ben
/
crater
mirror of https://github.com/earwig/crater
1
1
Fork 0
Code Releases 0 Activity
Browse Source

Merge branch 'feature/asm_instructions'

master
Ben Kurtovic 9 years ago
parent
3d968c12b5 764211a379
commit
d483afef51
13 changed files with 5327 additions and 1067 deletions
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. +6
    -1
      makefile
  2. +406
    -0
      scripts/update_asm_instructions.py
  3. +2
    -5
      src/assembler/errors.c
  4. +2
    -5
      src/assembler/errors.h
  5. +20
    -13
      src/assembler/inst_args.h
  6. +0
    -98
      src/assembler/inst_support.c
  7. +0
    -155
      src/assembler/inst_support.h
  8. +170
    -744
      src/assembler/instructions.c
  9. +2784
    -0
      src/assembler/instructions.inc.c
  10. +1859
    -0
      src/assembler/instructions.yml
  11. +76
    -44
      src/assembler/parse_util.c
  12. +1
    -1
      src/assembler/parse_util.h
  13. +1
    -1
      src/assembler/tokenizer.c

+ 6
- 1
makefile View File

@@ -12,6 +12,7 @@ CFLAGS = $(shell sdl2-config --cflags)
LIBS = $(shell sdl2-config --libs)
MKDIR = mkdir -p
RM = rm -rf
ASM_UP = scripts/update_asm_instructions.py

MODE = release
BNRY = $(PROGRAM)
@@ -23,7 +24,7 @@ DIRS = $(sort $(dir $(OBJS)))

ifdef DEBUG
BNRY := $(BNRY)$(DEVEXT)
FLAGS += -g -fsanitize=address -DDEBUG_MODE
FLAGS += -g -DDEBUG_MODE
MODE = debug
endif

@@ -47,6 +48,10 @@ $(BUILD)/$(MODE)/%.o: %.c

-include $(DEPS)

ASM_INST = $(SOURCES)/assembler/instructions
$(ASM_INST).inc.c: $(ASM_INST).yml $(ASM_UP)
python $(ASM_UP)

test: test-all test-z80 test-asm test-dasm

test-all:


+ 406
- 0
scripts/update_asm_instructions.py View File

@@ -0,0 +1,406 @@
#!/usr/bin/env python
# -*- coding: utf-8 -*-

# Copyright (C) 2014-2015 Ben Kurtovic <ben.kurtovic@gmail.com>
# Released under the terms of the MIT License. See LICENSE for details.

"""
This script generates 'src/assembler/instructions.inc.c' from
'src/assembler/instructions.yml'. It should be run automatically by make
when the latter is modified, but can also be run manually.
"""

from __future__ import print_function

from itertools import product
import re
import time

SOURCE = "src/assembler/instructions.yml"
DEST = "src/assembler/instructions.inc.c"

ENCODING = "utf8"
TAB = " " * 4

try:
import yaml
except ImportError:
print("Error: PyYAML is required (https://pypi.python.org/pypi/PyYAML)\n"
"If you don't want to rebuild {0}, do:\n`make -t {0}`".format(DEST))
exit(1)

re_date = re.compile(r"^(\s*@AUTOGEN_DATE\s*)(.*?)$", re.M)
re_inst = re.compile(
r"(/\* @AUTOGEN_INST_BLOCK_START \*/\n*)(.*?)"
r"(\n*/\* @AUTOGEN_INST_BLOCK_END \*/)", re.S)
re_lookup = re.compile(
r"(/\* @AUTOGEN_LOOKUP_BLOCK_START \*/\n*)(.*?)"
r"(\n*/\* @AUTOGEN_LOOKUP_BLOCK_END \*/)", re.S)

class Instruction(object):
"""
Represent a single ASM instruction mnemonic.
"""
ARG_TYPES = {
"register": "AT_REGISTER",
"immediate": "AT_IMMEDIATE",
"indirect": "AT_INDIRECT",
"indexed": "AT_INDEXED",
"condition": "AT_CONDITION",
"port": "AT_PORT"
}
PSEUDO_TYPES = {
"indirect_hl_or_indexed": ["AT_INDIRECT", "AT_INDEXED"]
}

def __init__(self, name, data):
self._name = name
self._data = data
self._has_optional_args = False

def _get_arg_parse_mask(self, num):
"""
Return the appropriate mask to parse_args() for the num-th argument.
"""
types = set()
optional = False
for case in self._data["cases"]:
if num < len(case["type"]):
atype = case["type"][num]
if atype in self.ARG_TYPES:
types.add(self.ARG_TYPES[atype])
else:
types.update(self.PSEUDO_TYPES[atype])
else:
optional = True

if not types:
return "AT_NONE"
if optional:
types.add("AT_OPTIONAL")
self._has_optional_args = True
return "|".join(sorted(types))

def _handle_return(self, ret, indent=1):
"""
Return code to handle an instruction return statement.
"""
data = ", ".join("0x%02X" % byte if isinstance(byte, int) else byte
for byte in ret)
return TAB * indent + "INST_RETURN({0}, {1})".format(len(ret), data)

def _build_case_type_check(self, args):
"""
Return the test part of an if statement for an instruction case.
"""
conds = ["INST_TYPE({0}) == {1}".format(i, self.ARG_TYPES[cond])
for i, cond in enumerate(args)]
check = " && ".join(conds)

if self._has_optional_args:
return "INST_NARGS == {0} && ".format(len(args)) + check
return check

def _build_register_check(self, num, cond):
"""
Return an expression to check for a particular register value.
"""
return "INST_REG({0}) == REG_{1}".format(num, cond.upper())

def _build_immediate_check(self, num, cond):
"""
Return an expression to check for a particular immediate value.
"""
if "." in cond:
itype, value = cond.split(".", 1)
try:
value = int(value)
except ValueError:
value = int(value, 16)
vtype = "sval" if itype.upper() in ["S8", "REL"] else "uval"

test1 = "INST_IMM({0}).mask & IMM_{1}".format(num, itype.upper())
if (itype.upper() == "U16"):
test1 += " && !INST_IMM({0}).is_label".format(num)
test2 = "INST_IMM({0}).{1} == {2}".format(num, vtype, value)
return "({0} && {1})".format(test1, test2)

return "INST_IMM({0}).mask & IMM_{1}".format(num, cond.upper())

def _build_indirect_check(self, num, cond):
"""
Return an expression to check for a particular indirect value.
"""
if cond.startswith("reg."):
test1 = "INST_INDIRECT({0}).type == AT_REGISTER".format(num)
test2 = "INST_INDIRECT({0}).addr.reg == REG_{1}".format(
num, cond[len("reg."):].upper())
return "({0} && {1})".format(test1, test2)

if cond == "imm" or cond == "immediate":
return "INST_INDIRECT({0}).type == AT_IMMEDIATE".format(num)

err = "Unknown condition for indirect argument: {0}"
return RuntimeError(err.format(cond))

def _build_indexed_check(self, num, cond):
"""
Return an expression to check for a particular indexed value.
"""
raise RuntimeError("The indexed arg type does not support conditions")

def _build_condition_check(self, num, cond):
"""
Return an expression to check for a particular condition value.
"""
return "INST_COND({0}) == COND_{1}".format(num, cond.upper())

def _build_port_check(self, num, cond):
"""
Return an expression to check for a particular port value.
"""
if cond == "reg" or cond == "reg.c":
return "INST_PORT({0}).type == AT_REGISTER".format(num)
if cond == "imm" or cond == "immediate":
return "INST_PORT({0}).type == AT_IMMEDIATE".format(num)

err = "Unknown condition for port argument: {0}"
return RuntimeError(err.format(cond))

