Home | History | Annotate | Download | only in Scripts 
      1 #!/usr/bin/env python
      2 
      3 import struct
      4 import sys
      5 import StringIO
      6 
      7 import common_dump
      8 
      9 class Reader:
     10     def __init__(self, path):
     11         if path == "-":
     12             # Snarf all the data so we can seek.
     13             self.file = StringIO.StringIO(sys.stdin.read())
     14         else:
     15             self.file = open(path, "rb")
     16         self.isLSB = None
     17         self.is64Bit = None
     18         self.isN64 = False
     19 
     20     def seek(self, pos):
     21         self.file.seek(pos)
     22 
     23     def read(self, N):
     24         data = self.file.read(N)
     25         if len(data) != N:
     26             raise ValueError, "Out of data!"
     27         return data
     28 
     29     def read8(self):
     30         return (ord(self.read(1)), 8)
     31 
     32     def read16(self):
     33         return (struct.unpack('><'[self.isLSB] + 'H', self.read(2))[0], 16)
     34 
     35     def read32(self):
     36         return (struct.unpack('><'[self.isLSB] + 'I', self.read(4))[0], 32)
     37 
     38     def read64(self):
     39         return (struct.unpack('><'[self.isLSB] + 'Q', self.read(8))[0], 64)
     40 
     41     def readWord(self):
     42         if self.is64Bit:
     43             return self.read64()
     44         else:
     45             return self.read32()
     46 
     47 class StringTable:
     48     def __init__(self, strings):
     49        self.string_table = strings
     50 
     51     def __getitem__(self, index):
     52        end = self.string_table.index('\x00', index)
     53        return self.string_table[index:end]
     54 
     55 class Section:
     56     def __init__(self, f):
     57         self.sh_name = f.read32()
     58         self.sh_type = f.read32()
     59         self.sh_flags = f.readWord()
     60         self.sh_addr = f.readWord()
     61         self.sh_offset = f.readWord()
     62         self.sh_size = f.readWord()
     63         self.sh_link = f.read32()
     64         self.sh_info = f.read32()
     65         self.sh_addralign = f.readWord()
     66         self.sh_entsize = f.readWord()
     67 
     68     def dump(self, shstrtab, f, strtab, dumpdata):
     69         print "  (('sh_name', %s)" % common_dump.HexDump(self.sh_name), "# %r" % shstrtab[self.sh_name[0]]
     70         print "   ('sh_type', %s)" % common_dump.HexDump(self.sh_type)
     71         print "   ('sh_flags', %s)" % common_dump.HexDump(self.sh_flags)
     72         print "   ('sh_addr', %s)" % common_dump.HexDump(self.sh_addr)
     73         print "   ('sh_offset', %s)" % common_dump.HexDump(self.sh_offset)
     74         print "   ('sh_size', %s)" % common_dump.HexDump(self.sh_size)
     75         print "   ('sh_link', %s)" % common_dump.HexDump(self.sh_link)
     76         print "   ('sh_info', %s)" % common_dump.HexDump(self.sh_info)
     77         print "   ('sh_addralign', %s)" % common_dump.HexDump(self.sh_addralign)
     78         print "   ('sh_entsize', %s)" % common_dump.HexDump(self.sh_entsize)
     79         if self.sh_type[0] == 2: # SHT_SYMTAB
     80             print "   ('_symbols', ["
     81             dumpSymtab(f, self, strtab)
     82             print "   ])"
     83         elif self.sh_type[0] == 4 or self.sh_type[0] == 9: # SHT_RELA / SHT_REL
     84             print "   ('_relocations', ["
     85             dumpRel(f, self, self.sh_type[0] == 4)
     86             print "   ])"
     87         elif dumpdata:
     88             f.seek(self.sh_offset[0])
     89             if self.sh_type != 8: # != SHT_NOBITS
     90                 data = f.read(self.sh_size[0])
     91                 print "   ('_section_data', '%s')" % common_dump.dataToHex(data)
     92             else:
     93                 print "   ('_section_data', '')" 
     94         print "  ),"
     95 
     96 def dumpSymtab(f, section, strtab):
     97     entries = section.sh_size[0] // section.sh_entsize[0]
     98 
     99     for index in range(entries):
    100         f.seek(section.sh_offset[0] + index * section.sh_entsize[0])
    101         print "    # Symbol %s" % index
    102         name = f.read32()
    103         print "    (('st_name', %s)" % common_dump.HexDump(name), "# %r" % strtab[name[0]]
    104         if not f.is64Bit:
    105             print "     ('st_value', %s)" % common_dump.HexDump(f.read32())
    106             print "     ('st_size', %s)" % common_dump.HexDump(f.read32())
    107         st_info = f.read8()[0]
    108         st_bind = (st_info >> 4, 4)
    109         st_type = (st_info & 0xf, 4)
    110         print "     ('st_bind', %s)" % common_dump.HexDump(st_bind)
    111         print "     ('st_type', %s)" % common_dump.HexDump(st_type)
    112         print "     ('st_other', %s)" % common_dump.HexDump(f.read8())
    113         print "     ('st_shndx', %s)" % common_dump.HexDump(f.read16())
    114         if f.is64Bit:
    115             print "     ('st_value', %s)" % common_dump.HexDump(f.read64())
    116             print "     ('st_size', %s)" % common_dump.HexDump(f.read64())
    117         print "    ),"
    118 
    119 def dumpRel(f, section, dumprela = False):
    120     entries = section.sh_size[0] // section.sh_entsize[0]
    121 
    122     for index in range(entries):
    123         f.seek(section.sh_offset[0] + index * section.sh_entsize[0])
