Initial commit

3 files changed, 310 insertions, 0 deletions

diff --git a/.gitignore b/.gitignore
new file mode 100644
index 0000000..1dabaad
--- /dev/null
+++ b/.gitignore
@@ -0,0 +1,5 @@
+*.o
+main
+m68k-trasm
+cmake[-_]build*/
+build*/
diff --git a/Makefile b/Makefile
new file mode 100644
index 0000000..5b1111e
--- /dev/null
+++ b/Makefile
@@ -0,0 +1,26 @@
+# SPDX-License-Identifier: Unlicense
+
+WARNFLAGS = -Wall -Wextra -pedantic -Wlogical-op
+OPTFLAGS = -O2
+INCLUDES = lib
+_CFLAGS = $(CFLAGS) $(WARNFLAGS) $(addprefix -I,$(INCLUDES)) $(OPTFLAGS) -pipe -g
+_CXXFLAGS = $(CXXFLAGS) $(WARNFLAGS) $(addprefix -I,$(INCLUDES)) $(OPTFLAGS) -pipe -g
+LDSCRIPTS =
+_LDFLAGS = $(LDFLAGS) $(OPTFLAGS) $(addprefix -T,$(LDSCRIPTS))
+
+OBJECTS=main.o
+
+.PHONY: all
+all: m68k-trasm
+
+m68k-trasm: $(OBJECTS) $(LDSCRIPTS) Makefile
+	$(CC) -o $@ $(_LDFLAGS) $(OBJECTS)
+
+$(OBJECTS): Makefile
+
+%.o: %.c Makefile
+	$(CC) $(_CFLAGS) -c -o $@ $<
+
+clean:
+	rm -rfv m68k-trasm $(OBJECTS)
+
diff --git a/main.c b/main.c
new file mode 100644
index 0000000..3020862
--- /dev/null
+++ b/main.c
@@ -0,0 +1,279 @@
+/* SPDX-License-Identifier: Unlicense
+ *
+ * This program translates Sierra m68k assembly dialect to GNU AS m68k dialect.
+ *
+ * NOTE: Unicode is not supported, ASCII only.
+ */
+
+#include <assert.h>
+#include <errno.h>
+#include <stdbool.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdint.h>
+
+#define OK 0
+#define ERR 2
+
+enum tok_kind {
+    TK_NONE = 0,
+    TK_SPACE,
+    TK_TAB,
+    TK_CR,
+    TK_LF,
+    TK_DOT,
+    TK_COMMA,
+    TK_PLUS,
+    TK_MINUS,
+    TK_EQ,
+    TK_COLON,
+    TK_ASTERISK,
+    TK_STRING,
+    TK_ALNUM,
+    TK_NUMDEC,
+    TK_NUMOCT,
+    TK_NUMHEX,
+    TK_PARENL,
+    TK_PARENR,
+    TK_COMMENT_ASTERISK,
+    TK_COMMENT_SEMICOLON,
+};
+
+struct tok {
+    enum tok_kind kind;
+    size_t offset;
+    size_t length;
+};
+
+enum lex_error {
+    LE_NONE = 0,
+};
+
+struct lex {
+    // Input data buffer
+    FILE *parsed_input_stream;
+    char *parsed_input;
+    size_t parsed_input_size;
+    // Tokens table
+    FILE *tokbuf_stream;
+    struct tok *tokbuf;
+    size_t tokbuf_size;
+    // State
+    size_t offset;
+    enum lex_error error;
+};
+
+enum stmt_kind {
+    SK_NONE = 0,
+    SK_LABEL,
+    SK_INSTRUCTION,
+    SK_TEXT,
+    SK_DIR_FILE,
+    SK_DIR_TEXT,
+    SK_DIR_ALIGN,
+    SK_DIR_DEF_ENDEF,
+    SK_DIR_GLOBL,
+    SK_DIR_LINE,
+};
+
+enum opcode {
+    OPCODE_NONE,
+    OPCODE_NOP,
+};
+
+enum opsize {
+    OPSIZE_NONE = 0,
+    OPSIZE_S,
+    OPSIZE_B,
+    OPSIZE_W,
+    OPSIZE_L,
+};
+
+enum arg_kind {
+    ARG_NONE = 0,
+    ARG_DN,
+    ARG_AN,
+    ARG_AN_ADDR,
+    ARG_AN_ADDR_INCR,
+    ARG_AN_ADDR_DECR,
+    ARG_AN_ADDR_16,
+    ARG_AN_ADDR_8_XN,
+    ARG_ADDR_WORD,
+    ARG_ADDR_LONG,
+    ARG_ADDR_UNSPEC,
+    ARG_PC_ADDR_16,
+    ARG_PC_ADDR_8_XN,
+    ARG_IMMEDIATE,
+};
+
+struct arg_8_xn {
+    int8_t val;
+    int8_t an;
+    int8_t xi;
+};
+
+struct instruction {
+    enum opcode opcode;
+    enum opsize opsize;
+    enum arg_kind arg1_kind, arg2_kind;
+    union {
+        int32_t imm, addr;
+        struct arg_8_xn arg_8_xn; // For (d,An,Xi) and (d,PC,Xn)
+    } arg1, arg2;
+};
+
+struct def_endef {
+    size_t sym_id;
+    size_t tag_sym_id;
+    int32_t size;
+    int32_t storage_class;
+    int32_t type;
+};
+
+struct stmt {
+    enum stmt_kind kind;
+    union {
+        struct instruction instruction;
+        int32_t align;
+        size_t globl_sym_id;
+        size_t file_sym_id;
+    };
+    size_t first_token, num_tokens; // Statement tokens span, may be NULL
+    size_t comment_token;
+};
+
+struct symbol {
+    size_t offset; // Byte offset in continuous null terminated symbol buffer
+    // Instead of strcmp every item in symtab we can compare hashes and get O(N)
+    // for search.
