From 4e45256a069e7b9a3570f095be265e081f02c8f0 Mon Sep 17 00:00:00 2001 From: Daniel Dunbar Date: Fri, 10 Dec 2010 06:19:45 +0000 Subject: macho-dump: Switch to C++ macho-dump tool. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@121466 91177308-0d34-0410-b5e6-96231b3b80d8 --- test/Scripts/macho-dump | 294 ------------------------------------------- test/Scripts/macho-dump.bat | 7 -- test/Scripts/macho-dumpx | 294 +++++++++++++++++++++++++++++++++++++++++++ test/Scripts/macho-dumpx.bat | 7 ++ tools/macho-dump/Makefile | 2 +- 5 files changed, 302 insertions(+), 302 deletions(-) delete mode 100755 test/Scripts/macho-dump delete mode 100644 test/Scripts/macho-dump.bat create mode 100755 test/Scripts/macho-dumpx create mode 100644 test/Scripts/macho-dumpx.bat diff --git a/test/Scripts/macho-dump b/test/Scripts/macho-dump deleted file mode 100755 index 71e06d837b..0000000000 --- a/test/Scripts/macho-dump +++ /dev/null @@ -1,294 +0,0 @@ -#!/usr/bin/env python - -import struct -import sys -import StringIO - -import common_dump - -class Reader: - def __init__(self, path): - if path == '-': - # Snarf all the data so we can seek. - self.file = StringIO.StringIO(sys.stdin.read()) - else: - self.file = open(path,'rb') - self.isLSB = None - self.is64Bit = None - - self.string_table = None - - def tell(self): - return self.file.tell() - - def seek(self, pos): - self.file.seek(pos) - - def read(self, N): - data = self.file.read(N) - if len(data) != N: - raise ValueError,"Out of data!" - return data - - def read8(self): - return ord(self.read(1)) - - def read16(self): - return struct.unpack('><'[self.isLSB] + 'H', self.read(2))[0] - - def read32(self): - # Force to 32-bit, if possible; otherwise these might be long ints on a - # big-endian platform. FIXME: Why??? - Value = struct.unpack('><'[self.isLSB] + 'I', self.read(4))[0] - return int(Value) - - def read64(self): - Value = struct.unpack('><'[self.isLSB] + 'Q', self.read(8))[0] - if Value == int(Value): - Value = int(Value) - return Value - - def registerStringTable(self, strings): - if self.string_table is not None: - raise ValueError,"%s: warning: multiple string tables" % sys.argv[0] - - self.string_table = strings - - def getString(self, index): - if self.string_table is None: - raise ValueError,"%s: warning: no string table registered" % sys.argv[0] - - end = self.string_table.index('\x00', index) - return self.string_table[index:end] - -def dumpmacho(path, opts): - f = Reader(path) - - magic = f.read(4) - if magic == '\xFE\xED\xFA\xCE': - f.isLSB, f.is64Bit = False, False - elif magic == '\xCE\xFA\xED\xFE': - f.isLSB, f.is64Bit = True, False - elif magic == '\xFE\xED\xFA\xCF': - f.isLSB, f.is64Bit = False, True - elif magic == '\xCF\xFA\xED\xFE': - f.isLSB, f.is64Bit = True, True - else: - raise ValueError,"Not a Mach-O object file: %r (bad magic)" % path - - print "('cputype', %r)" % f.read32() - print "('cpusubtype', %r)" % f.read32() - filetype = f.read32() - print "('filetype', %r)" % filetype - - numLoadCommands = f.read32() - print "('num_load_commands', %r)" % numLoadCommands - - loadCommandsSize = f.read32() - print "('load_commands_size', %r)" % loadCommandsSize - - print "('flag', %r)" % f.read32() - - if f.is64Bit: - print "('reserved', %r)" % f.read32() - - start = f.tell() - - print "('load_commands', [" - for i in range(numLoadCommands): - dumpLoadCommand(f, i, opts) - print "])" - - if f.tell() - start != loadCommandsSize: - raise ValueError,"%s: warning: invalid load commands size: %r" % ( - sys.argv[0], loadCommandsSize) - -def dumpLoadCommand(f, i, opts): - start = f.tell() - - print " # Load