diff --git a/src/Makefile b/src/Makefile index 7106e44..2fe219c 100755 --- a/src/Makefile +++ b/src/Makefile @@ -9,7 +9,7 @@ CFLAGS ?= -std=c17 -g\ all: assemble emulate -assemble: assemble.o +assemble: assemble.o parser.o fileio.o emulate: emulate.o clean: diff --git a/src/a64instruction.h b/src/a64instruction/a64instruction.h similarity index 100% rename from src/a64instruction.h rename to src/a64instruction/a64instruction.h diff --git a/src/a64instruction_Branch.h b/src/a64instruction/a64instruction_Branch.h similarity index 97% rename from src/a64instruction_Branch.h rename to src/a64instruction/a64instruction_Branch.h index b732d6d..2e525bb 100644 --- a/src/a64instruction_Branch.h +++ b/src/a64instruction/a64instruction_Branch.h @@ -1,6 +1,6 @@ #include #include "a64instruction_global.h" -#include "global.h" +#include "../global.h" typedef enum { a64inst_UNCONDITIONAL = 0, diff --git a/src/a64instruction_DP.h b/src/a64instruction/a64instruction_DP.h similarity index 100% rename from src/a64instruction_DP.h rename to src/a64instruction/a64instruction_DP.h diff --git a/src/a64instruction_DPImmediate.h b/src/a64instruction/a64instruction_DPImmediate.h similarity index 100% rename from src/a64instruction_DPImmediate.h rename to src/a64instruction/a64instruction_DPImmediate.h diff --git a/src/a64instruction_DPRegister.h b/src/a64instruction/a64instruction_DPRegister.h similarity index 100% rename from src/a64instruction_DPRegister.h rename to src/a64instruction/a64instruction_DPRegister.h diff --git a/src/a64instruction_Directive.h b/src/a64instruction/a64instruction_Directive.h similarity index 71% rename from src/a64instruction_Directive.h rename to src/a64instruction/a64instruction_Directive.h index da36624..5506b80 100644 --- a/src/a64instruction_Directive.h +++ b/src/a64instruction/a64instruction_Directive.h @@ -1,4 +1,4 @@ -#include "global.h" +#include "../global.h" typedef struct { word value; diff --git a/src/a64instruction_Label.h b/src/a64instruction/a64instruction_Label.h similarity index 100% rename from src/a64instruction_Label.h rename to src/a64instruction/a64instruction_Label.h diff --git a/src/a64instruction_SingleTransfer.h b/src/a64instruction/a64instruction_SingleTransfer.h similarity index 98% rename from src/a64instruction_SingleTransfer.h rename to src/a64instruction/a64instruction_SingleTransfer.h index f661116..cbd7529 100644 --- a/src/a64instruction_SingleTransfer.h +++ b/src/a64instruction/a64instruction_SingleTransfer.h @@ -1,6 +1,6 @@ #include #include "a64instruction_global.h" -#include "global.h" +#include "../global.h" typedef enum { a64inst_SINGLE_TRANSFER_SINGLE_DATA_TRANSFER = 1, diff --git a/src/a64instruction_global.h b/src/a64instruction/a64instruction_global.h similarity index 100% rename from src/a64instruction_global.h rename to src/a64instruction/a64instruction_global.h diff --git a/src/assemble.c b/src/assemble.c index ae760c0..ef1b06d 100755 --- a/src/assemble.c +++ b/src/assemble.c @@ -1,7 +1,35 @@ #include #include +#include "a64instruction/a64instruction.h" +#include "parser.h" +#include "fileio.h" +#include "parser.h" +#include "twopassassembly.c" int main(int argc, char **argv) { + // Check the arguments + if (argc < 3) { + fprintf(stderr, "Error: A source file and an object output file are required. Syntax: ./assemble "); + return EXIT_FAILURE; + } + + // Load the source file into memory + int lineCount = countLines(argv[1]); + char **source = readAssemblyFile(argv[1], lineCount); + + // Parse the source file + a64inst_instruction *instructions = parse(source, lineCount); + // First Pass: Create the symbol table + st *table = firstPass(instructions, lineCount); + + // Second Pass: Assemble the instructions + word *binary = secondPass(instructions, lineCount, table); // 1000 is just a temp fix. + + // Write the binary to the output file + writeBinaryFile(binary, argv[2], lineCount); // 1000 is just a temp fix. + + /* TODO: FREE MEMORY!! */ + return EXIT_SUCCESS; } diff --git a/src/decode.h b/src/decode.h index 132130e..d509ae1 100644 --- a/src/decode.h +++ b/src/decode.h @@ -1,5 +1,5 @@ #include "global.h" -#include "a64instruction.h" +#include "a64instruction/a64instruction.h" #define HALT_WORD 0x8a000000 diff --git a/src/emulate.c b/src/emulate.c old mode 100755 new mode 100644 index 82245e9..4c7f42e --- a/src/emulate.c +++ b/src/emulate.c @@ -1,6 +1,6 @@ #include #include -#include "a64instruction.h" +#include "a64instruction/a64instruction.h" #include "emulator.h" #include "fileio.h" #include "global.h" @@ -8,6 +8,11 @@ #include "decode.h" #include "execute.h" +int main(int arg, char **argv){ + return EXIT_SUCCESS; +} + +/* extern a64inst_instruction *decode(word w); int main(int argc, char **argv) { @@ -59,3 +64,4 @@ int main(int argc, char **argv) { return EXIT_SUCCESS; } +*/ diff --git a/src/execute.c b/src/execute.c deleted file mode 100644 index fc436af..0000000 --- a/src/execute.c +++ /dev/null @@ -1,448 +0,0 @@ -#include -#include -#include "execute.h" -#include "print.h" - -// Defines the