initial commit

.gitignore

@@ -0,0 +1 @@
+build/

.gitmodules

@@ -0,0 +1,3 @@
+[submodule "tinyriscv"]
+	path = tinyriscv
+	url = git@github.com:inixyz/tinyriscv.git

Makefile

@@ -0,0 +1,46 @@
+RISCV_CC ?= riscv64-unknown-elf-gcc
+RISCV_ARCH_FLAGS ?= -march=rv32im -mabi=ilp32
+HOST_CC ?= gcc
+HOST_CFLAGS ?= -Wall -Wextra -O3 -g
+HOST_LDFLAGS ?= -Wl,-s
+
+EMU_SRC_FILES := $(wildcard src/emulated/*.c)
+EMU_OBJ_FILES := $(EMU_SRC_FILES:src/emulated/%.c=build/%.riscv.o)
+EMU_OBJ := build/emulated.riscv.elf
+
+HOST_SRC := src/main.c src/elf.c
+HOST_EXE := build/main.exe
+
+# Disassembly
+RISCV_OBJDUMP ?= $(patsubst %-gcc,%-objdump,$(RISCV_CC))
+_DISAS_TARGETS := $(filter disas-%,$(MAKECMDGOALS))
+_DISAS_FILES   := $(if $(_DISAS_TARGETS),$(filter %.elf %.o,$(MAKECMDGOALS)))
+
+
+
+.PHONY: all emulated-riscv host clean disas-risc
+
+all: $(HOST_EXE) $(EMU_OBJ)
+
+emulated-riscv: $(EMU_OBJ)
+
+host: $(HOST_EXE)
+
+$(HOST_EXE): $(HOST_SRC) tinyriscv/src/tinyriscv.h src/config.h src/elf.h | build
+	$(HOST_CC) $(HOST_CFLAGS) $(HOST_LDFLAGS) -I src -I tinyriscv/src $(HOST_SRC) -o $@
+
+build/%.riscv.o: src/emulated/%.c | build
+	$(RISCV_CC) $(RISCV_ARCH_FLAGS) -c $< -o $@
+
+$(EMU_OBJ): $(EMU_OBJ_FILES) | build
+	$(RISCV_CC) $(RISCV_ARCH_FLAGS) -nostdlib -Wl,-e,entry -Wl,-Ttext=0 $(EMU_OBJ_FILES) -o $@
+
+build:
+	mkdir -p $@
+
+clean:
+	rm -rf build
+
+disas-risc: $(_DISAS_FILES)
+	$(RISCV_OBJDUMP) -d $^
+

src/config.h

@@ -0,0 +1,5 @@
+#pragma once
+
+#define EMU_MEM_SIZE 4096u
+#define EMU_PROGRAM_PATH "build/emulated.riscv.elf"
+#define EMU_ENTRY_SYMBOL "entry"

