From 6038363d2facd5c8a79d9c4920ca5b3cf70bef95 Mon Sep 17 00:00:00 2001 From: Aryadev Chavali Date: Mon, 16 Oct 2023 01:06:49 +0100 Subject: Added functionality to read and write instruction bytecode Uses some bit hacks to quickly check what data type an opcode may have by shifting down to units then casting it to a data_type_t. Not very well tested yet, we'll need to see now. --- Makefile | 2 +- src/base.h | 8 ++++ src/inst.c | 146 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ src/inst.h | 10 +++++ 4 files changed, 165 insertions(+), 1 deletion(-) create mode 100644 src/inst.c diff --git a/Makefile b/Makefile index 566f41a..a0106b3 100644 --- a/Makefile +++ b/Makefile @@ -7,7 +7,7 @@ OUT=ovm.out SRC=src DIST=build -CODE=$(addprefix $(SRC)/, darr.c runtime.c main.c) +CODE=$(addprefix $(SRC)/, darr.c inst.c runtime.c main.c) OBJECTS=$(CODE:$(SRC)/%.c=$(DIST)/%.o) DEPS=$(OBJECTS:%.o=%.d) diff --git a/src/base.h b/src/base.h index 0e822d5..670aa81 100644 --- a/src/base.h +++ b/src/base.h @@ -35,6 +35,14 @@ typedef union f64 as_float; } data_t; +typedef enum +{ + DATA_TYPE_NIL = 0, + DATA_TYPE_BYTE = 1, + DATA_TYPE_WORD = 3, + DATA_TYPE_FLOAT = 5, +} data_type_t; + #define DBYTE(BYTE) ((data_t){.as_byte = (BYTE)}) #define DWORD(WORD) ((data_t){.as_word = (WORD)}) #define DFLOAT(FLOAT) ((data_t){.as_float = (FLOAT)}) diff --git a/src/inst.c b/src/inst.c new file mode 100644 index 0000000..49d35d3 --- /dev/null +++ b/src/inst.c @@ -0,0 +1,146 @@ +/* Copyright (C) 2023 Aryadev Chavali + + * You may distribute and modify this code under the terms of the + * GPLv2 license. You should have received a copy of the GPLv2 + * license with this file. If not, please write to: + * aryadev@aryadevchavali.com. + + * Created: 2023-10-15 + * Author: Aryadev Chavali + * Description: Implementation of bytecode for instructions + */ + +#include +#include + +#include "./inst.h" + +data_type_t get_opcode_data_type(opcode_t opcode) +{ + data_type_t type = DATA_TYPE_NIL; + if (OPCODE_IS_TYPE(opcode, OP_TYPE_PUSH)) + type = opcode >> 1; + else if (OPCODE_IS_TYPE(opcode, OP_TYPE_PUSH_REGISTER)) + type = opcode >> 2; + else if (OPCODE_IS_TYPE(opcode, OP_TYPE_POP)) + type = opcode >> 3; + else if (OPCODE_IS_TYPE(opcode, OP_TYPE_MOV)) + type = opcode >> 4; + return type; +} + +size_t inst_bytecode_size(inst_t inst) +{ + size_t size = 1; // for opcode + if (OPCODE_IS_TYPE(inst.opcode, OP_TYPE_PUSH)) + { + if (inst.opcode == OP_PUSH_BYTE) + ++size; + else if (inst.opcode == OP_PUSH_WORD) + size += sizeof(word); + else if (inst.opcode == OP_PUSH_FLOAT) + size += sizeof(f64); + } + else if (OPCODE_IS_TYPE(inst.opcode, OP_TYPE_PUSH_REGISTER)) + ++size; + else if (OPCODE_IS_TYPE(inst.opcode, OP_TYPE_POP)) + { + // No operand or register so leave as is + } + else if (OPCODE_IS_TYPE(inst.opcode, OP_TYPE_MOV)) + { + if (inst.opcode == OP_MOV_BYTE) + ++size; + else if (inst.opcode == OP_MOV_WORD) + size += sizeof(word); + else if (inst.opcode == OP_MOV_FLOAT) + size += sizeof(f64); + // For the register + ++size; + } + return size; +} + +void inst_write_bytecode(inst_t inst, darr_t *darr) +{ + // Append opcode + darr_append_byte(darr, inst.opcode); + // Then append 0 or more operands + data_type_t to_append = DATA_TYPE_NIL; + if (OPCODE_IS_TYPE(inst.opcode, OP_TYPE_PUSH)) + to_append = (data_type_t)inst.opcode; + else if (OPCODE_IS_TYPE(inst.opcode, OP_TYPE_PUSH_REGISTER) || + OPCODE_IS_TYPE(inst.opcode, OP_TYPE_MOV)) + to_append = DATA_TYPE_BYTE; + + switch (to_append) + { + case DATA_TYPE_NIL: + break; + case DATA_TYPE_BYTE: + darr_append_byte(darr, inst.operand.as_byte); + break; + case DATA_TYPE_WORD: + darr_append_bytes(darr, (byte *)&inst.operand.as_word, sizeof(word)); + break; + case DATA_TYPE_FLOAT: + darr_append_bytes(darr, (byte *)&inst.operand.as_float, sizeof(f64)); + break; + } +} + +data_t read_type_from_darr(darr_t *darr, data_type_t type) +{ + switch (type) + { + case DATA_TYPE_NIL: + break; + case DATA_TYPE_BYTE: + if (darr->used >= darr->available) + // TODO: Error (darr has no space left) + return DBYTE(0); + return DBYTE(darr->data[darr->used++]); + break; + case DATA_TYPE_WORD: + if (darr->used + sizeof(word) >= darr->available) + // TODO: Error (darr has no space left) + return DWORD(0); + word w = 0; + memcpy(&w, darr->data + darr->used, sizeof(w)); + darr->used += sizeof(w); + return DWORD(w); + break; + case DATA_TYPE_FLOAT: + if (darr->used + sizeof(word) >= darr->available) + // TODO: Error (darr has no space left) + return DWORD(0); + f64 f = 0; + memcpy(&f, darr->data + darr->used, sizeof(f)); + darr->used += sizeof(f); + return DFLOAT(f); + break; + } + // TODO: Error (unrecognised type) + return DBYTE(0); +} + +inst_t inst_read_bytecode(darr_t *darr) +{ + if (darr->used >= darr->available) + return (inst_t){0}; + inst_t inst = {0}; + opcode_t opcode = darr->data[darr->used++]; + if (opcode > OP_HALT) + // Translate to NOOP + return inst; + // Read operands + if (OPCODE_IS_TYPE(opcode, OP_TYPE_PUSH)) + inst.operand = read_type_from_darr(darr, get_opcode_data_type(opcode)); + // Read register (as a byte) + if (OPCODE_IS_TYPE(opcode, OP_TYPE_PUSH_REGISTER) || + OPCODE_IS_TYPE(opcode, OP_TYPE_MOV)) + inst.operand = read_type_from_darr(darr, DATA_TYPE_BYTE); + // Otherwise opcode doesn't take operands + + return inst; +} diff --git a/src/inst.h b/src/inst.h index 3c77a86..7fed8e1 100644 --- a/src/inst.h +++ b/src/inst.h @@ -13,7 +13,10 @@ #ifndef INST_H #define INST_H +#include + #include "./base.h" +#include "./darr.h" typedef enum { @@ -57,6 +60,13 @@ typedef struct data_t operand; } inst_t; +size_t inst_bytecode_size(inst_t); +void inst_write_bytecode(inst_t, darr_t *); +// Here the dynamic array is a preloaded buffer of bytes, where +// darr.available is the number of overall bytes and used is the +// cursor (where we are in the buffer). +inst_t inst_read_bytecode(darr_t *); + #define INST_BPUSH(BYTE) \ ((inst_t){.opcode = OP_PUSH_BYTE, .operand = DBYTE(BYTE)}) -- cgit v1.2.3-13-gbd6f