maximum value that can be held in a register -#define MAX_REG_VAL ((1 << DWORD_BITS) - 1) - -// The number of bits to shift the immediate value in an arithmetic immediate data processing -// instruction if the shift flag is enabled. -#define DPI_ARITHM_SHIFT 12 - -// The number of bits to shift the immediate value in a wide move immediate data processing -// instruction if the shift flag is enabled. -#define DPI_WIDEMOV_SHIFT 16 - -// Prototypes -void execute_SDT(Machine *state, a64inst_instruction *inst); -void execute_Branch(Machine *state, a64inst_instruction *inst); -void executeMultiply(Machine *state, a64inst_instruction *inst); - -// Return maximum of two dwords -static dword max(dword a, dword b) { - return a > b ? a : b; -} - -// Truncate a given value to the size of a word or dword depending on the register type -static dword truncateValue(dword value, a64inst_regType regType) { - if (regType == a64inst_X) { - return value; - } else { - return (word)value; - //return value & (dword)(((dword)1 << WORD_BITS) - 1); - } -} - -// Sign extend a given value to a 64-bit signed integer given the number of bits -static int64_t signExtend(dword value, unsigned int n) { - if (n == 0 || n >= 64) { - // If n_bits is 0 or greater than or equal to 64, return the value as is - return (int64_t)value; - } - - uint64_t sign_bit_mask = (uint64_t)1 << (n - 1); - - // Mask to isolate the n-bit value - uint64_t n_bit_mask = (sign_bit_mask << 1) - 1; - - // Check if the sign bit is set - if (value & sign_bit_mask) { - // Sign bit is set, extend the sign - return (int64_t)(value | ~n_bit_mask); - } else { - // Sign bit is not set, return the value as is - return (int64_t)(value & n_bit_mask); - } -} - -// Read from processor register, ensuring that a valid register specifier is given -// and accounting for the case where the zero register is accessed. Truncate -// the 32 most significant bits stored in the R register when reading W register. -static dword readRegister(Machine *state, a64inst_regSpecifier reg, a64inst_regType regType) { - assert(reg <= REGISTER_COUNT); - if (reg == ZERO_REGISTER) { - return 0; - } else { - return truncateValue(state->registers[reg], regType); - } -} - -// TODO: - -// Write to a processor register, ensuring that a valid register specifier is given -// and truncating the value being written when it can't fit in the specified register -static void writeRegister(Machine *state, a64inst_regSpecifier reg, a64inst_regType regType, dword value) { - assert(reg <= REGISTER_COUNT); - if (reg != ZERO_REGISTER) { - state->registers[reg] = truncateValue(value, regType); - } -} - -// Returns the position of the MSB of the given register type -inline static dword getMSBPos(a64inst_regType regType) { - return (regType ? DWORD_BITS : WORD_BITS) - 1; -} - -// Returns the MSB of the given value assuming it's of the size stored in the given register type -inline static uint8_t getMSB(dword value, a64inst_regType regType) { - return value >> getMSBPos(regType); -} - -// Updates N and Z condition codes given the machine and a result value -static void updateCondNZ(Machine *state, dword result, a64inst_regType regType) { - state->conditionCodes.Negative = getMSB(result, regType); - state->conditionCodes.Zero = result == 0; -} - -// Execute a data processing immediate instruction -static void executeDPImmediate(Machine *state, a64inst_instruction *inst) { - assert(inst->type == a64inst_DPIMMEDIATE); - - a64inst_regType regType = inst->data.DPImmediateData.regType; - a64inst_regSpecifier dest = inst->data.DPImmediateData.dest; - switch(inst->data.DPImmediateData.DPIOpType) { - - // Execute an arithmetic immediate data processing instruction - case a64inst_DPI_ARITHM:; - - // If shift flag is enabled, logical left shift by the number of bits specified by the architecture - dword arithmImm = inst->data.DPImmediateData.processOpData.arithmData.immediate; - dword srcVal = state->registers[inst->data.DPImmediateData.processOpData.arithmData.src]; - if (inst->data.DPImmediateData.processOpData.arithmData.shiftImmediate) { - arithmImm = truncateValue(arithmImm << DPI_ARITHM_SHIFT, regType); - } - - switch(inst->data.DPImmediateData.processOp) { - - dword result; - case(a64inst_ADDS): - result = srcVal + arithmImm; - writeRegister(state, dest, regType, result); - - updateCondNZ(state, result, regType); - state->conditionCodes.Overflow = max(srcVal, arithmImm) > result; - state->conditionCodes.Carry = state->conditionCodes.Overflow; - break; - - case(a64inst_ADD): - writeRegister(state, dest, regType, srcVal + arithmImm); - break; - - case(a64inst_SUBS): - result = srcVal - arithmImm; - writeRegister(state, dest, regType, result); - - updateCondNZ(state, result, regType); - state->conditionCodes.Overflow = srcVal < result; - state->conditionCodes.Carry = state->conditionCodes.Overflow; - break; - - case(a64inst_SUB): - writeRegister(state, dest, regType, srcVal - arithmImm); - break; - - // Unknown opcode detected! - default: - fprintf(stderr, "Unknown opcode detected in a DPI arithmetic instruction!