src/elf.c

@@ -0,0 +1,346 @@
+#include "elf.h"
+#include "config.h"
+
+#include <stdint.h>
+#include <stdio.h>
+#include <errno.h>
+#include <string.h>
+#include <stdlib.h>
+
+
+static int read_file_bytes(const char* path, uint8_t** out_buf, size_t* out_size) {
+    FILE* f = fopen(path, "rb");
+    if (!f) {
+        fprintf(stderr, "Failed to open object file: %s (%s)\n", path, strerror(errno));
+        return LOAD_ERR_OPEN;
+    }
+
+    if (fseek(f, 0, SEEK_END) != 0) {
+        fclose(f);
+        fprintf(stderr, "Failed to seek object file: %s\n", path);
+        return LOAD_ERR_SMALL_FILE;
+    }
+
+    long file_size = ftell(f);
+    if (file_size < 0) {
+        fclose(f);
+        fprintf(stderr, "Failed to read object file size: %s\n", path);
+        return LOAD_ERR_SMALL_FILE;
+    }
+
+    if (fseek(f, 0, SEEK_SET) != 0) {
+        fclose(f);
+        fprintf(stderr, "Failed to rewind object file: %s\n", path);
+        return LOAD_ERR_SMALL_FILE;
+    }
+
+    uint8_t* file_buf = (uint8_t*)malloc((size_t)file_size);
+    if (!file_buf) {
+        fclose(f);
+        fprintf(stderr, "Failed to allocate %ld bytes for object file\n", file_size);
+        return LOAD_ERR_OOM;
+    }
+
+    size_t read_count = fread(file_buf, 1, (size_t)file_size, f);
+    fclose(f);
+
+    if (read_count != (size_t)file_size) {
+        free(file_buf);
+        fprintf(stderr, "Object file too small or unreadable: %s\n", path);
+        return LOAD_ERR_SMALL_FILE;
+    }
+
+    *out_buf = file_buf;
+    *out_size = read_count;
+    return LOAD_OK;
+}
+
+
+uint16_t read_u16_le(const uint8_t* data) {
+    return (uint16_t)data[0] | ((uint16_t)data[1] << 8);
+}
+
+uint32_t read_u32_le(const uint8_t* data) {
+    return (uint32_t)data[0]
+        | ((uint32_t)data[1] << 8)
+        | ((uint32_t)data[2] << 16)
+        | ((uint32_t)data[3] << 24);
+}
+
+uint32_t align_up(uint32_t value, uint32_t align) {
+    if (align == 0u) {
+        return value;
+    }
+    return (value + align - 1u) & ~(align - 1u);
+}
+
+int load_elf_object_sections(const char* path, uint8_t* memory, uint32_t mem_size) {
+    uint8_t* file_buf = NULL;
+    size_t bytes = 0u;
+    int file_status = read_file_bytes(path, &file_buf, &bytes);
+    if (file_status != LOAD_OK) {
+        return file_status;
+    }
+
+    if (bytes < 52u) {
+        fprintf(stderr, "Object file too small or unreadable: %s\n", path);
+        free(file_buf);
+        return LOAD_ERR_SMALL_FILE;
+    }
+
+    if (file_buf[0] != ELF_MAGIC_0 || file_buf[1] != ELF_MAGIC_1
+        || file_buf[2] != ELF_MAGIC_2 || file_buf[3] != ELF_MAGIC_3) {
+        fprintf(stderr, "Not an ELF file: %s\n", path);
+        free(file_buf);
+        return LOAD_ERR_NOT_ELF;
+    }
+
+    if (file_buf[4] != ELFCLASS32 || file_buf[5] != ELFDATA2LSB) {
+        fprintf(stderr, "Unsupported ELF format in %s (need 32-bit little-endian)\n", path);
+        free(file_buf);
+        return LOAD_ERR_UNSUPPORTED_FORMAT;
+    }
+
+    uint32_t e_shoff = read_u32_le(file_buf + 32);
+    uint16_t e_shentsize = read_u16_le(file_buf + 46);
+    uint16_t e_shnum = read_u16_le(file_buf + 48);
+
+    if (e_shoff >= bytes || e_shentsize == 0u) {
+        fprintf(stderr, "Invalid section header table in %s\n", path);
+        free(file_buf);
+        return LOAD_ERR_INVALID_SHT;
+    }
+
+    if ((uint32_t)e_shoff + (uint32_t)e_shentsize * (uint32_t)e_shnum > bytes) {
+        fprintf(stderr, "Section headers out of file bounds in %s\n", path);
+        free(file_buf);
+        return LOAD_ERR_SHT_OUT_OF_BOUNDS;
+    }
+
+    uint32_t cursor = 0u;
+
+    for (uint16_t i = 0; i < e_shnum; ++i) {
+        const uint8_t* sh = file_buf + e_shoff + (uint32_t)i * e_shentsize;
+
+        uint32_t sh_type = read_u32_le(sh + 4);
+        uint32_t sh_flags = read_u32_le(sh + 8);
+        uint32_t sh_addr = read_u32_le(sh + 12);
+        uint32_t sh_offset = read_u32_le(sh + 16);
+        uint32_t sh_size = read_u32_le(sh + 20);
+        uint32_t sh_addralign = read_u32_le(sh + 32);
+
+        if ((sh_flags & SHF_ALLOC) == 0u || sh_size == 0u) {
+            continue;
+        }
+
+        uint32_t dst = (sh_addr != 0u) ? sh_addr : align_up(cursor, sh_addralign);
+
+        if (dst >= mem_size || sh_size > mem_size - dst) {
+            fprintf(stderr, "Section %u does not fit into emulator memory\n", (unsigned)i);
+            free(file_buf);
+            return LOAD_ERR_SECTION_TOO_LARGE;
+        }
+
+        if (sh_type == SHT_NOBITS) {
+            memset(memory + dst, 0, sh_size);
+        } else {
+            if (sh_offset >= bytes || sh_size > (uint32_t)bytes - sh_offset) {
+                fprintf(stderr, "Section %u exceeds object file bounds\n", (unsigned)i);
+                free(file_buf);
+                return LOAD_ERR_SECTION_FILE_BOUNDS;
+            }
+            memcpy(memory + dst, file_buf + sh_offset, sh_size);
+        }
+
+        if (sh_addr == 0u) {
+            cursor = dst + sh_size;
+        }
+    }
+
+    free(file_buf);
+    return LOAD_OK;
+}
+
+int resolve_elf_symbol_memory_offset(const char* path, const char* symbol_name,
+    uint32_t mem_size, uint32_t* out_offset) {
+    uint8_t* file_buf = NULL;
+    size_t bytes = 0u;
+    int file_status = read_file_bytes(path, &file_buf, &bytes);
+    if (file_status != LOAD_OK) {
+        return file_status;
+    }
+
+    if (bytes < 52u) {
+        free(file_buf);
+        return LOAD_ERR_SMALL_FILE;
+    }
