Mach-Oオブジェクト形式を生成してみる

struct mach_header_64 {
	uint32_t	magic;		/* mach magic number identifier */
	int32_t		cputype;	/* cpu specifier */
	int32_t		cpusubtype;	/* machine specifier */
	uint32_t	filetype;	/* type of file */
	uint32_t	ncmds;		/* number of load commands */
	uint32_t	sizeofcmds;	/* the size of all the load commands */
	uint32_t	flags;		/* flags */
	uint32_t	reserved;	/* reserved */
};

struct load_command {
	uint32_t cmd;		/* type of load command */
	uint32_t cmdsize;	/* total size of command in bytes */
};

#include <mach-o/loader.h>
#include <mach-o/nlist.h>
#include <mach-o/reloc.h>
#include <mach-o/arm64/reloc.h>
#include <string.h>
#include <unistd.h>

int main(void) {
    // マシンコード (aarch64)
    static const uint32_t CODE[10] = {
        0xa9bf7bfd,    // stp     fp, lr, [sp, #-16]!
        // write(1, msg, 14);
        0xd28001c2,    // mov     x2, #14
        0x90000001,    // adrp    x1, msg@PAGE
        0x91000021,    // add     x1, x1, msg@PAGEOFF
        0x52800020,    // mov     w0, #1
        0x94000000,    // bl      _write
        // return 0;
        0x2a1f03e0,    // mov     w0, wzr
        0xa8c17bfd,    // ldp     fp, lr, [sp], #16
        0xd65f03c0,    // ret
    };

    // 読み取り専用データ
    static const char RODATA[16] = "Hello, world!\n";  // msg

    // 文字列テーブル
    static const char strtab[] =
        "\0"
        "_main\0"
        "msg\0"
        "_write\0";
    int null_nameofs = 0;
    int main_nameofs = null_nameofs + strlen("") + 1;
    int msg_nameofs = main_nameofs + strlen("_main") + 1;
    int write_nameofs = msg_nameofs + strlen("msg") + 1;

    // ヘッダ
    struct mach_header_64 header;
    // ロードコマンド
    struct segment_command_64 segmentcmd;
    struct build_version_command buildversioncmd;
    struct symtab_command symtabcmd;
    // セクション
    struct section_64 sections[2];
    // シンボル
    struct nlist_64 symbols[3];
    // 再配置情報
    struct relocation_info relocs[3];

    // ファイル内のオフセット位置を計算
    uint32_t size_of_cmds = sizeof(segmentcmd) + sizeof(sections) + sizeof(buildversioncmd) + sizeof(symtabcmd);
    uint64_t text_start_addr = 0;
    uint64_t text_start_off = sizeof(header) + size_of_cmds;
    uint64_t rodata_start_addr = sizeof(CODE);
    uint64_t rodata_start_off = text_start_off + sizeof(CODE);
    uint64_t reloc_start_off = rodata_start_off + sizeof(RODATA);
    uint64_t symbol_start_off = reloc_start_off + sizeof(relocs);
    uint64_t str_start_off = symbol_start_off + sizeof(symbols);

    sections[0] = (struct section_64){
        .sectname = "__text",
        .segname = "__TEXT",
        .addr = text_start_addr,
        .size = sizeof(CODE),
        .offset = text_start_off,
        .align = 2,  // 2^2
        .reloff = reloc_start_off,
        .nreloc = sizeof(relocs) / sizeof(*relocs),
        .flags = S_ATTR_PURE_INSTRUCTIONS | S_ATTR_SOME_INSTRUCTIONS,
    };
    sections[1] = (struct section_64){
        .sectname = "__const",
        .segname = "__DATA",
        .addr = rodata_start_addr,
        .size = sizeof(RODATA),
        .offset = rodata_start_off,
        .align = 0,  // 2^0
        .reloff = 0,
        .nreloc = 0,
        .flags = 0,
    };

    // シンボル
    symbols[0] = (struct nlist_64){
        .n_un = { main_nameofs },  // _main
        .n_type = N_SECT | N_EXT,
        .n_sect = 1,
        .n_desc = 0,
        .n_value = text_start_addr,
    };
    symbols[1] = (struct nlist_64){
        .n_un = { msg_nameofs },  // msg
        .n_type = N_SECT,
        .n_sect = 2,
        .n_desc = 0,
        .n_value = rodata_start_addr,
    };
    symbols[2] = (struct nlist_64){
        .n_un = { write_nameofs },  // _write
        .n_type = N_EXT,
        .n_sect = 0,
        .n_desc = 0,
        .n_value = 0,
    };

    // 再配置情報
    relocs[0] = (struct relocation_info){
        .r_address = 0x000008,
        .r_symbolnum = 1,  // msg
        .r_pcrel = 1,
        .r_length = 2,
        .r_extern = 1,
        .r_type = ARM64_RELOC_PAGE21,
    };
    relocs[1] = (struct relocation_info){
        .r_address = 0x00000c,
        .r_symbolnum = 1,  // msg
        .r_pcrel = 0,
        .r_length = 2,
        .r_extern = 1,
        .r_type = ARM64_RELOC_PAGEOFF12,
    };
    relocs[2] = (struct relocation_info){
        .r_address = 0x000014,
        .r_symbolnum = 2,  // _write
        .r_pcrel = 1,
        .r_length = 2,
        .r_extern = 1,
        .r_type = ARM64_RELOC_BRANCH26,
    };

    // 内容構築

    header = (struct mach_header_64){
        .magic = MH_MAGIC_64,
        .cputype = CPU_TYPE_ARM64,
        .cpusubtype = 0,
        .filetype = MH_OBJECT,
        .ncmds = 3,
        .sizeofcmds = size_of_cmds,
        .flags = MH_SUBSECTIONS_VIA_SYMBOLS,
    };
    segmentcmd = (struct segment_command_64){
        .cmd = LC_SEGMENT_64,
        .cmdsize = sizeof(segmentcmd) + sizeof(sections),
        .segname = "",
        .vmaddr = 0,
        .vmsize = sizeof(CODE) + sizeof(RODATA),
        .fileoff = text_start_off,
        .filesize = sizeof(CODE) + sizeof(RODATA),
        .maxprot = 7,   // rwx
        .initprot = 7,  // rwx
        .nsects = sizeof(sections) / sizeof(*sections),
        .flags = 0,
    };
    buildversioncmd = (struct build_version_command){
        .cmd = LC_BUILD_VERSION,
        .cmdsize = sizeof(buildversioncmd),
        .platform = PLATFORM_MACOS,
        .minos = 0x000e0000,  // 14.0.0
        .sdk = 0x000e0500,    // 14.5.0
        .ntools = 0,
    };
    symtabcmd = (struct symtab_command){
        .cmd = LC_SYMTAB,
        .cmdsize = sizeof(symtabcmd),
        .symoff = symbol_start_off,
        .nsyms = sizeof(symbols) / sizeof(*symbols),
        .stroff = str_start_off,
        .strsize = sizeof(strtab),
    };

    // 標準出力へ書き出し

    int fd = STDOUT_FILENO;
    write(fd, &header, sizeof(header));
    write(fd, &segmentcmd, sizeof(segmentcmd));
    write(fd, sections, sizeof(sections));
    write(fd, &buildversioncmd, sizeof(buildversioncmd));
    write(fd, &symtabcmd, sizeof(symtabcmd));
    write(fd, &CODE, sizeof(CODE));
    write(fd, &RODATA, sizeof(RODATA));
    write(fd, &relocs, sizeof(relocs));
    write(fd, &symbols, sizeof(symbols));
    write(fd, &strtab, sizeof(strtab));

    return 0;
}