\n"); - break; - } - break; - - // Execute a wide move immediate data processing instruction - case a64inst_DPI_WIDEMOV:; - uint8_t shiftScalar = inst->data.DPImmediateData.processOpData.wideMovData.shiftScalar; - dword wideMovImm = inst->data.DPImmediateData.processOpData.wideMovData.immediate; - - // NOTE: Not checking that shiftScalar has valid value for 32bit registers. Possibly add explicit error. - //printf("%x\n", wideMovImm << (shiftScalar * DPI_WIDEMOV_SHIFT) & ); - wideMovImm = truncateValue(wideMovImm << (shiftScalar * DPI_WIDEMOV_SHIFT), regType); - switch(inst->data.DPImmediateData.processOp) { - - case(a64inst_MOVN): - writeRegister(state, dest, regType, ~wideMovImm); - break; - - case(a64inst_MOVZ): - writeRegister(state, dest, regType, wideMovImm); - break; - - case(a64inst_MOVK):; - dword result = readRegister(state, dest, regType); - result = (result & ~(((1lu << DPI_WIDEMOV_SHIFT) - 1) << shiftScalar * DPI_WIDEMOV_SHIFT)) | wideMovImm; - writeRegister(state, dest, regType, result); - break; - - default: - fprintf(stderr, "Unknown opcode detected in a DPI wide move instruction!\n"); - break; - } - break; - - // Unknown instruction detected! - default: - fprintf(stderr, "Attempting to execute instruction with unknown DPI operand type!\n"); - break; - } -} - -// Execute a data processing register instruction -static void executeDPRegister(Machine *state, a64inst_instruction *inst) { - assert(inst->type == a64inst_DPREGISTER); - - a64inst_regType regType = inst->data.DPRegisterData.regType; - a64inst_regSpecifier dest = inst->data.DPRegisterData.dest; - dword src1Val = readRegister(state, inst->data.DPRegisterData.src1, regType); - dword src2Val = readRegister(state, inst->data.DPRegisterData.src2, regType); - - switch(inst->data.DPRegisterData.DPROpType) { - - // Execute an arithmetic or logic register data processing instruction - case a64inst_DPR_ARITHMLOGIC:; - - // Apply shift to value held in second register - a64inst_DPRegister_ArithmLogicData *arithmLogicData = &inst->data.DPRegisterData.processOpData.arithmLogicData; - uint8_t shiftAmount = arithmLogicData->shiftAmount; - switch(arithmLogicData->shiftType) { - - case a64inst_LSL: - src2Val = truncateValue(src2Val << shiftAmount, regType); - break; - - case a64inst_LSR: - src2Val = truncateValue(src2Val >> shiftAmount, regType); - break; - - case a64inst_ASR: - if (regType == a64inst_X) { - src2Val = truncateValue((int64_t)src2Val >> shiftAmount, regType); - } else { - src2Val = truncateValue((int32_t)src2Val >> shiftAmount, regType); - } - break; - - case a64inst_ROR: - if (arithmLogicData->type != a64inst_DPR_LOGIC) { - fprintf(stderr, "Attempting to perform ROR shift on non-logic register data processing instruction!\n"); - } - src2Val = truncateValue(src2Val >> shiftAmount | src2Val << (getMSBPos(regType) - shiftAmount), regType); - break; - - default: - fprintf(stderr, "Attempting to execute arithmetic/logic register data processing instruction with invalid shift type!\n"); - break; - } - - // Negate second operand if negShiftedSrc2 flag is enabled - if (arithmLogicData->negShiftedSrc2) { - src2Val = truncateValue(~src2Val, regType); - } - - dword result; - switch(arithmLogicData->type) { - - case a64inst_DPR_ARITHM: - switch(inst->data.DPRegisterData.processOp) { - - case(a64inst_ADDS): - result = src1Val + src2Val; - writeRegister(state, dest, regType, result); - - updateCondNZ(state, result, regType); - state->conditionCodes.Overflow = max(src1Val, src2Val) > result; - state->conditionCodes.Carry = state->conditionCodes.Overflow; - break; - - case(a64inst_ADD): - writeRegister(state, dest, regType, src1Val + src2Val); - break; - - case(a64inst_SUBS): - result = src1Val - src2Val; - writeRegister(state, dest, regType, result); - - updateCondNZ(state, result, regType); - state->conditionCodes.Overflow = getMSB(src1Val, regType) != getMSB(src2Val, regType) && getMSB(src1Val, regType) != getMSB(result, regType); - state->conditionCodes.Carry = src1Val >= src2Val; - break; - - case(a64inst_SUB): - writeRegister(state, dest, regType, src1Val - src2Val); - break; - - // Unknown opcode detected! - default: - fprintf(stderr, "Unknown opcode detected in a DPI arithmetic instruction!\n"); - break; - } - break; - - case a64inst_DPR_LOGIC: - switch(inst->data.DPRegisterData.processOp) { - - case a64inst_AND: - writeRegister(state, dest, regType, src1Val & src2Val); - break; - - case a64inst_OR: - writeRegister(state, dest, regType, src1Val | src2Val); - break; - - case a64inst_XOR: - writeRegister(state, dest, regType, src1Val ^ src2Val); - break; - - case a64inst_AND_FLAGGED:; - result = src1Val & src2Val; - writeRegister(state, dest, regType, result); - state->conditionCodes.Overflow = 0; - state->conditionCodes.Carry = 0; - updateCondNZ(state, result, regType); - break; - } - break; - - default: - fprintf(stderr, "Attempting to execute an instruction with an unknown DPR arithmetic or logic subtype!