tools/cygprofile/symbol_extractor.py

   1 #!/usr/bin/python
   2 # Copyright 2015 The Chromium Authors. All rights reserved.
   3 # Use of this source code is governed by a BSD-style license that can be
   4 # found in the LICENSE file.
   5
   6 """Utilities to get and manipulate symbols from a binary."""
   7
   8 import collections
   9 import logging
  10 import os
  11 import re
  12 import subprocess
  13 import sys
  14
  15 import cygprofile_utils
  16
  17 sys.path.insert(
  18     0, os.path.join(os.path.dirname(__file__), os.pardir, os.pardir,
  19                     'third_party', 'android_platform', 'development',
  20                     'scripts'))
  21 import symbol
  22
  23 _MAX_WARNINGS_TO_PRINT = 200
  24
  25 SymbolInfo = collections.namedtuple('SymbolInfo', ('name', 'offset', 'size',
  26                                                    'section'))
  27
  28 def SetArchitecture(arch):
  29   """Set the architecture for binaries to be symbolized."""
  30   symbol.ARCH = arch
  31
  32
  33 def _FromObjdumpLine(line):
  34   """Create a SymbolInfo by parsing a properly formatted objdump output line.
  35
  36   Args:
  37     line: line from objdump
  38
  39   Returns:
  40     An instance of SymbolInfo if the line represents a symbol, None otherwise.
  41   """
  42   # All of the symbol lines we care about are in the form
  43   # 0000000000  g    F   .text.foo     000000000 [.hidden] foo
  44   # where g (global) might also be l (local) or w (weak).
  45   parts = line.split()
  46   if len(parts) < 6 or parts[2] != 'F':
  47     return None
  48
  49   assert len(parts) == 6 or (len(parts) == 7 and parts[5] == '.hidden')
  50   accepted_scopes = set(['g', 'l', 'w'])
  51   assert parts[1] in accepted_scopes
  52
  53   offset = int(parts[0], 16)
  54   section = parts[3]
  55   size = int(parts[4], 16)
  56   name = parts[-1].rstrip('\n')
  57   assert re.match('^[a-zA-Z0-9_.]+$', name)
  58   return SymbolInfo(name=name, offset=offset, section=section, size=size)
  59
  60
  61 def _SymbolInfosFromStream(objdump_lines):
  62   """Parses the output of objdump, and get all the symbols from a binary.
  63
  64   Args:
  65     objdump_lines: An iterable of lines
  66
  67   Returns:
  68     A list of SymbolInfo.
  69   """
  70   symbol_infos = []
  71   for line in objdump_lines:
  72     symbol_info = _FromObjdumpLine(line)
  73     if symbol_info is not None:
  74       symbol_infos.append(symbol_info)
  75   return symbol_infos
  76
  77
  78 def SymbolInfosFromBinary(binary_filename):
  79   """Runs objdump to get all the symbols from a binary.
  80
  81   Args:
  82     binary_filename: path to the binary.
  83
  84   Returns:
  85     A list of SymbolInfo from the binary.
  86   """
  87   command = (symbol.ToolPath('objdump'), '-t', '-w', binary_filename)
  88   p = subprocess.Popen(command, shell=False, stdout=subprocess.PIPE)
  89   try:
  90     result = _SymbolInfosFromStream(p.stdout)
  91     return result
  92   finally:
  93     p.wait()
  94
  95
  96 def GroupSymbolInfosByOffset(symbol_infos):
  97   """Create a dict {offset: [symbol_info1, ...], ...}.
  98
  99   As several symbols can be at the same offset, this is a 1-to-many
 100   relationship.
 101
 102   Args:
 103     symbol_infos: iterable of SymbolInfo instances
 104
 105   Returns:
 106     a dict {offset: [symbol_info1, ...], ...}
 107   """
 108   offset_to_symbol_infos = collections.defaultdict(list)
 109   for symbol_info in symbol_infos:
 110     offset_to_symbol_infos[symbol_info.offset].append(symbol_info)
 111   return dict(offset_to_symbol_infos)
 112
 113 def GroupSymbolInfosByName(symbol_infos):
 114   """Create a dict {name: [symbol_info1, ...], ...}.
 115
 116   A symbol can have several offsets, this is a 1-to-many relationship.
 117
 118   Args:
 119     symbol_infos: iterable of SymbolInfo instances
 120
 121   Returns:
 122     a dict {name: [symbol_info1, ...], ...}
 123   """
 124   name_to_symbol_infos = collections.defaultdict(list)
 125   for symbol_info in symbol_infos:
 126     name_to_symbol_infos[symbol_info.name].append(symbol_info)
 127   return dict(name_to_symbol_infos)
 128
 129 def CreateNameToSymbolInfo(symbol_infos):
 130   """Create a dict {name: symbol_info, ...}.
 131
 132   Args:
 133     symbol_infos: iterable of SymbolInfo instances
 134
 135   Returns:
 136     a dict {name: symbol_info, ...}
 137     If a symbol name corresponds to more than one symbol_info, the symbol_info
 138     with the lowest offset is chosen.
 139   """
 140   #TODO(azarchs): move the functionality in this method into check_orderfile.
 141   symbol_infos_by_name = {}
 142   warnings = cygprofile_utils.WarningCollector(_MAX_WARNINGS_TO_PRINT)
 143   for infos in GroupSymbolInfosByName(symbol_infos).itervalues():
 144     first_symbol_info = min(infos, key=lambda x:x.offset)
 145     symbol_infos_by_name[first_symbol_info.name] = first_symbol_info
 146     if len(infos) > 1:
 147       warnings.Write('Symbol %s appears at %d offsets: %s' %
 148                      (first_symbol_info.name,
 149                       len(infos),
 150                       ','.join([hex(x.offset) for x in infos])))
 151   warnings.WriteEnd('symbols at multiple offsets.')
 152   return symbol_infos_by_name