From 857cb72a94c366998ae6bf21df9072d3343d0769 Mon Sep 17 00:00:00 2001 From: Dale Weiler Date: Fri, 20 Apr 2012 00:44:44 -0400 Subject: [PATCH] Starting assembly --- assembler.c | 222 ++++++++++++---------------------------------------- code.c | 24 ++++-- gmqcc.h | 73 +++++++++++++++++ 3 files changed, 141 insertions(+), 178 deletions(-) diff --git a/assembler.c b/assembler.c index e64fb9c..6a4ec46 100644 --- a/assembler.c +++ b/assembler.c @@ -21,84 +21,6 @@ * SOFTWARE. */ #include "gmqcc.h" -/* - * This is the assembler, gmqas, this is being implemented because I'm - * not exactly sure how codegen would work for the C compiler as of yet - * and also I plan to allow inline assembly for the compiler. - */ -static const struct { - const char *m; /* menomic */ - const size_t o; /* operands */ - const size_t l; /* menomic len */ -} const asm_instr[] = { - [INSTR_DONE] = { "DONE" , 1, 4 }, - [INSTR_MUL_F] = { "MUL_F" , 3, 5 }, - [INSTR_MUL_V] = { "MUL_V" , 3, 5 }, - [INSTR_MUL_FV] = { "MUL_FV" , 3, 6 }, - [INSTR_MUL_VF] = { "MUL_VF" , 3, 6 }, - [INSTR_DIV_F] = { "DIV" , 0, 3 }, - [INSTR_ADD_F] = { "ADD_F" , 3, 5 }, - [INSTR_ADD_V] = { "ADD_V" , 3, 5 }, - [INSTR_SUB_F] = { "SUB_F" , 3, 5 }, - [INSTR_SUB_V] = { "DUB_V" , 3, 5 }, - [INSTR_EQ_F] = { "EQ_F" , 0, 4 }, - [INSTR_EQ_V] = { "EQ_V" , 0, 4 }, - [INSTR_EQ_S] = { "EQ_S" , 0, 4 }, - [INSTR_EQ_E] = { "EQ_E" , 0, 4 }, - [INSTR_EQ_FNC] = { "ES_FNC" , 0, 6 }, - [INSTR_NE_F] = { "NE_F" , 0, 4 }, - [INSTR_NE_V] = { "NE_V" , 0, 4 }, - [INSTR_NE_S] = { "NE_S" , 0, 4 }, - [INSTR_NE_E] = { "NE_E" , 0, 4 }, - [INSTR_NE_FNC] = { "NE_FNC" , 0, 6 }, - [INSTR_LE] = { "LE" , 0, 2 }, - [INSTR_GE] = { "GE" , 0, 2 }, - [INSTR_LT] = { "LT" , 0, 2 }, - [INSTR_GT] = { "GT" , 0, 2 }, - [INSTR_LOAD_F] = { "FIELD_F" , 0, 7 }, - [INSTR_LOAD_V] = { "FIELD_V" , 0, 7 }, - [INSTR_LOAD_S] = { "FIELD_S" , 0, 7 }, - [INSTR_LOAD_ENT] = { "FIELD_ENT" , 0, 9 }, - [INSTR_LOAD_FLD] = { "FIELD_FLD" , 0, 9 }, - [INSTR_LOAD_FNC] = { "FIELD_FNC" , 0, 9 }, - [INSTR_ADDRESS] = { "ADDRESS" , 0, 7 }, - [INSTR_STORE_F] = { "STORE_F" , 0, 7 }, - [INSTR_STORE_V] = { "STORE_V" , 0, 7 }, - [INSTR_STORE_S] = { "STORE_S" , 0, 7 }, - [INSTR_STORE_ENT] = { "STORE_ENT" , 0, 9 }, - [INSTR_STORE_FLD] = { "STORE_FLD" , 0, 9 }, - [INSTR_STORE_FNC] = { "STORE_FNC" , 0, 9 }, - [INSTR_STOREP_F] = { "STOREP_F" , 0, 8 }, - [INSTR_STOREP_V] = { "STOREP_V" , 0, 8 }, - [INSTR_STOREP_S] = { "STOREP_S" , 0, 8 }, - [INSTR_STOREP_ENT] = { "STOREP_ENT", 0, 10}, - [INSTR_STOREP_FLD] = { "STOREP_FLD", 0, 10}, - [INSTR_STOREP_FNC] = { "STOREP_FNC", 0, 10}, - [INSTR_RETURN] = { "RETURN" , 0, 6 }, - [INSTR_NOT_F] = { "NOT_F" , 0, 5 }, - [INSTR_NOT_V] = { "NOT_V" , 0, 5 }, - [INSTR_NOT_S] = { "NOT_S" , 0, 5 }, - [INSTR_NOT_ENT] = { "NOT_ENT" , 0, 7 }, - [INSTR_NOT_FNC] = { "NOT_FNC" , 0, 7 }, - [INSTR_IF] = { "IF" , 0, 2 }, - [INSTR_IFNOT] = { "IFNOT" , 0, 5 }, - [INSTR_CALL0] = { "CALL0" , 0, 5 }, - [INSTR_CALL1] = { "CALL1" , 0, 5 }, - [INSTR_CALL2] = { "CALL2" , 0, 5 }, - [INSTR_CALL3] = { "CALL3" , 0, 5 }, - [INSTR_CALL4] = { "CALL4" , 0, 5 }, - [INSTR_CALL5] = { "CALL5" , 0, 5 }, - [INSTR_CALL6] = { "CALL6" , 0, 5 }, - [INSTR_CALL7] = { "CALL7" , 0, 5 }, - [INSTR_CALL8] = { "CALL8" , 0, 5 }, - [INSTR_STATE] = { "STATE" , 0, 5 }, - [INSTR_GOTO] = { "GOTO" , 0, 4 }, - [INSTR_AND] = { "AND" , 0, 3 }, - [INSTR_OR] = { "OR" , 0, 2 }, - [INSTR_BITAND] = { "BITAND" , 0, 6 }, - [INSTR_BITOR] = { "BITOR" , 0, 5 } -}; - /* * Some assembler keywords not part of the opcodes above: these are * for creating functions, or constants. @@ -162,6 +84,31 @@ void asm_clear() { mem_d(assembly_constants_data); } +int asm_parsetype(const char *key, char **skip, long line) { + size_t keylen = strlen(key); + if (!strncmp(key, *skip, keylen)) { + if ((*skip)[keylen] != ':'){ + printf("%li: Missing `:` after decltype\n", line); + exit(1); + } + *skip += keylen+1; + while (**skip == ' ' || **skip == '\t') + (*skip)++; + + if (!isalpha(**skip)) { + printf("%li: Invalid identififer: %s\n", line, *skip); + exit(1); + } else { + assembly_constants_add((globals) { + .name = util_strdup("empty"), + .offset = code_globals_elements + }); + return 1; + } + } + return 0; +} + void asm_parse(FILE *fp) { char *data = NULL; char *skip = NULL; @@ -169,116 +116,44 @@ void asm_parse(FILE *fp) { size_t size = 0; /* size of line */ asm_state state = ASM_NULL; - while ((data = asm_getline(&size, fp)) != NULL) { - skip = data; - asm_skipwhite(skip); - asm_rmnewline(skip, &size); - - #define DECLTYPE(X, CODE) \ - if (!strncmp(X, skip, strlen(X))) { \ - if (skip[strlen(X)] != ':') { \ - printf("%li: Missing `:` after decltype\n",line); \ - exit (1); \ - } \ - skip += strlen(X)+1; \ - asm_skipwhite(skip); \ - if(!isalpha(*skip)) { \ - printf("%li: Invalid identififer: %s\n", line, skip); \ - exit (1); \ - } else { \ - size_t offset_code = code_statements_elements+1; \ - size_t offset_chars = code_chars_elements +1; \ - size_t offset_globals = code_globals_elements +1; \ - size_t offset_functions = code_functions_elements +1; \ - size_t offset_fields = code_fields_elements +1; \ - size_t offset_defs = code_defs_elements +1; \ - CODE \ - /* silent unused warnings */ \ - (void)offset_code; \ - (void)offset_chars; \ - (void)offset_globals; \ - (void)offset_functions; \ - (void)offset_fields; \ - (void)offset_defs; \ - assembly_constants_add((globals){ \ - .name = util_strdup("empty"), \ - .offset = offset_globals \ - }); \ - } \ - goto end; \ - } + while ((data = skip = asm_getline(&size, fp)) != NULL) { + /* remove any whitespace at start */ + while (*skip == ' ' || *skip == '\t') + skip++; + /* remove newline at end of string */ + *(skip+*(&size)-1) = '\0'; - /* FLOAT */ - DECLTYPE(asm_keys[0], { - //util_debug("ASM", "Constant FLOAT\n"); - code_defs_add((prog_section_def){ - .type = TYPE_FLOAT, - .offset = offset_globals, /* global table */ - .name = offset_chars /* string table TODO */ - }); - float f = 0; /*TODO*/ - code_globals_add(*(int*)&f); - - }); - /* VECTOR */ - DECLTYPE(asm_keys[1], { - //util_debug("ASM", "Constant VECTOR\n"); - code_defs_add((prog_section_def){ - .type = TYPE_FLOAT, - .offset = offset_globals, /* global table */ - .name = offset_chars /* string table TODO */ - }); - float f1 = 0; - float f2 = 0; - float f3 = 0; - code_globals_add(*(int*)&f1); - code_globals_add(*(int*)&f2); - code_globals_add(*(int*)&f3); - }); - /* ENTITY */ DECLTYPE(asm_keys[2], {util_debug("ASM", "Constant ENTITY\n");}); - /* FIELD */ DECLTYPE(asm_keys[3], {util_debug("ASM", "Constant FIELD\n");}); - /* STRING */ - DECLTYPE(asm_keys[4], { - //util_debug("ASM", "Constant STRING\n"); - code_defs_add((prog_section_def){ - .type = TYPE_STRING, - .offset = offset_globals, /* offset to offset in string table (for data)*/ - .name = offset_chars /* location of name in string table (for name)*/ - }); - code_strings_add("hello world"); - }); - /* FUNCTION */ - DECLTYPE(asm_keys[5], { - //util_debug("ASM", "Constant FUNCTION\n"); - /* TODO: parse */ + if (asm_parsetype(asm_keys[5], &skip, line)) { if (state != ASM_NULL) { printf("%li: Error unfinished function block, expected DONE or RETURN\n", line); goto end; } - state = ASM_FUNCTION; code_defs_add((prog_section_def){ .type = TYPE_VOID, - .offset = offset_globals, - .name = offset_chars + .offset = code_globals_elements, + .name = code_chars_elements }); - code_globals_add(offset_functions); - code_functions_add((prog_section_function){ - .entry = offset_code, + code_globals_add(code_functions_elements); + code_functions_add((prog_section_function) { + .entry = code_statements_elements, .firstlocal = 0, .locals = 0, .profile = 0, - .name = offset_chars, + .name = code_chars_elements, .file = 0, .nargs = 0, .argsize = {0} }); - }); - + code_strings_add(skip); + }; + + #if 0 /* if we make it this far then we have statements */ { size_t i = 0; /* counter */ size_t o = 0; /* operands */ + size_t c = 0; /* copy */ char *t = NULL; /* token */ /* @@ -300,6 +175,7 @@ void asm_parse(FILE *fp) { } /* parse the statement */ + c = i; o = asm_instr[i].o; /* operands */ skip += asm_instr[i].l; /* skip instruction */ t = strtok(skip, " ,"); @@ -310,10 +186,6 @@ void asm_parse(FILE *fp) { i ++; } - // util_debug("ASM", "Operand 1: %s\n", util_strrnl(op[0])); - // util_debug("ASM", "Operand 2: %s\n", util_strrnl(op[1])); - // util_debug("ASM", "Operand 3: %s\n", util_strrnl(op[2])); - /* check */ if (i != o) { printf("not enough operands, expected: %li, got %li\n", o, i); @@ -321,12 +193,16 @@ void asm_parse(FILE *fp) { /* TODO: hashtable value LOAD .... etc */ code_statements_add((prog_section_statement){ - i, {0}, {0}, {0} + c, + { atof(op[0]) }, + { atof(op[1]) }, + { atof(op[2]) } }); goto end; } } } + #endif /* if we made it this far something is wrong */ if (*skip != '\0') diff --git a/code.c b/code.c index 4abd99b..674eb1a 100644 --- a/code.c +++ b/code.c @@ -164,7 +164,7 @@ void code_write() { fwrite(code_globals_data, 1, sizeof(int) *code_globals_elements, fp); fwrite(code_chars_data, 1, 1 *code_chars_elements, fp); - util_debug("GEN","header:\n"); + util_debug("GEN","HEADER:\n"); util_debug("GEN"," version: = %d\n", code_header.version ); util_debug("GEN"," crc16: = %d\n", code_header.crc16 ); util_debug("GEN"," entfield: = %d\n", code_header.entfield); @@ -175,10 +175,11 @@ void code_write() { util_debug("GEN"," globals = {.offset = % 8d, .length = % 8d}\n", code_header.globals .offset, code_header.globals .length); util_debug("GEN"," strings = {.offset = % 8d, .length = % 8d}\n", code_header.strings .offset, code_header.strings .length); - /* WRITE out all functions */ - util_debug("GEN", "functions:\n"); - size_t i = 0; + /* FUNCTIONS */ + util_debug("GEN", "FUNCTIONS:\n"); + size_t i = 1; for (; i < code_functions_elements; i++) { + size_t j = code_functions_data[i].entry; util_debug("GEN", " {.entry =% 5d, .firstlocal =% 5d, .locals =% 5d, .profile =% 5d, .name =% 5d, .file =% 5d, .nargs =% 5d, .argsize =%0X }\n", code_functions_data[i].entry, code_functions_data[i].firstlocal, @@ -189,6 +190,20 @@ void code_write() { code_functions_data[i].nargs, *((int32_t*)&code_functions_data[i].argsize) ); + util_debug("GEN", " NAME: %s\n", &code_chars_data[code_functions_data[i].name]); + util_debug("GEN", " CODE:\n"); + for (;;) { + if (code_statements_data[j].opcode != INSTR_DONE && + code_statements_data[j].opcode != INSTR_RETURN) + util_debug("GEN", " %s {0x%05d,0x%05d,0x%05d}\n", + asm_instr[code_statements_data[j].opcode].m, + code_statements_data[j].s1, + code_statements_data[j].s2, + code_statements_data[j].s3 + ); + else break; + j++; + } } mem_d(code_statements_data); @@ -197,6 +212,5 @@ void code_write() { mem_d(code_functions_data); mem_d(code_globals_data); mem_d(code_chars_data); - fclose(fp); } diff --git a/gmqcc.h b/gmqcc.h index 7243665..f1004eb 100644 --- a/gmqcc.h +++ b/gmqcc.h @@ -395,6 +395,79 @@ void code_init (); //=================================================================== //========================= assembler.c ============================= //=================================================================== +static const struct { + const char *m; /* menomic */ + const size_t o; /* operands */ + const size_t l; /* menomic len */ +} const asm_instr[] = { + [INSTR_DONE] = { "DONE" , 1, 4 }, + [INSTR_MUL_F] = { "MUL_F" , 3, 5 }, + [INSTR_MUL_V] = { "MUL_V" , 3, 5 }, + [INSTR_MUL_FV] = { "MUL_FV" , 3, 6 }, + [INSTR_MUL_VF] = { "MUL_VF" , 3, 6 }, + [INSTR_DIV_F] = { "DIV" , 0, 3 }, + [INSTR_ADD_F] = { "ADD_F" , 3, 5 }, + [INSTR_ADD_V] = { "ADD_V" , 3, 5 }, + [INSTR_SUB_F] = { "SUB_F" , 3, 5 }, + [INSTR_SUB_V] = { "DUB_V" , 3, 5 }, + [INSTR_EQ_F] = { "EQ_F" , 0, 4 }, + [INSTR_EQ_V] = { "EQ_V" , 0, 4 }, + [INSTR_EQ_S] = { "EQ_S" , 0, 4 }, + [INSTR_EQ_E] = { "EQ_E" , 0, 4 }, + [INSTR_EQ_FNC] = { "ES_FNC" , 0, 6 }, + [INSTR_NE_F] = { "NE_F" , 0, 4 }, + [INSTR_NE_V] = { "NE_V" , 0, 4 }, + [INSTR_NE_S] = { "NE_S" , 0, 4 }, + [INSTR_NE_E] = { "NE_E" , 0, 4 }, + [INSTR_NE_FNC] = { "NE_FNC" , 0, 6 }, + [INSTR_LE] = { "LE" , 0, 2 }, + [INSTR_GE] = { "GE" , 0, 2 }, + [INSTR_LT] = { "LT" , 0, 2 }, + [INSTR_GT] = { "GT" , 0, 2 }, + [INSTR_LOAD_F] = { "FIELD_F" , 0, 7 }, + [INSTR_LOAD_V] = { "FIELD_V" , 0, 7 }, + [INSTR_LOAD_S] = { "FIELD_S" , 0, 7 }, + [INSTR_LOAD_ENT] = { "FIELD_ENT" , 0, 9 }, + [INSTR_LOAD_FLD] = { "FIELD_FLD" , 0, 9 }, + [INSTR_LOAD_FNC] = { "FIELD_FNC" , 0, 9 }, + [INSTR_ADDRESS] = { "ADDRESS" , 0, 7 }, + [INSTR_STORE_F] = { "STORE_F" , 0, 7 }, + [INSTR_STORE_V] = { "STORE_V" , 0, 7 }, + [INSTR_STORE_S] = { "STORE_S" , 0, 7 }, + [INSTR_STORE_ENT] = { "STORE_ENT" , 0, 9 }, + [INSTR_STORE_FLD] = { "STORE_FLD" , 0, 9 }, + [INSTR_STORE_FNC] = { "STORE_FNC" , 0, 9 }, + [INSTR_STOREP_F] = { "STOREP_F" , 0, 8 }, + [INSTR_STOREP_V] = { "STOREP_V" , 0, 8 }, + [INSTR_STOREP_S] = { "STOREP_S" , 0, 8 }, + [INSTR_STOREP_ENT] = { "STOREP_ENT", 0, 10}, + [INSTR_STOREP_FLD] = { "STOREP_FLD", 0, 10}, + [INSTR_STOREP_FNC] = { "STOREP_FNC", 0, 10}, + [INSTR_RETURN] = { "RETURN" , 0, 6 }, + [INSTR_NOT_F] = { "NOT_F" , 0, 5 }, + [INSTR_NOT_V] = { "NOT_V" , 0, 5 }, + [INSTR_NOT_S] = { "NOT_S" , 0, 5 }, + [INSTR_NOT_ENT] = { "NOT_ENT" , 0, 7 }, + [INSTR_NOT_FNC] = { "NOT_FNC" , 0, 7 }, + [INSTR_IF] = { "IF" , 0, 2 }, + [INSTR_IFNOT] = { "IFNOT" , 0, 5 }, + [INSTR_CALL0] = { "CALL0" , 0, 5 }, + [INSTR_CALL1] = { "CALL1" , 0, 5 }, + [INSTR_CALL2] = { "CALL2" , 0, 5 }, + [INSTR_CALL3] = { "CALL3" , 0, 5 }, + [INSTR_CALL4] = { "CALL4" , 0, 5 }, + [INSTR_CALL5] = { "CALL5" , 0, 5 }, + [INSTR_CALL6] = { "CALL6" , 0, 5 }, + [INSTR_CALL7] = { "CALL7" , 0, 5 }, + [INSTR_CALL8] = { "CALL8" , 0, 5 }, + [INSTR_STATE] = { "STATE" , 0, 5 }, + [INSTR_GOTO] = { "GOTO" , 0, 4 }, + [INSTR_AND] = { "AND" , 0, 3 }, + [INSTR_OR] = { "OR" , 0, 2 }, + [INSTR_BITAND] = { "BITAND" , 0, 6 }, + [INSTR_BITOR] = { "BITOR" , 0, 5 } +}; + void asm_init (const char *, FILE **); void asm_close(FILE *); void asm_parse(FILE *); -- 2.39.2