Created
November 22, 2021 19:13
-
-
Save moyix/7345bb494ed0cd90b9c8725f03098642 to your computer and use it in GitHub Desktop.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
#!/usr/bin/env python3 | |
import sys | |
import struct | |
from pprint import pprint | |
# Parser for ELF executables | |
def parse_elf_header(f): | |
# Parse ELF header | |
f.seek(0) | |
magic = f.read(4) | |
if magic != b'\x7fELF': | |
raise Exception('Not an ELF file') | |
ei_class = f.read(1) | |
ei_data = f.read(1) | |
ei_version = f.read(1) | |
ei_osabi = f.read(1) | |
ei_abiversion = f.read(1) | |
pad = f.read(7) | |
e_type = f.read(2) | |
e_machine = f.read(2) | |
e_version = f.read(4) | |
if ei_class == b'\x02': | |
e_entry = f.read(8) | |
e_phoff = f.read(8) | |
e_shoff = f.read(8) | |
else: | |
e_entry = f.read(4) | |
e_phoff = f.read(4) | |
e_shoff = f.read(4) | |
e_flags = f.read(4) | |
e_ehsize = f.read(2) | |
e_phentsize = f.read(2) | |
e_phnum = f.read(2) | |
e_shentsize = f.read(2) | |
e_shnum = f.read(2) | |
e_shstrndx = f.read(2) | |
ei_class = int.from_bytes(ei_class, byteorder='little') | |
ei_data = int.from_bytes(ei_data, byteorder='little') | |
ei_version = int.from_bytes(ei_version, byteorder='little') | |
ei_osabi = int.from_bytes(ei_osabi, byteorder='little') | |
ei_abiversion = int.from_bytes(ei_abiversion, byteorder='little') | |
e_type = int.from_bytes(e_type, byteorder='little') | |
e_machine = int.from_bytes(e_machine, byteorder='little') | |
e_version = int.from_bytes(e_version, byteorder='little') | |
e_entry = int.from_bytes(e_entry, byteorder='little') | |
e_phoff = int.from_bytes(e_phoff, byteorder='little') | |
e_shoff = int.from_bytes(e_shoff, byteorder='little') | |
e_flags = int.from_bytes(e_flags, byteorder='little') | |
e_ehsize = int.from_bytes(e_ehsize, byteorder='little') | |
e_phentsize = int.from_bytes(e_phentsize, byteorder='little') | |
e_phnum = int.from_bytes(e_phnum, byteorder='little') | |
e_shentsize = int.from_bytes(e_shentsize, byteorder='little') | |
e_shnum = int.from_bytes(e_shnum, byteorder='little') | |
e_shstrndx = int.from_bytes(e_shstrndx, byteorder='little') | |
return { | |
'magic': magic, | |
'ei_class': ei_class, | |
'ei_data': ei_data, | |
'ei_version': ei_version, | |
'ei_osabi': ei_osabi, | |
'ei_abiversion': ei_abiversion, | |
'e_type': e_type, | |
'e_machine': e_machine, | |
'e_version': e_version, | |
'e_entry': e_entry, | |
'e_phoff': e_phoff, | |
'e_shoff': e_shoff, | |
'e_flags': e_flags, | |
'e_ehsize': e_ehsize, | |
'e_phentsize': e_phentsize, | |
'e_phnum': e_phnum, | |
'e_shentsize': e_shentsize, | |
'e_shnum': e_shnum, | |
'e_shstrndx': e_shstrndx | |
} | |
def parse_program_headers(f, header): | |
# Parse program header | |
f.seek(header['e_phoff']) | |
ph_headers = [] | |
for i in range(header['e_phnum']): | |
ph_header = {} | |
if header['ei_class'] == 2: | |
ph_header['p_type'] = f.read(4) | |
ph_header['p_flags'] = f.read(4) | |
ph_header['p_offset'] = f.read(8) | |
ph_header['p_vaddr'] = f.read(8) | |
ph_header['p_paddr'] = f.read(8) | |
ph_header['p_filesz'] = f.read(8) | |
ph_header['p_memsz'] = f.read(8) | |
ph_header['p_align'] = f.read(8) | |
else: | |
ph_header['p_type'] = f.read(4) | |
ph_header['p_offset'] = f.read(4) | |
ph_header['p_vaddr'] = f.read(4) | |
ph_header['p_paddr'] = f.read(4) | |
ph_header['p_filesz'] = f.read(4) | |
ph_header['p_memsz'] = f.read(4) | |
ph_header['p_flags'] = f.read(4) | |
ph_header['p_align'] = f.read(4) | |
# Convert to ints | |
for k in ph_header: | |