Command %r" % i - cmd = f.read32() - print " (('command', %r)" % cmd - cmdSize = f.read32() - print " ('size', %r)" % cmdSize - - if cmd == 1: - dumpSegmentLoadCommand(f, opts, False) - elif cmd == 2: - dumpSymtabCommand(f, opts) - elif cmd == 11: - dumpDysymtabCommand(f, opts) - elif cmd == 25: - dumpSegmentLoadCommand(f, opts, True) - elif cmd == 27: - import uuid - print " ('uuid', %s)" % uuid.UUID(bytes=f.read(16)) - else: - print >>sys.stderr,"%s: warning: unknown load command: %r" % ( - sys.argv[0], cmd) - f.read(cmdSize - 8) - print " )," - - if f.tell() - start != cmdSize: - raise ValueError,"%s: warning: invalid load command size: %r" % ( - sys.argv[0], cmdSize) - -def dumpSegmentLoadCommand(f, opts, is64Bit): - print " ('segment_name', %r)" % f.read(16) - if is64Bit: - print " ('vm_addr', %r)" % f.read64() - print " ('vm_size', %r)" % f.read64() - print " ('file_offset', %r)" % f.read64() - print " ('file_size', %r)" % f.read64() - else: - print " ('vm_addr', %r)" % f.read32() - print " ('vm_size', %r)" % f.read32() - print " ('file_offset', %r)" % f.read32() - print " ('file_size', %r)" % f.read32() - print " ('maxprot', %r)" % f.read32() - print " ('initprot', %r)" % f.read32() - numSections = f.read32() - print " ('num_sections', %r)" % numSections - print " ('flags', %r)" % f.read32() - - print " ('sections', [" - for i in range(numSections): - dumpSection(f, i, opts, is64Bit) - print " ])" - -def dumpSymtabCommand(f, opts): - symoff = f.read32() - print " ('symoff', %r)" % symoff - nsyms = f.read32() - print " ('nsyms', %r)" % nsyms - stroff = f.read32() - print " ('stroff', %r)" % stroff - strsize = f.read32() - print " ('strsize', %r)" % strsize - - prev_pos = f.tell() - - f.seek(stroff) - string_data = f.read(strsize) - print " ('_string_data', %r)" % string_data - - f.registerStringTable(string_data) - - f.seek(symoff) - print " ('_symbols', [" - for i in range(nsyms): - dumpNlist32(f, i, opts) - print " ])" - - f.seek(prev_pos) - -def dumpNlist32(f, i, opts): - print " # Symbol %r" % i - n_strx = f.read32() - print " (('n_strx', %r)" % n_strx - n_type = f.read8() - print " ('n_type', %#x)" % n_type - n_sect = f.read8() - print " ('n_sect', %r)" % n_sect - n_desc = f.read16() - print " ('n_desc', %r)" % n_desc - if f.is64Bit: - n_value = f.read64() - print " ('n_value', %r)" % n_value - else: - n_value = f.read32() - print " ('n_value', %r)" % n_value - print " ('_string', %r)" % f.getString(n_strx) - print " )," - -def dumpDysymtabCommand(f, opts): - print " ('ilocalsym', %r)" % f.read32() - print " ('nlocalsym', %r)" % f.read32() - print " ('iextdefsym', %r)" % f.read32() - print " ('nextdefsym', %r)" % f.read32() - print " ('iundefsym', %r)" % f.read32() - print " ('nundefsym', %r)" % f.read32() - print " ('tocoff', %r)" % f.read32() - print " ('ntoc', %r)" % f.read32() - print " ('modtaboff', %r)" % f.read32() - print " ('nmodtab', %r)" % f.read32() - print " ('extrefsymoff', %r)" % f.read32() - print " ('nextrefsyms', %r)" % f.read32() - indirectsymoff = f.read32() - print " ('indirectsymoff', %r)" % indirectsymoff - nindirectsyms = f.read32() - print " ('nindirectsyms', %r)" % nindirectsyms - print " ('extreloff', %r)" % f.read32() - print " ('nextrel', %r)" % f.read32() - print " ('locreloff', %r)" % f.read32() - print " ('nlocrel', %r)" % f.read32() - - prev_pos = f.tell() - - f.seek(indirectsymoff) - print " ('_indirect_symbols', [" - for i in range(nindirectsyms): - print " # Indirect Symbol %r" % i - print " (('symbol_index', %#x),)," % f.read32() - print " ])" - - f.seek(prev_pos) - -def dumpSection(f, i, opts, is64Bit): - print " # Section %r" % i - print " (('section_name', %r)" % f.read(16) - print " ('segment_name', %r)" % f.read(16) - if is64Bit: - print " ('address', %r)" % f.read64() - size = f.read64() - print " ('size', %r)" % size - else: - print " ('address', %r)" % f.read32() - size = f.read32() - print " ('size', %r)" % size - offset = f.read32() - print " ('offset', %r)" % offset - print " ('alignment', %r)" % f.read32() - reloc_offset = f.read32() - print " ('reloc_offset', %r)" % reloc_offset - num_reloc = f.read32() - print " ('num_reloc', %r)" % num_reloc - print " ('flags', %#x)" % f.read32() - print " ('reserved1', %r)" % f.read32() - print " ('reserved2', %r)" % f.read32() - if is64Bit: - print " ('reserved3', %r)" % f.read32() - print " )," - - prev_pos = f.tell() - - f.seek(reloc_offset) - print " ('_relocations', [" - for i in range(num_reloc): - print " # Relocation %r" % i - print " (('word-0', %#x)," % f.read32() - print " ('word-1', %#x))," % f.read32() - print " ])" - - if opts.dumpSectionData: - f.seek(offset) - print " ('_section_data', '%s')" % common_dump.dataToHex(f.read(size)) - - f.seek(prev_pos) - -def main(): - from optparse import OptionParser, OptionGroup - parser = OptionParser("usage: %prog [options] {files}") - parser.add_option("", "--dump-section-data", dest="dumpSectionData", - help="Dump the contents of sections", - action="store_true", default=False) - (opts, args) = parser.parse_args() - - if not args: - args.append('-') - - for arg in args: - dumpmacho(arg, opts) - -if __name__ == '__main__': - main() diff --git a/test/Scripts/macho-dump.bat b/test/Scripts/macho-dump.bat deleted file mode 100644 index 81484f67d7..0000000000 --- a/test/Scripts/macho-dump.bat +++ /dev/null @@ -1,7 +0,0 @@ -@echo off - -@rem We need to set -u to treat stdin as binary. Python 3 has support for doing -@rem this in code, but I haven't found a way to do this in 2.6 yet. - -%PYTHON_EXECUTABLE% -u %LLVM_SRC_ROOT%\test\Scripts\macho-dump %1 %2 %3 %4 %5 %6 %7 %8 %9 - diff --git a/test/Scripts/macho-dumpx b/test/Scripts/macho-dumpx new file mode 100755 index 0000000000..71e06d837b --- /dev/null +++ b/test/Scripts/macho-dumpx @@ -0,0 +1,294 @@ +#!/usr/bin/env python + +import struct +import sys +import StringIO + +import common_dump + +class Reader: + def __init__(self, path): + if path == '-': + # Snarf all the data so we can seek. + self.file = StringIO.StringIO(sys.stdin.read()) + else: + self.file = open(path,'rb') + self.isLSB = None + self.is64Bit = None + + self.string_table = None + + def tell(self): + return self.file.tell() + + def seek(self, pos): + self.file.seek(pos) + + def read(self, N): + data = self.file.read(N) + if len(data) != N: + raise ValueError,"Out of data!" + return data + + def read8(self): + return ord(self.read(1)) + + def read16(self): + return struct.unpack('><'[self.isLSB] + 'H', self.read(2))[0] + + def read32(self): + # Force to 32-bit, if possible; otherwise these might be long ints on a + # big-endian platform. FIXME: Why??? + Value = struct.unpack('><'[self.isLSB] + 'I', self.read(4))[0] + return int(Value) + + def read64(self): + Value = struct.unpack('><'[self.isLSB] + 'Q', self.read(8))[0] + if Value == int(Value): + Value = int(Value) + return Value + + def registerStringTable(self, strings): + if self.string_table is not None: + raise ValueError,"%s: warning: multiple string tables" % sys.argv[0] + + self.string_table = strings + + def getString(self, index): + if self.string_table is None: + raise ValueError,"%s: warning: no string table registered" % sys.argv[0] + + end = self.string_table.index('\x00', index) + return