SS,
|
|
"sp": pin.SP,
|
|
}
|
|
|
|
OP2_STR_MAP = {
|
|
"mov": ASM.MOV,
|
|
"add": ASM.ADD,
|
|
"sub": ASM.SUB,
|
|
"cmp": ASM.CMP,
|
|
"and": ASM.AND,
|
|
"or": ASM.OR,
|
|
"xor": ASM.XOR,
|
|
}
|
|
|
|
OP1_STR_MAP = {
|
|
"inc": ASM.INC,
|
|
"dec": ASM.DEC,
|
|
"not": ASM.NOT,
|
|
"jmp": ASM.JMP,
|
|
"jo": ASM.JO,
|
|
"jno": ASM.JNO,
|
|
"jz": ASM.JZ,
|
|
"jnz": ASM.JNZ,
|
|
"jp": ASM.JP,
|
|
"jnp": ASM.JNP,
|
|
"push": ASM.PUSH,
|
|
"pop": ASM.POP,
|
|
"call": ASM.CALL,
|
|
"int": ASM.INT,
|
|
}
|
|
|
|
OP0_STR_MAP = {
|
|
"nop": ASM.NOP,
|
|
"hlt": ASM.HLT,
|
|
"ret": ASM.RET,
|
|
"iret": ASM.IRET,
|
|
"sti": ASM.STI,
|
|
"cli": ASM.CLI,
|
|
}
|
|
|
|
OP2_VALUE_SET = set(OP2_STR_MAP.values())
|
|
OP1_VALUE_SET = set(OP1_STR_MAP.values())
|
|
OP0_VALUE_SET = set(OP0_STR_MAP.values())
|
|
|
|
|
|
class Code:
|
|
TYPE_CODE = 1
|
|
TYPE_LABLE = 2
|
|
def __init__(self, number, source):
|
|
self.number: int = number # 源文件中行号
|
|
self.source: str = source # 源代码
|
|
self._op_str = self._src = self._dst = None
|
|
|
|
self.code_type = ...
|
|
self.name = ...
|
|
self.code_from_bin_addr = ... # 对应bin文件中的位置
|
|
|
|
self.prepare_source()
|
|
|
|
def __repr__(self):
|
|
return f"[{self.number}-{self.code_from_bin_addr}] - op: {self._op_str}, dst: {self._dst}, src: {self._src} - '{self.source}'"
|
|
|
|
def prepare_source(self):
|
|
if self.source.endswith(":"):
|
|
self.code_type = self.TYPE_LABLE
|
|
self.name = self.source.strip(":")
|
|
return
|
|
|
|
self.code_type = self.TYPE_CODE
|
|
|
|
tup = self.source.split(",")
|
|
if len(tup) > 2:
|
|
raise Exception("指令格式错误")
|
|
|
|
if len(tup) == 2:
|
|
self._src = tup[1].strip()
|
|
|
|
tup = re.split(r" +", tup[0].strip())
|
|
|
|
if len(tup) > 2:
|
|
raise Exception("指令格式错误")
|
|
|
|
if len(tup) == 2:
|
|
self._dst = tup[1].strip()
|
|
|
|
self._op_str = tup[0]
|
|
|
|
@property
|
|
def op(self):
|
|
# 根据 mov, add 等指令得到对应的操作二进制码
|
|
if self._op_str in OP2_STR_MAP:
|
|
op = OP2_STR_MAP[self._op_str]
|
|
elif self._op_str in OP1_STR_MAP:
|
|
op = OP1_STR_MAP[self._op_str]
|
|
else:
|
|
op = OP0_STR_MAP[self._op_str]
|
|
return op
|
|
|
|
def get_am(self, addr: str):
|
|
# 得到目的操作数和源操作数, 返回(寻址方式, 地址)
|
|
global marks
|
|
if addr in marks:
|
|
return pin.AM_INS, marks[addr].code_from_bin_addr * 3 # 我们所有的指令都是占了3个字节: ir dst src
|
|
if addr is None:
|
|
return None, None
|
|
if addr in ALLOW_REGISTER_MAP: # 如果是寄存器寻址, 且必须是允许的寄存器
|
|
return pin.AM_REG, ALLOW_REGISTER_MAP[addr]
|
|
if re.match(r"^[0-9]+$", addr): # 如果是立即数
|
|
return pin.AM_INS, int(addr)
|
|
if re.match(r"^0[Xx][0-9A-Za-z]+$", addr): # 如果是16进制立即数
|
|
return pin.AM_INS, int(addr, 16)
|
|
if re.match(r"^\[[0-9]+\]$", addr): # 如果是立即数
|
|
return pin.AM_DIR, int(addr[1:-1])
|
|
if re.match(r"^\[0[Xx][0-9A-Za-z]+\]$", addr): # 如果是16进制立即数
|
|
return pin.AM_DIR, int(addr[1:-1], 16)
|
|
if re.match(r"^\[.+\]$", addr): # 如果是 [a]
|
|
reg = addr[1:-1]
|
|
if reg in ALLOW_REGISTER_MAP:
|
|
return pin.AM_RAM, ALLOW_REGISTER_MAP[reg]
|
|
|
|
raise Exception(f"操作数 地址错误: {self}")
|
|
|
|
def compile_code(self):
|
|
amd, dst = self.get_am(self._dst) # 目的操作数
|
|
ams, src = self.get_am(self._src) # 源操作数
|
|
# print(amd, dst, ams, src)
|
|
|
|
if src and (amd, ams) not in ASM.INSTRUCTIONS[2][self.op]:
|
|
raise Exception(f"操作数 地址错误: {self}")
|
|
elif src is None and dst and amd not in ASM.INSTRUCTIONS[1][self.op]:
|
|
raise Exception(f"操作数 地址错误: {self}")
|
|
elif src is None and dst is None and self.op not in ASM.INSTRUCTIONS[0]:
|
|
raise Exception(f"操作数 地址错误: {self}")
|
|
|
|
amd = amd or 0
|
|
dst = dst or 0
|
|
ams = ams or 0
|
|
src = src or 0
|
|
|
|
# 根据op, 构建 ir寄存器
|
|
if self.op in OP2_VALUE_SET:
|
|
ir = self.op | amd << 2 | ams
|
|
elif self.op in OP1_VALUE_SET:
|
|
ir = self.op | amd
|
|
else:
|
|
ir = self.op
|
|
return ir, dst, src
|
|
|
|
|
|
def compile_program(input_file: str):
|
|
codes = []
|
|
with open(input_file, "r", encoding="utf8") as f:
|
|
lines = f.readlines()
|
|
for idx, line in enumerate(lines):
|
|
source = line.strip().split(";")[0] # 去掉空格和注释
|
|
if len(source) == 0:
|
|
continue
|
|
code = Code(idx + 1, source)
|
|
codes.append(code)
|
|
|
|
# 保底的
|
|
codes.append(Code(idx + 1 + 1, "hlt"))
|
|
|
|
ret = []
|
|
marks = {}
|
|
current_code = ...
|
|
for code in reversed(codes): # 从后往前
|
|
if code.code_type == Code.TYPE_CODE:
|
|
current_code = code
|
|
ret.insert(0, code)
|
|
continue
|
|
if code.code_type == Code.TYPE_LABLE:
|
|
marks[code.name] = current_code
|
|
continue
|
|
|
|
raise Exception("语法之外的 错误")
|
|
|
|
# 更新代码在bin中的索引
|
|
for idx, code in enumerate(ret):
|
|
code.code_from_bin_addr = idx
|
|
|
|
return ret, marks
|
|
|
|
|
|
if __name__ == '__main__':
|
|
data_width = 8
|
|
addr_width = 16
|
|
input_file = "./test_mov.asm"
|
|
output_file = "test_mov.bin"
|
|
|
|
program_bin_data = []
|
|
ret_codes, marks = compile_program(input_file)
|
|
for code in ret_codes:
|
|
values = code.compile_code()
|
|
for value in values:
|
|
program_bin_data.append(value)
|
|
|
|
print(get_hex_str(program_bin_data, data_width / 4))
|
|
save_hex_str(get_hex_str(program_bin_data, data_width / 4), 8, "program.bin") |