ph_header[k] = int.from_bytes(ph_header[k], byteorder='little') | |
ph_headers.append(ph_header) | |
return ph_headers | |
def parse_section_headers(f, header): | |
# Parse section header | |
f.seek(header['e_shoff']) | |
sh_headers = [] | |
for i in range(header['e_shnum']): | |
sh_header = {} | |
if header['ei_class'] == 2: | |
sh_header['sh_name'] = f.read(4) | |
sh_header['sh_type'] = f.read(4) | |
sh_header['sh_flags'] = f.read(8) | |
sh_header['sh_addr'] = f.read(8) | |
sh_header['sh_offset'] = f.read(8) | |
sh_header['sh_size'] = f.read(8) | |
sh_header['sh_link'] = f.read(4) | |
sh_header['sh_info'] = f.read(4) | |
sh_header['sh_addralign'] = f.read(8) | |
sh_header['sh_entsize'] = f.read(8) | |
else: | |
sh_header['sh_name'] = f.read(4) | |
sh_header['sh_type'] = f.read(4) | |
sh_header['sh_flags'] = f.read(4) | |
sh_header['sh_addr'] = f.read(4) | |
sh_header['sh_offset'] = f.read(4) | |
sh_header['sh_size'] = f.read(4) | |
sh_header['sh_link'] = f.read(4) | |
sh_header['sh_info'] = f.read(4) | |
sh_header['sh_addralign'] = f.read(4) | |
sh_header['sh_entsize'] = f.read(4) | |
# Convert to ints | |
for k in sh_header: | |
sh_header[k] = int.from_bytes(sh_header[k], byteorder='little') | |
sh_headers.append(sh_header) | |
return sh_headers | |
# Only care about i386, x86_64, ARM, and ARM64 | |
elf_machine_types = { | |
0x03: 'i386', | |
0x3e: 'x86_64', | |
0x28: 'ARM', | |
0xB7: 'ARM64' | |
} | |
e_type_names = { | |
0x00: 'ET_NONE', | |
0x01: 'ET_REL', | |
0x02: 'ET_EXEC', | |
0x03: 'ET_DYN', | |
0x04: 'ET_CORE', | |
0xfe00: 'ET_LOPROC', | |
0xffff: 'ET_HIPROC' | |
} | |
def print_elf_header(header): | |
print('ELF Header:') | |
print(' Magic: ' + header['magic'].hex()) | |
print(' Class: ' + 'ELF64' if header['ei_class'] == 2 else 'ELF32') | |
print(' Data: ' + 'LE' if header['ei_data'] == 1 else 'BE') | |
print(' Version: ' + str(header['ei_version'])) | |
print(' OS/ABI: ' + str(header['ei_osabi'])) | |
print(' ABI Version: ' + str(header['ei_abiversion'])) | |
print(' Type: ' + e_type_names[header['e_type']]) | |
print(' Machine: ' + elf_machine_types[header['e_machine']]) | |
print(' Version: ' + str(header['e_version'])) | |
print(' Entry point: ' + hex(header['e_entry'])) | |
print(' Program header offset: ' + hex(header['e_phoff'])) | |
print(' Section header offset: ' + hex(header['e_shoff'])) | |
print(' Flags: ' + hex(header['e_flags'])) | |
print(' Size of this header: ' + hex(header['e_ehsize'])) | |
print(' Size of program header: ' + hex(header['e_phentsize'])) | |
print(' Number of program headers: ' + hex(header['e_phnum'])) | |
print(' Size of section header: ' + hex(header['e_shentsize'])) | |
print(' Number of section headers: ' + hex(header['e_shnum'])) | |
print(' Section header string table index: ' + hex(header['e_shstrndx'])) | |
program_header_names = { | |
0x00: 'PT_NULL', | |
0x01: 'PT_LOAD', | |
0x02: 'PT_DYNAMIC', | |
0x03: 'PT_INTERP', | |
0x04: 'PT_NOTE', | |
0x05: 'PT_SHLIB', | |
0x06: 'PT_PHDR', | |
0x07: 'PT_TLS', | |
0x60000000: 'PT_LOOS', | |
0x6fffffff: 'PT_HIOS', | |
0x70000000: 'PT_LOPROC', | |
0x7fffffff: 'PT_HIPROC', | |
} | |
program_header_names_reverse = {v: k for k, v in program_header_names.items()} | |
program_header_names[program_header_names_reverse['PT_LOOS'] + 0x474e550] = 'PT_GNU_EH_FRAME' | |
program_header_names[program_header_names_reverse['PT_LOOS'] + 0x474e551] = 'PT_GNU_STACK' | |
program_header_names[program_header_names_reverse['PT_LOOS'] + 0x474e552] = 'PT_GNU_RELRO' | |