+
+    if (file_buf[0] != ELF_MAGIC_0 || file_buf[1] != ELF_MAGIC_1
+        || file_buf[2] != ELF_MAGIC_2 || file_buf[3] != ELF_MAGIC_3) {
+        free(file_buf);
+        return LOAD_ERR_NOT_ELF;
+    }
+
+    if (file_buf[4] != ELFCLASS32 || file_buf[5] != ELFDATA2LSB) {
+        free(file_buf);
+        return LOAD_ERR_UNSUPPORTED_FORMAT;
+    }
+
+    uint32_t e_shoff = read_u32_le(file_buf + 32);
+    uint16_t e_shentsize = read_u16_le(file_buf + 46);
+    uint16_t e_shnum = read_u16_le(file_buf + 48);
+
+    if (e_shoff >= bytes || e_shentsize == 0u) {
+        free(file_buf);
+        return LOAD_ERR_INVALID_SHT;
+    }
+
+    if ((uint32_t)e_shoff + (uint32_t)e_shentsize * (uint32_t)e_shnum > bytes) {
+        free(file_buf);
+        return LOAD_ERR_SHT_OUT_OF_BOUNDS;
+    }
+
+    uint32_t* section_runtime = (uint32_t*)calloc(e_shnum, sizeof(uint32_t));
+    uint32_t* section_addr = (uint32_t*)calloc(e_shnum, sizeof(uint32_t));
+    uint8_t* section_loaded = (uint8_t*)calloc(e_shnum, sizeof(uint8_t));
+    if (!section_runtime || !section_addr || !section_loaded) {
+        free(section_runtime);
+        free(section_addr);
+        free(section_loaded);
+        free(file_buf);
+        return LOAD_ERR_OOM;
+    }
+
+    uint32_t cursor = 0u;
+    for (uint16_t i = 0; i < e_shnum; ++i) {
+        const uint8_t* sh = file_buf + e_shoff + (uint32_t)i * e_shentsize;
+        uint32_t sh_flags = read_u32_le(sh + 8);
+        uint32_t sh_addr = read_u32_le(sh + 12);
+        uint32_t sh_size = read_u32_le(sh + 20);
+        uint32_t sh_addralign = read_u32_le(sh + 32);
+
+        if ((sh_flags & SHF_ALLOC) == 0u || sh_size == 0u) {
+            continue;
+        }
+
+        uint32_t dst = (sh_addr != 0u) ? sh_addr : align_up(cursor, sh_addralign);
+        if (dst >= mem_size || sh_size > mem_size - dst) {
+            free(section_runtime);
+            free(section_addr);
+            free(section_loaded);
+            free(file_buf);
+            return LOAD_ERR_SECTION_TOO_LARGE;
+        }
+
+        section_runtime[i] = dst;
+        section_addr[i] = sh_addr;
+        section_loaded[i] = 1u;
+
+        if (sh_addr == 0u) {
+            cursor = dst + sh_size;
+        }
+    }
+
+    int found_symtab = 0;
+    int status = LOAD_ERR_SYMBOL_NOT_FOUND;
+
+    for (uint16_t i = 0; i < e_shnum; ++i) {
+        const uint8_t* sh = file_buf + e_shoff + (uint32_t)i * e_shentsize;
+        uint32_t sh_type = read_u32_le(sh + 4);
+        uint32_t sh_offset = read_u32_le(sh + 16);
+        uint32_t sh_size = read_u32_le(sh + 20);
+        uint32_t sh_link = read_u32_le(sh + 24);
+        uint32_t sh_entsize = read_u32_le(sh + 36);
+
+        if (sh_type != SHT_SYMTAB) {
+            continue;
+        }
+
+        found_symtab = 1;
+
+        if (sh_link >= e_shnum || sh_offset >= bytes || sh_size > (uint32_t)bytes - sh_offset) {
+            status = LOAD_ERR_STRTAB_INVALID;
+            break;
+        }
+
+        const uint8_t* strtab_sh = file_buf + e_shoff + sh_link * (uint32_t)e_shentsize;
+        uint32_t strtab_offset = read_u32_le(strtab_sh + 16);
+        uint32_t strtab_size = read_u32_le(strtab_sh + 20);
+
+        if (strtab_offset >= bytes || strtab_size > (uint32_t)bytes - strtab_offset) {
+            status = LOAD_ERR_STRTAB_INVALID;
+            break;
+        }
+
+        if (sh_entsize == 0u) {
+            sh_entsize = 16u;
+        }
+
+        if (sh_entsize < 16u) {
+            status = LOAD_ERR_STRTAB_INVALID;
+            break;
+        }
+
+        uint32_t count = sh_size / sh_entsize;
+        const uint8_t* symtab = file_buf + sh_offset;
+        const char* strtab = (const char*)(file_buf + strtab_offset);
+
+        for (uint32_t s = 0u; s < count; ++s) {
+            const uint8_t* sym = symtab + s * sh_entsize;
+            uint32_t st_name = read_u32_le(sym + 0);
+            uint32_t st_value = read_u32_le(sym + 4);
+            uint16_t st_shndx = read_u16_le(sym + 14);
+
+            if (st_name >= strtab_size) {
+                continue;
+            }
+
+            const char* name = strtab + st_name;
+            if (strcmp(name, symbol_name) != 0) {
+                continue;
+            }
+
+            if (st_shndx == 0u) {
+                continue;
+            }
+
+            if (st_shndx >= e_shnum || !section_loaded[st_shndx]) {
+                status = LOAD_ERR_SYMBOL_INVALID_SECTION;
+                break;
+            }
+
+            uint32_t section_base_runtime = section_runtime[st_shndx];
+            uint32_t section_base_elf = section_addr[st_shndx];
+            if (st_value < section_base_elf) {
+                status = LOAD_ERR_SYMBOL_INVALID_SECTION;
+                break;
+            }
+
+            uint32_t symbol_offset = section_base_runtime + (st_value - section_base_elf);
+            if (symbol_offset >= mem_size) {
+                status = LOAD_ERR_SYMBOL_OUT_OF_RANGE;
+                break;
+            }
+
+            *out_offset = symbol_offset;
+            status = LOAD_OK;
+            break;
+        }
+
+        if (status == LOAD_OK || status == LOAD_ERR_SYMBOL_INVALID_SECTION
+            || status == LOAD_ERR_SYMBOL_OUT_OF_RANGE) {
+            break;
+        }
+    }
+
+    if (!found_symtab) {
+        status = LOAD_ERR_SYMTAB_MISSING;
+    }
+
+    free(section_runtime);
+    free(section_addr);
+    free(section_loaded);
+    free(file_buf);
+    return status;
+}
+