+    uint32_t hash;
+};
+
+struct pars {
+    struct lex *lex;
+    // Statement table
+    FILE *stmttab_stream;
+    struct stmt *stmttab;
+    size_t stmttab_size;
+    // Symbol table
+    FILE *symtab_stream;
+    struct sym *symtab;
+    size_t symtab_size;
+    // Symbol buffer for symbol table
+    FILE *symbuf_stream;
+    char *symbuf;
+    size_t symbuf_size;
+};
+
+struct assem {
+    struct pars *pars;
+};
+
+static int lex_init(struct lex *const self)
+{
+    (void) self;
+    return OK;
+}
+
+static int lex_next(struct lex *const self)
+{
+    (void) self;
+    return OK;
+}
+
+static void lex_destroy(struct lex *const self)
+{
+    (void) self;
+}
+
+static int pars_init(struct pars *const self, struct lex *const lex)
+{
+    (void) self;
+    (void) lex;
+    return OK;
+}
+
+static int pars_run(struct pars *const self, FILE *const stream)
+{
+    (void) self;
+    (void) stream;
+    return OK;
+}
+
+static void pars_destroy(struct pars *const self)
+{
+    (void) self;
+}
+
+static int assem_init(struct assem *const self, struct pars *const pars)
+{
+    (void) self;
+    (void) pars;
+    return OK;
+}
+
+static int assem_resolve(struct assem *const self)
+{
+    (void) self;
+    return OK;
+}
+
+static int assem_emit(struct assem *const self, FILE *const stream)
+{
+    (void) self;
+    (void) stream;
+    return OK;
+}
+
+static void assem_destroy(struct assem *const self)
+{
+    (void) self;
+}
+
+int main(const int argc, char *const argv[])
+{
+    // No fucks given about arguments for now
+    (void)argc;
+    (void)argv;
+    struct lex lex;
+    struct pars pars;
+    if (OK != lex_init(&lex)) {
+        return EXIT_FAILURE;
+    }
+    // Chain lexer and parser
+    if (OK != pars_init(&pars, &lex)) {
+        lex_destroy(&lex);
+        return EXIT_FAILURE;
+    }
+    // Parse assembly program text (tokenize + parse)
+    if (OK != pars_run(&pars, stdin)) {
+        pars_destroy(&pars);
+        lex_destroy(&lex);
+        return EXIT_FAILURE;
+    }
+    struct assem assem;
+    // Allocate adn populate code table and metadata table from parsed data
+    if (OK != assem_init(&assem, &pars)) {
+        pars_destroy(&pars);
+        lex_destroy(&lex);
+        return EXIT_FAILURE;
+    }
+    // Resolve all ambiguities
+    if (OK != assem_resolve(&assem)) {
+        assem_destroy(&assem);
+        pars_destroy(&pars);
+        lex_destroy(&lex);
+        return EXIT_FAILURE;
+    }
+    // Emit unambiguous assembly language program text for specified dialect
+    // (currently m68k GNU AS only is supported)
+    if (OK != assem_emit(&assem, stdout)) {
+        assem_destroy(&assem);
+        pars_destroy(&pars);
+        lex_destroy(&lex);
+        return EXIT_FAILURE;
+    }
+    assem_destroy(&assem);
+    pars_destroy(&pars);
+    lex_destroy(&lex);
+}