program_header_names[program_header_names_reverse['PT_LOOS'] + 0x474e553] = 'PT_GNU_PROPERTY' | |
program_header_names[program_header_names_reverse['PT_LOOS'] + 0x5a3dbe6] = 'PT_OPENBSD_RANDOMIZE' | |
program_header_names[program_header_names_reverse['PT_LOOS'] + 0x5a3dbe7] = 'PT_OPENBSD_WXNEEDED' | |
program_header_names[program_header_names_reverse['PT_LOOS'] + 0x5a41be6] = 'PT_OPENBSD_BOOTDATA' | |
def print_program_header(ph_header): | |
print('Program Header:') | |
if ph_header['p_type'] in program_header_names: | |
print(' Type: ' + program_header_names[ph_header['p_type']]) | |
else: | |
print(' Type: ' + hex(ph_header['p_type'])) | |
print(' Offset: ' + hex(ph_header['p_offset'])) | |
print(' Virtual Address: ' + hex(ph_header['p_vaddr'])) | |
print(' Physical Address: ' + hex(ph_header['p_paddr'])) | |
print(' File Size: ' + hex(ph_header['p_filesz'])) | |
print(' Memory Size: ' + hex(ph_header['p_memsz'])) | |
print(' Flags: ' + hex(ph_header['p_flags'])) | |
print(' Alignment: ' + hex(ph_header['p_align'])) | |
def parse_string_table(f, header, sh_headers): | |
# Parse string table | |
sh_strtab = sh_headers[header['e_shstrndx']] | |
f.seek(sh_strtab['sh_offset']) | |
string_table = f.read(sh_strtab['sh_size']) | |
return string_table | |
section_header_types = { | |
0x00: 'SHT_NULL', | |
0x01: 'SHT_PROGBITS', | |
0x02: 'SHT_SYMTAB', | |
0x03: 'SHT_STRTAB', | |
0x04: 'SHT_RELA', | |
0x05: 'SHT_HASH', | |
0x06: 'SHT_DYNAMIC', | |
0x07: 'SHT_NOTE', | |
0x08: 'SHT_NOBITS', | |
0x09: 'SHT_REL', | |
0x0a: 'SHT_SHLIB', | |
0x0b: 'SHT_DYNSYM', | |
0x0c: 'SHT_INIT_ARRAY', | |
0x0d: 'SHT_FINI_ARRAY', | |
0x0e: 'SHT_PREINIT_ARRAY', | |
0x0f: 'SHT_GROUP', | |
0x10: 'SHT_SYMTAB_SHNDX', | |
0x11: 'SHT_NUM', | |
0x12: 'SHT_LOOS', | |
0x13: 'SHT_GNU_ATTRIBUTES', | |
0x14: 'SHT_GNU_HASH', | |
0x15: 'SHT_GNU_LIBLIST', | |
0x16: 'SHT_CHECKSUM', | |
0x17: 'SHT_LOSUNW', | |
0x18: 'SHT_SUNW_move', | |
0x19: 'SHT_SUNW_COMDAT', | |
0x1a: 'SHT_SUNW_syminfo', | |
0x1b: 'SHT_GNU_verdef', | |
0x1c: 'SHT_GNU_verneed', | |
0x1d: 'SHT_GNU_versym', | |
0x1e: 'SHT_HISUNW', | |
0x1f: 'SHT_HIOS', | |
0x40: 'SHT_LOPROC', | |
0x7f: 'SHT_HIPROC', | |
0x80: 'SHT_LOUSER', | |
0xff: 'SHT_HIUSER' | |
} | |
def print_section_header(sh_header, string_table): | |
print('Section Header:') | |
print(' Name: ' + string_table[sh_header['sh_name']:].decode('utf-8').split('\x00')[0]) | |
if sh_header['sh_type'] in section_header_types: | |
print(' Type: ' + section_header_types[sh_header['sh_type']]) | |
else: | |
print(' Type: ' + hex(sh_header['sh_type'])) | |
print(' Flags: ' + hex(sh_header['sh_flags'])) | |
print(' Address: ' + hex(sh_header['sh_addr'])) | |
print(' Offset: ' + hex(sh_header['sh_offset'])) | |
print(' Size: ' + hex(sh_header['sh_size'])) | |
print(' Link: ' + hex(sh_header['sh_link'])) | |
print(' Info: ' + hex(sh_header['sh_info'])) | |
print(' Address alignment: ' + hex(sh_header['sh_addralign'])) | |
print(' Entry size: ' + hex(sh_header['sh_entsize'])) | |
elf_file = open(sys.argv[1], 'rb') | |
header = parse_elf_header(elf_file) | |
print_elf_header(header) | |
ph_headers = parse_program_headers(elf_file, header) | |
for ph_header in ph_headers: | |
print_program_header(ph_header) | |
sh_headers = parse_section_headers(elf_file, header) | |
string_table = parse_string_table(elf_file, header, sh_headers) | |
for sh_header in sh_headers: | |
print_section_header(sh_header, string_table) | |
elf_file.close() |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment