Add a bitmask generator for the new instruction handler

macro.

This program will help with generating the bitmasks for
the instruction handler macro.

1 files changed, 565 insertions, 0 deletions

diff --git a/opcode-bitmask-gen.c b/opcode-bitmask-gen.c
new file mode 100644
index 0000000..304809c
--- /dev/null
+++ b/opcode-bitmask-gen.c
@@ -0,0 +1,565 @@
+#include "opcode-bitmask-gen.h"
+#include <ctype.h>
+#include <stdarg.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+#define AM_ORTHO (AM_ORTHO | AM_ORTHO2)
+#define AM_ADDR (AM_ZM | AM_ZMX | AM_ZMY | \
+		 AM_IND | AM_INDX | AM_INDY | \
+		 AM_ABS | AM_ABX | AM_ABY | \
+		 AM_AIND | AM_AINDX | AM_AINDY | \
+		 AM_INDX2 | AM_ZM2)
+
+void vprint_reason(const char *reason, va_list ap) {
+	if (reason != NULL && isprint(reason[0])) {
+		printf("Reason: ");
+		vprintf(reason, ap);
+	}
+}
+
+void print_reason(const char *reason, ...) {
+	va_list ap;
+	va_start(ap, reason);
+	vprint_reason(reason, ap);
+	va_end(ap);
+}
+
+void failed(const char *name, const char *reason, ...) {
+	va_list ap;
+	va_start(ap, reason);
+	printf("%s failed. ", name);
+	vprint_reason(reason, ap);
+	putchar('\n');
+	va_end(ap);
+}
+
+int vformat_len(const char *fmt, va_list ap) {
+	return vsnprintf(NULL, 0, fmt, ap);
+}
+
+int format_len(const char *fmt, ...) {
+	int len = 0;
+	va_list ap;
+	va_start(ap, fmt);
+	len = vformat_len(fmt, ap);
+	va_end(ap);
+	return len;
+}
+
+#if 0
+int log2_int(uint32_t n) {
+	int j = 0;
+
+	for (int i = 4; i >= 0; i--) {
+		int tmp = 1 << i;
+		j |= ((n >> j) >= (uint32_t)1 << tmp) ? tmp : 0;
+	}
+	return j;
+}
+#else
+int log2_int(uint32_t n) {
+	return (n > 1) ? 1+log2_int(n/2) : 0;
+}
+#endif
+
+uint32_t get_msb(uint32_t n) {
+	return 1 << log2_int(n);
+}
+
+int get_addr_mode_type(uint32_t addr_mode) {
+	return addr_mode_type[log2_int(addr_mode)];
+}
+
+uint8_t get_ind(int instr, int addr_mode) {
+	uint8_t base_idx = 0;
+	uint8_t offset = 0;
+	switch (instr) {
+		case CMP: base_idx = CMP_IND; break;
+		case CPB: base_idx = CPB_IND; break;
+		case JMP: base_idx = JMP_IND; break;
+		case JSR: base_idx = JSR_IND; break;
+		case LDA: base_idx = LDA_IND; break;
+		case LDB: base_idx = LDB_IND; break;
+		case LDX: base_idx = LDX_IND; break;
+		case LDY: base_idx = LDY_IND; break;
+		case STA: base_idx = STA_IND; break;
+		case STB: base_idx = STB_IND; break;
+		case STX: base_idx = STX_IND; break;
+		case STY: base_idx = STY_IND; break;
+	}
+	switch (addr_mode) {
+		case IND : offset  = 0; break;
+		case INDY: offset += 1; break;
+		case INDX: offset += 2; break;
+	}
+	return base_idx + offset;
+}
+
+uint8_t get_eind(int instr) {
+	switch (instr) {
+		case DEC: return DEC_EIND;
+		case INC: return INC_EIND;
+		case STY: return STY_EIND;
+		case STA: return STA_EIND;
+		case STB: return STB_EIND;
+		case LDX: return LDX_EIND;
+		case STX: return STX_EIND;
+		case CPB: return CPB_EIND;
+		case CPX: return CPX_EIND;
+		case CPY: return CPY_EIND;
+	}
+	return 0xFF;
+}
+
+uint8_t get_ext_ortho(int instr) {
+	switch (instr) {
+		case LEA: return OP_LEA;
+		case PEA: return OP_PEA;
+		case ADD: return OP_ADD;
+		case SUB: return OP_SUB;
+		case NOT: return OP_NOT;
+		case CLZ: return OP_CLZ;
+		case CLO: return OP_CLO;