_SUBCASE_LOOKUP_TABLE = {
"register": _build_register_check,
"immediate": _build_immediate_check,
"indirect": _build_indirect_check,
"indexed": _build_indexed_check,
"condition": _build_condition_check,
"port": _build_port_check
}

def _build_subcase_check(self, types, conds):
"""
Return the test part of an if statement for an instruction subcase.
"""
conds = [self._SUBCASE_LOOKUP_TABLE[types[i]](self, i, cond)
for i, cond in enumerate(conds) if cond != "_"]
return " && ".join(conds)

def _iter_permutations(self, types, conds):
"""
Iterate over all permutations of the given subcase conditions.
"""
def split(typ, cond):
if "|" in cond:
splits = [split(typ, c) for c in cond.split("|")]
merged = [choice for s in splits for choice in s]
if len(merged) != len(set(merged)):
msg = "Repeated conditions for {0}: {1}"
raise RuntimeError(msg.format(typ, cond))
return merged
if typ == "register":
if cond == "i":
return ["ix", "iy"]
if cond == "ih":
return ["ixh", "iyh"]
if cond == "il":
return ["ixl", "iyl"]
return [cond]

splits = [split(typ, cond) for typ, cond in zip(types, conds)]
num = max(len(cond) for cond in splits)

if any(1 < len(cond) < num for cond in splits):
msg = "Invalid condition permutations: {0}"
raise RuntimeError(msg.format(conds))

choices = [cond * num if len(cond) == 1 else cond for cond in splits]
return zip(*choices)

def _adapt_return(self, types, conds, ret):
"""
Return a modified byte list to accomodate for prefixes and immediates.
"""
ret = ret[:]
for i, byte in enumerate(ret):
if not isinstance(byte, int):
if byte == "u8":
try:
index = types.index("immediate")
imm = "INST_IMM({0})".format(index)
except ValueError:
index = types.index("port")
imm = "INST_PORT({0}).port.imm".format(index)
ret[i] = imm + ".uval"

elif byte == "u16":
if i < len(ret) - 1:
raise RuntimeError("U16 return byte must be last")
try:
index = types.index("immediate")
imm = "INST_IMM({0})".format(index)
except ValueError:
indir = types.index("indirect")
if not conds[indir].startswith("imm"):
msg = "Passing non-immediate indirect as immediate"
raise RuntimeError(msg)
imm = "INST_INDIRECT({0}).addr.imm".format(indir)
ret[i] = "INST_IMM_U16_B1({0})".format(imm)
ret.append("INST_IMM_U16_B2({0})".format(imm))
break

elif byte == "rel":
index = types.index("immediate")
ret[i] = "INST_IMM({0}).sval - 2".format(index)

else:
msg = "Unsupported return byte: {0}"
raise RuntimeError(msg.format(byte))

for i, cond in enumerate(conds):
if types[i] == "register" and cond[0] == "i":
prefix = "INST_I{0}_PREFIX".format(cond[1].upper())
if ret[0] != prefix:
ret.insert(0, prefix)
elif types[i] == "indexed":
ret.insert(0, "INST_INDEX_PREFIX({0})".format(i))
ret.insert(2, "INST_INDEX({0}).offset".format(i))

return ret

def _handle_null_case(self, case):
"""
Return code to handle an instruction case that takes no arguments.
"""
return [
TAB + "if (INST_NARGS == 0) {",
self._handle_return(case["return"], 2),
TAB + "}"
]

def _handle_pseudo_case(self, pseudo, case):
"""
Return code to handle an instruction pseudo-case.

Pseudo-cases are cases that have pseudo-types as arguments. This means
they are expanded to cover multiple "real" argument types.
"""
index = case["type"].index(pseudo)

if pseudo == "indirect_hl_or_indexed":
case["type"][index] = "indexed"
indexed = self._handle_case(case)

case["type"][index] = "indirect"
indirect = self._handle_case(case)
base_cond = self._build_case_type_check(case["type"])
hl_reg = TAB * 3 + self._build_indirect_check(index, "reg.hl")
indirect[0] = TAB + "if ({0} &&\n{1}) {{".format(base_cond, hl_reg)

return indirect + indexed

raise RuntimeError("Unknown pseudo-type: {0}".format(pseudo))

def _handle_case(self, case):
"""
Return code to handle an instruction case.
"""
ctype = case["type"]
if not ctype:
return self._handle_null_case(case)

for pseudo in self.PSEUDO_TYPES:
if pseudo in ctype:
return self._handle_pseudo_case(pseudo, case)

lines = []
cond = self._build_case_type_check(ctype)
lines.append(TAB + "if ({0}) {{".format(cond))

subcases = [(perm, sub["return"]) for sub in case["cases"]
for perm in self._iter_permutations(ctype, sub["cond"])]
for cond, ret in subcases:
check = self._build_subcase_check(ctype, cond)
ret = self._adapt_return(ctype, cond, ret)
if check:
lines.append(TAB * 2 + "if ({0})".format(check))
lines.append(self._handle_return(ret, 3))
else:
lines.append(self._handle_return(ret, 2))
break # Unconditional subcase
else:
lines.append(TAB * 2 + "INST_ERROR(ARG_VALUE)")

lines.append(TAB + "}")
return lines

def render(self):
"""
Convert data for an individual instruction into a C parse function.
"""
lines = []

if self._data["args"]:
lines.append("{tab}INST_TAKES_ARGS(\n{tab2}{0},\n{tab2}{1},"
"\n{tab2}{2}\n{tab})".format(
self._get_arg_parse_mask(0), self._get_arg_parse_mask(1),
self._get_arg_parse_mask(2), tab=TAB, tab2=TAB * 2))
else:
lines.append(TAB + "INST_TAKES_NO_ARGS")

if "return" in self._data:
lines.append(self._handle_return(self._data["return"]))
elif "cases" in self._data:
for case in self._data["cases"]:
lines.extend(self._handle_case(case))
lines.append(TAB + "INST_ERROR(ARG_TYPE)")
else:
msg = "Missing return or case block for {0} instruction"
raise RuntimeError(msg.format(self._name))

contents = "\n".join(lines)
return "INST_FUNC({0})\n{{\n{1}\n}}".format(self._name, contents)


def build_inst_block(data):
"""
Return the instruction parser block, given instruction data.
"""
return "\n\n".join(
Instruction(k, v).render() for k, v in sorted(data.items()))

def build_lookup_block(data):
"""
Return the instruction lookup block, given instruction data.
"""
macro = TAB + "HANDLE({0})"
return "\n".join(macro.format(inst) for inst in sorted(data.keys()))

def process(template, data):
"""
Return C code generated from a source template and instruction data.
"""
inst_block = build_inst_block(data)
lookup_block = build_lookup_block(data)
date = time.asctime(time.gmtime())

result = re_date.sub(r"\1{0} UTC".format(date), template)
result = re_inst.sub(r"\1{0}\3".format(inst_block), result)
result = re_lookup.sub(r"\1{0}\3".format(lookup_block), result)
return result

def main():
"""
Main script entry point.
"""
with open(SOURCE, "r") as fp:
text = fp.read().decode(ENCODING)
with open(DEST, "r") as fp:
template = fp.read().decode(ENCODING)

data = yaml.load(text)
result = process(template, data)

with open(DEST, "w") as fp:
fp.write(result.encode(ENCODING))

if __name__ == "__main__":
main()

+ 2
- 5
src/assembler/errors.c View File

@@ -53,11 +53,8 @@ static const char *error_descs[] = {
[ED_PS_TOO_FEW_ARGS] = "too few arguments for opcode",
[ED_PS_TOO_MANY_ARGS] = "too many arguments for opcode",
[ED_PS_ARG_SYNTAX] = "invalid syntax in argument(s)",
[ED_PS_ARG0_TYPE] = "invalid type for first argument",
[ED_PS_ARG0_BAD_REG] = "unsupported register as first argument",
[ED_PS_ARG1_TYPE] = "invalid type for second argument",
[ED_PS_ARG1_BAD_REG] = "unsupported register as second argument",
[ED_PS_ARG1_RANGE] = "second argument out of range"
[ED_PS_ARG_TYPE] = "invalid argument type",
[ED_PS_ARG_VALUE] = "invalid value for argument"
};

/* Internal structs */


+ 2
- 5
src/assembler/errors.h View File

@@ -52,11 +52,8 @@ typedef enum {
ED_PS_TOO_FEW_ARGS,
ED_PS_TOO_MANY_ARGS,
ED_PS_ARG_SYNTAX,
ED_PS_ARG0_TYPE,
ED_PS_ARG0_BAD_REG,
ED_PS_ARG1_TYPE,
ED_PS_ARG1_BAD_REG,
ED_PS_ARG1_RANGE
ED_PS_ARG_TYPE,
ED_PS_ARG_VALUE
} ASMErrorDesc;

/* Structs */


+ 20
- 13
src/assembler/inst_args.h View File

@@ -8,12 +8,14 @@
#define MAX_SYMBOL_SIZE 256

typedef enum {
AT_REGISTER,
AT_IMMEDIATE,
AT_INDIRECT,
AT_INDEXED,
AT_LABEL,
AT_CONDITION
AT_NONE = 0x00,
AT_OPTIONAL = 0x01,
AT_REGISTER = 0x02,
AT_IMMEDIATE = 0x04,
AT_INDIRECT = 0x08,
AT_INDEXED = 0x10,
AT_CONDITION = 0x20,
AT_PORT = 0x40
} ASMArgType;

typedef enum {
@@ -35,20 +37,17 @@ typedef enum {

typedef struct {
ASMArgImmType mask;
bool is_label;
uint16_t uval;
int16_t sval;
char label[MAX_SYMBOL_SIZE];
} ASMArgImmediate;

typedef struct {
char text[MAX_SYMBOL_SIZE];
} ASMArgLabel;

typedef struct {
ASMArgType type;
union {
ASMArgRegister reg;
ASMArgImmediate imm;
ASMArgLabel label;
} addr;
} ASMArgIndirect;

@@ -58,7 +57,7 @@ typedef struct {
} ASMArgIndexed;

typedef enum {
COND_NZ, COND_N, COND_NC, COND_C, COND_PO, COND_PE, COND_P, COND_M
COND_NZ, COND_Z, COND_NC, COND_C, COND_PO, COND_PE, COND_P, COND_M
} ASMArgCondition;

typedef struct {
@@ -66,9 +65,17 @@ typedef struct {
union {
ASMArgRegister reg;
ASMArgImmediate imm;
} port;
} ASMArgPort;

typedef struct {
ASMArgType type;
union {
ASMArgRegister reg;
ASMArgImmediate imm;
ASMArgIndirect indirect;
ASMArgIndexed index;
ASMArgLabel label;
ASMArgCondition cond;
ASMArgPort port;
} data;
} ASMInstArg;

+ 0
- 98
src/assembler/inst_support.c View File

@@ -1,98 +0,0 @@
/* Copyright (C) 2014-2015 Ben Kurtovic <ben.kurtovic@gmail.com>
Released under the terms of the MIT License. See LICENSE for details. */

#include "inst_support.h"

/* Macro used by parse_arg() */

#define TRY_PARSER(func, argtype, field) \
if (argparse_##func(&arg->data.field, info)) { \
arg->type = argtype; \
return ED_NONE; \
}

/*
Fill an instruction's byte array with the given data.

This internal function is only called for instructions longer than four
bytes (of which there is only one: the fake emulator debugging/testing
opcode with mnemonic "emu"), so it does not get used in normal situations.

Return the value of the last byte inserted, for compatibility with the
INST_SETn_ family of macros.
*/
uint8_t fill_bytes_variadic(uint8_t *bytes, size_t len, ...)
{
va_list vargs;
va_start(vargs, len);
for (size_t i = 0; i < len; i++)
bytes[i] = va_arg(vargs, unsigned);
va_end(vargs);
return bytes[len - 1];
}

/*
Parse a single instruction argument into an ASMInstArg object.

Return ED_NONE (0) on success or an error code on failure.
*/
static ASMErrorDesc parse_arg(
ASMInstArg *arg, const char *str, size_t size, ASMDefineTable *deftable)
{
ASMArgParseInfo info = {.arg = str, .size = size, .deftable = deftable};
TRY_PARSER(register, AT_REGISTER, reg)
TRY_PARSER(immediate, AT_IMMEDIATE, imm)
TRY_PARSER(indirect, AT_INDIRECT, indirect)
TRY_PARSER(indexed, AT_INDEXED, index)
TRY_PARSER(condition, AT_CONDITION, cond)
TRY_PARSER(label, AT_LABEL, label)
return ED_PS_ARG_SYNTAX;
}

/*
Parse an argument string into ASMInstArg objects.

Return ED_NONE (0) on success or an error code on failure.
*/
ASMErrorDesc parse_args(
ASMInstArg args[3], size_t *nargs, ASMArgParseInfo ap_info)
{
ASMErrorDesc err;
ASMDefineTable *dt = ap_info.deftable;
const char *str = ap_info.arg;
size_t size = ap_info.size, start = 0, i = 0;

while (i < size) {
char c = str[i];
if (c == ',') {
if (i == start)
return ED_PS_ARG_SYNTAX;
if ((err = parse_arg(&args[*nargs], str + start, i - start, dt)))
return err;
(*nargs)++;

i++;
if (i < size && str[i] == ' ')
i++;
start = i;
if (i == size)
return ED_PS_ARG_SYNTAX;
if (*nargs >= 3)
return ED_PS_TOO_MANY_ARGS;
} else {
if ((c >= 'a' && c <= 'z') || (c >= '0' && c <= '9') ||
c == ' ' || c == '+' || c == '-' || c == '(' || c == ')' ||
c == '$' || c == '_' || c == '.')
i++;
else
return ED_PS_ARG_SYNTAX;
}
}

if (i > start) {
if ((err = parse_arg(&args[*nargs], str + start, i - start, dt)))
return err;
(*nargs)++;
}
return ED_NONE;
}

+ 0
- 155
src/assembler/inst_support.h View File

@@ -1,155 +0,0 @@
/* Copyright (C) 2014-2015 Ben Kurtovic <ben.kurtovic@gmail.com>
Released under the terms of the MIT License. See LICENSE for details. */

#pragma once

#include <stdarg.h>
#include <stdlib.h>

#include "errors.h"
#include "inst_args.h"
#include "parse_util.h"
#include "../util.h"

/* Helper macros for get_inst_parser() */

#define JOIN_(a, b, c, d) ((uint32_t) ((a << 24) + (b << 16) + (c << 8) + d))

#define DISPATCH_(s, z) ( \
(z) == 2 ? JOIN_(s[0], s[1], 0x00, 0x00) : \
(z) == 3 ? JOIN_(s[0], s[1], s[2], 0x00) : \
JOIN_(s[0], s[1], s[2], s[3])) \

#define MAKE_CMP_(s) DISPATCH_(s, sizeof(s) / sizeof(char) - 1)

#define HANDLE(m) if (key == MAKE_CMP_(#m)) return parse_inst_##m;

/* Internal helper macros */

#define INST_ALLOC_(len) \
*length = len; \
*bytes = cr_malloc(sizeof(uint8_t) * (len));

#define INST_SET_(b, val) ((*bytes)[b] = val)
#define INST_SET1_(b1) INST_SET_(0, b1)
#define INST_SET2_(b1, b2) INST_SET1_(b1), INST_SET_(1, b2)
#define INST_SET3_(b1, b2, b3) INST_SET2_(b1, b2), INST_SET_(2, b3)
#define INST_SET4_(b1, b2, b3, b4) INST_SET3_(b1, b2, b3), INST_SET_(3, b4)

#define INST_DISPATCH_(a, b, c, d, target, ...) target

#define INST_FILL_BYTES_(len, ...) \
((len > 4) ? fill_bytes_variadic(*bytes, len, __VA_ARGS__) : \
INST_DISPATCH_(__VA_ARGS__, INST_SET4_, INST_SET3_, INST_SET2_, \
INST_SET1_, __VA_ARGS__)(__VA_ARGS__));

#define INST_IX_PREFIX_ 0xDD
#define INST_IY_PREFIX_ 0xFD

#define INST_PREFIX_(reg) \
(((reg) == REG_IX || (reg) == REG_IXH || (reg) == REG_IXL) ? \
INST_IX_PREFIX_ : INST_IY_PREFIX_)

#define INST_RETURN_WITH_SYMBOL_(len, label, ...) { \
*symbol = cr_strdup(label.text); \
INST_ALLOC_(len) \
INST_FILL_BYTES_(len - 2, __VA_ARGS__) \
return ED_NONE; \
}

/* Essential/basic helper macros */

#define INST_FUNC(mnemonic) \
static ASMErrorDesc parse_inst_##mnemonic( \
uint8_t **bytes, size_t *length, char **symbol, ASMArgParseInfo ap_info) \

#define INST_ERROR(desc) return ED_PS_##desc;

#define INST_TAKES_NO_ARGS \
if (ap_info.arg) \
INST_ERROR(TOO_MANY_ARGS)

#define INST_TAKES_ARGS(lo, hi) \
if (!ap_info.arg) \
INST_ERROR(TOO_FEW_ARGS) \
ASMInstArg args[3]; \
size_t nargs = 0; \
ASMErrorDesc err = parse_args(args, &nargs, ap_info); \
if (err) \
return err; \
if (nargs < lo) \
INST_ERROR(TOO_FEW_ARGS) \
if (nargs > hi) \
INST_ERROR(TOO_MANY_ARGS)

#define INST_NARGS nargs
#define INST_TYPE(n) args[n].type
#define INST_REG(n) args[n].data.reg
#define INST_IMM(n) args[n].data.imm
#define INST_INDIRECT(n) args[n].data.indirect
#define INST_INDEX(n) args[n].data.index
#define INST_LABEL(n) args[n].data.label
#define INST_COND(n) args[n].data.cond

#define INST_RETURN(len, ...) { \
(void) symbol; \
INST_ALLOC_(len) \
INST_FILL_BYTES_(len, __VA_ARGS__) \
return ED_NONE; \
}

/* Convenience macros */

#define INST_FORCE_TYPE(n, t) { \
if (INST_TYPE(n) != t) \
INST_ERROR(ARG##n##_TYPE) \
}

#define INST_CHECK_IMM(n, m) { \
if (!(INST_IMM(n).mask & (m))) \
INST_ERROR(ARG##n##_RANGE) \
}

#define INST_REG_ONLY(n, reg) { \
if (INST_TYPE(n) != AT_REGISTER) \
INST_ERROR(ARG##n##_TYPE) \
if (INST_REG(n) != reg) \
INST_ERROR(ARG##n##_BAD_REG) \
}

#define INST_INDIRECT_HL_ONLY(n) { \
if (INST_INDIRECT(n).type != AT_REGISTER) \
INST_ERROR(ARG##n##_TYPE) \
if (INST_INDIRECT(n).addr.reg != REG_HL) \
INST_ERROR(ARG##n##_BAD_REG) \
}

#define INST_INDEX_PREFIX(n) INST_PREFIX_(INST_INDEX(n).reg)

#define INST_INDEX_BYTES(n, b) \
INST_INDEX_PREFIX(n), b, INST_INDEX(n).offset

#define INST_INDIRECT_IMM(n) \
INST_INDIRECT(n).addr.imm.uval >> 8, \
INST_INDIRECT(n).addr.imm.uval & 0xFF

#define INST_RETURN_INDIRECT_LABEL(n, len, ...) \
INST_RETURN_WITH_SYMBOL_(len, INST_INDIRECT(n).addr.label, __VA_ARGS__)

#define INST_HANDLE_MAIN_8_BIT_REGS(base) \
case REG_A: INST_RETURN(1, base + 7) \
case REG_B: INST_RETURN(1, base) \
case REG_C: INST_RETURN(1, base + 1) \
case REG_D: INST_RETURN(1, base + 2) \
case REG_E: INST_RETURN(1, base + 3) \
case REG_H: INST_RETURN(1, base + 4) \
case REG_L: INST_RETURN(1, base + 5) \
case REG_IXH: INST_RETURN(2, INST_IX_PREFIX_, base + 4) \
case REG_IXL: INST_RETURN(2, INST_IX_PREFIX_, base + 5) \
case REG_IYH: INST_RETURN(2, INST_IY_PREFIX_, base + 4) \
case REG_IYL: INST_RETURN(2, INST_IY_PREFIX_, base + 5) \

/* Internal functions */

uint8_t fill_bytes_variadic(uint8_t*, size_t, ...);
ASMErrorDesc parse_args(ASMInstArg args[3], size_t*, ASMArgParseInfo);

+ 170
- 744
src/assembler/instructions.c View File

@@ -1,705 +1,203 @@
/* Copyright (C) 2014-2015 Ben Kurtovic <ben.kurtovic@gmail.com>
Released under the terms of the MIT License. See LICENSE for details. */

#include "instructions.h"
#include "inst_support.h"

/* Instruction parser functions */

INST_FUNC(adc)
{
INST_TAKES_ARGS(2, 2)
INST_FORCE_TYPE(0, AT_REGISTER)
switch (INST_REG(0)) {
case REG_A:
switch (INST_TYPE(1)) {
case AT_REGISTER:
switch (INST_REG(1)) {
INST_HANDLE_MAIN_8_BIT_REGS(0x88)
default: INST_ERROR(ARG1_BAD_REG)
}
case AT_IMMEDIATE:
INST_CHECK_IMM(1, IMM_U8)
INST_RETURN(2, 0xCE, INST_IMM(1).uval)
case AT_INDIRECT:
INST_INDIRECT_HL_ONLY(1)
INST_RETURN(1, 0x8E)
case AT_INDEXED:
INST_RETURN(3, INST_INDEX_BYTES(1, 0x8E))
default:
INST_ERROR(ARG1_TYPE)
}
case REG_HL:
INST_FORCE_TYPE(1, AT_REGISTER)
switch (INST_REG(1)) {
case REG_BC: INST_RETURN(2, 0xED, 0x4A)
case REG_DE: INST_RETURN(2, 0xED, 0x5A)
case REG_HL: INST_RETURN(2, 0xED, 0x6A)
case REG_SP: INST_RETURN(2, 0xED, 0x7A)
default: INST_ERROR(ARG1_BAD_REG)
}
default:
INST_ERROR(ARG0_TYPE)
}
}

INST_FUNC(add)
{
INST_TAKES_ARGS(2, 2)
INST_FORCE_TYPE(0, AT_REGISTER)
switch (INST_REG(0)) {
case REG_A:
switch (INST_TYPE(1)) {
case AT_REGISTER:
switch (INST_REG(1)) {
INST_HANDLE_MAIN_8_BIT_REGS(0x80)
default: INST_ERROR(ARG1_BAD_REG)
}
case AT_IMMEDIATE:
INST_CHECK_IMM(1, IMM_U8)
INST_RETURN(2, 0xC6, INST_IMM(1).uval)
case AT_INDIRECT:
INST_INDIRECT_HL_ONLY(1)
INST_RETURN(1, 0x86)
case AT_INDEXED:
INST_RETURN(3, INST_INDEX_BYTES(1, 0x86))
default:
INST_ERROR(ARG1_TYPE)
}
case REG_HL:
INST_FORCE_TYPE(1, AT_REGISTER)
switch (INST_REG(1)) {
case REG_BC: INST_RETURN(1, 0x09)
case REG_DE: INST_RETURN(1, 0x19)
case REG_HL: INST_RETURN(1, 0x29)
case REG_SP: INST_RETURN(1, 0x39)
default: INST_ERROR(ARG1_BAD_REG)
}
case REG_IX:
case REG_IY:
INST_FORCE_TYPE(1, AT_REGISTER)
switch (INST_REG(1)) {
case REG_BC: INST_RETURN(2, INST_INDEX_PREFIX(1), 0x09)
case REG_DE: INST_RETURN(2, INST_INDEX_PREFIX(1), 0x19)
case REG_IX:
case REG_IY:
if (INST_REG(0) != INST_REG(1))
INST_ERROR(ARG1_BAD_REG)
INST_RETURN(2, INST_INDEX_PREFIX(1), 0x29)
case REG_SP: INST_RETURN(2, INST_INDEX_PREFIX(1), 0x39)
default: INST_ERROR(ARG1_BAD_REG)
}
default:
INST_ERROR(ARG0_TYPE)
}
}

INST_FUNC(and)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}

INST_FUNC(bit)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}

INST_FUNC(call)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}

INST_FUNC(ccf)
{
INST_TAKES_NO_ARGS
INST_RETURN(1, 0x3F)
}

INST_FUNC(cp)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}

INST_FUNC(cpd)
{
INST_TAKES_NO_ARGS
INST_RETURN(2, 0xED, 0xA9)
}

INST_FUNC(cpdr)
{
INST_TAKES_NO_ARGS
INST_RETURN(2, 0xED, 0xB9)
}

INST_FUNC(cpi)
{
INST_TAKES_NO_ARGS
INST_RETURN(2, 0xED, 0xA1)
}

INST_FUNC(cpir)
{
INST_TAKES_NO_ARGS
INST_RETURN(2, 0xED, 0xB1)
}

INST_FUNC(cpl)
{
INST_TAKES_NO_ARGS
INST_RETURN(1, 0x2F)
}

INST_FUNC(daa)
{
INST_TAKES_NO_ARGS
INST_RETURN(1, 0x27)
}

INST_FUNC(dec)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}

INST_FUNC(di)
{
INST_TAKES_NO_ARGS
INST_RETURN(1, 0xF3)
}
#include <stdarg.h>
#include <stdlib.h>

INST_FUNC(djnz)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}
#include "instructions.h"
#include "inst_args.h"
#include "../util.h"

INST_FUNC(ei)
{
INST_TAKES_NO_ARGS
INST_RETURN(1, 0xFB)
}
/* Helper macros for get_inst_parser() and lookup_parser() */

INST_FUNC(ex)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}
#define JOIN(a, b, c, d) ((uint32_t) ((a << 24) + (b << 16) + (c << 8) + d))

INST_FUNC(exx)
{
INST_TAKES_NO_ARGS
INST_RETURN(1, 0xD9)
}
#define DISPATCH_(s, z) ( \
(z) == 2 ? JOIN(s[0], s[1], 0x00, 0x00) : \
(z) == 3 ? JOIN(s[0], s[1], s[2], 0x00) : \
JOIN(s[0], s[1], s[2], s[3])) \

INST_FUNC(halt)
{
INST_TAKES_NO_ARGS
INST_RETURN(1, 0x76)
}
#define MAKE_CMP_(s) DISPATCH_(s, sizeof(s) / sizeof(char) - 1)

INST_FUNC(im)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}
#define HANDLE(m) if (key == MAKE_CMP_(#m)) return parse_inst_##m;

INST_FUNC(in)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}
/* Helper macro for parse_arg() */

INST_FUNC(inc)
{
INST_TAKES_ARGS(1, 1)
switch (INST_TYPE(0)) {
case AT_REGISTER:
switch (INST_REG(0)) {
case REG_A: INST_RETURN(1, 0x3C)
case REG_B: INST_RETURN(1, 0x04)
case REG_C: INST_RETURN(1, 0x0C)
case REG_D: INST_RETURN(1, 0x14)
case REG_E: INST_RETURN(1, 0x1C)
case REG_H: INST_RETURN(1, 0x24)
case REG_L: INST_RETURN(1, 0x2C)
case REG_BC: INST_RETURN(1, 0x03)
case REG_DE: INST_RETURN(1, 0x13)
case REG_HL: INST_RETURN(1, 0x23)
case REG_SP: INST_RETURN(1, 0x33)
case REG_IX: INST_RETURN(2, 0xDD, 0x23)
case REG_IY: INST_RETURN(2, 0xFD, 0x23)
case REG_IXH: INST_RETURN(2, 0xDD, 0x2C)
case REG_IXL: INST_RETURN(2, 0xFD, 0x2C)
case REG_IYH: INST_RETURN(2, 0xDD, 0x2C)
case REG_IYL: INST_RETURN(2, 0xFD, 0x2C)
default: INST_ERROR(ARG0_BAD_REG)
}
case AT_INDIRECT:
INST_INDIRECT_HL_ONLY(0)
INST_RETURN(1, 0x34)
case AT_INDEXED:
INST_RETURN(3, INST_INDEX_BYTES(0, 0x34))
default:
INST_ERROR(ARG0_TYPE)
#define TRY_PARSER(func, argtype, field) \
if (mask & argtype && argparse_##func(&arg->data.field, info)) { \
arg->type = argtype; \
return ED_NONE; \
}
}

INST_FUNC(ind)
{
INST_TAKES_NO_ARGS
INST_RETURN(2, 0xED, 0xAA)
}

INST_FUNC(indr)
{
INST_TAKES_NO_ARGS
INST_RETURN(2, 0xED, 0xBA)
}

INST_FUNC(ini)
{
INST_TAKES_NO_ARGS
INST_RETURN(2, 0xED, 0xA2)
}

INST_FUNC(inir)
{
INST_TAKES_NO_ARGS
INST_RETURN(2, 0xED, 0xB2)
}

INST_FUNC(jp)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}

INST_FUNC(jr)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}

INST_FUNC(ld)
{
uint8_t base;
INST_TAKES_ARGS(2, 2)
switch (INST_TYPE(0)) {
case AT_REGISTER:
switch (INST_REG(0)) {
case REG_A:
switch (INST_TYPE(1)) {
case AT_REGISTER:
switch (INST_REG(1)) {
INST_HANDLE_MAIN_8_BIT_REGS(0x78)
case REG_I: INST_RETURN(2, 0xED, 0x57)
case REG_R: INST_RETURN(2, 0xED, 0x5F)
default: INST_ERROR(ARG1_BAD_REG)
}
case AT_IMMEDIATE:
INST_CHECK_IMM(1, IMM_U8)
INST_RETURN(2, 0x3E, INST_IMM(1).uval)
case AT_INDIRECT:
switch (INST_INDIRECT(1).type) {
case AT_REGISTER:
switch (INST_INDIRECT(1).addr.reg) {
case REG_BC: INST_RETURN(1, 0x0A)
case REG_DE: INST_RETURN(1, 0x1A)
case REG_HL: INST_RETURN(1, 0x7E)
default: INST_ERROR(ARG0_BAD_REG)
}
case AT_IMMEDIATE:
INST_RETURN(3, 0x3A, INST_INDIRECT_IMM(1))
case AT_LABEL:
INST_RETURN_INDIRECT_LABEL(1, 3, 0x3A)
default:
INST_ERROR(ARG1_TYPE)
}
case AT_INDEXED:
INST_RETURN(3, INST_INDEX_BYTES(1, 0x7E))
default:
INST_ERROR(ARG1_TYPE)
}
case REG_B:
base = 0x00;
case REG_C:
base = 0x08;
case REG_D:
base = 0x10;
case REG_E:
base = 0x18;
switch (INST_TYPE(1)) {
case AT_REGISTER:
switch (INST_REG(1)) {
INST_HANDLE_MAIN_8_BIT_REGS(base + 0x40)
default: INST_ERROR(ARG1_BAD_REG)
}
case AT_IMMEDIATE:
INST_CHECK_IMM(1, IMM_U8)
INST_RETURN(2, base + 0x06, INST_IMM(1).uval)
case AT_INDIRECT:
INST_INDIRECT_HL_ONLY(1)
INST_RETURN(1, base + 0x46)
case AT_INDEXED:
INST_RETURN(3, INST_INDEX_BYTES(1, base + 0x46))
default:
INST_ERROR(ARG1_TYPE)
}
case REG_H: // TODO (11 cases)
case REG_L: // TODO (11 cases)
case REG_I:
INST_REG_ONLY(1, REG_A)
INST_RETURN(2, 0xED, 0x47)
case REG_R:
INST_REG_ONLY(1, REG_A)
INST_RETURN(2, 0xED, 0x4F)
case REG_BC: // TODO ( 2 cases)
case REG_DE: // TODO ( 2 cases)
case REG_HL: // TODO ( 3 cases)
case REG_IX: // TODO ( 2 cases)
case REG_IY: // TODO ( 2 cases)
case REG_SP: // TODO ( 5 cases)
case REG_IXH: // TODO ( 8 cases)
case REG_IXL: // TODO ( 8 cases)
case REG_IYH: // TODO ( 8 cases)
case REG_IYL: // TODO ( 8 cases)
default: INST_ERROR(ARG0_BAD_REG)
}
case AT_INDIRECT:
switch (INST_INDIRECT(0).type) {
case AT_REGISTER:
switch (INST_INDIRECT(0).addr.reg) {
case REG_BC: // TODO (1 case )
case REG_DE: // TODO (1 case )
case REG_HL: // TODO (8 cases)
default: INST_ERROR(ARG0_BAD_REG)
}
case AT_IMMEDIATE:
// TODO (8 cases)
case AT_LABEL:
// TODO (same 8 cases)
default:
INST_ERROR(ARG0_TYPE)
}
case AT_INDEXED:
// TODO (16 cases)
default:
INST_ERROR(ARG0_TYPE)
/* Internal helper macros */

#define INST_ALLOC_(len) \
*length = len; \
*bytes = cr_malloc(sizeof(uint8_t) * (len));

#define INST_SET_(b, val) ((*bytes)[b] = val)
#define INST_SET1_(b1) INST_SET_(0, b1)
#define INST_SET2_(b1, b2) INST_SET1_(b1), INST_SET_(1, b2)
#define INST_SET3_(b1, b2, b3) INST_SET2_(b1, b2), INST_SET_(2, b3)
#define INST_SET4_(b1, b2, b3, b4) INST_SET3_(b1, b2, b3), INST_SET_(3, b4)

#define INST_DISPATCH_(a, b, c, d, target, ...) target

#define INST_FILL_BYTES_(len, ...) \
((len > 4) ? fill_bytes_variadic(*bytes, len, __VA_ARGS__) : \
INST_DISPATCH_(__VA_ARGS__, INST_SET4_, INST_SET3_, INST_SET2_, \
INST_SET1_, __VA_ARGS__)(__VA_ARGS__));

#define INST_PREFIX_(reg) \
(((reg) == REG_IX || (reg) == REG_IXH || (reg) == REG_IXL) ? \
INST_IX_PREFIX : INST_IY_PREFIX)

/* Helper macros for instruction parsers */

#define INST_FUNC(mnemonic) \
static ASMErrorDesc parse_inst_##mnemonic( \
uint8_t **bytes, size_t *length, char **symbol, ASMArgParseInfo ap_info) \

#define INST_ERROR(desc) return ED_PS_##desc;

#define INST_TAKES_NO_ARGS \
if (ap_info.arg) \
INST_ERROR(TOO_MANY_ARGS)

#define INST_TAKES_ARGS(a0, a1, a2) \
ASMInstArg args[3]; \
size_t nargs; \
ASMArgType masks[] = {a0, a1, a2}; \
ASMErrorDesc err = parse_args(args, &nargs, ap_info, masks); \
if (err) \
return err; \

#define INST_NARGS nargs
#define INST_TYPE(n) args[n].type
#define INST_REG(n) args[n].data.reg
#define INST_IMM(n) args[n].data.imm
#define INST_INDIRECT(n) args[n].data.indirect
#define INST_INDEX(n) args[n].data.index
#define INST_COND(n) args[n].data.cond
#define INST_PORT(n) args[n].data.port

#define INST_IX_PREFIX 0xDD
#define INST_IY_PREFIX 0xFD

#define INST_RETURN(len, ...) { \
(void) symbol; \
INST_ALLOC_(len) \
INST_FILL_BYTES_(len, __VA_ARGS__) \
return ED_NONE; \
}
}

INST_FUNC(ldd)
{
INST_TAKES_NO_ARGS
INST_RETURN(2, 0xED, 0xA8)
}

INST_FUNC(lddr)
{
INST_TAKES_NO_ARGS
INST_RETURN(2, 0xED, 0xB8)
}

INST_FUNC(ldi)
{
INST_TAKES_NO_ARGS
INST_RETURN(2, 0xED, 0xA0)
}

INST_FUNC(ldir)
{
INST_TAKES_NO_ARGS
INST_RETURN(2, 0xED, 0xB0)
}

INST_FUNC(neg)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}

INST_FUNC(nop)
{
INST_TAKES_NO_ARGS
INST_RETURN(1, 0x00)
}

INST_FUNC(or)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}

INST_FUNC(otdr)
{
INST_TAKES_NO_ARGS
INST_RETURN(2, 0xED, 0xBB)
}

INST_FUNC(otir)
{
INST_TAKES_NO_ARGS
INST_RETURN(2, 0xED, 0xB3)
}

INST_FUNC(out)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}

INST_FUNC(outd)
{
INST_TAKES_NO_ARGS
INST_RETURN(2, 0xED, 0xAB)
}

INST_FUNC(outi)
{
INST_TAKES_NO_ARGS
INST_RETURN(2, 0xED, 0xA3)
}

INST_FUNC(pop)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}

INST_FUNC(push)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}

INST_FUNC(res)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}

INST_FUNC(ret)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}

INST_FUNC(reti)
{
INST_TAKES_NO_ARGS
INST_RETURN(2, 0xED, 0x4D)
}

INST_FUNC(retn)
{
INST_TAKES_NO_ARGS
INST_RETURN(2, 0xED, 0x45)
}

INST_FUNC(rl)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}

INST_FUNC(rla)
{
INST_TAKES_NO_ARGS
INST_RETURN(1, 0x17)
}

INST_FUNC(rlc)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}
#define INST_IMM_U16_B1(imm) \
((imm).is_label ? (*symbol = cr_strdup((imm).label), 0) : (imm).uval >> 8)
#define INST_IMM_U16_B2(imm) \
((imm).is_label ? 0 : (imm).uval & 0xFF)

INST_FUNC(rlca)
{
INST_TAKES_NO_ARGS
INST_RETURN(1, 0x07)
}
#define INST_INDEX_PREFIX(n) INST_PREFIX_(INST_INDEX(n).reg)

INST_FUNC(rld)
{
INST_TAKES_NO_ARGS
INST_RETURN(2, 0xED, 0x6F)
}
/* ----------------------------- END WORK BLOCK ---------------------------- */

INST_FUNC(rr)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}

INST_FUNC(rra)
{
INST_TAKES_NO_ARGS
INST_RETURN(1, 0x1F)
}

INST_FUNC(rrc)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}

INST_FUNC(rrca)
{
INST_TAKES_NO_ARGS
INST_RETURN(1, 0x0F)
}

INST_FUNC(rrd)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}

INST_FUNC(rst)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}

INST_FUNC(sbc)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}

INST_FUNC(scf)
{
INST_TAKES_NO_ARGS
INST_RETURN(1, 0x37)
}
/*
Fill an instruction's byte array with the given data.

INST_FUNC(set)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}
This internal function is only called for instructions longer than four
bytes (of which there is only one: the fake emulator debugging/testing
opcode with mnemonic "emu"), so it does not get used in normal situations.

INST_FUNC(sl1)
Return the value of the last byte inserted, for compatibility with the
INST_SETn_ family of macros.
*/
static uint8_t fill_bytes_variadic(uint8_t *bytes, size_t len, ...)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
va_list vargs;
va_start(vargs, len);
for (size_t i = 0; i < len; i++)
bytes[i] = va_arg(vargs, unsigned);
va_end(vargs);
return bytes[len - 1];
}

INST_FUNC(sla)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}
/*
Parse a single instruction argument into an ASMInstArg object.

INST_FUNC(sll)
Return ED_NONE (0) on success or an error code on failure.
*/
static ASMErrorDesc parse_arg(
ASMInstArg *arg, const char *str, size_t size, ASMDefineTable *deftable,
ASMArgType mask)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
ASMArgParseInfo info = {.arg = str, .size = size, .deftable = deftable};
TRY_PARSER(register, AT_REGISTER, reg)
TRY_PARSER(condition, AT_CONDITION, cond)
TRY_PARSER(indexed, AT_INDEXED, index)
TRY_PARSER(indirect, AT_INDIRECT, indirect)
TRY_PARSER(port, AT_PORT, port)
TRY_PARSER(immediate, AT_IMMEDIATE, imm)
return ED_PS_ARG_SYNTAX;
}

INST_FUNC(sls)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}
/*
Parse an argument string into ASMInstArg objects.

INST_FUNC(sra)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}
Return ED_NONE (0) on success or an error code on failure.
*/
static ASMErrorDesc parse_args(
ASMInstArg args[3], size_t *nargs, ASMArgParseInfo ap_info,
ASMArgType masks[3])
{
ASMErrorDesc err;
ASMDefineTable *dt = ap_info.deftable;
const char *str = ap_info.arg;
size_t size = ap_info.size, start = 0, i = 0, n = 0;

while (i < size) {
char c = str[i];
if (c == ',') {
if (i == start)
return ED_PS_ARG_SYNTAX;
if (masks[n] == AT_NONE)
return ED_PS_TOO_MANY_ARGS;
err = parse_arg(&args[n], str + start, i - start, dt, masks[n]);
if (err)
return err;
n++;

i++;
if (i < size && str[i] == ' ')
i++;
start = i;
if (i == size)
return ED_PS_ARG_SYNTAX;
if (n >= 3)
return ED_PS_TOO_MANY_ARGS;
} else {
if ((c >= 'a' && c <= 'z') || (c >= '0' && c <= '9') ||
c == ' ' || c == '+' || c == '-' || c == '(' || c == ')' ||
c == '$' || c == '_' || c == '.')
i++;
else
return ED_PS_ARG_SYNTAX;
}
}

INST_FUNC(srl)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}
if (i > start) {
if (masks[n] == AT_NONE)
return ED_PS_TOO_MANY_ARGS;
if ((err = parse_arg(&args[n], str + start, i - start, dt, masks[n])))
return err;
n++;
}

INST_FUNC(sub)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
if (n < 3 && masks[n] != AT_NONE && !(masks[n] & AT_OPTIONAL))
return ED_PS_TOO_FEW_ARGS;
*nargs = n;
return ED_NONE;
}

INST_FUNC(xor)
{
// TODO
INST_TAKES_NO_ARGS
INST_ERROR(ARG_SYNTAX)
INST_RETURN(1, 0xFF)
}
#include "instructions.inc.c"

/*
Return the relevant ASMInstParser function for a given mnemonic.
@@ -710,78 +208,6 @@ ASMInstParser get_inst_parser(char mstr[MAX_MNEMONIC_SIZE])
{
// Exploit the fact that we can store the entire mnemonic string as a
// single 32-bit value to do fast lookups:
uint32_t key = (mstr[0] << 24) + (mstr[1] << 16) + (mstr[2] << 8) + mstr[3];

HANDLE(adc)
HANDLE(add)
HANDLE(and)
HANDLE(bit)
HANDLE(call)
HANDLE(ccf)
HANDLE(cp)
HANDLE(cpd)
HANDLE(cpdr)
HANDLE(cpi)
HANDLE(cpir)
HANDLE(cpl)
HANDLE(daa)
HANDLE(dec)
HANDLE(di)
HANDLE(djnz)
HANDLE(ei)
HANDLE(ex)
HANDLE(exx)
HANDLE(halt)
HANDLE(im)
HANDLE(in)
HANDLE(inc)
HANDLE(ind)
HANDLE(indr)
HANDLE(ini)
HANDLE(inir)
HANDLE(jp)
HANDLE(jr)
HANDLE(ld)
HANDLE(ldd)
HANDLE(lddr)
HANDLE(ldi)
HANDLE(ldir)
HANDLE(neg)
HANDLE(nop)
HANDLE(or)
HANDLE(otdr)
HANDLE(otir)
HANDLE(out)
HANDLE(outd)
HANDLE(outi)
HANDLE(pop)
HANDLE(push)
HANDLE(res)
HANDLE(ret)
HANDLE(reti)
HANDLE(retn)
HANDLE(rl)
HANDLE(rla)
HANDLE(rlc)
HANDLE(rlca)
HANDLE(rld)
HANDLE(rr)
HANDLE(rra)
HANDLE(rrc)
HANDLE(rrca)
HANDLE(rrd)
HANDLE(rst)
HANDLE(sbc)
HANDLE(scf)
HANDLE(set)
HANDLE(sl1)
HANDLE(sla)
HANDLE(sll)
HANDLE(sls)
HANDLE(sra)
HANDLE(srl)
HANDLE(sub)
HANDLE(xor)

return NULL;
uint32_t key = JOIN(mstr[0], mstr[1], mstr[2], mstr[3]);
return lookup_parser(key);
}

+ 2784
- 0
src/assembler/instructions.inc.c
File diff suppressed because it is too large
View File


+ 1859
- 0
src/assembler/instructions.yml
File diff suppressed because it is too large
View File


+ 76
- 44
src/assembler/parse_util.c View File

@@ -57,29 +57,31 @@ static bool adjust_for_indirection(ASMArgParseInfo *ap_info)
/*
Calculate the mask field for an ASMArgImmediate based on its uval/sval.
*/
static void calculate_immediate_mask(ASMArgImmediate *imm)
static ASMArgImmType calculate_immediate_mask(uint32_t uval, bool negative)
{
imm->mask = 0;
if (imm->sval < 0) {
if (imm->sval >= INT8_MIN)
imm->mask |= IMM_S8;
if (imm->sval >= INT8_MIN + 2)
imm->mask |= IMM_REL;
ASMArgImmType mask = 0;
if (negative && uval != 0) {
int32_t sval = -uval;
if (sval >= INT8_MIN)
mask |= IMM_S8;
if (sval >= INT8_MIN + 2)
mask |= IMM_REL;
} else {
imm->mask = IMM_U16;
if (imm->uval <= UINT8_MAX)
imm->mask |= IMM_U8;
if (imm->uval <= INT8_MAX)
imm->mask |= IMM_S8;
if (imm->uval <= INT8_MAX + 2)
imm->mask |= IMM_REL;
if (imm->uval <= 7)
imm->mask |= IMM_BIT;
if (!(imm->uval & ~0x38))
imm->mask |= IMM_RST;
if (imm->uval <= 2)
imm->mask |= IMM_IM;
mask = IMM_U16;
if (uval <= UINT8_MAX)
mask |= IMM_U8;
if (uval <= INT8_MAX)
mask |= IMM_S8;
if (uval <= INT8_MAX + 2)
mask |= IMM_REL;
if (uval <= 7)
mask |= IMM_BIT;
if (!(uval & ~0x38))
mask |= IMM_RST;
if (uval <= 2)
mask |= IMM_IM;
}
return mask;
}

/*
@@ -300,7 +302,7 @@ bool argparse_condition(ASMArgCondition *result, ASMArgParseInfo ai)
switch (ai.size) {
case 1:
switch (buf[0]) {
case 'n': return (*result = COND_N), true;
case 'z': return (*result = COND_Z), true;
case 'c': return (*result = COND_C), true;
case 'p': return (*result = COND_P), true;
case 'm': return (*result = COND_M), true;
@@ -319,6 +321,26 @@ bool argparse_condition(ASMArgCondition *result, ASMArgParseInfo ai)
}

/*
Read in a label immediate argument and store it in *result.
*/
static bool argparse_imm_label(ASMArgImmediate *result, ASMArgParseInfo ai)
{
if (ai.size >= MAX_SYMBOL_SIZE)
return false;

for (const char *i = ai.arg; i < ai.arg + ai.size; i++) {
if (!is_valid_symbol_char(*i, i == ai.arg))
return false;
}

result->mask = IMM_U16;
result->is_label = true;
strncpy(result->label, ai.arg, ai.size);
result->label[ai.size] = '\0';
return true;
}

/*
Read in an immediate argument and store it in *result.
*/
bool argparse_immediate(ASMArgImmediate *result, ASMArgParseInfo ai)
@@ -326,10 +348,11 @@ bool argparse_immediate(ASMArgImmediate *result, ASMArgParseInfo ai)
if (ai.size <= 0)
return false;

bool negative = false;
bool negative = false, modifiers = false;
ssize_t i = 0;

while (ai.arg[i] == '-' || ai.arg[i] == '+' || ai.arg[i] == ' ') {
modifiers = true;
if (ai.arg[i] == '-')
negative = !negative;
if (++i >= ai.size)
@@ -341,9 +364,10 @@ bool argparse_immediate(ASMArgImmediate *result, ASMArgParseInfo ai)
const ASMDefine *define = asm_deftable_find(ai.deftable, ai.arg, ai.size);
if (define) {
if (negative) {
result->is_label = false;
result->uval = define->value.uval;
result->sval = -define->value.sval;
calculate_immediate_mask(result);
result->mask = calculate_immediate_mask(result->uval, true);
} else {
*result = define->value;
}
@@ -351,16 +375,23 @@ bool argparse_immediate(ASMArgImmediate *result, ASMArgParseInfo ai)
}

uint32_t uval;
if (!parse_uint32_t(&uval, ai.arg, ai.size) || uval > UINT16_MAX)
if (!parse_uint32_t(&uval, ai.arg, ai.size)) {
if (!modifiers && argparse_imm_label(result, ai))
return true;
return false;
}

if (uval > UINT16_MAX)
return false;

int32_t sval = negative ? -uval : uval;
if (sval < INT16_MIN)
return false;

result->is_label = false;
result->uval = uval;
result->sval = sval;
calculate_immediate_mask(result);
result->mask = calculate_immediate_mask(uval, negative);
return true;
}

@@ -374,9 +405,9 @@ bool argparse_indirect(ASMArgIndirect *result, ASMArgParseInfo ai)

ASMArgRegister reg;
ASMArgImmediate imm;
ASMArgLabel label;
if (argparse_register(&reg, ai)) {
if (reg == REG_BC || reg == REG_DE || reg == REG_HL) {
if (reg == REG_BC || reg == REG_DE || reg == REG_HL ||
reg == REG_IX || reg == REG_IY) {
result->type = AT_REGISTER;
result->addr.reg = reg;
return true;
@@ -387,10 +418,6 @@ bool argparse_indirect(ASMArgIndirect *result, ASMArgParseInfo ai)
result->addr.imm = imm;
return true;
}
} else if (argparse_label(&label, ai)) {
result->type = AT_LABEL;
result->addr.label = label;
return true;
}
return false;
}
@@ -433,24 +460,29 @@ bool argparse_indexed(ASMArgIndexed *result, ASMArgParseInfo ai)
}

/*
Read in a label argument and store it in *result.
Read in a port argument and store it in *result.
*/
bool argparse_label(ASMArgLabel *result, ASMArgParseInfo ai)
bool argparse_port(ASMArgPort *result, ASMArgParseInfo ai)
{
if (ai.size <= 0 || ai.size >= MAX_SYMBOL_SIZE)
if (ai.size < 3 || !adjust_for_indirection(&ai))
return false;

for (const char *i = ai.arg; i < ai.arg + ai.size; i++) {
if (!is_valid_symbol_char(*i, i == ai.arg))
return false;
ASMArgRegister reg;
ASMArgImmediate imm;
if (argparse_register(&reg, ai)) {
if (reg == REG_C) {
result->type = AT_REGISTER;
result->port.reg = reg;
return true;
}
} else if (argparse_immediate(&imm, ai)) {
if (imm.mask & IMM_U8) {
result->type = AT_IMMEDIATE;
result->port.imm = imm;
return true;
}
}

if (asm_deftable_find(ai.deftable, ai.arg, ai.size))
return false;

strncpy(result->text, ai.arg, ai.size);
result->text[ai.size] = '\0';
return true;
return false;
}

/*


+ 1
- 1
src/assembler/parse_util.h View File

@@ -33,7 +33,7 @@ bool argparse_condition(ASMArgCondition*, ASMArgParseInfo);
bool argparse_immediate(ASMArgImmediate*, ASMArgParseInfo);
bool argparse_indirect(ASMArgIndirect*, ASMArgParseInfo);
bool argparse_indexed(ASMArgIndexed*, ASMArgParseInfo);
bool argparse_label(ASMArgLabel*, ASMArgParseInfo);
bool argparse_port(ASMArgPort*, ASMArgParseInfo);

/* Preprocessor directive parsers */
bool dparse_bool(bool*, const ASMLine*, const char*);


+ 1
- 1
src/assembler/tokenizer.c View File

@@ -155,7 +155,7 @@ static ErrorInfo* handle_define_directive(
ASMArgImmediate imm;
ASMArgParseInfo info = {
.arg = line->data + i, .size = line->length - i, .deftable = deftab};
if (!argparse_immediate(&imm, info))
if (!argparse_immediate(&imm, info) || imm.is_label)
return error_info_create(line, ET_PREPROC, ED_PP_BAD_ARG);

ASMDefine *define = cr_malloc(sizeof(ASMDefine));


Write Preview
Loading…
Cancel
Save