    124         print "    # Relocation %s" % index
    125         print "    (('r_offset', %s)" % common_dump.HexDump(f.readWord())
    126 
    127         if f.isN64:
    128             r_sym =   f.read32()
    129             r_ssym =  f.read8()
    130             r_type3 = f.read8()
    131             r_type2 = f.read8()
    132             r_type =  f.read8()
    133             print "     ('r_sym', %s)" % common_dump.HexDump(r_sym)
    134             print "     ('r_ssym', %s)" % common_dump.HexDump(r_ssym)
    135             print "     ('r_type3', %s)" % common_dump.HexDump(r_type3)
    136             print "     ('r_type2', %s)" % common_dump.HexDump(r_type2)
    137             print "     ('r_type', %s)" % common_dump.HexDump(r_type)
    138         else:
    139             r_info = f.readWord()[0]
    140             if f.is64Bit:
    141                 r_sym = (r_info >> 32, 32)
    142                 r_type = (r_info & 0xffffffff, 32)
    143             else:
    144                 r_sym = (r_info >> 8, 24)
    145                 r_type = (r_info & 0xff, 8)
    146             print "     ('r_sym', %s)" % common_dump.HexDump(r_sym)
    147             print "     ('r_type', %s)" % common_dump.HexDump(r_type)
    148         if dumprela:
    149             print "     ('r_addend', %s)" % common_dump.HexDump(f.readWord())
    150         print "    ),"
    151 
    152 def dumpELF(path, opts):
    153     f = Reader(path)
    154 
    155     magic = f.read(4)
    156     assert magic == '\x7FELF'
    157 
    158     fileclass = f.read8()
    159     if fileclass[0] == 1: # ELFCLASS32
    160         f.is64Bit = False
    161     elif fileclass[0] == 2: # ELFCLASS64
    162         f.is64Bit = True
    163     else:
    164         raise ValueError, "Unknown file class %s" % common_dump.HexDump(fileclass)
    165     print "('e_indent[EI_CLASS]', %s)" % common_dump.HexDump(fileclass)
    166 
    167     byteordering = f.read8()
    168     if byteordering[0] == 1: # ELFDATA2LSB
    169         f.isLSB = True
    170     elif byteordering[0] == 2: # ELFDATA2MSB
    171         f.isLSB = False
    172     else:
    173         raise ValueError, "Unknown byte ordering %s" % common_dump.HexDump(byteordering)
    174     print "('e_indent[EI_DATA]', %s)" % common_dump.HexDump(byteordering)
    175 
    176     print "('e_indent[EI_VERSION]', %s)" % common_dump.HexDump(f.read8())
    177     print "('e_indent[EI_OSABI]', %s)" % common_dump.HexDump(f.read8())
    178     print "('e_indent[EI_ABIVERSION]', %s)" % common_dump.HexDump(f.read8())
    179 
    180     f.seek(16) # Seek to end of e_ident.
    181 
    182     print "('e_type', %s)" % common_dump.HexDump(f.read16())
    183 
    184     # Does any other architecture use N64?
    185     e_machine = f.read16()
    186     if e_machine[0] == 0x0008 and f.is64Bit: # EM_MIPS && 64 bit
    187         f.isN64 = True 
    188     
    189     print "('e_machine', %s)" % common_dump.HexDump(e_machine)
    190     print "('e_version', %s)" % common_dump.HexDump(f.read32())
    191     print "('e_entry', %s)" % common_dump.HexDump(f.readWord())
    192     print "('e_phoff', %s)" % common_dump.HexDump(f.readWord())
    193     e_shoff = f.readWord()
    194     print "('e_shoff', %s)" % common_dump.HexDump(e_shoff)
    195     print "('e_flags', %s)" % common_dump.HexDump(f.read32())
    196     print "('e_ehsize', %s)" % common_dump.HexDump(f.read16())
    197     print "('e_phentsize', %s)" % common_dump.HexDump(f.read16())
    198     print "('e_phnum', %s)" % common_dump.HexDump(f.read16())
    199     e_shentsize = f.read16()
    200     print "('e_shentsize', %s)" % common_dump.HexDump(e_shentsize)
    201     e_shnum = f.read16()
    202     print "('e_shnum', %s)" % common_dump.HexDump(e_shnum)
    203     e_shstrndx = f.read16()
    204     print "('e_shstrndx', %s)" % common_dump.HexDump(e_shstrndx)
    205 
    206     # Read all section headers
    207     sections = []
    208     for index in range(e_shnum[0]):
    209         f.seek(e_shoff[0] + index * e_shentsize[0])
    210         s = Section(f)
    211         sections.append(s)
    212 
    213     # Read .shstrtab so we can resolve section names
    214     f.seek(sections[e_shstrndx[0]].sh_offset[0])
    215     shstrtab = StringTable(f.read(sections[e_shstrndx[0]].sh_size[0]))
    216 
    217     # Get the symbol string table
    218     strtab = None
    219     for section in sections:
    220         if shstrtab[section.sh_name[0]] == ".strtab":
    221             f.seek(section.sh_offset[0])
    222             strtab = StringTable(f.read(section.sh_size[0]))
    223             break
    224 
    225     print "('_sections', ["
    226     for index in range(e_shnum[0]):
    227         print "  # Section %s" % index
    228         sections[index].dump(shstrtab, f, strtab, opts.dumpSectionData)
    229     print "])"
    230 
    231 if __name__ == "__main__":
    232     from optparse import OptionParser, OptionGroup
    233     parser = OptionParser("usage: %prog [options] {files}")
    234     parser.add_option("", "--dump-section-data", dest="dumpSectionData",
    235                       help="Dump the contents of sections",
    236                       action="store_true", default=False)
    237     (opts, args) = parser.parse_args()
    238 
    239     if not args:
    240         args.append('-')
    241 
    242     for arg in args:
    243         dumpELF(arg, opts)
    244