+		case SWP: return OP_SWP;
+		case PCN: return OP_PCN;
+	}
+	return 0xFF;
+}
+
+uint8_t get_opcode(const instruction *instr, int addr_mode, int inst_ext, int ext, int inst_id) {
+	const uint8_t opcode = instr->op;
+	const uint32_t addr_mode_mask = instr->am;
+	addr_mode = (ext == ORTHO) ? ORTHO_EXT : addr_mode;
+	if (inst_ext < ext) {
+		switch (ext) {
+			case EXT:
+				if (addr_mode != EIND) {
+					return 0xFF;
+				}
+				break;
+			case ORTHO:
+				if (addr_mode != ORTHO_EXT) {
+					return 0xFF;
+				}
+				break;
+		}
+	}
+	switch (addr_mode) {
+		case ZM:
+			if (addr_mode_mask & AM_ZM) {
+				return opcode + 0x04;
+			} else if (addr_mode_mask & AM_ZM2) {
+				/* Add 0x20 to opcode if this is JMP zm, or JSR zm. */
+				return opcode + 0x20;
+			}
+			break;
+		case ZMX:
+			if (addr_mode_mask & AM_ZMX) {
+				/* Add 0x06 to opcode if we're using the base ISA, and 0x54 if we're using the base extension. */
+				return opcode + ((ext == BASE) ? 0x06 : 0x54);
+			}
+			break;
+		case ZMY:
+			if (addr_mode_mask & AM_ZMY) {
+				return opcode + 0x14;
+			}
+			break;
+		case INDX:
+			if (addr_mode_mask & AM_INDX) {
+				/* Add 0x16 to opcode if we're using the base ISA, and 0x94 if we're using the base extension. */
+				return opcode + ((ext == BASE) ? 0x16 : 0x94);
+			}
+			/* Falls Through. */
+		case IND:
+		case INDY:
+			/* Are we using the base ISA? */
+			if (ext == BASE && (addr_mode_mask & AM_IND|AM_INDY|AM_INDX2)) {
+				/* Get opcode from lookup table. */
+				return ind_ops[get_ind(inst_id, addr_mode)];
+			/* Are we using the base extension? */
+			} else if (ext == 1) {
+				/* Add 0x44 to opcode if addressing mode is just indirect, and 0x84 otherwise. */
+				return opcode + ((addr_mode == IND) ? 0x44 : 0x84);
+			}
+			break;
+		case ABS:
+			if (addr_mode_mask & AM_ABS) {
+				return opcode + 0x10;
+			}
+			break;
+		case ABSX:
+			if (addr_mode_mask & AM_ABX) {
+				return opcode + 0x50;
+			}
+			break;
+		case ABSY:
+			if (addr_mode_mask & AM_ABY) {
+				return opcode;
+			}
+			break;
+		case AIND:
+			if (addr_mode_mask & AM_AIND) {
+				return opcode + 0x40;
+			}
+			break;
+		case AINDX:
+			if (addr_mode_mask & AM_AINDX) {
+				return opcode + 0x90;
+			}
+			break;
+		case AINDY:
+			if (addr_mode_mask & AM_AINDY) {
+				return opcode + 0x80;
+			}
+			break;
+		case EIND:
+			if (ext < ORTHO && (addr_mode_mask & AM_EIND|AM_EIND2)) {
+				const int eind_type = ((addr_mode_mask & AM_EIND2) != 0);
+				switch (eind_type) {
+					case 0: return (ext == EXT) ? opcode+0x14 : opcode+0x10;
+					case 1: return (ext == EXT) ? opcode+0x01 : eind_base_ops[get_eind(inst_id)];
+				}
+			}
+			break;
+		case ORTHO_EXT:
+			if (addr_mode_mask & AM_ORTHO|AM_ORTHO2) {
+				return (inst_ext == EXT) ? ext_ortho_ops[get_ext_ortho(inst_id)] : opcode;
+			}
+			break;
+		case IMM:
+			if (addr_mode_mask & AM_IMM) {
+				return opcode;
+			}
+			break;
+		case IMPL:
+			if (addr_mode_mask & AM_IMPL) {
+				return opcode;
+			}
+			break;
+		case BREG:
+			if (addr_mode_mask & AM_BREG) {
+				return opcode + 0x14;
+			}
+			break;
+		default: break;
+	}
+	return 0xFF;
+}
+
+void set_opcode_bit(uint32_t *mask, int opcode) {