\n"); - break; - } - break; - - // Execute a multiply register data processing instruction - case a64inst_DPR_MULTIPLY: - break; - - // Unknown instruction detected! - default: - fprintf(stderr, "Attempting to execute instruction with unknown DPR operand type!\n"); - break; - } -} - -void execute(Machine *state, a64inst_instruction *inst) { - - switch (inst->type) { - - // Halt the program - case a64inst_HALT: - break; - - // Execute a data processing immediate instruction - case a64inst_DPIMMEDIATE: - executeDPImmediate(state, inst); - break; - - // Execute a branch instruction - case a64inst_BRANCH: - execute_Branch(state, inst); - break; - - // Execute a data processing register instruction - case a64inst_DPREGISTER: - if (inst->data.DPRegisterData.DPROpType == a64inst_DPR_MULTIPLY) - executeMultiply(state, inst); - else - executeDPRegister(state, inst); - break; - - case a64inst_SINGLETRANSFER: - execute_SDT(state, inst); - break; - - // Unknown instruction - default: - break; - } - -} - -void execute_SDT(Machine *state, a64inst_instruction *inst) { - word address; - bool isLoad; - if (inst->data.SingleTransferData.SingleTransferOpType == a64inst_SINGLE_TRANSFER_LOAD_LITERAL) { - // Load Literal - isLoad = true; - address = state->pc + inst->data.SingleTransferData.processOpData.loadLiteralData.offset * 4; - } else { - address = state->registers[inst->data.SingleTransferData.processOpData.singleDataTransferData.base]; - isLoad = inst->data.SingleTransferData.processOpData.singleDataTransferData.transferType == a64inst_LOAD; - switch (inst->data.SingleTransferData.processOpData.singleDataTransferData.addressingMode) { - case a64inst_UNSIGNED_OFFSET: - address += inst->data.SingleTransferData.processOpData.singleDataTransferData.a64inst_addressingModeData.unsignedOffset * (inst->data.SingleTransferData.regType == a64inst_W ? 4 : 8); - break; - case a64inst_REGISTER_OFFSET: - address += state->registers[inst->data.SingleTransferData.processOpData.singleDataTransferData.a64inst_addressingModeData.offsetReg]; - break; - case a64inst_PRE_INDEXED: - address += inst->data.SingleTransferData.processOpData.singleDataTransferData.a64inst_addressingModeData.indexedOffset; - state->registers[inst->data.SingleTransferData.processOpData.singleDataTransferData.base] = address; - break; - case a64inst_POST_INDEXED: - state->registers[inst->data.SingleTransferData.processOpData.singleDataTransferData.base] = address + inst->data.SingleTransferData.processOpData.singleDataTransferData.a64inst_addressingModeData.indexedOffset; - break; - } - } - - if (isLoad) { - if (inst->data.SingleTransferData.regType == a64inst_W) { - // 32 bit access - state->registers[inst->data.SingleTransferData.target] = readWord(state->memory, address); - } else { - state->registers[inst->data.SingleTransferData.target] = readDoubleWord(state->memory, address); - } - } else { - *(word *)(state->memory + address) = state->registers[inst->data.SingleTransferData.target]; - - // Update base register if post indexed - if (inst->data.SingleTransferData.processOpData.singleDataTransferData.addressingMode == a64inst_POST_INDEXED) { - writeRegister(state, inst->data.SingleTransferData.processOpData.singleDataTransferData.base, inst->data.SingleTransferData.regType == a64inst_W, address + inst->data.SingleTransferData.processOpData.singleDataTransferData.a64inst_addressingModeData.indexedOffset); - } - } - -} - -static bool isConditionMet(Machine* state, a64inst_ConditionType cond) { - switch(cond) { - case EQ: - return state->conditionCodes.Zero; - case NE: - return !state->conditionCodes.Zero; - case GE: - return state->conditionCodes.Negative == state->conditionCodes.Overflow; - case LT: - return state->conditionCodes.Negative != state->conditionCodes.Overflow; - case GT: - return !state->conditionCodes.Zero && (state->conditionCodes.Negative == state->conditionCodes.Overflow); - case LE: - return state->conditionCodes.Zero || (state->conditionCodes.Negative != state->conditionCodes.Overflow); - case AL: - return true; - default: - fprintf(stderr, "Unknown condition specified!\n"); - exit(1); - } -} - -void execute_Branch(Machine *state, a64inst_instruction *inst) { - switch (inst->data.BranchData.BranchType) { - case a64inst_UNCONDITIONAL: - state->pc += signExtend(inst->data.BranchData.processOpData.unconditionalData.unconditionalOffset * 4, 26); - break; - - case a64inst_REGISTER: - state->pc = state->registers[inst->data.BranchData.processOpData.registerData.src]; - break; - - case a64inst_CONDITIONAL: - if (isConditionMet(state, inst->data.BranchData.processOpData.conditionalData.cond)) { - state->pc += signExtend(inst->data.BranchData.processOpData.conditionalData.offset * 4, 19); - } - break; - } -} - -void executeMultiply(Machine *state, a64inst_instruction *inst) { - dword product = state->registers[inst->data.DPRegisterData.src1] * state->registers[inst->data.DPRegisterData.src2]; - dword sum = readRegister(state, inst->data.DPRegisterData.processOpData.multiplydata.summand, inst->data.DPRegisterData.regType) + (inst->data.DPRegisterData.processOpData.multiplydata.negProd ? -product : product); - writeRegister(state, inst->data.DPRegisterData.dest, inst->data.DPRegisterData.regType, sum); -} diff --git a/src/execute.h b/src/execute.h index fcf39ec..8b691e6 100644 --- a/src/execute.h +++ b/src/execute.h @@ -1,6 +1,6 @@ #ifndef __EXECUTE__ #define __EXECUTE__ -#include "a64instruction.h" +#include "a64instruction/a64instruction.h" #include "emulator.h" void execute(Machine *state, a64inst_instruction *inst); diff --git a/src/fileio.c b/src/fileio.c index 1dcdd77..f2f47fd 100644 --- a/src/fileio.c +++ b/src/fileio.c @@ -1,48 +1,90 @@ -#include -#include #include -#include "fileio.h" #include "global.h" +#include "fileio.h" -/* Loads a binary file located at filePath to memory, taking up a block of exactly memorySize bytes, - and returns the starting address of the data. If memorySize is insufficient to store the entire file, - an appropriate error is reported. Excess memory is set to 0 bit values. */ +#define MAX_ASM_LINE_LENGTH 300 -byte *fileio_loadBin(const char *filePath, size_t memorySize) { - FILE *file = fopen(filePath, "rb"); +int isValidFileFormat(char filename[], char expectedExtension[]){ + char *pointLoc = strrchr(filename, '.'); + + if(pointLoc != NULL){ + if(strcmp(pointLoc, expectedExtension)==0){ + return(1); + } + } + return(0); +} + +void writeBinaryFile(word instrs[], char outputFile[], int numInstrs) { + FILE *fp = fopen(outputFile, "wb"); + if (fp == NULL) { + fprintf(stderr, "Error: Could not open file %s\n", outputFile); + exit(EXIT_FAILURE); + } + + fwrite(instrs, sizeof(word), numInstrs, fp); + fclose(fp); +} + +int countLines(char *filename) { + FILE *file = fopen(filename, "r"); if (file == NULL) { - fprintf(stderr, "Couldn't open %s!\n", filePath); + fprintf(stderr, "Error: Could not read file %s\n", filename); exit(EXIT_FAILURE); } - byte *fileData = malloc(memorySize); - if (fileData == NULL) { - fprintf(stderr, "Ran out of memory attempting to load %s!\n", filePath); + int count = 0; + char c; + + while ((c = fgetc(file)) != EOF) { + if (c == '\n') { + count++; + } + } + + return count; +} + +char **readAssemblyFile(char filename[], int lineCount) { + FILE *fp = fopen(filename, "r"); + if (fp == NULL) { + fprintf(stderr, "Error: Could not read file %s\n", filename); exit(EXIT_FAILURE); } - // Loop while reading from the file yields data. Only terminates if EOF is reached or ERROR occurs. - // Explicitly deal with attempting to write too much data to memory block, rather than allow segfault. - const size_t byteCount = memorySize/sizeof(byte); - int i = 0; - while (fread(fileData + i, sizeof(byte), 1, file)) { - if (i >= byteCount) { - fprintf(stderr, "Attempting to load binary %s to memory of smaller size %zu!\n", filePath, memorySize); + char **lines = malloc(sizeof(char *) * lineCount + 1); + if (lines == NULL) { + fprintf(stderr, "Error: Could not allocate memory to store the assembly lines"); + exit(EXIT_FAILURE); + } + + rewind(fp); // Back to the beginning of the file. + + char buffer[MAX_ASM_LINE_LENGTH]; + int currentLine = 0; + + while (fgets(buffer, MAX_ASM_LINE_LENGTH, fp) != NULL) { + if (buffer[strlen(buffer) - 1] != '\n') { + // It was actually longer than the maximum. + // NOTE: I believe this must mean that this is a malformed line, so throw an error. + fprintf(stderr, "Error: Line %d in the file %s is too long\n", currentLine, filename); exit(EXIT_FAILURE); } - i++; - } + lines[currentLine] = malloc(strlen(buffer) + 1); + if (lines[currentLine] == NULL) { + fprintf(stderr, "Error: Could not allocate memory to store the assembly line"); + exit(EXIT_FAILURE); + } - if (ferror(file)) { - fprintf(stderr, "Encountered error attempting to read %s!\n", filePath); + strcpy(lines[currentLine], buffer); + currentLine++; + } + + if (ferror(fp)) { + fprintf(stderr, "Error: Could not read file %s", filename); exit(EXIT_FAILURE); } - assert(fclose(file) != EOF); - // If part of memory block was left uninitialized, initialize it to zero. - if (i < byteCount) { - memset(fileData + i, 0, (byteCount - i) * sizeof(byte)); - } - return fileData; + return lines; } diff --git a/src/fileio.h b/src/fileio.h index a2d4262..3a509ca 100644 --- a/src/fileio.h +++ b/src/fileio.h @@ -1,9 +1,13 @@ #ifndef __FILEIO__ #define __FILEIO__ +#include #include #include "global.h" #define EXIT_FAILURE 1 -extern byte *fileio_loadBin(const char *filePath, size_t memorySize); +char **readAssemblyFile(char filename[], int lineCount); +void writeBinaryFile(word instrs[], char outputFile[], int numInstrs); +int countLines(char *filename); + #endif diff --git a/src/parser.c b/src/parser.c index 8b98cb0..31b5ffc 100644 --- a/src/parser.c +++ b/src/parser.c @@ -1,8 +1,10 @@ +#include #include +#include #include +#include #include "parser.h" - -#include "a64instruction.h" +#include "a64instruction/a64instruction.h" //takes input string, read from asm file and returns //input as an a64 instruction @@ -11,79 +13,111 @@ // - use string matching to get opcode, and operands (DONE) // - check operand count (DONE) // - match opcode to a64 struct types (DONE) -// - count operands and match type/values -// - generate final a64inst and return +// - count operands and match type/values (DONE) +// - generate final a64inst and return (TODO: DP instrs) +// - ASK ABOUT OFFSET CALCULATION +// - CREATE FUNC TO TIDY UP OPERANDS IN DP + +int isOperandRegister(char *operand){ + return((strcmp(&(operand[0]), "x")==0) || (strcmp(&(operand[0]), "w")==0)); +} //calculate offsets from string -void calcluateAddressFormat(a64inst_instruction *instr, char *operandList[]){ +void calcluateAddressFormat(a64inst_instruction *instr, char *operandList[], int numOperands){ + char *endptr; + uint8_t base = strtol(&(operandList[1][2]), &endptr, 10); + instr->data.SingleTransferData.processOpData.singleDataTransferData.base = base; - if(strcmp(operandList[2][strlen(operandList[1])-1], "!")==0){ - instr->data.processOpData.addressingMode = a64inst_PRE_INDEXED; - } else if(strcmp(operandList[1][strlen(operandList[0])-1], "]") == 0) { + if(strcmp(&(operandList[2][strlen(operandList[1])-1]), "!")==0){ + instr->data.SingleTransferData.processOpData.singleDataTransferData.addressingMode = a64inst_PRE_INDEXED; + instr->data.SingleTransferData.processOpData.singleDataTransferData.a64inst_addressingModeData.indexedOffset = strtol(&(operandList[2][1]), &endptr, 10); + } else if(strcmp(&(operandList[1][strlen(operandList[0])-1]), "]") == 0) { //post-indexed - instr->data.processOpData.addressingMode = a64inst_POST_INDEXED; - } else if( (strcmp(operandList[2][strlen(operandList[1])-1], "x") == 0) - || (strcmp(operandList[2][strlen(operandList[1])-1], "w") == 0)){ + instr->data.SingleTransferData.processOpData.singleDataTransferData.addressingMode = a64inst_POST_INDEXED; + instr->data.SingleTransferData.processOpData.singleDataTransferData.a64inst_addressingModeData.indexedOffset = strtol(&(operandList[2][1]), &endptr, 10); + } else if( (isOperandRegister(&(operandList[2][0])) == 1) + || (isOperandRegister(&(operandList[2][0])) == 1)){ //register - instr->data.processOpData.addressingMode = a64inst_REGISTER_OFFSET; + instr->data.SingleTransferData.processOpData.singleDataTransferData.addressingMode = a64inst_REGISTER_OFFSET; + instr->data.SingleTransferData.processOpData.singleDataTransferData.a64inst_addressingModeData.offsetReg = strtol(&(operandList[2][1]), &endptr, 10); } else { - instr->data.processOpData.addressingMode = a64inst_UNSIGNED_OFFSET; + instr->data.SingleTransferData.processOpData.singleDataTransferData.addressingMode = a64inst_UNSIGNED_OFFSET; + if(numOperands==3){ + int offset = strtol(&(operandList[2][1]), &endptr, 10); + if(instr->data.SingleTransferData.regType == 1){ + instr->data.SingleTransferData.processOpData.singleDataTransferData.a64inst_addressingModeData.unsignedOffset = offset/8; + } else { + instr->data.SingleTransferData.processOpData.singleDataTransferData.a64inst_addressingModeData.unsignedOffset = offset/4; + } + } } } -void generateLoadStoreOperands(a64inst_instruction *instr, char *opcode, char *operandList[]){ - switch(instr->data.type){ +void generateLoadStoreOperands(a64inst_instruction *instr, char *opcode, char *operandList[], int numOperands){ + switch(instr->type){ case a64inst_SINGLETRANSFER: - if(strcmp(operandList[0][0], "x")==0){ + if(strcmp(&(operandList[0][0]), "x")==0){ //x-register - instr->data.regType = 1; + instr->data.SingleTransferData.regType = 1; } else { - instr->data.regType = 0; + instr->data.SingleTransferData.regType = 0; } char *endptr; - instr->data.target = strtol(operandList[0][0]+1, endptr, 2); - calcluateAddressFormat(instr, operandList); + instr->data.SingleTransferData.target = strtol(&(operandList[0][0])+1, &endptr, 10); + calcluateAddressFormat(instr, operandList, numOperands); break; case a64inst_LOADLITERAL: break; + default: + break; } } void generateBranchOperands(a64inst_instruction *instr, char* opcode, char *operandList[]){ - switch(instr->data.BranchType){ + char *endptr; + switch(instr->data.BranchData.BranchType){ case a64inst_UNCONDITIONAL: //define and sign extend immediate offset //use symbol table + printf("unconditional"); break; case a64inst_REGISTER: - char *endptr; - instr->data.processOpData.src = strtol(operandList[0] + 1, endptr, 2) + instr->data.BranchData.processOpData.registerData.src = strtol(operandList[0] + 1, &endptr, 10); break; case a64inst_CONDITIONAL: - char* condition = strtok(strdup(opcode), "b."); - condition = strtok(NULL, ""); - if(strcmp(condition, "eq")==0){ - instr->data.processOpData.cond = EQ; - } else if (strcmp(condition, "ne")==0){ - instr->data.processOpData.cond = NE; - } else if (strcmp(condition, "ge")==0){ - instr->data.processOpData.cond = GE; - } else if (strcmp(condition, "lt")==0){ - instr->data.processOpData.cond = LT; - } else if (strcmp(condition, "gt")==0){ - instr->data.processOpData.cond = GT; - } else if (strcmp(condition, "le")==0){ - instr->data.processOpData.cond = LE; - } else if (srtcmp(condition, "al")==0){ - instr->data.processOpData.cond = AL; + { + char *condition = NULL; + condition = strcpy(condition, opcode); + condition += 2; + if(strcmp(condition, "eq")==0){ + instr->data.BranchData.processOpData.conditionalData.cond = EQ; + } else if (strcmp(condition, "ne")==0){ + instr->data.BranchData.processOpData.conditionalData.cond = NE; + } else if (strcmp(condition, "ge")==0){ + instr->data.BranchData.processOpData.conditionalData.cond = GE; + } else if (strcmp(condition, "lt")==0){ + instr->data.BranchData.processOpData.conditionalData.cond = LT; + } else if (strcmp(condition, "gt")==0){ + instr->data.BranchData.processOpData.conditionalData.cond = GT; + } else if (strcmp(condition, "le")==0){ + instr->data.BranchData.processOpData.conditionalData.cond = LE; + } else if (strcmp(condition, "al")==0){ + instr->data.BranchData.processOpData.conditionalData.cond = AL; + } + break; + //calculate offset from symbol table. } - break; - //calculate offset from symbol table. } } -void classifyOpcode(char* opcode, a64inst_instruction *instr, char *operandList[]){ +int classifyDPInst(char *operandList[]){ + return(isOperandRegister(operandList[0]) && + isOperandRegister(operandList[1]) && + isOperandRegister(operandList[2])); +} + +void classifyOpcode(char* opcode, a64inst_instruction *instr, char *operandList[], int numOperands){ int isUnconditional = strcmp(opcode, "b"); int isRegister = strcmp(opcode, "br"); int isLoad = strcmp(opcode, "ldr"); @@ -99,7 +133,6 @@ void classifyOpcode(char* opcode, a64inst_instruction *instr, char *operandList[ instr->data.BranchData.BranchType = a64inst_REGISTER; } else { instr->data.BranchData.BranchType = a64inst_CONDITIONAL; - //instr->data.branchData.processOpData.cond = {remove first two chars of opcode} } generateBranchOperands(instr, opcode, operandList); } else if(isLoad == 0 || isStore == 0){ @@ -108,20 +141,29 @@ void classifyOpcode(char* opcode, a64inst_instruction *instr, char *operandList[ if( *address == '['){ //type is register instr->type = a64inst_SINGLETRANSFER; - instr->data.SingleTransferOpType = a64inst_SINGLE_TRANSFER_SINGLE_DATA_TRANSFER; + instr->data.SingleTransferData.SingleTransferOpType = a64inst_SINGLE_TRANSFER_SINGLE_DATA_TRANSFER; if(isLoad == 0){ - instr->data.processOpData.transferType = a64inst_LOAD; + instr->data.SingleTransferData.processOpData.singleDataTransferData.transferType = a64inst_LOAD; } else { - instr->data.processOpData.transferType = a64inst_STORE; + instr->data.SingleTransferData.processOpData.singleDataTransferData.transferType = a64inst_STORE; } } else { instr->type = a64inst_LOADLITERAL; - //instr->data.processOpData.offset = {} to be defined by symbol table + if(operandList[0][0] =='#'){ + //offset is immediate + char *immOffset = NULL; + immOffset = strcpy(immOffset, operandList[0]); + immOffset++; + char *endptr = NULL; + int offset = strtol(immOffset, &endptr, 10); + instr->data.SingleTransferData.processOpData.loadLiteralData.offset = offset; + } else { + //offset is literal, use symbol table and calculate difference + } } } else { - int numOperands = sizeof(operandList) / sizeof(operandList[0]) - if(numOperands==3){ + if(classifyDPInst(operandList)){ instr->type = a64inst_DPREGISTER; } else { instr->type = a64inst_DPIMMEDIATE; @@ -130,34 +172,38 @@ void classifyOpcode(char* opcode, a64inst_instruction *instr, char *operandList[ } } -char *tokeniseOperands(char* str, int operandCount, char *operands[]){ - char *operandsDupe = strdup(str); +void tokeniseOperands(char* str, int *operandCount, char *operands[], int *numOperands){ + assert(str != NULL); + char operandsDupe[strlen(str)+1]; + strcpy(operandsDupe, str); char *operand = strtok(operandsDupe, OPERAND_DELIMITER); operands[0] = operand; while (operand != NULL){ - operandCount++; + *operandCount = *(operandCount)+1; operand = strtok(NULL, OPERAND_DELIMITER); - operands[operandCount] = operand; + operands[*(operandCount)] = operand; } + *(numOperands) = *(operandCount)+1; } //takes inputted assembly line and returns a //pointer to an abstract representation of the instruction -a64inst_instruction *parser(char asmLine[]){ - a64inst_instruction *instr = malloc(sizeof(a64inst_instruction)); +void parser_instruction(char asmLine[], a64inst_instruction *instr) { + int numOperands = 0; if (instr == NULL){ exit(EXIT_FAILURE); } if(strcmp(asmLine, HALT_ASM_CMD) == 0){ instr->type = a64inst_HALT; - return(instr); + return; } //"opcode operand1, {operand2}, ..." //duplicated as strtok modifies the input string - char *stringptr = strdup(asmLine); + char stringptr[strlen(asmLine) + 1]; + strcpy(stringptr, asmLine); char *opcode = strtok(stringptr, " "); char *operands = strtok(NULL, ""); @@ -166,22 +212,58 @@ a64inst_instruction *parser(char asmLine[]){ //type is directive instr->type = a64inst_DIRECTIVE; - } else if(strcmp(opcode[strlen(opcode)-1], ":") == 0) { + } else if(opcode[strlen(opcode)-1]== ':') { //type is label //add to symbol table instr->type = a64inst_LABEL; - char *opcodeCpy = strdup(opcode); + char *opcodeCpy = NULL; + opcodeCpy = strcpy(opcodeCpy, opcode); char *labelData = strtok(opcodeCpy, ":"); - instr->data.label = labelData; + instr->data.LabelData.label = labelData; } else { //type is instruction int operandCount = 0; - const char *operandList[4]; - tokeniseOperands(operands, &operandCount, operandList); + char *operandList[4]; + //generate list of operands + tokeniseOperands(operands, &operandCount, operandList, &numOperands); + //categorise instruction type from opcode and operands + classifyOpcode(opcode, instr, operandList, operandCount); + //define struct values according to operands and type + switch(instr->type){ + case a64inst_BRANCH: + generateBranchOperands(instr, opcode, operandList); + break; + case a64inst_SINGLETRANSFER: + generateLoadStoreOperands(instr, opcode, operandList, numOperands); + break; + case a64inst_LOADLITERAL: + generateLoadStoreOperands(instr, opcode, operandList, numOperands); + break; + case a64inst_DPREGISTER: + //generate DP operands; + break; + case a64inst_DPIMMEDIATE: + //generate DP operands; + break; + default: + printf("INVALID INSTRUCTION"); + break; + } } - return(instr); - } +// Takes an array of strings, each string representing an assembly instruction. +// Returns an array of a64inst_instruction pointers, each representing an instruction. +a64inst_instruction *parse(char **asmLines, int lineCount) { + a64inst_instruction *instructions = malloc(sizeof(a64inst_instruction) * lineCount); + + int i = 0; + while (asmLines[i] != NULL) { + parser_instruction(asmLines[i], &instructions[i]); + i++; + } + + return instructions; +} diff --git a/src/parser.h b/src/parser.h index 5c3a461..2d7f382 100644 --- a/src/parser.h +++ b/src/parser.h @@ -1,5 +1,6 @@ -#ifndef __PARSERCONSTS__ -#define __PARSERCONSTS__ +#include "a64instruction/a64instruction.h" + #define OPERAND_DELIMITER ", " #define HALT_ASM_CMD "and x0, x0, x0" -#endif \ No newline at end of file + +a64inst_instruction *parse(char **asmLines, int lineCount); diff --git a/src/symboltable.h b/src/symboltable.c similarity index 92% rename from src/symboltable.h rename to src/symboltable.c index cd2037c..a4c99e1 100644 --- a/src/symboltable.h +++ b/src/symboltable.c @@ -1,10 +1,9 @@ #include typedef struct st st; +typedef struct node node; // forward declaration - - -typedef struct { +typedef struct node { const void* key; void* value; node* prev; @@ -29,11 +28,6 @@ void st_add(st table, void* key, void* value) { } } -// returns the pointer to key of the specified node, or null, if it does not exist -void* st_search(st table, void* key) { - return nodeSearch(table.head, key); -} - void* nodeSearch(node* n, void* key) { if (n != NULL) { if ((*n).key == key) { @@ -46,4 +40,9 @@ void* nodeSearch(node* n, void* key) { else { return NULL; } -} \ No newline at end of file +} + +// returns the pointer to key of the specified node, or null, if it does not exist +void* st_search(st table, void* key) { + return nodeSearch(table.head, key); +} diff --git a/src/twopassassembly.c b/src/twopassassembly.c index cda97ec..b4ecdec 100644 --- a/src/twopassassembly.c +++ b/src/twopassassembly.c @@ -1,33 +1,34 @@ -# include "global.h" -# include "a64instruction.h" -# include "symboltable.h" -//generates assembled code based on two pass assembly method +#include "global.h" +#include "a64instruction/a64instruction.h" +#include "symboltable.c" +#include +#include +// Generates assembled code based on the two-pass assembly method -word assembleBranch(a64inst_instruction *instr){ +word assembleBranch(a64inst_instruction *instr) { word binInstr = 0; - binInstr += (5^28); //101 start of branch instr - switch (instr->data.BranchData.BranchType) - { + binInstr += (5 << 28); // 101 start of branch instr + switch (instr->data.BranchData.BranchType) { case a64inst_UNCONDITIONAL: - //000101 - //25-0: sign extended simm26 - binInstr += instr->data.processOpData.unconditionalOffset; + // 000101 + // 25-0: sign extended simm26 + binInstr += instr->data.BranchData.processOpData.unconditionalData.unconditionalOffset; break; case a64inst_REGISTER: - //10000 - //11111 - //000000 - //9-5: address from register - //0000 - binInstr += ((instr->processOpData.src)^5); + // 10000 + // 11111 + // 000000 + // 9-5: address from register + // 0000 + binInstr += ((instr->data.BranchData.processOpData.registerData.src) << 5); break; case a64inst_CONDITIONAL: - // 01010100 + // 01010100 // 25-5: sign extended offset // 4-0: 0{condition} - binInstr += ((instr->processOpData.offset)^5); - binInstr += instr->processOpData.cond; + binInstr += ((instr->data.BranchData.processOpData.conditionalData.offset) << 5); + binInstr += instr->data.BranchData.processOpData.conditionalData.cond; break; default: break; @@ -35,43 +36,43 @@ word assembleBranch(a64inst_instruction *instr){ return binInstr; } -st* firstPass(a64inst_instruction instrs[], int numInstrs){ - //TODO: +st* firstPass(a64inst_instruction instrs[], int numInstrs) { + // TODO: // -iterate over instructions, adding to symbol table // create symbol table and map labels to addresses/lines - struct st table; - for(int i=0; i