self.string_table[index:end] + +def dumpmacho(path, opts): + f = Reader(path) + + magic = f.read(4) + if magic == '\xFE\xED\xFA\xCE': + f.isLSB, f.is64Bit = False, False + elif magic == '\xCE\xFA\xED\xFE': + f.isLSB, f.is64Bit = True, False + elif magic == '\xFE\xED\xFA\xCF': + f.isLSB, f.is64Bit = False, True + elif magic == '\xCF\xFA\xED\xFE': + f.isLSB, f.is64Bit = True, True + else: + raise ValueError,"Not a Mach-O object file: %r (bad magic)" % path + + print "('cputype', %r)" % f.read32() + print "('cpusubtype', %r)" % f.read32() + filetype = f.read32() + print "('filetype', %r)" % filetype + + numLoadCommands = f.read32() + print "('num_load_commands', %r)" % numLoadCommands + + loadCommandsSize = f.read32() + print "('load_commands_size', %r)" % loadCommandsSize + + print "('flag', %r)" % f.read32() + + if f.is64Bit: + print "('reserved', %r)" % f.read32() + + start = f.tell() + + print "('load_commands', [" + for i in range(numLoadCommands): + dumpLoadCommand(f, i, opts) + print "])" + + if f.tell() - start != loadCommandsSize: + raise ValueError,"%s: warning: invalid load commands size: %r" % ( + sys.argv[0], loadCommandsSize) + +def dumpLoadCommand(f, i, opts): + start = f.tell() + + print " # Load Command %r" % i + cmd = f.read32() + print " (('command', %r)" % cmd + cmdSize = f.read32() + print " ('size', %r)" % cmdSize + + if cmd == 1: + dumpSegmentLoadCommand(f, opts, False) + elif cmd == 2: + dumpSymtabCommand(f, opts) + elif cmd == 11: + dumpDysymtabCommand(f, opts) + elif cmd == 25: + dumpSegmentLoadCommand(f, opts, True) + elif cmd == 27: + import uuid + print " ('uuid', %s)" % uuid.UUID(bytes=f.read(16)) + else: + print >>sys.stderr,"%s: warning: unknown load command: %r" % ( + sys.argv[0], cmd) + f.read(cmdSize - 8) + print " )," + + if f.tell() - start != cmdSize: + raise ValueError,"%s: warning: invalid load command size: %r" % ( + sys.argv[0], cmdSize) + +def dumpSegmentLoadCommand(f, opts, is64Bit): + print " ('segment_name', %r)" % f.read(16) + if is64Bit: + print " ('vm_addr', %r)" % f.read64() + print " ('vm_size', %r)" % f.read64() + print " ('file_offset', %r)" % f.read64() + print " ('file_size', %r)" % f.read64() + else: + print " ('vm_addr', %r)" % f.read32() + print " ('vm_size', %r)" % f.read32() + print " ('file_offset', %r)" % f.read32() + print " ('file_size', %r)" % f.read32() + print " ('maxprot', %r)" % f.read32() + print " ('initprot', %r)" % f.read32() + numSections = f.read32() + print " ('num_sections', %r)" % numSections + print " ('flags', %r)" % f.read32() + + print " ('sections', [" + for i in range(numSections): + dumpSection(f, i, opts, is64Bit) + print " ])" + +def dumpSymtabCommand(f, opts): + symoff = f.read32() + print " ('symoff', %r)" % symoff + nsyms = f.read32() + print " ('nsyms', %r)" % nsyms + stroff = f.read32() + print " ('stroff', %r)" % stroff + strsize = f.read32() + print " ('strsize', %r)" % strsize + + prev_pos = f.tell() + + f.seek(stroff) + string_data = f.read(strsize) + print " ('_string_data', %r)" % string_data + + f.registerStringTable(string_data) + + f.seek(symoff) + print " ('_symbols', [" + for i in range(nsyms): + dumpNlist32(f, i, opts) + print " ])" + + f.seek(prev_pos) + +def dumpNlist32(f, i, opts): + print " # Symbol %r" % i + n_strx = f.read32() + print " (('n_strx', %r)" % n_strx + n_type = f.read8() + print " ('n_type', %#x)" % n_type + n_sect = f.read8() + print " ('n_sect', %r)" % n_sect + n_desc = f.read16() + print " ('n_desc', %r)" % n_desc + if f.is64Bit: + n_value = f.read64() + print " ('n_value', %r)" % n_value + else: + n_value = f.read32() + print " ('n_value', %r)" % n_value + print " ('_string', %r)" % f.getString(n_strx) + print " )," + +def dumpDysymtabCommand(f, opts): + print " ('ilocalsym', %r)" % f.read32() + print " ('nlocalsym', %r)" % f.read32() + print " ('iextdefsym', %r)" % f.read32() + print " ('nextdefsym', %r)" % f.read32() + print " ('iundefsym', %r)" % f.read32() + print " ('nundefsym', %r)" % f.read32() + print " ('tocoff', %r)" % f.read32() + print " ('ntoc', %r)" % f.read32() + print " ('modtaboff', %r)" % f.read32() + print " ('nmodtab', %r)" % f.read32() + print " ('extrefsymoff', %r)" % f.read32() + print " ('nextrefsyms', %r)" % f.read32() + indirectsymoff = f.read32() + print " ('indirectsymoff', %r)" % indirectsymoff + nindirectsyms = f.read32() + print " ('nindirectsyms', %r)" % nindirectsyms + print " ('extreloff', %r)" % f.read32() + print " ('nextrel', %r)" % f.read32() + print " ('locreloff', %r)" % f.read32() + print " ('nlocrel', %r)" % f.read32() + + prev_pos = f.tell() + + f.seek(indirectsymoff) + print " ('_indirect_symbols', [" + for i in range(nindirectsyms): + print " # Indirect Symbol %r" % i + print " (('symbol_index', %#x),)," % f.read32() + print " ])" + + f.seek(prev_pos) + +def dumpSection(f, i, opts, is64Bit): + print " # Section %r" % i + print " (('section_name', %r)" % f.read(16) + print " ('segment_name', %r)" % f.read(16) + if is64Bit: + print " ('address', %r)" % f.read64() + size = f.read64() + print " ('size', %r)" % size + else: + print " ('address', %r)" % f.read32() + size = f.read32() + print " ('size', %r)" % size + offset = f.read32() + print " ('offset', %r)" % offset + print " ('alignment', %r)" % f.read32() + reloc_offset = f.read32() + print " ('reloc_offset', %r)" % reloc_offset + num_reloc = f.read32() + print " ('num_reloc', %r)" % num_reloc + print " ('flags', %#x)" % f.read32() + print " ('reserved1', %r)" % f.read32() + print " ('reserved2', %r)" % f.read32() + if is64Bit: + print " ('reserved3', %r)" % f.read32() + print " )," + + prev_pos = f.tell() + + f.seek(reloc_offset) + print " ('_relocations', [" + for i in range(num_reloc): + print " # Relocation %r" % i + print " (('word-0', %#x)," % f.read32() + print " ('word-1', %#x))," % f.read32() + print " ])" + + if opts.dumpSectionData: + f.seek(offset) + print " ('_section_data', '%s')" % common_dump.dataToHex(f.read(size)) + + f.seek(prev_pos) + +def main(): + from optparse import OptionParser, OptionGroup + parser = OptionParser("usage: %prog [options] {files}") + parser.add_option("", "--dump-section-data", dest="dumpSectionData", + help="Dump the contents of sections", + action="store_true", default=False) + (opts, args) = parser.parse_args() + + if not args: + args.append('-') + + for arg in args: + dumpmacho(arg, opts) + +if __name__ == '__main__': + main() diff --git a/test/Scripts/macho-dumpx.bat b/test/Scripts/macho-dumpx.bat new file mode 100644 index 0000000000..81484f67d7 --- /dev/null +++ b/test/Scripts/macho-dumpx.bat @@ -0,0 +1,7 @@ +@echo off + +@rem We need to set -u to treat stdin as binary. Python 3 has support for doing +@rem this in code, but I haven't found a way to do this in 2.6 yet. + +%PYTHON_EXECUTABLE% -u %LLVM_SRC_ROOT%\test\Scripts\macho-dump %1 %2 %3 %4 %5 %6 %7 %8 %9 + diff --git a/tools/macho-dump/Makefile b/tools/macho-dump/Makefile index 7f965f36c4..638015e928 100644 --- a/tools/macho-dump/Makefile +++ b/tools/macho-dump/Makefile @@ -8,7 +8,7 @@ ##===----------------------------------------------------------------------===## LEVEL = ../.. -TOOLNAME = macho-dumpx +TOOLNAME = macho-dump # This tool has no plugins, optimize startup time. TOOL_NO_EXPORTS = 1 -- cgit v1.2.3