+	const uint32_t o8 = opcode / 32, o8m = 1 << (opcode % 32);
+	mask[o8] |= o8m;
+}
+
+void set_ortho_opcode_bits(uint32_t *mask, int opcode, int is_1op, int is_cc) {
+	for (int i = 0; i < ((is_cc) ? 8 : 1); ++i, opcode += (is_cc) ? 0x20 : 0) {
+		for (int j = 0; j < 4-is_1op; j += 1+is_1op) {
+			set_opcode_bit(mask, opcode | (j << 3));
+		}
+	}
+}
+
+void search_inst_table_entry(const instruction *inst, uint32_t *mask, uint32_t addr_modes, int inst_ext, int ext, uint8_t inst_id) {
+	addr_modes &= inst->am;
+	switch (ext) {
+		case BASE: addr_modes &= ~(AM_ORTHO|AM_ORTHO2|AM_EIND|AM_EIND2); break;
+		case EXT : addr_modes &= (inst_ext == BASE) ? (AM_EIND|AM_EIND2) : ~(AM_ORTHO|AM_ORTHO2); break;
+	}
+	for (uint32_t i = addr_modes; i; i &= ~get_msb(i)) {
+		const uint8_t opcode = get_opcode(inst, get_addr_mode_type(i), inst_ext, ext, inst_id);
+		const uint32_t o8 = opcode / 32, o8m = 1 << (opcode % 32);
+		if (opcode == 0xFF) {
+			continue;
+		}
+
+		if (ext == ORTHO) {
+			const int is_1op = (inst->am & AM_ORTHO2) == AM_ORTHO2;
+			const int is_cc = (inst_ext == ORTHO && inst_id == SET);
+			set_ortho_opcode_bits(mask, opcode, is_1op, is_cc);
+		} else {
+			set_opcode_bit(mask, opcode);
+		}
+
+		/*printf("table: %i, i: %2i, j: 0x%08X, mask: {", table, i, j);
+		for (int k = mask_size-1; k >= 0; --k) {
+			printf("0x%08X%s", mask[k], (k) ? ", " : "}\n");
+		}*/
+	}
+}
+
+int print_mask(uint32_t addr_modes, instruction_id *inst_ids, int ext) {
+	char *str = calloc(format_len("t(0, 0x%08X, 0x%08X, 0x%08X, 0x%08X, 0x%08X, 0x%08X, 0x%08X, 0x%08X)\n", 0, 1, 2, 3, 4, 5, 6, 7) + 1, sizeof(char));
+	char *tmp = str;
+	uint32_t mask[8] = {};
+	const size_t mask_size = sizeof(mask) / sizeof(uint32_t);
+	const uint32_t max_addr_mode_mask = -(uint32_t)1 >> (32-AMT_LEN);
+	addr_modes = (addr_modes & ~max_addr_mode_mask) ? max_addr_mode_mask : addr_modes;
+
+	if (ext > ORTHO) {
+		failed("print_mask()", "Supplied extension is invalid, because it's out of the intended bounds.");
+		free(str);
+		tmp = NULL;
+		str = NULL;
+		return 0;
+	}
+
+	if (inst_ids == NULL) {
+		for (int table = 0; table <= ext; ++table) {
+			int max_ops = 0;
+			const uint32_t addr_mode_mask = (ext && table < ext) ? ((ext == EXT) ? AM_EIND|AM_EIND2 : AM_ORTHO|AM_ORTHO2) : addr_modes;
+			const instruction *inst_table;
+
+			switch (ext) {
+				case BASE: inst_table = inst, max_ops = OPNUM; break;
+				case EXT: inst_table = ext_inst, max_ops = EXT_OPNUM; break;
+				case ORTHO: inst_table = ortho_inst, max_ops = ORTHO_OPNUM; break;
+			}
+
+			for (int i = 0; i < max_ops; ++i) {
+				if (inst_table[i].am & addr_mode_mask) {
+					search_inst_table_entry(&inst_table[i], mask, addr_modes, table, ext, i);
+					#if 0
+					for (uint32_t j = addr_modes; j; j &= ~get_msb(j)) {
+						const uint8_t opcode = get_opcode(&inst_table[i], get_addr_mode_type(j), table, ext, i);
+						const uint32_t o8 = opcode / 32, o8m = 1 << (opcode % 32);
+						if (opcode == 0xFF) {
+							continue;
+						}
+
+						if (ext == 2) {
+							const int is_1op = (inst[i].am & AM_ORTHO2) == AM_ORTHO2;
+							const int is_cc = (table == ORTHO && i == SET);
+							set_ortho_opcode_bits(mask, opcode, is_1op, is_cc);
+						} else {