src/elf.h

@@ -0,0 +1,41 @@
+#pragma once
+
+#define ELF_MAGIC_0 0x7f
+#define ELF_MAGIC_1 'E'
+#define ELF_MAGIC_2 'L'
+#define ELF_MAGIC_3 'F'
+
+#define ELFCLASS32 1
+#define ELFDATA2LSB 1
+
+#define SHT_NOBITS 8u
+#define SHT_SYMTAB 2u
+#define SHF_ALLOC 0x2u
+
+#include <stdint.h>
+
+enum {
+    LOAD_OK = 1,
+    LOAD_ERR_OPEN = -1,
+    LOAD_ERR_SMALL_FILE = -2,
+    LOAD_ERR_NOT_ELF = -3,
+    LOAD_ERR_UNSUPPORTED_FORMAT = -4,
+    LOAD_ERR_INVALID_SHT = -5,
+    LOAD_ERR_SHT_OUT_OF_BOUNDS = -6,
+    LOAD_ERR_SECTION_TOO_LARGE = -7,
+    LOAD_ERR_SECTION_FILE_BOUNDS = -8,
+    LOAD_ERR_OOM = -9,
+    LOAD_ERR_SYMTAB_MISSING = -10,
+    LOAD_ERR_STRTAB_INVALID = -11,
+    LOAD_ERR_SYMBOL_NOT_FOUND = -12,
+    LOAD_ERR_SYMBOL_INVALID_SECTION = -13,
+    LOAD_ERR_SYMBOL_OUT_OF_RANGE = -14
+};
+
+uint16_t read_u16_le(const uint8_t* data);
+uint32_t read_u32_le(const uint8_t* data);
+uint32_t align_up(uint32_t value, uint32_t align);
+int load_elf_object_sections(const char* path, uint8_t* memory, uint32_t mem_size);
+int resolve_elf_symbol_memory_offset(const char* path, const char* symbol_name,
+    uint32_t mem_size, uint32_t* out_offset);
+