+							set_opcode_bit(mask, opcode);
+						}
+
+						/*printf("table: %i, i: %2i, j: 0x%08X, mask: {", table, i, j);
+						for (int k = mask_size-1; k >= 0; --k) {
+							printf("0x%08X%s", mask[k], (k) ? ", " : "}\n");
+						}*/
+					}
+					#endif
+				}
+			}
+		}
+	} else {
+		for (int i = 0; inst_ids[i].ext >= BASE; ++i) {
+			const uint32_t addr_mode_mask = (ext && inst_ids[i].ext < ext) ? ((ext == EXT) ? AM_EIND|AM_EIND2 : AM_ORTHO|AM_ORTHO2) : addr_modes;
+			const instruction *inst_entry;
+			switch (inst_ids[i].ext) {
+				case BASE : inst_entry = &inst[inst_ids[i].id]; break;
+				case EXT  : inst_entry = &ext_inst[inst_ids[i].id]; break;
+				case ORTHO: inst_entry = &ortho_inst[inst_ids[i].id]; break;
+			}
+
+			if (inst->am & addr_mode_mask) {
+				search_inst_table_entry(inst_entry, mask, addr_modes, inst_ids[i].ext, ext, inst_ids[i].id);
+			}
+		}
+	}
+
+	tmp += sprintf(tmp, "t(0, ");
+	for (int i = mask_size-1; i >= 0; --i) {
+		tmp += sprintf(tmp, "0x%08X%s", mask[i], (i) ? ", " : ")\n");
+	}
+
+	printf("%s", str);
+
+	free(str);
+	tmp = NULL;
+	str = NULL;
+	return 1;
+}
+
+int get_delm_span(const char *str, const char *delm) {
+	const int inv_span_len = strcspn(str, delm);
+	return inv_span_len + strspn(&str[inv_span_len], delm);
+}
+
+int get_csv_len(const char *str) {
+	int ret = 0;
+	for (int i = 0; str[i] != '\0'; i += get_delm_span(&str[i], ", "), ++ret);
+	return ret;
+}
+
+void free_csv_list(char **list) {
+	for (int i = 0; list[i] != NULL; ++i) {
+		free(list[i]);
+		list[i] = NULL;
+	}
+}
+
+char **parse_csv_list(const char *str) {
+	const char *delm = ", ";
+	char **list = calloc(get_csv_len(str)+1, sizeof(char *));
+	for (int i = 0, j = 0; str[i] != '\0'; i += get_delm_span(&str[i], delm), ++j) {
+		const int len = strcspn(&str[i], delm);
+		list[j] = calloc(len+1, sizeof(char));
+		memcpy(list[j], &str[i], len);
+	}
+	return list;
+}
+
+void set_extension(int *ext, int ext2, int force, const char *reason, ...) {
+	int dummy = -1;
+	ext = (ext != NULL) ? ext : &dummy;
+	if (*ext < ext2 || *ext < BASE || force) {
+		va_list ap;
+		printf("Switching extension from %s, to %s. ", (*ext < BASE || *ext >= EXT_LEN) ? ext_names[*ext] : "auto", ext_names[ext2]);
+		va_start(ap, reason);
+		vprint_reason(reason, ap);
+		va_end(ap);
+		*ext = ext2;
+	}
+}
+
+uint32_t get_addr_mode_bits(const char **list, int *ext) {
+	uint32_t addr_modes = 0;
+	for (int i = 0; list[i] != NULL; ++i) {
+		const char *str = list[i];
+		for (int j = 0; j < AMT_LEN; ++j) {
+			if (!strcasecmp(str, adrmode[j])) {
+				//printf("i: %i, str: %s, adrmode: %s\n", i, str, adrmode[i]);
+				addr_modes |= (1 << j);
+			}
+		}
+	}
+	if (addr_modes & (AM_ORTHO|AM_ORTHO2|AM_EIND|AM_EIND2|AM_AIND|AM_AINDX|AM_AINDY)) {
+		const int ext2 = (addr_modes & AM_ORTHO) ? ORTHO : EXT;
+		const char *addr_mode_reason = (ext2 == ORTHO) ? "was set to ortho" : "is a base extension addressing mode";
+		set_extension(ext, ext2, 1, "The addressing mode %s.\n", addr_mode_reason);
+	}
+	return addr_modes;
+}
+
+instruction_id *get_instruction_ids(const char **list, int *ext) {
+	int has_failed = 0;
+	char *fail_reason = NULL;
+	instruction_id *inst_ids = NULL;
+	int list_len = 0;
+	if (list != NULL) {
+		for (list_len = 0; list[list_len] != NULL; ++list_len);
+
+		if (list_len == 0) {
+			has_failed = 1;
+			fail_reason = "Instruction list length is zero.";
+		} else {
+			inst_ids = calloc(list_len+1, sizeof(instruction_id));
+			memset(inst_ids, -1, sizeof(instruction_id)*(list_len+1));
+		}
+	} else {
+		has_failed = 1;
+		fail_reason = "Instruction list is NULL.";
+	}
+
+	if (has_failed) {
+		failed("get_instruction_ids()", "%s", fail_reason);
+		return NULL;
+	}
+
+	for (int i = 0; list[i] != NULL; ++i) {
+		for (int j = 0; j < EXT_LEN && inst_ids[i].ext < 0; ++j) {
+			const char **table;
+			switch (j) {
+				case BASE : table = mne; break;
+				case EXT  : table = ext_mne; break;
+				case ORTHO: table = ortho_mne; break;
+			}
+
+			for (int k = 0; table[k] != NULL; ++k) {
+				if (!strcasecmp(list[i], table[k])) {
+					char *ext_str = NULL;
+					inst_ids[i] = (instruction_id){.ext = j, .id = k};
+					switch (j) {
+						case BASE : ext_str = "Base ISA"; break;
+						case EXT  : ext_str = "Base Extension"; break;
+						case ORTHO: ext_str = "Ortho Extension"; break;
+					}
+					set_extension(ext, j, 0, "The instruction is from the %s.\n", ext_str);
+					break;
+				}
+			}
+		}
+	}
+	return inst_ids;
+}
+
+int get_extension_id(const char *str) {
+	for (int i = 0; i < EXT_LEN; ++i) {
+		if (!strcasecmp(str, ext_names[i])) {
+			return i;
+		}
+	}
+	return -1;
+}
+
+void usage(const char *name) {
+	printf("Usage: %s -a <addr_mode>[,<addr_mode> ...] -i <instruction>[,<instruction> ...] -e <isa_extension>\n", name);
+}
+
+options *get_options(int argc, char **argv) {
+	options *opts = calloc(1, sizeof(options));
+	for (int i = 1; i < argc; ++i) {
+		if (argv[i][0] == '-') {
+			const char option = argv[i++][1];
+			if (option == 'a' || option == 'i') {
+				const char **list = (const char **)parse_csv_list(argv[i]);
+				switch (option) {
+					case 'a': opts->addr_modes = get_addr_mode_bits(list, &opts->ext); break;
+					case 'i': opts->instr = get_instruction_ids(list, &opts->ext); break;
+				}
+				opts->addr_modes = (opts->instr != NULL && opts->addr_modes == 0) ? -1 : opts->addr_modes;
+				free_csv_list((char **)list);
+			} else if (option == 'e') {
+				const int ext = get_extension_id(argv[i]);
+				opts->ext = (opts->ext < ext) ? ext : opts->ext;
+				opts->addr_modes = (opts->addr_modes == 0) ? -1 : opts->addr_modes;
+			} else {
+				failed(argv[0], "Invalid option \"-%c\".", option);
+				usage(argv[0]);
+				free(opts);
+				return NULL;
+			}
+		}
+	}
+	return opts;
+}
+
+options *parse_args(int argc, char **argv) {
+	if (argc == 1) {
+		failed(argv[0], "No arguments.");
+		usage(argv[0]);
+		return NULL;
+	}
+
+	return get_options(argc, argv);
+}
+
+int main(int argc, char **argv) {
+	int ret;
+	char *reason = NULL;
+	options *opts = parse_args(argc, argv);
+	if (opts == NULL) {
+		return 1;
+	}
+	if (!opts->addr_modes) {
+		reason = "Addressing mode mask is zero.";
+		ret = 1;
+	} else {
+		ret = !print_mask(opts->addr_modes, opts->instr, opts->ext);
+	}
+
+	if (ret != 0) {
+		if (reason != NULL) {
+			failed(argv[0], reason);
+			usage(argv[0]);
+		}
+	}
+
+	free(opts->instr);
+	free(opts);
+	return ret;
+}