src/emulated/custom_instructions.h

@@ -0,0 +1,15 @@
+#pragma once
+
+
+#define trunc_xor(a, imm12)                             \
+    __extension__ ({                                    \
+        int32_t _result;                                \
+        asm volatile (                                  \
+            ".insn i 0x0B, 0x00, %0, %1, %2"            \
+            : "=r" (_result)                            \
+            : "r"  ((int32_t)(a)),                      \
+              "I"  (imm12)                              \
+        );                                              \
+        _result;                                        \
+    })
+

src/emulated/function.c

@@ -0,0 +1,13 @@
+
+#include "function.h"
+
+#include "custom_instructions.h"
+#include <stdint.h>
+
+
+
+int32_t operation(int32_t a) {
+    return trunc_xor(a, 3);
+}
+
+

src/emulated/function.h

@@ -0,0 +1,6 @@
+#pragma once
+
+#include <stdint.h>
+
+int32_t operation(int32_t a);
+

src/emulated/program.c

@@ -0,0 +1,10 @@
+#include <stdint.h>
+
+#include "function.h"
+#include "custom_instructions.h"
+
+
+uint8_t entry(uint8_t a) {
+    return operation(a);
+}
+

src/main.c

@@ -0,0 +1,75 @@
+#define TINYRISCV_M
+
+#include "tinyriscv.h"
+#include "elf.h"
+#include "config.h"
+
+#include <errno.h>
+#include <stdio.h>
+#include <string.h>
+
+
+#define TEST_A 5u
+#define TEST_B 7u
+#define MAX_STEPS 1024u
+
+
+int main(){
+    uint8_t memory[EMU_MEM_SIZE]; //4kb of internal memory
+    tinyriscv_core cpu;   //create the emulation core
+    memset(&cpu, 0, sizeof(cpu));
+    cpu.mem = memory;  //use the uint8_t array as the memory
+    cpu.mem_size = EMU_MEM_SIZE;
+    memset(memory, 0, sizeof(memory));
+
+    int load_status = load_elf_object_sections(EMU_PROGRAM_PATH, memory, cpu.mem_size);
+    if (load_status != LOAD_OK) {
+        fprintf(stderr, "ELF load failed with code %d for '%s'\n", load_status, EMU_PROGRAM_PATH);
+        fflush(stderr);
+        printf("ELF load failed with code %d\n", load_status);
+        fflush(stdout);
+        return 1;
+    }
+
+    uint32_t entry_offset = 0u;
+    int symbol_status = resolve_elf_symbol_memory_offset(
+        EMU_PROGRAM_PATH, EMU_ENTRY_SYMBOL, cpu.mem_size, &entry_offset);
+    if (symbol_status != LOAD_OK) {
+        fprintf(stderr, "Entry symbol '%s' resolve failed with code %d for '%s'\n",
+            EMU_ENTRY_SYMBOL, symbol_status, EMU_PROGRAM_PATH);
+        return 2;
+    }
+
+    tinyriscv_init(&cpu); //initialise core for execution
+    cpu.pc = tinyriscv_MEM_OFFSET + entry_offset;
+
+    // Call operation(a, b) using RV32 calling convention:
+    // a0=x10, return value in a0, return address in ra=x1.
+    cpu.x[10] = TEST_A;
+    cpu.x[1] = tinyriscv_MEM_OFFSET + cpu.mem_size;
+
+    uint32_t steps = 0u;
+    while (tinyriscv_valid_step(&cpu) && steps < MAX_STEPS) {
+        tinyriscv_step(&cpu);
+        ++steps;
+    }
+
+    if (steps == MAX_STEPS) {
+        fprintf(stderr, "Execution did not halt within %u steps\n", MAX_STEPS);
+        return 2;
+    }
+
+    uint32_t actual = (uint32_t)cpu.x[10];
+    uint32_t expected = ((TEST_A << 3) ^ 0xFF) >> 3;
+    printf("operation(%u) -> %u (expected %u)\n",
+        (unsigned)TEST_A,
+        (unsigned)actual, (unsigned)expected);
+
+    if (actual != expected) {
+        fprintf(stderr, "Verification failed\n");
+        return 3;
+    }
+
+    return 0;
+}
+